About this Author
DBL%20Hendrix%20small.png College chemistry, 1983

Derek Lowe The 2002 Model

Dbl%20new%20portrait%20B%26W.png After 10 years of blogging. . .

Derek Lowe, an Arkansan by birth, got his BA from Hendrix College and his PhD in organic chemistry from Duke before spending time in Germany on a Humboldt Fellowship on his post-doc. He's worked for several major pharmaceutical companies since 1989 on drug discovery projects against schizophrenia, Alzheimer's, diabetes, osteoporosis and other diseases. To contact Derek email him directly: Twitter: Dereklowe

Chemistry and Drug Data: Drugbank
Chempedia Lab
Synthetic Pages
Organic Chemistry Portal
Not Voodoo

Chemistry and Pharma Blogs:
Org Prep Daily
The Haystack
A New Merck, Reviewed
Liberal Arts Chemistry
Electron Pusher
All Things Metathesis
C&E News Blogs
Chemiotics II
Chemical Space
Noel O'Blog
In Vivo Blog
Terra Sigilatta
BBSRC/Douglas Kell
Realizations in Biostatistics
ChemSpider Blog
Organic Chem - Education & Industry
Pharma Strategy Blog
No Name No Slogan
Practical Fragments
The Curious Wavefunction
Natural Product Man
Fragment Literature
Chemistry World Blog
Synthetic Nature
Chemistry Blog
Synthesizing Ideas
Eye on FDA
Chemical Forums
Symyx Blog
Sceptical Chymist
Lamentations on Chemistry
Computational Organic Chemistry
Mining Drugs
Henry Rzepa

Science Blogs and News:
Bad Science
The Loom
Uncertain Principles
Fierce Biotech
Blogs for Industry
Omics! Omics!
Young Female Scientist
Notional Slurry
Nobel Intent
SciTech Daily
Science Blog
Gene Expression (I)
Gene Expression (II)
Adventures in Ethics and Science
Transterrestrial Musings
Slashdot Science
Cosmic Variance
Biology News Net

Medical Blogs
DB's Medical Rants
Science-Based Medicine
Respectful Insolence
Diabetes Mine

Economics and Business
Marginal Revolution
The Volokh Conspiracy
Knowledge Problem

Politics / Current Events
Virginia Postrel
Belmont Club
Mickey Kaus

Belles Lettres
Uncouth Reflections
Arts and Letters Daily
In the Pipeline: Don't miss Derek Lowe's excellent commentary on drug discovery and the pharma industry in general at In the Pipeline

In the Pipeline

« A Trial Too Far | Main | The Novartis Way »

January 24, 2005

What We Are Pleased to Call State of the Art

Email This Entry

Posted by Derek

So what are these cancer animal models that I was speaking of so poorly? On the face of it, they actually seem like they'd be pretty good, other than being rather disgusting. (That said, it's important to keep in mind that they're not as disgusting as watching people die from cancer when you could be doing something about it.)

What you do is take human cancer cell lines and implant them in a mouse, a process called xenografting. When they form a tumor, then you treat the mouse with your drug candidate and see if the growth rate slow, stalls, or reverses compared to untreated controls. Sounds pretty simple.

But the complications show up very quickly as you look closer. For one thing, these human cancer cells are often cell lines that have been propagated for a while in vitro, and there's room to wonder about how much they've changed since their days as primary tissue. There's also the issue of the number of different cancer cell types you could use - hundreds, thousands, more? We know what tissue they came from, and we know some of the biochemical differences between them, but nowhere near all. Not even most of the important differences, if you ask me, since we don't even know what some of those important differences are yet.

What we have are characterizations like "Cell line such-and-such, non-small cell lung cancer, resistant," or "colon, slow-growing, responds to everything." Each cell line has its own reputation. At least the fact that these reputations are pretty constant gives you some confidence that we're all talking about roughly the same cells, which is no small thing. (More than once in the history of cellular research, people have realized that cell lines which were all supposed to be the same thing had drifted apart.)

Another level of difficulty is that these things are implanted, rather than growing in situ in the tissue of interest. Any cell biologist will tell you that the matrix a cell grows in is one of the fundamental variables of cell culture. Now, once the tumor has formed, the cells are surrounded by other cancer cells, which is closer to the real situation. But they're still being vascularized by mouse blood vessels, which obviously respond to mouse signals and carry mouse blood. That's the fundamental animal model problem, and it's a tough one.

