For roughly 20 years, scientists have been working to engineer a
virus that will attack cancer. The basic idea is sound, and every few
years there have been some promising-looking results, with tumors
shrinking dramatically in response to an infection. But the viruses
never seem to go beyond small trials, and the companies making them
always seem to focus on different things.
Over the weekend, Nature Medicine described some further
promising results, this time with a somewhat different approach to
ensuring that the virus leads to the death of cancer cells: if the virus
doesn't kill the cells directly, it revs up the immune system to attack
them. It's not clear this result will make it to a clinic, but it
provides a good opportunity to review the general approach of treating
cancer with viruses.
 The basic idea is to leverage decades of work on some common viruses.
This research has identified a variety of mutations keeping viruses
from growing in normal cells. It means that if you inject the virus into
a healthy individual, it won't be able to infect any of their cells.
 But cancer cells are different, as they carry a series of mutations
of their own. In some cases, these mutations compensate for the problems
in the virus. To give one example, the p53 protein normally induces
aberrant cells to undergo an orderly death called apoptosis. It also
helps shut down the growth of viruses in a cell, which is why some
viruses encode a protein that inhibits p53. Cancer cells tend to damage
or eliminate their copies of p53 so that it doesn't cause them to
So imagine a virus with its p53 inhibitor deleted. It can't grow in
normal cells since they have p53 around, but it can grow in cancer
cells, which have eliminated their p53. The net result should be a
cancer-killing virus. (A great idea, but this is one of the viruses that
got dropped after preliminary trials.)
In the new trial, the virus in question takes a similar approach. The
virus, vaccinia (a relative of smallpox used for vaccines), carries a
gene that is essential for it to make copies of itself. Researchers have
engineered a version without that gene, ensuring it can't grow in
normal cells (which have their equivalent of the gene shut down). Cancer
cells need to reactivate the gene, meaning they present a hospitable
environment for the mutant virus.
But the researchers added another trick by inserting a gene for a
molecule that helps recruit immune cells (the awkwardly named
granulocyte-macrophage colony-stimulating factor, or GM-CSF). The immune
system plays an important role in controlling cancer, but it doesn't
always generate a full-scale response to cancer. By adding GM-CSF, the
virus should help bring immune cells to the site of the cancer and
activate them, creating a more aggressive immune response to any cells
that survive viral infection.
The study here was simply checking the tolerance for two different
doses of the virus. In general, the virus was tolerated well. Most
subjects reported a short bout of flu-like symptoms, but only one
subject out of 30 had a more severe response.
However, the tumors did respond. Based on placebo-controlled trials,
the average survival time of patients like the ones in the trial would
have been expected to be about two to four months. Instead, the low-dose
group had a survival time of nearly seven months; for the higher dose
group, that number went up to over a year. Two of those treated were
still alive after more than two years. Imaging of tumors showed lots of
dead cells, and tests of the immune system indicate the virus had
generated a robust response.
Will this virus finally make it out of clinical trials? Right now,
the company that makes it is still engaged in phase 2 trials (like this
one), which are designed to identify the treatment used for a
large-scale effectiveness trial.  There are still hurdles to overcome,
but the promising results described here indicate the company behind it
is very interested in seeing this through.