Desarrollo de un APS

1_{Department of Pharmacology, Schering-Plough Research Institute, PO BOX 20, 5342}

CC Oss, The Netherlands; 2_{reMYND NV, Bio-incubator, Leuven Arenberg Research}

The VitotoxTM _{and RadarScreen assays were evaluated as early screens for bacterial}

mutagenicity and in vitro mammalian genotoxicity, respectively. The VitotoxTM _{assay is}

a bacterial reporter assay in Salmonella typhimurium based on the SOS response, and it contains a luciferase gene under control of the recN promoter. The RadarScreen assay is a RAD54 promoter-linked β-galactosidase reporter assay in yeast.

Recently an ECVAM workgroup defi ned a list of 20 genotoxic and 42 non-genotoxic compounds [Mutat. Res 653 (2008) 99-108], that can be used for the validation and/ or optimization of in vitro genotoxicity assays. In the present study, this compound set was used for the validation of the assays. Moreover, an additional set of 192 compounds was used to broaden this validation study. The compounds of this additional set can be classifi ed as non-genotoxins and genotoxins and consists of both in-house and reference

compounds. In case of the ECVAM compound list, the results from the VitotoxTM _and

RadarScreen assays were compared to the genotoxic/non-genotoxic classifi cation of the compounds in this list. In case of the additionally tested compounds, the results

of the VitotoxTM _{and RadarScreen assays were compared, respectively, to bacterial}

mutagenicity (Ames) or in vitro mammalian genotoxicity data obtained in-house or from the literature.

The validation with respect to the ECVAM compound list resulted in a sensitivity

for both the VitotoxTM _{and RadarScreen assay of 70% (14/20). If both assays were}

combined the sensitivity increased to 85% (17/20). Both tests also gave a low number of

false positive results. The specifi city of the VitotoxTM _{and RadarScreen assay was 93%}

(39/42) and 83% (35/42), respectively. This resulted in a predictivity of the VitotoxTM

and RadarScreen assay of 85% (53/62) and 79% (49/62), respectively. In case both tests

were combined the specifi city and the predictivity of the VitotoxTM _{and RadarScreen}

assay appeared 81% (34/42) and 82% (51/62), respectively.

The results from the additional list of 192 compounds confi rmed the results found

with the ECVAM compound list. The results from the VitotoxTM _{assay showed a high}

correlation with Ames results of 91% (132/145). Subsequently, the RadarScreen assay had a correlation with in vitro mammalian genotoxicity of 76% (93/123). The specifi city

of the VitotoxTM _{assay was 94% (90/96) for Ames results and that of the RadarScreen assay}

was 74% (34/46) for in vitro mammalian genotoxicity. Moreover, the sensitivity of the

VitotoxTM _{and RadarScreen assays were 86% (42/49) and 77% (59/77), respectively.}

Implementation of the VitotoxTM _{and RadarScreen assays in the early research phase}

of drug development can lead to fast deselection for genotoxicity. It is expected that this application will reduce the number of compounds that have a positive score in the regulatory Ames and in vitro mammalian genotoxicity tests.

Introduction

Screening on genotoxicity and immediate deselection of genotoxic compounds in the early research phase of drug development can improve the success rate of new chemical entities. In the early research phase, medium- or high-throughput toxicity assays are a prerequisite as the number of compounds is high while the available amount of compound is limited.

The regulatory genotoxicity tests like the Ames, micronucleus (MN), chromosome aberration (CA), mouse lymphoma assay (MLA), and sister chromatid exchange (SCE) tests are all relatively time-consuming and laborious and therefore, at least in their present format less applicable as an early screen.

The VitotoxTM _{and RadarScreen are two simple assays that may be useful for early}

genotoxicity screening. The VitotoxTM _{assay is a bacterial reporter assay based on the}

SOS response system; it contains a luciferase gene under control of the recN promoter [1]. Normally, this recN promoter is strongly repressed, but in the presence of a DNA- damaging genotoxic compound, the RecA regulator protein recognizes the resultant free ends or mismatches in DNA. This initiates a cascade of reactions known as the SOS response that cleaves the LexA repressor and de-represses the strong recN promoter, which leads to an increased luciferase expression. This increase can be quantifi ed by measurement of luminescence after addition of luciferin.

