Curing Cancer
When President Richard Nixon initiated the War on Cancer in 1972, we did not know very much about the causes of cancer, although many Americans were dying of it. In the 25 years since then, we have learned a lot. We now have a clear idea of how cells make the decision to divide, and of how damage to this decision-making process leads to cancer. This knowledge, obtained in many dozens of laboratories, is at last providing us with the weapons we need to attempt to cure cancer. While not yet in the marketplace, a variety of therapies are being tested that offer real promise.

How a Cell Decides to Divide
The new cancer therapies depend critically on what we have learned about how normal cells divide. While the process is complex, involving interactions of dozens of proteins within the cell, its general character turns out to be quite straightforward. A cell divides when it receives a signal to do so (often a command issued by another cell) or when it grows large enough to trigger division. The key to understanding the decision to divide is to follow the information as it moves from the surface of the cell inward towards the nucleus where the decision is taken. A variety of proteins are involved in this journey, first receiving a "divide" signal, then passing it through the cell interior to the nucleus. As a group, the genes encoding these proteins cause cancer if changed by mutation to become more active. Like stepping on the accelerator of a car, increasing the activity of these genes amplifies the "Lets divide!" signal, causing the cell to divide repeatedly.

Once it reaches the nucleus, the signal overrides the cell's normal division suppressors. A second way to get cancer is to blow away these division suppressors, called "tumor suppressors." Like letting up on the brakes of a speeding car, decreasing the activity of tumor suppressors speeds up the cell division process.

The New Molecular Therapies
Potential cancer therapies are being developed on many fronts. Many of the most exciting possibilities, which we can loosely call molecular therapies, focus on different stages of the cell's decision-making process.

1. Receiving the signal to divide. The first step in the decision process is the reception of a "divide" signal, usually a small protein called a growth factor released from a neighboring cell. The growth factor is received by a special receptor on the cell surface whose shape fits the growth factor like a hand fits a glove. Mutations that increase the number of receptors on the cell surface amplify the division signal and so lead to cancer. Over 20% of breast cancers prove to overproduce the receptor for epidermal growth factor (EGF).

Therapies directed at this stage of the decision process utilize the human immune system to attack cancer cells. Special protein molecules called "monoclonal antibodies," created by genetic engineering, are the theraputic agents. These monoclonal antibodies are designed to seek out and stick to EGF. Like waving a red flag, the presence of the monoclonal antibody then calls down attack by the immune system on the cell with EGF. Because the cancer cells of breast cancer patients overproduce EGF, they are killed preferrentially. Genentech's monoclonal antibody, called "anti-HER2," has given promising results in initial tests: of 43 advanced breast cancer patients where chemotherapy had failed, 14 stabilized and 5 went into remission when treated with anti-HER2.

2. The relay switch. The second step in the decision process is the passage of the signal into the cell's interior, a semiliquid space called the cytoplasm. This is carried out in normal cells by a protein called Ras that acts as a relay switch. When growth factor binds to a receptor like EGF, the adjacent Ras protein acts like it has been "goosed," contorting into a new shape. This new shape is chemically active, and initiates a chain of reactions that passes the "divide" signal inward toward the nucleus. Mutated forms of the Ras protein behave like a relay switch stuck in the "ON" position, continually instructing the cell to divide when it should not. 30% of all cancers have a mutant form of Ras.

Therapies directed at this stage of the decision process take advantage of the fact that normal Ras proteins are inactive when made. Only after it has been modified by a special enzyme with the jaw-breaking name farnesyl transferase does Ras protein become able to function as a relay switch. Drugs that inhibit farnesyl transferase offer great promise as anticancer therapies. In tests on animals, farnesyl transferase inhibitors induce the regression of tumors and prevent the formation of new ones.

3. Amplifying the signal. The third step in the decision process is the amplification of the signal within the cytoplasm. Just as a TV signal needs to be amplified in order to be received at a distance, so a "divide" signal must be amplified if it is to reach the nucleus at the interior of the cell, a very long journey at a molecular scale. Cells use an ingenious trick to amplify the signal. Ras, when "ON," acts as a enzyme, a protein kinase. It activates other protein kinases that in their turn activate still others. The trick is that once a protein kinase enzyme is activated, it goes to work like a demon, activating hoards of others every second! And each and every one it activates behaves the same way too, activating still more, in a cascade of ever-widening effect. At each stage of the relay, the signal is amplified thousands-fold. Mutations in any of the protein kinases can dangerously increase the already amplified signal and lead to cancer. 5% of all cancers, for example, have a mutant form of the protein kinase Src.

