In 1971, Richard Nixon signed a bill that launched the American “war on cancer.”  That war has sent millions of mice to their deaths. Survival has improved for some cancers; not so much for others. The War on Cancer is still on, and mice remain its conscripts.

Recently, Ted Kennedy died of a malignant brain tumor, probably a glioblastoma, about 15 months after his diagnosis. Forty years into our war on cancer, and despite advances on many fronts, the prognosis for malignant glioblastoma patients remains very poor: about one year from diagnosis to death is common.  

And so the research continues: in human subjects, in cultured cells, and in animal “models” including dogs, rats, and overwhelmingly, mice.

For most studies of brain cancer in people, you would start by screening people diagnosed with brain cancer. Headaches, seizures, vision loss, loss of coordination or other neurologic symptoms lead to screening, such as by MRI, which leads to a diagnosis. As cancers go, they’re not the most numerous, roughly 2% of all American cancer deaths, with roughly one-fourth of those brain cancers being malignant glioblastomas. The NIH lists hundreds of human clinical trials for glioblastoma on their website: With enough human cases, you can assign some to one treatment, some to another, and follow their progress. But how do you decide which treatments to try in those human guinea pigs? How do you know which seem promising enough and safe enough to go forward in human trials?

I call your attention to the mice (and other animals) enlisted as “models” in the cancer laboratories. Without mouse data, very few treatments get to that next step of human testing.

Here are 3 ways to get a population of mice with brain cancers that could then be used in research on better diagnostics, better treatments, better preventions.

First, the search for animal models has always relied on close observation of spontaneous disease in mouse colonies, and so you MIGHT find a natural case of brain cancer in your mice. We call this is a “spontaneous” model, when an animal shows up with a condition you want to study. This has historically been a significant source of laboratory animal ‘models’ for many cancers, endocrine disorders, immune deficiencies and other diseases, but not a good source of animals for brain tumor studies. Even the most brain tumor-prone strain of mouse, the VM/Dk mouse, has an incidence of less than 2% (and of only one type of astrocyte tumor, not necessarily a glioblastoma). Relying on these mice would mean screening thousands of animals to get enough subjects for a single experiment. It would be one thing if that simply meant looking in the cage and picking out the rare animal with a brain tumor, but that would not be the case: they look so normal in those early stages that you would need to screen animals with something like an MRI, and possibly more than once per animal, to identify your study animals. Thousands of mouse MRIs to get a few dozen study subjects is too inefficient for most research projects.

More typically, mouse brain cancer studies start by intentionally causing the brain cancers into previously healthy animals. If you inject tumor cells into mice under the right conditions, those tumors will grow as if they were the mouse’s own cancer.  For some studies, you can inject brain tumor cells under the mouse’s skin in a site on the side where they won’t cause the mouse much problem until they grow really large. For most brain cancer studies, you will inject those tumor cells directly into the mouse’ brain. That’s a major surgical procedure. You anesthetize the mouse, placing him or her in a stereotactic head-holder, make a skin incision and drill a 1 mm hole in the top of the skull. Through this hole, you insert a needle and inject cancer cells. Withdraw the needle, close the surgical site, and monitor your mouse for signs of tumors. You’ve launched your experiment.

A third and new approach: Salk Institute scientists published a genetic technique this spring to engineer mice with spontaneous tumors of the glia. Rather than insert dozens or hundreds of tumor cells into an immune-suppressed mouse, they insert cancer-forming genes into the brains of normal mice, to more closely mimic what probably happens in people (that is – genetic changes in a handful of cells lead to cancers).   For more on this, see

Once you establish a population of mice with brain tumors, you can then proceed to your studies of the best ways to image them. You can try out chemotherapies that will target them, deliver cell toxins directly to them, turn off their blood supply, irradiate them – whatever your experiment is trying to develop.

My focus in this blog is on these mice --- so in my next couple of postings I’ll get into more detail about a vet’s-eye view of the animal welfare hotspots in these kinds of studies – where I see the greatest potential for animal suffering and therefore, the greatest need for animal welfare interventions.