A Model Dilemma

Cell-based toxicogenomics studies are emerging as an alternative to animal testing. But for now, animal models are the necessary choice.

The use of animals in laboratory testing is one of the unfortunate realities of science. No one relishes the thought of suffering and death of animals, but their use in research has been necessary to protect the lives and health of human beings. Toxicology is one of the most animal-intensive fields of research, and, in recent years, there has been a great deal of enthusiasm for replacing these studies with cell-based toxicogenomic studies. The technologies needed for this transition are just coming to maturity, making animal-free toxicology a plausible option. However, there are a number of reasons why it is still necessary to use animal models.

Cost comparisons between animal toxicology testing and cell-based toxicogenomics are some of the most difficult and most contentious comparisons that can be made between the two options. The main reason for this is that it is basically an “apples to oranges” comparison.

Currently, a typical Investigational New Drug (IND) application includes both types of data, and advocates of cell-based toxicogenomics correctly point out that their portion of the case for an investigational new drug is cheaper. Conventional wisdom pegs the cost of a simple, two-week in vivo tox screen (in animals) at around $20,000. A comparable cell-based screen is quite a bit more economical, and can be completed earlier in the development pipeline, saving even more money.

But is a cell-based screen really a replacement for an animal screen? That’s not entirely clear. The FDA requires animal studies to be carried out in addition to any cell-based studies. There’s a good chance that the cost of that cell-based study would increase substantially if the burden of the entire toxicology program were on it. Additionally, many cell-based screens are designed to complement animal studies. Some researchers are interested in modeling rat physiology, not human, so they can prioritize compounds before they go into animals. This is an excellent strategy for minimizing money wasted on compounds that are destined to fail, but the costs of such a study have little to do with the theoretical costs of a toxicology screen using only cell-based toxicogenomic models and no animals.

The devil is in the details
Toxicogenomic assays can make amazingly accurate predictions of a toxicology risk profile. Unfortunately, there are many special situations where it is hard to imagine that a cell-based model could ever match an in vivo model for accuracy.

One such case is drug metabolism. If a drug undergoes first-pass metabolism in the liver, and then a metabolite exhibits a cardiac toxicity, this becomes much more difficult to model in vitro. It’s not impossible, because metabolic activity can be retained in hepatic cell cultures, for example. But it does add an additional layer of complexity if it becomes necessary to accurately model metabolism in the body, and then also model all of the toxic effects of all of the metabolites on all of the organs and tissues of the body. Toxic effects may also be indirect, meaning that they do not affect gene expression, but affect functionality of structures and molecules in the cell. This would be difficult to observe with any genomic strategy.

Gene expression in a Petri dish may be significantly different from gene expression in a live animal, under any given circumstance. A greater understanding of systems biology may eventually correct for these changes, but at the present time, it’s not possible to guarantee that the gene expression profile of an isolated cell line is really going to be the same as gene expression in the intact animal.

Murty Chegalvala, PhD, is program director for Immunology Services at Covance (Denver, Pa.). Among other services, Covance does in vivo animal testing in the area of vaccines and immunochemistry. Says Chegalvala: “In my opinion, you would not be able to replace the information from whole-animal model testing. Cell-based models are a lot more biased because they are isolated systems.”

Choosing the best model
There is a good deal more subtlety to choosing the right animal model than many people realize. The industry has accumulated decades’ worth of experience in selecting and using animal models, and exploring their strengths and weaknesses. Cell-based toxicogenomic modeling is a young field, and can not offer the same depth of experience or versatility.

George Lathrop, DVM, director of Laboratory Animal Resources at Southern Research Institute (Birmingham, Ala.), is responsible for the humane treatment of the animals at the facility, including minimizing the total number of animals used and making sure that animals receive pain relief, if possible. “For the last hundred years, we’ve used mammals, rodents, and non-human primates to develop different non-human platforms. Dogs and pigs for cardiac drugs … for skin and dermatology, rabbits … sometimes it does take testing and massaging of the model. You start with what is known.”

“Massaging” a model may involve altering the animal’s diet, giving supplements (although not for toxicology), or selecting animals of a certain age or sex. Animals are also available as genetic knock-outs, knock-ins, or transgenic models for specific purposes. Disease states can be induced, chemically or surgically, that realistically model human disease. All animal models have shortcomings in representing the human body, but these shortcomings are exceedingly well understood.

Ultimately, the best model for the development of human therapeutics is the human. In order to conform to the traditions and mores of the larger society, it is still unfortunately necessary to carry out toxicology testing on animals.

About the Author
Catherine Shaffer is a freelance science writer specializing in biotechnology and related disciplines with a background in laboratory research in the pharmaceutical industry.

This article was published in Drug Discovery & Development magazine: Vol. 12, No. 3, March, 2009, pp. 14-15.

---

Filed Under: Genomics/Proteomics

 

---