In the past several years, a few investigators have undertaken new approaches to the problem of trying to understand how drugs attack cells and how they may be utilized more effectively in killing cancer while sparing normal cells. Through studies of cell kinetics, scientists have been comparing the growth characteristics of populations of cancer and normal cells and determining the relative effects of known cancer drugs on them. These studies have been carried out in tissue culture and in animal systems, such as mice bearing an advanced transplanted lymphoma which hopefully would serve as a "model" for the development of a rationale for the curative treatment of human leukemia.

Scientists can gain insight into the effects of drugs on cancerous and normal cells by comparing the growth characteristics of each type. The exponential curve illustrated above represents the theoretical rate of tumor cell growth, assuming that the cells are identical and non-differentiating. Each cell divides into 2 cells once each generation time, so that by 10 generations a single cell proliferates into 1,000 (10^2) cells.

Alkylating agents, such as nitrogen mustard, attack cells in all phases of the division cycle; antimetabolites, such as methotrexate, attack cells only in the S phase; the action of vincristine, a mitotic inhibitor, seems to occur in the G 2 phase, just prior to mitosis. Thus, the toxic effects of drugs appear related to the different phases of the cell division cycle. Normal cells have periods when drugs do not affect them; but tumor cells are not protected in this way from the toxicity of drugs. This information suggests a way of producing selective toxicity; that is, increasing the destruction of a tumor cell population without increasing the destruction of normal cells.

## Cancer Cell Populations

Calculations based on laboratory experiments suggested that the simplest situation for cell growth would hypothetically be one in which a tumor consists of identical, actively proliferating nondifferentiating cells. Each tumor cell divides into two cells once each generation time (the time for one entire division cycle) and the generation time is approximately the same for all cells in the population. The sequence of doubling is such that by 10 generations a single cell proliferates into 1,000 (10^3) cells. The growth of many cells following the same behavior will follow the same rule and will result in a thousandfold increase every 10 generations. Such growth is called exponential or logarithmic, and its pattern can be predictably charted as a smooth curve.

It was also estimated that to be detected a tumor must have a volume of at least 1 cubic centimeter (1x1x1 centimeter) consisting of roughly 10^9 cells (1 billion or 1,000,000,000); the estimate for a critical volume, that is, one that would kill a patient, is 1,000 cubic centimeters (10 x 10 x 10 cm.), or roughly 1012 cells (1 trillion). Since these values are a thousandfold (103) apart, their separation in time represents 10 generations of growth.

It has been calculated that the best therapies today can reduce the lymphocytic leukemic cell population in acute leukemia of childhood from 1012 to 106. At this point the disease is undetectable and the patient is in "complete remission." The goal of chemotherapy, then, is to reduce the last million leukemic cells (106) to zero, since scientists, in general, assume that 100 percent cell kill of leukemic cells is necessary to achieve cure. This assumption is based on a study showing that a single leukemic cell implanted in a mouse could multiply and eventually cause the death of the animal.

The practical problems are many. Quantitative methods for measuring the number of tumor cells remaining after treatment are being sought. At the present time, the number of cancer cells ordinarily cannot be measured below the 109 level, and they can be determined only indirectly by the duration of unmain-tained remission. Also, investigators are recognizing that diameter, volume, or weight is not a satisfactory end point for measuring the response of a solid tumor to treatment, since the size of a tumor seems to be related to the proportion of viable and dividing cells and dead and dying cells. After an effective treatment the size of the tumor may increase slightly, even though a large fraction of the tumor cells may have been killed. Sometimes, after cells have been exposed to a lethal dose of a drug they can go through one more mitosis and actually show an increase in volume, before the tumor decreases in volume and then regrowth begins.

Other problems relate to the fact that tumors are not uniform in growth or composition. Experimental solid tumors and mouse leukemia cells grow exponentially at first, but then the rate of increase slows down. In addition, after treatment with drugs some remaining tumor cells may be different kinetically; for example, they may be in a prolonged resting phase, and may therefore be less susceptible to attack by drugs.

Significant information has been obtained indicating that a judicious choice and spacing of drugs may produce selective toxicity: it is possible to increase doses of some drugs to increase destruction of a tumor cell population without increasing normal cell kill. This selective toxicity appears related to the effects of the drugs on different stages of the cell division cycle. In general, alkylating agents, such as nitrogen mustard, act on cells in all phases of the generation cycle and also act on some resting cells. Antimetabolites, such as methotrexate, attack cells only in the S (DNA synthesis) phase of the generation cycle and do not act on resting cells. Vincristine, a mitotic inhibitor, has not yet been as thoroughly studied; its action seems to occur in the G2 phase, just prior to mitosis.