The Two Sides of Steroids

Major grant will search for separation of benefits and risks
Share Share This
May 19, 2015

The "fight or flight" response blamed for some of our modern ills was a great thing, back in the day.

This automatic physiological response to stress pumps the hormone cortisol through your veins, rapidly shutting down your immune system and redirecting resources to your muscles and brain to give you the extra energy to fight or outrun the unseen threat.

It surely enabled our ancestors to hunt their dinners without becoming dinner themselves.

Modern medicine has harnessed the immune-blocking features of cortisol and synthetic derivatives known as corticosteroids to treat a variety of human ailments, from poison ivy to rheumatoid arthritis.

But the healing power of these drugs comes with a price so high that many people liken it to making a deal with the devil.

Being in a drug-induced state of stressed survival creates a host of new problems, such as trouble sleeping, elevated blood sugar levels, weight gain, weakened bones, and high blood pressure.

For decades, researchers have studied the two sides of the corticosteroid response, looking for any way to separate the good from the bad so that patients with chronic inflammatory or autoimmune diseases could enjoy the benefits of treatment without suffering from unwanted metabolic side effects. Though there has been some progress, no one has come close to accomplishing that feat.

Now Timothy Reddy and his colleagues at the Duke Center for Genomic and Computational Biology have a $5.9 million grant from the National Institutes of Health to chart how the human genome responds to corticosteroid treatment. The researchers will conduct thousands of separate experiments to determine what and when genes are turned on and off, and how this process is controlled.

“The hope is that we will find the right combination of knobs or buttons we can manipulate to constrain or reprogram the response in human cells,” said Reddy, an assistant professor with training in both genomics and computer science. “We have the genomic engineering tools we need to make this possible, but as far as I know no one has done that before, or even attempted it.”

Reddy’s laboratory has already spent years studying the molecular response to steroids, albeit on a smaller scale. What he has uncovered is not a case of one set of genes simply being turned on and another being turned off. Instead, some genes are activated within the first hour of treatment, then repressed a few hours later as another set is activated, and so on, oscillating back and forth.

“We are seeing this incredible diversity of dynamics, which suggests that these long-term effects are downstream echoes of that initial burst of activity,” said Reddy. “What we will be doing now with this grant is massively expanding on that initial work, to go as deep as we can, throwing every experiment that we know how to do at this problem so we can gather as many viewpoints as possible.”

Many of those experiments will capture information on the “knobs and buttons” that regulate gene expression. These regulatory elements, as they are called, exert their influence a number of different ways. Some of them keep the strands of DNA packaged into tight little coils, hiding them from the machinery responsible for reading and translating the genetic code. Others recruit special proteins called transcription factors, which bind to specific sites around a gene and mark it as reading material.

Greg Crawford

Reddy has teamed up with genomics researcher Greg Crawford to measure these elements, as well as the genes that are ultimately turned on or off, every hour for the first 12 hours of steroid treatment in cultured human lung cells. The researchers chose these cells because they are particularly relevant to the millions of asthma patients who take corticosteroids by an inhaler.

Over the course of three years, the researchers will conduct more than 2500 experiments, resulting in what Reddy calls the most extensive and detailed study of the genome-wide effects of a hormone ever produced. However, they expect the 150 terabytes of data they collect will only give a glimpse of the bigger picture.

“The human genome has roughly 20,000 genes, and we are focusing on a subset of about 4,000 that are responsive to corticosteroids,” said Reddy. “When you start thinking about all of the different ways that you can get combinatorial regulation of thousands of genes, you would have to take a thousand and square it, then square it again and again, and you would quickly end up with a number that exceeds the grains of sand. There is no way you could be completely comprehensive and measure every possible regulatory state.”

Instead, the researchers are going to search the data they generate for patterns or rules that could be applied across the entire system.

Alex Hartemink
Alex Hartemink

Computational biologists and statisticians Alex Hartemink and Barbara Engelhardt will then use computer modeling to paint a picture of the entire network of regulatory knobs and buttons. The computer model will be able to generate predictions, forecasting what would happen if the researchers go in and disconnect any part of that network, or if they create new linkages that didn’t exist before.


Barbara Engelhardt
Barbara Engelhardt

“One of the most exciting parts of this project is we will be able to choose our experiments in a statistically rigorous way rather than willy-nilly,” said Reddy. “You could do a thousand experiments and learn something new, but then you have to pick the thousand-and-first experiment. I could sit down and powwow with a couple of people and come up with a reasonable strategy. But instead, we are going to let the computer predict which experiment will tell us the most about whether we are right or wrong, or which will give us the most new information about the patterns and rules.”

To test their predictions, the researchers will rewrite the genome of these cultured cells, manipulating regulatory elements one by one to see how the patterns or rules change.

Though it may sound farfetched to be able to exert such precise control over genes, genomic engineer Charles Gersbach has a variety of tools already available and under development for changing the sequence and structure of the human genome.

His favorite technology of late is CRISPR/Cas9, based on an enzyme that bacteria use to selectively slice up the DNA of invading pathogens. Researchers have commandeered this defense system for their own purposes, turning it into a kind of molecular scalpel that can snip and replace specific DNA sequences. The revolutionary technique has enabled scientists to introduce mutations they would like to study in model organisms, or to fix broken genes that cause inherited disorders like muscular dystrophy.

Recently, Gersbach came up with yet another application of this technology. He disabled the DNA-cutting property of CRISPR/Cas9 and used it as a vehicle to deliver molecules that help unwind or wind up DNA at specific spots in the double helix, ultimately controlling whether genes in the vicinity are activated or repressed.

“It is amazing to us to think how far we have come with this technology," Gersbach said. "Five years ago, we wouldn’t even have dreamed of being able to undertake this type of project. Between the advances in genomics and sequencing, and the advances in genome engineering techniques like the CRISPR/Cas9 system, we can now not only change the sequence but change practically any property of the genome at will.”

Gersbach will unleash his technology on the same human lung cell line that Reddy used to make his initial experiments. His tiny genome editing machines will skitter along the DNA, scanning for preprogrammed sequences and then adding or taking away regulatory elements to try to change the molecular response to steroids. The process will be iterative, with Gersbach manipulating the cells several different ways, Reddy and Crawford measuring the response, and Hartemink and Englehardt modeling the next steps.

Hopefully, each step will bring them closer and closer to separating the anti-inflammatory properties of steroids from their damaging side effects.

In the end, the researchers may find that the two genetic trajectories are inextricably linked, and that that they can’t get one row of dominoes to fall without setting off the other.

But even if they don’t get the result they expected, Reddy believes that their work will still lead to a greater understanding of the dynamics of the human genome. They plan to make all of their data publicly available in the hopes that other scientists can use the information they have gathered to launch their own investigations. No matter the outcome, Reddy and his team will continue to explore their own questions, and the more challenging the better.

“We did an experiment the other day that gave us exactly the result I expected and I was really bummed,” said Reddy. “Because if it didn’t do that, it would have been fascinating. I would much rather try something ambitious that has the potential to uncover new knowledge than back up an existing theory.”