During the past few years, a new view of cancer has emerged which suggests that the cancer cell itself, with all its genetic changes, is not the only culprit in the progression to metastatic tumor growth. This view of cancer has been best put forward in a review by Cold Spring Harbor Laboratory alumnus Douglas Hanahan and Robert Weinberg. They pointed out that for a tumor to develop to the stage where it is a clinical problem, the tumor must have a number of acquired characteristics. These include changes that allow the cancer cells themselves to produce and receive growth signals, to avoid being killed by a process called apoptosis, to proliferate with limitless potential, to attract a blood supply to sustain the increase in cell mass, and to invade and escape from the surrounding tissue. I would add another acquired characteristic that of escaping from the body's immune surveillance that almost certainly helps in suppressing tumor growth, but of which little is known.

There are many ways of collecting the set of acquired tumor characteristics, and clearly accumulating genetic changes in the cancer cell is the primary driving force. But as Hanahan and Weinberg point out, cells that surround the cancer, such as invading immune system cells and the surrounding "normal" tissue, can provide many of the factors necessary to sustain and even change a cancer cell. Importantly, this new view of cancer progression offers a new way of think-ing about cancer therapy. If the acquired characteristics are necessary to create a clinically dangerous tumor, then attacking one of them should provide new hope for cancer treatment. Attacking two different acquired characteristics might provide a benefit greater than the sum of the two alone, and so on. We are just entering an era where this thinking is being tested in the development of new cancer therapies.

The most common method of treating cancers now is to treat the tumor with agents that cause catastrophic damage to cells, such as chemotherapy with DNA intercalating drugs, with drugs that damage the apparatus that ensures accurate segregation of chromosomes, or with radiation that both damages DNA and causes chromosome mis-segregation. These methods take advantage of the loss in cancer cells of the normal response to such external stresses, allowing the cancer cell to proliferate and eventually die due to the catastrophic accumulation of damage. But as we well know, such therapies are toxic to normal cells as well, and the window between killing cancer cells and normal proliferating cells is often all too small. Moreover, as clearly shown by Scott Lowe and his colleagues here at CSHL, cancer mutations such as those in the p53 tumor suppressor pathway can cause resistance to such treatments. This is why cancer therapies must be multidimensional, attacking the tumor from different angles. These new approaches include targeting the proteins in a cancer cell that initiate the tumor to proliferate in the first place, inhibiting the blood supply by anti-angiogenesis therapy, and inducing an immune reaction to the tumor cells.

A priori, it might seem that the protein products of the genetic changes that occur in a cancer cell might not be good targets for cancer therapy because they usually occur early in the life of the cancer cell. Since many genetic changes occur during the life of cancer cells, there is no guarantee that inhibiting one oncogene product will be sufficient to inhibit growth of the cancer cell and even shrink the tumor. But recent results from both basic and clinical research suggest that there is hope for this approach.

Recent research from Michael Bishop, one of the pioneers of cancer genetics, and his colleagues suggests that the primary changes in a cancer cell may well be an Achilles' heel. They created tumor cells that overexpressed the c-myc gene under conditions where it could be turned off at will. In experimental animals, c-myc-dependent tumors arose with predictable frequencies, but interestingly, and for many people unexpectedly, when the c-myc gene was turned off, not only did the tumors stop growing, but the cancer cells died and the tumor regressed. Thus, although a cancer might require multiple genetic changes, targeting a single oncogene product might be sufficient for therapy.