Molecular Toxicology: Bridging Animal and Human Studies

Pam Contag, Ph.D., Bio Imaging Founder and Executive, New Ventures, Caliper Life Sciences

A major shift in molecular toxicology will be the search for biomarkers that you can connect preclinically in animals and humans to a specific disease process. We recently spoke to Caliper’s Pam Contag to get her perspective. The following contains excerpts from that interview.

Molecular toxicology is a relatively new field and one where people are very interested in novel technologies. The drawback is that any new approach has to be validated if it is going to be used in submissions to regulatory agencies, either in Europe or here, to the FDA. As a result, most of the changes we see in this field are not dramatic; rather, they are incremental.

Traditionally we think of biomarkers such as glucose and cholesterol, levels of which are markers of disease and its extent. But nowadays, we want to know more about all of a compound’s effects, and whether we’ll see the same things in humans as in our animal models. You want to know if you can have confidence in what you see in your animal models. Many drugs’ side effects are never noticed until post-launch, after many patients have taken the drug. New tools are necessary if we are going to get more information earlier. 

One of the reasons I love the merging of Xenogen and Caliper is that we can now offer excellent means to do comparisons and to distinguish between early stage events and later toxic effects. 

All the technologies that allow you to do a differential transcription and translation analysis – genomics, informatics, RNA expression technology, and proteomics – help us understand the level at which genes and proteins are expressed in a normal and disease state. Our interpretation of that information is complicated by the fact that there are so many changes going on in cells. Unless you do the total in vivo animal studies, it is really hard to pick out that key marker and know what pathway it is in, and then trace it back to the disease state or toxic effect. 

And ultimately none of this is useful unless you can connect that information in a meaningful way and then predict how the compound will behave in humans. 

As one of my toxicology professors liked to say, the problem facing all biologists is that once you get into an organism, you need to face the biochemical dragon. Toxicology is an extremely complex issue. Humans, animals, and cells all have redundant pathways. You have to think of it as a cascade effect. If you block one pathway, a domino effect occurs. The energy has to go somewhere. By shutting down a gene with a drug you can essentially push all the pathways in a different direction. Some are redundant, some act to protect the organism, some are toxic. We have to sort that out.

Lessons from Sepsis

What we all want now is the best animal model possible for the human disease we are studying. 

An important lesson for us all was what happened in sepsis during the late 1990s. So many candidate sepsis drugs failed at this time that it became almost impossible to get newer projects funded in this indication. 

We only had cell culture models for sepsis, and not everything we did in these cell lines translated to humans. We had the giant realization that not all sepsis occurs through TNF- (tumor necrosis factor alpha) or NF-K B (nuclear factor-kappa B). But even when we figured that out, we still did not have the diagnostics to figure out what was going on in a particular patient. This is a challenging indication already, because morbidity and mortality in sepsis occur within 72 hours.

If you try to target things that are important early in the pathway you may not interfere with what happens later. So we need to build our drugs for diseases like this in tandem with diagnostics – biomarkers that may themselves not even be drug targets but that tell you what’s going on and where you are in the disease’s trajectory.

For example, matrix metalloproteinases (MMPs) are elevated in certain diseases, but have nothing to do with the symptoms that you’d like to treat. So while they can be biomarkers in those cases, they are certainly not targets.

Next Steps with Animals

The next step in the evolution of toxicology is to get better information from animals. You should be able to get much more information from mice than the simple “tails up or tails down” analysis. Since the entire genome has been sequenced for black 6 mice, we can do a nice comparison between human and mouse genes. 

It is easy to be led astray, however. CYP450 genes between mice and humans are similar, but mice metabolize drugs differently than humans. Why is that? We need to ask those questions. One approach is to clone the human receptors and put them into mouse models. Then you ask whether the pathway of interest is induced or not, and what are the byproducts.

I’m not in favor of looking at phenotype of disease in an animal. The reasons animals die of their particular diseases and how they demonstrate toxicity is often not the same as what we see in people. Do not make the mistake of looking at whole animals for insights into human diseases. Instead you have to look at the pathway information. After all, we have never stopped a trial on a mouse because of headache. 

Right now, we tend to always measure much more extreme things in mice. We need to take a step back and look at a mouse, rat, etc. and use finer methods and tools that distinguish what’s happening in specific pathways.

For example, mice get tumors very easily. That complicates things when you are looking at drug side effects. But you don’t have to do away with mice. It is worth taking our time and taking a better look and seeing how we can use animal models better.

I think two important directions of the future will be immune molecules and the brain. Immune molecules can be great markers of things such as inflammation. And we are not so good at finding biomarkers in the brain, but there are many different diseases that increase or decrease neurotransmitter levels. We might be able to look carefully at what is going on in the brain and learn more about a whole range of conditions. 

The key is to not to stick with any one modality. I combine my business with a high-throughput in vitro technology. I think that going forward, it is going to be extremely necessary to bring technology to high throughput while connecting it. It will be a long time before we have in silico substitutes for animals. The low-hanging fruit is gone and now we have to go and develop new technology if we want more answers.

5-02- 07 - Biomarker Breakthroughs