Finally, these aren't any old mice. In order to get the cells to "take" when they're injected, these mice have a severely compromised immune system. They mostly have no thymus, for starters (and no hair, either, as a side issue.) Here's one - if you find hairless dog and cat breeds cute, you probably won't mind these guys, either. They don't make very good pets, though, because (as you'd imagine), they will catch every disease available, and likely as not die from it.

At bottom, these models are probably too permissive. As I mentioned the other day, they can make compounds like Iressa look just fine, when we now know that they confer no real benefit in humans. (If our market were nude mice with good health insurance, we'd be set, though, as would the mice.)

So what good are they, and are we really doing a good thing by running them? Well, it's hard to imagine that your compound is going to do any good in humans if it doesn't at least work in the nude mice, so they serve a screening function. It's true, though, that for some years now, if the compound hasn't worked in the mice it's never gotten to humans, so we don't have as many checks on that idea as we'd need to be sure of that assumption. But we see a lot of disconnects like Iressa, which argues for false positives being more of a problem than false negatives.

And I'm not sure how good the models are at rank-ordering compounds, either. I can justify their use as a pass/fail, but that's about it. We should be doing better, and people are trying to. And a lot more are trying behind closed doors - better animal models would simultaneously help large numbers of desperate patients and save the drug industry about a billion dollars. More on all this in another installment.

Comments (5) + TrackBacks (0) | Category: Cancer


1. DrZZ on January 25, 2005 9:25 AM writes...

One of the people that has put a lot of effort into trying to understand this problem is Heiner Fiebig. One of his latest papers is Fiebig HH, Maier A, Burger AM. Eur J Cancer. 2004 Apr;40(6):802-20.

NCI has also taken a look:
Johnson JI, et. al. Relationships between drug activity in NCI preclinical in vitro and in vivo models and early clinical trials. Br J Cancer. 2001 May 18;84(10):1424-31

I think the bottom line is that if you are careful and put a lot of resources into it, you can get SOME predictive value, but there is no real dispute that current models fall well short of providing a solid tool for identifiying major new anticancer therapies.

Permalink to Comment

2. steve on January 26, 2005 8:19 PM writes...

A naive question: has anyone considered instrumenting tumors in vivo in human cancer patients? I mean, sticking probes in to monitor cell behavior. At the least, you could get a sense of what the heck the chemo was actually doing in cells at the surface, below, etc. I realize that lots of cancer is hard to localize and get at physically, but aren't there some you could do this with? Of course, you normally wouldn't be able to use untested compounded on live patients, but looking at how approved compounds work or don't work more specifically might improve modelling overall.

Permalink to Comment

3. Derek Lowe on January 26, 2005 9:21 PM writes...

There's been an awful lot of work put into looking at the genomic (and expression) differences in different human tumors, but (as far as I know) that's all done with biopsy samples, not in vivo. It's true that we could probably get some interesting data (although the kinds of probes we have are pretty limited.) Could be quite uncomfortable for the patients involved, though. . .

Permalink to Comment

4. steve on January 30, 2005 12:02 AM writes...

Isn't that why God invented anaesthetics? Seriously, you'd need to develop really tiny sensors that could be implanted and be queried wirelessly or else very fine needle-type probes. And some serious drugs to make it bearable for the patient.

But I'd imagine it would be useful to know more about things like tumor metabolism at the surface and below the surface of the tumor, etc. It might tell you what happens to chemo drugs when they hit the cancer cells in the actual environment that counts. I'm not sure genetic probes are the first thing to shoot for, although that seems pretty important for eventual understanding.

Permalink to Comment

5. DrZZ on January 31, 2005 8:34 AM writes...

Esentially all of the work on in vivo monitoring of tumor characteristics uses some kind of imaging, not some kind of sensor placed in the tumor. Historically imaging has been concerned with detecting the tumor, but there is an increasing amount of development work going into designing agents that report on some kind of tumor biochemistry. For more information you might want to check out and

Permalink to Comment


Email this entry to:

Your email address:

Message (optional):

The Last Post
The GSK Layoffs Continue, By Proxy
The Move is Nigh
Another Alzheimer's IPO
Cutbacks at C&E News
Sanofi Pays to Get Back Into Oncology
An Irresponsible Statement About Curing Cancer
Oliver Sacks on Turning Back to Chemistry