Controlling The Code
DNA carries the instructions for assembling all of life’s essential building blocks. Logic would have it that decoding the human genome, achieved a decade ago, would set out a straight, if dauntingly steep, path for inventing new drugs: Determine the genetic mutation driving a disease and then find the right compound to overcome it.
But our fate is not sealed by DNA alone. Overlaying genetics is epigenetics, the elegant and enormously complex molecular movements that dictate how and when DNA goes to work. Researchers have long pondered the therapeutic potential of tapping into the epigenome. Early efforts yielded a handful of cancer drugs, but their discovery was more a product of brute force and luck than of a refined understanding of epigenetics.
Today, significant progress is being made toward understanding the biology of epigenetic processes, which control DNA expression without altering the genetic sequence. Encouraged by the discovery of clear links between epigenetic enzymes and specific cancers, drug companies have developed the confidence to invest substantially in the field. And although epigenetics is still in its infancy—scientists have yet to test a second-generation epigenetics-based drug in humans—the industry is excited about its potential.
The scientific case for pursuing epigenetics-based therapeutics has coalesced in the past five years. Evidence is mounting that targeting mutations in DNA is simply not enough to overcome many diseases. “It’s now becoming increasingly clear, at least in oncology, that the genome doesn’t work on its own,” says Dashyant Dhanak, head of GlaxoSmithKline’s cancer epigenetics discovery group. “There’s an important cross talk between the genome and epigenome.”
But the ability to confidently target epigenetic enzymes and proteins is a more recent development. The field was still nascent in 2008, when the epigenetics companies Constellation Pharmaceuticals and Epizyme launched. “Prior to that time, epigenetics had largely been the playground for very elite basic biochemists who were breaking apart the fundamentals of how chromatin was compacted and regulated,” says Mark A. Goldsmith, executive chairman of Constellation’s board of directors and until recently the biotech company’s president and chief executive officer.
Four epigenetic drugs that block first-generation targets are already approved to treat blood cancers: Celgene’s Vidaza and Eisai’s Dacogen inhibit the enzyme DNA methyltransferase, and Merck & Co.’s Zolinza and Celgene’s Istodax inhibit histone deacetylase. But the drugs lack selectivity, meaning clinicians have had to sort out how best to use them; nevertheless, by proving that epigenetic functions can be safely modulated, these drugs open the door to new avenues of cancer research. Companies now have their sights set on drugs for the next generation of targets.
To squeeze inside a cell, DNA winds around protein complexes containing histones. That spool-like structure, known as chromatin, loosens and unwinds to allow the instructions to be read when a protein needs to be made.
The timing and tempo of how chromatin opens or shuts access to a stretch of DNA are orchestrated by small chemical groups, or marks, such as methyl or acetyl. Epigenetic drugs target three categories of proteins: writers, the enzymes responsible for adding a chemical group to histone; erasers, the enzymes that remove a chemical group; and readers, the proteins that interpret the marks. Within each category are protein families or domains defined by where and what kind of mark is added, removed, or interpreted.
Drug companies are interested in understanding histone modifications and finding molecules to modulate them. It sounds straightforward, akin to blocking a protein kinase implicated in cancer. But epigenetic processes are rarely as simple as a one-step addition or subtraction of a chemical group. If scientists could zoom in and watch the enzymes in action, they would see a complicated, carefully controlled process. Without a magical microscope, drug discovery will require painstaking research to understand these fundamental processes.
A look at the EZH2 drug discovery program under way at Cambridge, Mass.-based Epizyme provides a flavor of the complexity—and the opportunity—of epigenetics. Epizyme’s activities around the epigenetic enzyme EZH2 were sparked by a puzzling finding published in early 2010 (Nat. Genet., DOI: 10.1038/ng.518).
A group of Canadian scientists reported that a subset of patients with non-Hodgkin’s lymphoma had a mutation in EZH2, a writer that adds up to three methyl groups on a specific lysine located on a specific histone in cells. Patients’ cells expressed equal amounts of the mutant and wild type, or normal, enzyme. But strangely, the researchers’ assays also showed the mutant enzyme was inactive. They concluded that the mutation caused EZH2 to stop working, spurring the onset of lymphoma.
“The Epizyme team read that paper very carefully and was perplexed by a couple aspects of these findings,” says Robert A. Copeland, Epizyme’s chief scientific officer. For example, they knew that previous studies linked tumor growth to increased activity of EZH2. “We decided to take a look at these mutants in much greater biochemical and biological detail.”