The results from the VitotoxTM _{assay correlate good with results from the Ames test}

[2]. Furthermore, for many compounds the lowest effective concentration (LEC) in the

VitotoxTM _{assay was 5-100 times lower than the LEC measured in the Ames test [2]. This}

higher sensitivity for several compounds may be explained by the fact that induction of DNA repair already occurs at lower compound concentrations than fi xation of a mutation. Furthermore, the Ames assay measures only one specifi c mutation per strain.

The principle of using the induction of the SOS response in bacteria as a measure for Ames positive results has been used in several assays. Several genes playing a role in the SOS response were used as reporter genes. The SOS chromotest which is one of the best known of these assays has been validated in several studies [3-5]. One of the largest validation studies was performed by Mersch-Sundermann et al. [3], who tested a set of 330 chemicals. The predictivity of the SOS chromotest for the Ames test was 86.4% in this study (sensitivity, 78.6%; specifi city, 100%). These results showed that the induction of the SOS response might be a good predictor for Ames results. This was the

reason that the VitotoxTM _{assay was evaluated as an early screen to detect Ames-positive}

compounds.

The RadarScreen assay is based on yeast strain SKAM4 that contains a RAD54 β-galactosidase reporter construct (reMYND, Leuven, Belgium). This β-galactosidase gene is under the transcriptional control of the RAD54 promoter and is expressed after

Compounds supplied to the growth medium can therefore be evaluated for genotoxicity by determining their effect on reporter gene expression. The quantity of β-galactosidase can be measured easily by the addition of its specifi c substrate D-luciferin-o-β- galactopyranoside (Promega, Madison, USA), which is cleaved into galactose and luciferin. The latter product can be measured luminometrically. The advantage of yeast compared with bacteria, is that clastogenic and aneugenic compounds causing chromosome aberrations, micronuclei formation and improper chromatid exchange may also be detected. The RAD54 promoter is also used in the GreenScreen assay from Gentronix [8, 9]. This GreenScreen assay is quite similar to the RadarScreen assay, however, it has a different read-out system with a Green Fluorescent Protein (GFP) gene under control of the RAD54 promoter. The advantage of the RadarScreen over the use of the GreenScreen assay is the sensitive read-out of luminescence. This even makes the use of S9 mixture for metabolic activation in a medium- or high-throughput assay setup possible. Furthermore, many (pharmaceutical) compounds can cause autofl uorescence and may interfere with the GFP measurement. Approximately 25% of the in-house compounds were missed due to these effects [10]. This can be prevented with a luminescent read-out.

A big concern with in vitro genotoxicity assays is the high number of (irrelevant) positive results compared with the in vivo results. For this reason the European Centre for the Validation of Alternative Methods (ECVAM) recently published a list of 62 compounds that can be used for the validation/optimization of (new) in vitro genotoxicity tests [11]. Based on the available data, the compounds were defi ned by an expert panel as genotoxic or non-genotoxic. The list of compounds was divided into 20 genotoxic and 42 non-genotoxic compounds. The latter group also contained 19 non-genotoxic compounds that often give false positive results in in vitro genotoxicity assays. In the present validation study the compounds of this ECVAM list were tested. Also an additional set of 192 compounds was tested to broaden the study in order to obtain a higher comfort level. The compounds in this additional set can be classifi ed as non-genotoxins or genotoxins, and consisted of both in-house and reference compounds. The genotoxins acted via diverse mechanisms and belong to the direct-acting genotoxins (i.e. methyl methane sulfonate (MMS) and 4-nitroquinolin-1-oxide (4NQO), cross-linking agents (e.g. cisplatin), topoisomerase inhibitors (e.g. doxorubicin and ellipticin), nucleotide/ DNA synthesis inhibitors (i.e. 5-fl uorouracil and methotrexate), reactive oxygen species generators (i.e. hydrogen peroxide) and aneugens (i.e. colchicine). Furthermore the compound list contains several steroidal compounds which have been reported as being positive in in vitro mammalian genotoxicity tests [12].

In case of the ECVAM compound list the results from the VitotoxTM _{and RadarScreen}

in this list. In case of the additionally tested compounds, the results of the VitotoxTM

assay were compared with Ames scores. Subsequently, the results from the RadarScreen assay were compared with in vitro mammalian genotoxicity data obtained in-house or

from the literature. These two comparisons were performed as the VitotoxTM _{assay is a}

pre-screen for bacterial mutagenicity (Ames test) and the RadarScreen assay is a pre- screen for in vitro mammalian genotoxicity.