CANCER CELLS

CANCER CELLS TO DIE
S I WAS CONTENDING WITH the debate over cancer funding in
the early 1970s, Charlotte Friend, a professor of microbiology at
Mount Sinai Medical School in New York, was experimenting with a
virus that caused leukemia in mice by inducing genetic mutations. To
better understand the process through which the virus transformed healthy blood
cells into abnormal ones, she decided to “superinfect” the blood cells with the
virus, that is, to force even more of it into the mouse cells.
To do this, Friend treated the cells with a solvent called dimethyl sulfoxide,
or DMSO, that made their membranes more permeable to large molecules.
DMSO is a by-product of pulp and paper production that at one time had a
reputation for remarkable healing properties when dabbed on the skin. But in
1965 the Food and Drug Administration (FDA) banned it for human use because
there was no proof of its effectiveness, and it was also a member of a class of
chemicals, called polar molecules, that are known to be toxic.
The DMSO was merely the first mechanical step in her research, but two or
three days after the cells had been treated with the solution, she observed
something startling: the mouse leukemia cells, which had been colorless before
being treated, were tinted pink and red.
In my own research at Columbia, I had continued my investigation into the
regulation of the expression of globin genes in red blood cells, and I had been
publishing papers about my findings. Globin or, more accurately, hemoglobin is
the vital protein in red blood cells that takes up oxygen when the cells circulate
through our lungs. When hemoglobin is formed, an iron molecule is inserted into
the globin protein, giving it its distinctive red color. Friend, having read some of
my papers, called me to discuss her observations of the DMSO experiment. She
got right to the point.
“What human proteins are red?” she asked.
I answered, “There’s only one.”
I told her it was hemoglobin. What Friend had discovered was that DMSO
had, in effect, switched on a genetic function in the leukemic mouse blood cells
that the disease had switched off. That is how the DMSO had induced the
infected mouse leukemia cells to produce hemoglobin. It was as though they had
gone from sick to healthy. This was new. She reported her findings in the
Proceedings of the National Academy of Sciences, in 1971, and in her paper
“Hemoglobin Synthesis in Murine Virus-Induced Leukemic Cells in Vitro:
Stimulation of Erythroid Differentiation by Dimethyl Sulfoxide,” she offered a
theory about the surprising reaction. “This action of dimethyl sulfoxide, which
was reversible, may represent the derepression of leukemic cells to permit their
maturation.”
I wondered whether this fascinating accident might further my studies into
how the expression of globin genes is controlled within cells. I asked Friend if
she could share some mouse leukemia cells and DMSO so I could reproduce her
experiment. She invited me to her laboratory at Mount Sinai Hospital at Madison
Avenue and East 101st Street to collect the samples.
This was the essence of serious science: having a plan and a goal, but being
willing to alter it at any time because of new information. Scientists need to be
not just receptive when an interesting accident happens, but ready to exploit the
opening and tease out the meaning of the unexpected phenomenon. Study,
planning, and discipline are critical, but the best scientists can improvise when
the unexpected takes over. As Nobel laureate Peter Medawar wrote, “The
prepared mind is essential to good science.” Sometimes the unexpected is a path
to failure, but at others it can be a gateway to discovery.
Excitedly, I returned to my lab with Friend’s samples and treated the mouse
cells with the DMSO. I waited. Nothing happened. The test tubes with the mouse
leukemia cells did not turn red. I repeated the procedure, but the DMSO was not
working. I had no choice but to call Friend and acknowledge my failure. She
said she knew.
Scientists are human, and it turned out she had given me the wrong cells. I
chalked it up to paranoia about my intentions, a test of sorts on her part. But she
agreed to provide the correct sample, and I was soon staring into a test tube
tinted red; about two-thirds of the cancer cells were suddenly in living color. I
was awed. It was like looking at a sort of biological crop circle—this was either
a pretty but meaningless curiosity or a profound, if opaque, message from deep
inside our cells.
That, it turned out, was just the start. Observing the red-tinted cells in a test
tube a few days later, I was at first mystified and then struck by another
observation. The DMSO had not just induced a moribund cellular function to
snap into action, but also stopped the cancer cells from growing. Leukemia
causes blood cells to multiply madly out of control, but DMSO had brought the
process to a halt in all the treated samples.
I turned over the possibility in my mind. Was I looking at a potential
treatment for cancer?
After decades of painstaking science, by the early 1970s we were beginning
to understand the behavior of cancer cells at the most fundamental level. A
normal cell undergoes a genetic mutation or, more often, several mutations that
transform it from a cog in the biological machine of the body, performing some
small but critical task, into a serial killer. The transformation disrupts the normal
processes that regulate cell behavior, and the cancer cells divide out of control,
refuse to die, and then gain mobility and the ability to flow into other parts of the
body, destroying healthy tissue and organs. But we still knew almost nothing
about how to manipulate or reverse those lethal genetic insurgencies.
Charlotte Friend discovered that adding the solvent DMSO to a test tube of mouse leukemia cells spurred
the production of hemoglobin. The test tube went from clear to red.