The team discovered the opposite of what the Canadian scientists concluded. EZH2 is actually overactive in the lymphoma patients. The wild type and mutant enzymes, it turns out, act in concert to jump-start cancer. The mutant on its own has a hard time putting the first methyl group onto the lysine. But if the wild type sticks on the first chemical mark—a job it does well—the mutant can swiftly step in with the second and third methyl groups, allowing cancer cells to proliferate.
Unraveling the biology in such detail opens the door to drug discovery programs. Whereas the original findings might have led researchers to look for ways to activate EZH2, Epizyme’s results show that treating the rare lymphoma would require a molecule that could block both the wild type and mutant enzymes. The company has since made significant progress in developing EZH2 inhibitors.
This kind of highly specific, genetics-driven approach to drug discovery is one reason big pharma and biotech companies are moving into epigenetics. And the list of cancers linked to an errant epigenetic enzyme seems to be ever-expanding.
“There’s now a genetic underpinning to show that tumors, as a way of gaining some kind of growth advantage, mutate these enzymes to either make them more or less active,” says Michael D. Varney, senior vice president of small-molecule drug discovery at Genentech. “That is one of the strongest pieces of biological evidence that if you had a target, you could go in and modulate it in some way.”
Genentech has taken a deep dive into epigenetics through a partnership with Constellation. The move, Varney says, was motivated by progress in understanding the fundamental biology of epigenetic pathways, which paves the way to the clinic, allowing firms to home in on the right patients on whom to test their drugs.
Moreover, epigenetics could be a way to access genomic targets that for decades have stymied medicinal chemists. GSK’s Dhanak points to the recent discovery that inhibiting members of the BET protein family—readers that interpret acetyl marks left on histones—can shut down the MYC oncogene, an enticing but, to date, elusive drug target. “The implications are clear: Those traditional oncogenes we’ve all been struggling for 50 years to target, maybe we can now go after them from an epigenetic point of view,” Dhanak says.
Armed with the genetic clues that make good starting points for drug development, companies are tackling epigenetics in a variety of ways. Some are placing bets on one family of enzymes, with the idea that a focused effort will yield the quickest successes. Others are working across the entire field to tap into therapeutic opportunities as the biology reveals itself.
Epizyme has chosen to build a strong foundation in one family of enzymes, writers called histone methyltransferases (HMTs). After spending its first few years sorting through the fundamental biology of HMTs, Epizyme is now using high-throughput screening of compounds and compound fragments to look for potential drugs that act against its enzyme targets.
Enzyme-mechanism-guided drug design also has proven fruitful for the team. After applying classic enzymology tools to dig into how the enzymes interact with their substrate, Epizyme scientists develop an idea of what the substrate and its product look like. “Then we use our chemical intuition to infer what intermediate species might look like and design compounds that would mimic those intermediates,” Copeland says.
The biotech firm used that approach to discover inhibitors of DOT1L, an HMT implicated in mixed-lineage leukemia. The first series of compounds was only modestly potent. But by making analogs and solving the crystal structure of the enzyme with various inhibitors bound to it, Epizyme researchers refined their understanding of what’s needed in a drug candidate and developed more potent inhibitors.
The company now has two compounds in preclinical testing, a DOT1L inhibitor and a molecule that blocks EZH2. The latter program is partnered with Eisai, a Japanese drugmaker that is bankrolling the R&D efforts through clinical studies.
GSK’s cancer epigenetics group is also focused on HMT inhibitors, both through internal efforts and a partnership established in early 2011 with Epizyme. Its most advanced compound, an EZH2 inhibitor, originated from a traditional high-throughput screen of its compound collection. “For other epigenetic targets in general, we’re not using one particular approach,” Dhanak says. In addition to high-throughput screening, the company is deploying fragment-based screens. And when the structural biology around a target is detailed enough, it designs molecules from scratch.
The British drug firm appears intent on being one of the first companies to put a second-generation epigenetics compound into the clinic. Earlier this month at the annual meeting of the American Association for Cancer Research, GSK gave a first peek at the structure of its EZH2 inhibitor. This marks the first time a second-generation epigenetics drug has been unveiled, but GSK has yet to publish the structure. “The intent is to push the program forward into the clinic with some urgency,” Dhanak says.
