�鶹ӰԺ

A major collaboration of Colorado institutions uses new technology to show, after more than 30 years and 50,000 papers on the subject, the direct targets of the gene p53, the most potent “tumor suppressor” gene. The finding is a strong step toward affecting the disease trajectories of nearly all cancer types.

The gene p53 is the most commonly inactivated gene in cancers, responsible for recognizing when a cell’s DNA is damaged and marking damaged cells for death. When p53 acts, cells are stopped or killed before they can survive, grow, replicate and cause cancer. As such, all cancers must deal with p53’s anti-tumor effects. Generally, there are two ways that cancer cells do this: by mutating p53 directly or by making a protein called MDM2 that stops p53 from functioning.

The current study, published in the journal eLife, explores cancer cells’ second strategy, namely these cells’ attempts to block p53 function by producing the protein MDM2. Now, significant work shows the promise of MDM2 inhibitors – the reasoning goes that turning off MDM2 should allow p53 to restart its anti-cancer activities.

“MDM2 inhibitors, which are through phase I human trials, effectively activate p53 but manage to kill only about one in twenty tumors. The question is why. What else is happening in these cancer cells that allow them to evade p53?” says Joaquín Espinosa, PhD, investigator at the University of Colorado Cancer Center, associate professor in the Department of Molecular, Cellular and Developmental Biology at CU �鶹ӰԺ, and the paper’s co-senior author.

The answer is in what are called “downstream” effects of this gene, because p53 doesn’t act against cancer alone. Instead, it is the master switch that sets in motion a cascade of genetic events that lead to the destruction of cancer cells. Until now, it was unclear exactly which other genes were directly activated by p53.

The imperfect knowledge of p53’s effects isn’t for lack of research interest. Thousands of papers explore p53’s targets and many genetic targets are previously known. Most of these studies determine genetic targets by measuring levels of RNA. Here is how it works:

When a gene is activated, it creates a protein. But between the gene and its protein product is the measurable step of RNA – the more gene-specific RNA, the more often a gene’s informational blueprint is carried to the cell’s manufacturing centers, and the more protein is eventually made. This is a major way genes regulate other genes: they affect the levels of gene-specific RNA. Researchers measure RNA to see which genes are being turned up or down by any other gene.

“But the problem is, measuring overall RNA levels is like looking in a huge bucket full of water – you see the water but you don’t really know where it came from. And imagine you are dripping water into this bucket – it takes a long time for those drips to create a measurable change in the overall water level,” Espinosa says. Also, it’s very difficult with traditional methods to tell whether increased RNA is a direct effect of a gene or whether more RNA in the bucket is a product of two- or three-steps removed signaling – p53 may activate another gene, which activates another, and down the line until the far downstream result is increased RNA.

“Instead, to measure the���徱���𳦳���genetic targets of p53, we measured not the water in the bucket, but the faucet dripping into it,” Espinosa says.

"Instead, to measure the direct genetic tarts of p53, we measured not the water in the bucket, but the faucet dripping into it," Espinosa says.

The technique is called GRO-Seq, or Global Run-On Sequencing, and it measures new RNA being created, not overall RNA levels.

“Many teams around the world have been getting cancer cells, treating them with MDM2 inhibitors and waiting hours and hours to see what genes turn on and then only imprecisely. GRO-Seq lets us do it in minutes and the discoveries are massive,” Espinosa says.

The discoveries also generate an astounding quantity of data. That’s because the technique requires counting tens of thousands of RNA molecules before and after p53 activation.

“What you get is terabytes of data in the form of short RNA sequences,” Espinosa says.

In addition to the Espinosa team made of molecular biologists specializing in p53, the experiment required designing algorithms to sort through the data – it��required��computational biologists driving a supercomputer. To address this issue, Espinosa��partnered with computational biologist Robin Dowell, DSc, at the University of Colorado at �鶹ӰԺ and the BioFrontiers Institute. Together Espinosa and Dowell co-mentored�� a postdoctoral fellow, Mary Allen, PhD, who was capable of doing both the molecular biological and computational aspects of the work.

“The data collection took a year and the computational analysis took a year and a half. Mary is a very rare type of scientist who could do both the bench work and the computational work,” Dowell says.

Allen explains that, in all, the experiment generated over 4 terabytes of information, requiring sometimes a week or more for the BioFrontiers supercomputing resource to process single queries.

In addition to many known p53 targets, the study described dozens of new genes directly regulated by p53. Further research will explore which of these genes are necessary for p53’s cancer-killing effect, how cancer cells evade these p53-activated genes, and how doctors may be able to affect cancer cells’ ability to stay safe from these genetic attempts at suppression.

In fact, the study already solves a major outstanding question about cancer cells that die or survive when faced with MDM2 inhibiting drugs. Remember: cancer cells that don’t directly mutate p53 manufacture MDM2, which blocks the gene’s function. Drugs inhibit MDM2, but then only 1/20 tumors die while the majority of remaining tumors arrest their growth but aren’t necessarily killed.

“Previously, using traditional technologies, people concluded that death genes activated by p53 simply take longer than arrest genes and so the action of the arrest genes superseded the action of the death genes,” Espinosa says. “But now we see that p53 activates death genes early on��in addition��to these arrest genes. Cancer cells get mixed messages and choose to arrest rather than die. That's good news: p53 is doing what we want it to do, but cells are protecting themselves from the death genes. So now we need to focus on understanding these protective mechanisms and how to shut them down.”

The technique of GRO-Seq may have additional, far-reaching applications. For example, the Dowell lab plans to find RNAs whose synthesis is changed by a third copy of chromosome 21 in Down Syndrome individuals.