Mutations happen all the time in human DNA, says Kristin Eckert, "or else there wouldn't be evolution."

In normal cells, mutagenesis is closely regulated. Eckert, an assistant professor of pathology at Penn State's Hershey College of Medicine, explains:

"There are checkpoints in the cell cycle. If the cell detects an abnormality in its DNA, it's like a stress response. A signal goes out that says 'Stop. Things aren't right.' Repair enzymes arrive. They make their repairs, and then things return to normal."

The continued health of the cell, then, comes down to a kind of ongoing race between the forward march of replication and the backing and correcting of repair. But what are the actual, biochemical workings of this system -- and what exactly changes when a cell becomes abnormal? What is it that gets fouled up?

Eckert's research into these questions focuses on a key player in mutagenesis: DNA polymerases.

Polymerases are the enzymes that carry out replication, the twinning of a cell's genetic material. When a cell divides, its DNA strands are pulled apart; a DNA polymerase comes along, reads the unique sequence of bases, and taps out an exact copy. You end up with double the DNA, one complement for each daughter cell.

"Imagine you're a typist keying in a phone book, where every space and character has meaning," Eckert says. "You have to enter all that information in a short amount of time as accurately as possible. That's what the job of a polymerase is like."

"These are very accurate enzymes," she adds. Still, sometimes the polymerase makes a typo. In normal cells, however, because the replication process is so tightly regulated, the mistake is cleaned up before it causes any trouble.

In some tumor cells, it is this saving regulation that has been lost. The tumor cells have lost functioning repair proteins, so after the polymerase goofs up, the error doesn't get rectified. Instead, it gets compounded. "Ordinary" mutations begin to multiply, spinning out of control. "This is one of the most exciting new findings in cancer research," Eckert says.

There's a hot theory that the linch-pin to all this is a regulator gene called p53: once it is removed, the result is unchecked mutation. Another theory suggests that the genes that encode for the polymerases themselves might somehow be altered. "But this hasn't been proven," Eckert says.

Part of the reason for the persistent uncertainty is that we dont really know very much about human DNA polymerases.

Most of what we do know, says Eckert, comes from studies of bacteria -- and that just isn't good enough. "There are many reasons to believe that human cells are very different, that they have novel mechanisms of making mutations that won't be seen in bacterial cells."

In humans, for starters, there are three types of DNA polymerase needed to replicate DNA, not one, as in the common bacteria E. coli. Each type, Eckert says, "is a very different enzyme," and each operates at a different phase of replication. There, however, "the biochemistry begins to get fuzzy."

A fourth human enzyme, dubbed polymerase-beta, is known to make more errors in replication than the others do. Pol-beta's errors, moreover, are of a particular type.

Most any gene, Eckert explains, contains stretches of repetition. In a simple repeat, one DNA base -- an A, say -- comes up several times in a row: AAAAA. When pol-Beta reads such a sequence, Eckert says, it tends to miss one of these A's -- "or else it'll add one." This quirk is known as a "frame-shift" error.

Human cells, Eckert continues, have lots of repetitive sequences -- more than bacteria cells do. Typically, these are two- and three-base repeats: AGAGAGAG, say, or TGCTGCTGC . . . . Typically, too, they occur in regions of so-called "junk" DNA, stretches that don't seem to code for a protein.

All of us have these di- and tri-repeats scattered throughout our DNA. Recently, though, biochemists made a tantalizing connection: Unusually long repeat sequences can be linked to genetic disease. People with Huntingdon's disease, for example, have been shown to possess extra-long sequences of a particular tri-repeat.

One implication of this finding, Eckert notes, is that junk DNA is not junk. "It's there for a purpose. We just aren't smart enough to have figured that purpose out yet."

The other implication, regarding polymerases, is that those pol-Beta frame-shift errors may be very important in the development of abnormal cells.

Here's the thinking. When a pol-Beta comes across a long stretch of repeats, its usual gaffes begin to mount -- maybe to the point of crossing some threshhold beyond which they can't be contained.

In order to test this theory, Eckert says, we first need to find out a lot more about the normal biochemistry of DNA polymerases. She is currently setting up a lab to do just that. In March, Eckert received a sizeable boost when she was awarded the Gertrude Elion Cancer Research Award by the American Association for Cancer Research.

Her first step, she says, will be to focus in on pol-Beta, to see how it interacts with "normal," undamaged DNA. Then she'll look at how it interacts with DNA that has been damaged by carcinogens.

"Most chemicals that cause cancer bind to the DNA," Eckert explains. "You end up with DNA with this foreign thing sticking off called an adduct."

When things are working well, repair enzymes erase these adducts before a polymerase can run across them. But again, the question is: What exactly happens when this pre-emptive system fails?

If Eckert can find the answer, she and others may someday find a way to keep mutagenesis under control, even after a cell is damaged.

"There's a possibility," she says, "that if we could slow or block the cascade of mutations, we could keep cancer in check, even before it becomes clinically detectable."

Kristin A. Eckert, Ph.D., is assistant professor of pathology in the College of Medicine, The Milton S. Hershey Medical Center, Pennsylvania State University, 500 University Drive, Hershey, PA 17033; 717-531-4065. Eckert received the Gertrude Elion Cancer Research Award from the American Association for Cancer Research in March 1995. She has received funding from the American Cancer Society, the Four Diamonds Fund, and from the Jake Gittlen Cancer Research Institute, Hershey, PA.