Therapies directed at this stage of the decision process initially employed protein kinase inhibitors. However, the cell uses nearly 1000 protein kinases for other important jobs in the cell, all with highly similar structures, so that generalized protein kinase inhibitors disrupted the activity of many of them, leading to undesirable side effects. A new approach is to prepare so-called "anti-sense RNA" directed specifically against Src or other cancer-inducing kinase mutations. The idea is that the src gene uses a sort of Xerox copy of itself to manufacture the Src protein (the "sense" RNA or messenger RNA), and a mirror image complementary copy of the sense RNA ("anti-sense RNA") will stick to it, gumming it up so it can't be used to make Src protein. The approach appears promising. In tissue culture, anti-sense RNAs inhibit the growth of cancer cells, and some also appear to block the growth of human tumors implanted in laboratory animals. Human clinical trials are underway.

4. Releasing the brake. The fourth step in the decision process is the removal of the "brake" the cell uses to restrain cell division. In healthy cells this brake, a tumor suppressor protein called pRB, blocks the activity of another protein called E2F. When free, E2F directs the cell to copy its DNA. Normal cell division is triggered to begin when pRB is inhibited, unleashing E2F. Mutations which destroy pRB release E2F from its control completely, leading to ceaseless cell division. 40% of all cancers have a defective form of pRB.

Therapies directed at this stage of the decision process are only now being attempted. They focus on drugs able to inhibit E2F, which should halt the growth of tumors arising from inactive pRB. Experiments in mice in which the E2F genes have been destroyed provide a model system to study such drugs, which are being actively investigated.

5. Checking that everything is ready. The most exciting of the new therapies focuses on the final step in the decision process, the mechanism used by the cell to ensure that it's DNA is undamaged and ready to divide. This job is carried out in healthy cells by a remarkable tumor suppressor protein called p53, sometimes called the "Guardian Angel" of the cell. p53 inspects the DNA, and when it detects damaged or foreign DNA it stops cell division and activates the cell's DNA repair systems. If the damage doesn't get repaired in a reasonable time, p53 pulls the plug, triggering events that kill the cell. In this way, mutations such as those that cause cancer are either repaired or the cells containing them eliminated. If p53 is itself destroyed by mutation, future damage accumulates unrepaired. Among this damage are mutations that lead to cancer, mutations that would have been repaired by a healthy p53. 50% of all cancers have a disabled p53. Fully 70%-80% of lung cancers have a mutant inactive p53 -- the chemical benzo (a) pyrene in cigarette smoke is a potent mutangen of the p53 gene.

Late last year a therapy was reported for cancers with a mutant p53. The story of its development spans 20 years. In 1977 virologists studying the human adenovirus (responsible for mild colds) discovered that adenovirus could not reproduce within human cells without a working copy of a virus gene dubbed E1B. What does the protein encoded by gene E1B do that is so crucial? Infecting human cells, the virus E1B binds to the human p53 protein! For over 10 years this finding wasn't appreciated, as little was known about p53. Only five years ago did researchers learn that p53 prevents cells from replicating damaged or foreign DNA. The infected human cell will not replicate the foreign adenovirus DNA unless p53 has first been neutralized. That is why adenovirus needs a working version of E1B -- to bind to p53 in the human cell and block this watchdog from shutting down DNA replication.

In 1992 Frank McCormick, a biochemist from ONYX Pharmaceuticals in Richmond, California, realized that because adenovirus must turn off p53 to replicate, an adenovirus without E1B could not disable p53 and would be unable to grow in healthy cells -- but it would grow just fine in cells lacking p53. What human cells lack p53? Cancer cells! If McCormick was right, adenovirus with disabled E1B would not grow in normal human cells, but would grow in and destroy, a wide range of cancer cells ( those with defective p53).

Initial studies reported in October 1996 have been very promising. In tissue culture the E1B-negative adenovirus does not grow in healthy skin cells, but does grow in a wide variety of tumor cells, including colon and lung cancer cells. When human tumor cells are introduced into mice lacking an immune system and allowed to produce substantial tumors, 60% of the tumors simply disappear when treated with E1B-deficient adenovirus, and do not reappear later. Initial human trials have been started.

While E1B-deficient adenovirus offers great promise as a therapy for a wide range of cancers, including most lung cancers, a significant technical hurdle remains. Recall that the initial animal tests were done on mice with no immune system. Humans have active immune systems. In cancer patients the adenovirus therapy may be neutralized by the patient's immune system before the virus has a chance to do any good, simply because most people have had adenovirus colds in the past and so can be expected to carry antibodies directed against adenoviruses. In these people, such antibodies might attack any E1B-deficient adenovirus introduced to fight cancer. Anticipating this problem, investigators are exploring alternative viruses that would not provoke an immune response.

Molecular therapies such as those described here are only part of a wave of potential treatments under development and clinical trial. The clinical trials will take years to complete, but by the turn of the century -- only three years away -- we can expect to greet a new mellenium in which cancer is becoming a curable disease.