I had been focusing on relatively narrow questions about the ways that genes
express their coded information in red blood cells. I had not yet approached
cancer directly because of its immense complexity. I was contributing, I
believed, to building a foundation for understanding the ways cancers work. But
now things seemed to have leaped ahead. The potential implications of my test
tube of mouse blood cells were unmistakable: I was looking at a chemical that
coaxed the monster back into a benign state and, even better, did not appear to
endanger healthy cells.
Chemically, DMSO is a simple molecule, represented schematically as:
Polar methyl
-O = S-(CH3)2
On the left is an oxygen molecule (O), which makes up what is known as a
“polar group,” linked to a sulfur molecule (S). Polar groups are chemically
reactive and in normal processes are known to attach to various molecules in the
body. On the right side of the DMSO molecule, the sulfur is linked to two
methyl groups (CH3)2, which can also be involved in various metabolic
reactions in healthy cells. So our first challenge was determining which side of
the DMSO molecule, the polar group or the methyl group, was stopping the
growth of cancer cells.
We conducted the obvious tests on the mouse leukemia cells, first
substituting other compounds for the DMSO that acted as methyl donors. None
worked. But when we substituted chemical compounds that were polar, the
growth of the leukemia cells was stopped cold, leading us to conclude that our
target was some kind of polar molecule. Because an industrial solvent like
DMSO is too toxic to use in humans, we had to find a safe polar compound that
produced the same results.
I pored over a long list of polar chemicals, but I was a medical doctor
steeped in molecular biology—the names meant nothing to me. Working with
my colleagues who had some background in chemistry, we chose a dozen and
ordered them from a supply house. Like magic, all the polar compounds induced
the mouse leukemia cells to produce hemoglobin and to stop growing. (I
explained these results in a paper published in a scientific journal in 1975. But
the phenomenon was so complex that it would take us another twenty years to
figure out that the polar compounds worked by blocking enzymes called histone
deacetylases.) These were promising insights, but we were still in the dark on the
ultimate test, whether the polar compounds could stop cancerous growths in
humans.
We did not let this delay our search for what might be a powerful anticancer
drug, and we moved on to the challenge of finding a different polar chemical
that could be tolerated in humans. I called my friend Roy Vagelos, head of
research at Merck, told him of my findings, and asked if he could give me a
small sample of polar chemicals that they had in their drug-development
pipeline. I was hoping that the chemists at Merck had made something that
would have the effect of DMSO without negative side effects. Roy generously
arranged to send me about two hundred compounds, though he kept the
proprietary chemical formulas confidential. None had the effect of DMSO.
But another happy accident opened the door to more effective approaches. A
talented postdoctoral fellow in my laboratory, Roberta Rubin, had previously
done a fellowship with Columbia chemistry professor Ronald Breslow, and she
suggested that we consult with him. He was immediately intrigued, starting a
collaboration that has lasted more than three decades. After I explained our
findings, he sketched out the structure of several polar compounds on his
blackboard. If one polar group on a chemical like DMSO was good, he argued,
two might be better. So he proposed fabricating compounds with polar groups at
both ends, separated by anywhere from one to eight carbon molecules (C), and
then testing them one at a time.
We determined in the lab that the optimal effect was obtained with a
compound in which the two polar groups were separated by six carbon (C6)
atoms. The chemical was called hexamethylene bisacetamide, or HMBA. It was
fifty times more active than DMSO in turning the mouse leukemia cells red, and
it was not toxic to normal cells in a test tube.
The next step was to begin testing HMBA on various cancer cells in test
tubes. We obtained more than sixty different human cancer cell lines that were
kept at the National Cancer Institute precisely for the sort of experiment we were
conducting. The HMBA arrested the growth in those human cancer cells as well.
We were excited by the promise of what we had found and thought we were
closing in on an understanding of how these compounds stopped cancers.
In refining the DMSO solvent, we found the optimal anticancer benefits came from a chemical compound
in which two polar groups were separated by six carbon atoms, called HMBA.
The real question, of course, was how the chemical would work outside of
test tubes. Would HMBA stop a human tumor from growing? We decided to
take the first preliminary step toward finding out: we would try the HMBA in
live mice with the experimentally induced leukemia. The efforts failed. The drug
had no therapeutic effect on the leukemia; the mice died in six to eight weeks
whether or not they received HMBA.
In an article we wrote in 1979 for the Annual Review of Biochemistry, we
described how we had found a potentially “new and provocative mechanism” for
arresting cancer but acknowledged that the mechanism by which it worked was a
mystery. We also did not understand why the HMBA worked in test tubes but
failed in stopping the cancer in mice.
We had run into the inscrutability and evasiveness of the disease. Each new
problem we faced had led to a string of new molecular conundrums that taxed
our scientific understanding of the cell. But by 1980 we knew that we had
stopped the unlimited proliferation of cancer cells in test tubes with a chemical
that did not kill healthy cells.