Search This Blog

Tuesday, July 25, 2006

Curing Cancer

A common form of biotech R&D is finding something that binds to a drug target and seeing what affect that has on the disease. Antibodies are very popular because they bind to proteins. They were designed to do so by God (insert your god here). The immune system makes antibodies against foreign proteins, they bind to the foreign protein and the rest of the immune system removes the antibody and what it is binding to.

One approach to curing cancer was to make an antibody against denatured collagen. The concept was that it will bind to areas who tissue is breaking down. If a cancer is invading a tissue the antibody will go to that area. Good idea so far. The leap of faith in this research was that denatured collagen held a criptic binding site for angiogenisis factors. There is no evidence of this. The way around it was to simply inject the antibody into a mouse with a tumor and see if it prevented the tumor from growing. A tumor needs blood vessels to provide oxygen and carry away its waste. If you inhibit angiogenisis then the tumor will shrivel up.

Another protein target to prevent tumor grown was VEGF (vascular endothelial growth factor). The antibody against this target binds to the target and it prevents tumors from growing in mice. This I have witnessed first hand. As a control for the anti-angiogenisis antibodies we had, we used the anti-VEGF antibody. In this group of mice the most you could see was a red swelling where the tumor cells had been injected. In all other groups tumors had formed, of all different shapes and sizes. Our antibody didn't come close to the results of the anti-VEGF antibody.

The concept has been proven, in mice at least. An antibody can be developed and used as a drug. Rituxan, (anti-VEGF) does not work as well in humans but it is for sale for cancer treatment. The anti-collagen antibody is in preclinical development, just where it was when I left the project over four years ago. Each year a presentation is given at a biotech seminar on how well it reduces tumors. 57% in nude mice, 67% in female Balbc mice. The truth is, it doesn't work at all. It didn't work in laboratory assays such as adhesion and cell migration assays. It didn't work in mice. A graduate student showed us how to measure the tumors to get "better data". After measuring and getting the numbers she wanted she made the mistake to ask if I had any questions. I handed her one of the mice she had already measured. She was immediately embarrassed. "Which one is this?" she asked. "If I told you it would take away from the experiment of testing the reproducibility of what you just demonstrated," I answered. She took a shot. She was off by 30%. Imagine how the measurements of entire groups of mice are treated. Then think about the presentation to the big time scientists at the biotech conferences. Any questions?

By the time it's all said and done, we still do not have a good anti-cancer drug. We have a ton of molecules that bind to something that may be involved in cancer. How do we know when our drugs reach their targets? How much is needed and where? Biotech is swinging for homeruns by finding targets and finding something that binds to them. Peptides, antibodies, enzymes, phage, viruses, small molecules and other sythetic compounds. One thing can be said about an antibody however. It binds to a protein but it also is part of a system that protects us from harm. It does not work alone. Binding alone is not enough.

No comments: