<b>A Combined Experimental/Predictive Approach To Toxicity</b><br/><br/>

A Combined Experimental/Predictive Approach To Toxicity

by Robert Milczarek, John M. Prewitt, IV, James H. Kelly, Debra L. Toburen, and David A. Evans

Introduction

Figure 1. Standard toxins tested in ACTIVTox behave as predicted by primary hepatocytes. When the IC50 for a series of known hepatotoxins are tested in ACTIVTox via LDH release, cells die at the concentrations predicted by primary hepatocytes. IC50s are calculated directly in MDL Assay Explorer.

Figure 2. Aflatoxin B1 is toxic in ACTIVTox but not in HepG2. Aflatoxin B1, a standard hepatotoxin, is toxic in ACTIVTox (blue squares) at the same concentrations at which it shows toxicity in primary hepatocytes. HepG2, a widely used human hepatoblastoma cell line (3) is not killed even at the highest concentrations tested here since it is unable to metabolize aflatoxin to the toxic epoxide. HepG2 was cultured in both DMEM (green circles) as recommended by the ATCC, and in ACTIVTox medium (red triangles).

Liver toxicity is a common cause of failure in new drug development.⁽¹⁾ A system for predictive hepatotoxicology that could be used both in the design of new screening campaigns and then in optimizing leads could dramatically reduce the time and cost of drug development. Any such system consists of three essential parts: a reliable human hepatotoxicity model, a database of compounds that have been screened in the system, and software to collate the data.

ACTIVTox (Amphioxus Cell Technologies, Inc., Houston, TX) is a hepatocyte based system coupling a human liver cell line with simple, rugged assays for hepatotoxicity, such as LDH release. Elsevier MDL software provides a series of modules, such as MDL Assay Explorer, that can be used to manage screening campaigns as well as visualize the results. Together, Elsevier MDL and Amphioxus Cell Technologies are creating a large database of compounds tested in a variety of toxicity assays that will provide a framework in which to examine the structural basis of liver toxicity. This tool has the unique attribute that proprietary compounds can be screened in the same experimental system and added to the database, creating a customized environment for lead optimization.

Aims

To demonstrate the utility of a combined experimental/predictive approach to toxicity. One of the major limitations of current predictive software is that the data on which the model is built are inaccessible to the user. The Elsevier MDL system encourages incorporation of user generated data.
To demonstrate the utility of MDL Assay Explorer and ACTIVTox in this application. Both systems are designed for high-throughput and simplicity.

Method

ACTIVTox cells were cultured in serum containing medium. Drugs were dissolved as 100× concentrates in DMSO and diluted into medium such that DMSO was 1% at the highest concentration of drug. Compounds were incubated with cells in 96 well plates in standard CO₂ incubators for 48 hours before assay. All compounds were tested in triplicate. As a direct measure of toxicity, aliquots of the supernatant were taken for lactate dehydrogenase assay using Promega (Madison, WI) CytoTox96 kits. For cyp1A induction, ethoxyresorufin was added directly to the medium at a concentration of 10 μM.⁽²⁾ After one hour, the reaction was quantitated using a fluorescence plate reader.

Results

CYP1A Induction. A common toxicology question is: Does my compound interact with other drugs by inducing the drug metabolizing enzymes? By screening compounds for cyp1A induction, and employing MDL Assay Explorer, a quick answer can be found. As part of our database construction, we are screening the Sigma LOPAC1280 collection (Sigma-Aldrich, St. Louis, MO). Cyp1A induction is a simple homogenous fluorescence assay. The data shown here examined compounds for induction at 10 μM in triplicate.

MDL Assay Explorer collects the data from the readers, finds the mean and standard deviation, and plots the data. By applying selection criteria, in this case, fold induction <15, the compound information and structure can be immediately displayed. This selection criteria identified a number of known and unknown inducers of cyp1A.

Figure 3. Results were generated with the ACTIVTox cyp1A assay system and the raw data captured and calculated in MDL Assay Explorer.

Figure 4. MDL Assay Explorer supports the key data collection and
analysis steps from screening to in vivo and provides a central
organization for laboratory data management.

Figure 5. Using MDL ChemBio AE and MDL Report Manager to examine
compounds expected to behave as cyp1A
inducers.

Conclusion

The ACTIVTox provides a high-throughput, hepatocyte-based tool for predictive toxicology, while MDL Assay Explorer provides a framework for collection and visualization of screening data coupled directly to pertinent chemical information. The assays combine with MDL Assay Explorer and other Elsevier MDL software to create a database of information that can be used throughout the discovery process to identify and eliminate hepatotoxicity.

About the authors

Robert Milczarek, John M. Prewitt, IV, and James H. Kelly are with Amphioxus Cell Technologies in Houston, TX; Debra L. Toburen and David A. Evans are with Elsevier MDL in San Ramon, CA.

More information is available from the respective companies, or from: Elsevier MDL925-543-5400www.mdl.com

References

1. Maggs, J.L., Pirmohamed, M., Williams, D.P. The role of metabolic activation in drug-induced hepatotoxicity. Ann. Rev. Pharmacol. Toxicol. 45:177-202 (2005).
2. Kelly, J.H., Sussman, N.L. A fluorescent cell based assay for cytochrome P450 isozyme 1A2 induction and inhibition. J. Biomol. Screen. 5:249-254 (2000).
3. Knowles, B.B., Howe, C.C., Aden, D.P. Human hepatocellular carcinoma cell lines secrete the major plasma proteins and hepatitis B surface antigen. Science 209:497-499 (1980).