Like GSK and Epizyme, AstraZeneca is devoting its epigenetics efforts to HMT inhibitors. The firm, which has been working on epigenetics for the past five years at its labs in Waltham, Mass., decided that narrowing down to one class of enzymes would increase its chances of success. The ability to find safe and potent drugs against these targets was not well established, says Kevin Webster, AstraZeneca’s executive director of cancer bioscience, so “we chose to focus our chemical efforts in one area to see if we could crack it open.”
AstraZeneca scientists have taken multiple approaches to finding HMT inhibitors. “We did carry out high-throughput screening, but honestly, I don’t think any of us thought that would be the ultimate answer to opening up the space, just because our libraries were not built on the back of enzymes like this,” Webster says. Enough HMT structures have been solved that company researchers are able to use comparative biology to map the sites where substrates interact with enzymes and design molecules from scratch, Webster says.
The company also uses fragment-based screening to identify starting points for HMT inhibitors. The combination of high-throughput screening, structure-based drug design, and fragment-based screening eventually “allowed us to start to generate some tool compounds in this area,” Webster says. Tool compounds are highly selective chemical probes that help researchers determine the biological function of a given target. Last year, the company reported one such compound, AZ505, an inhibitor of the lysine methyltransferase SMYD2 (Structure, DOI: 10.1016/j.str.2011.06.011).
Meanwhile, other companies, such as Constellation, are trying to address a larger swath of the epigenetics landscape. “The strategy we pursued from the very beginning was to take a very broad approach to this field of chromatin biology or epigenetics—not to pick a small number of targets and simply place our bets, win or lose, but to really deepen and broaden our knowledge of all targets,” Constellation’s Goldsmith says.
Like its competitors, Constellation is using myriad screening methods to search for compounds. The Cambridge-based firm is also using some unusual tools as it tries to uncover the best opportunities in epigenetics.
In 2009, the company embedded a postdoctoral fellow in the labs of Benjamin A. Garcia, a chemist at Princeton University with expertise in using mass spectrometry to characterize histone modifications. That collaboration prompted Constellation to hire a dedicated scientist to expand on the technology, which allows the firm to monitor changes in multiple chromatin marks after an enzyme interacts with a compound.
The pursuit of knowledge across the entire field of epigenetics is part of what attracted Genentech to Constellation. Earlier this year, the biotech giant paid Constellation $95 million to jointly develop cancer drugs based on epigenetic targets; the deal also includes an option for Genentech to buy Constellation later on.
“Getting into a partnership with them allows us to catch up to where the frontrunners are,” Genentech’s Varney says. Meanwhile, Genentech can complement Constellation’s painstaking efforts to define the biology around epigenetic enzymes with the “biological tools to go in and sort out which of these targets are real and which are not ultimately going to yield drugs,” he adds.
CellCentric, one of the first biotech firms launched to tap the potential of the epigenome, is also exploring multiple classes of drugs. The Cambridge, England-based company operates virtually, relying on relationships with academic principal investigators and contract research organizations to advance its pipeline.
“We’ve put together a network of PIs, experts within the epigenetics field to identify potential drug targets,” says CellCentric’s scientific director, Anthony Brown. CellCentric scientists winnow the targets to decide which ones make the most sense to pursue, factoring in the biological argument for the target, the existence of a clear path to the clinic, and the target’s competitive advantage in the marketplace.
Currently, CellCentric is pushing forward four preclinical programs across multiple targets, including HMTs and demethylases, Brown says.
Forma Therapeutics, a recent arrival to the epigenetics field, is also taking a sweeping approach in its drug discovery efforts. The Watertown, Mass.-based biotech company launched its epigenetics program just last September, but the subsequent seven months have been busy: Already, the firm has screened a dozen epigenetic targets, or about two in every family, says CEO Steven Tregay.
The company’s computational and structure-guided approach to drug discovery drove its swift move into the field, Tregay says. Forma scientists use fragment-based technology to generate maps of the surfaces of target proteins. The maps guide drug design by showing both commonalities across family members and differences that enable selectivity in drug design, Tregay explains. The effort also gives the company a good sense of whether the targets are “druggable,” meaning compounds can be designed to interact with them.
From there, the company uses high-throughput screening and some medicinal chemistry elbow grease to generate tool compounds that can provide information about the intricacies of a given target’s function. Forma has spun off its epigenetics program into a new company that will be ripe for partnering or acquisition, Tregay says.
Regardless of specific strategies, all companies acknowledge that epigenetics still suffers from enormous knowledge gaps, both in the biology of targets and in the chemistry of compounds that modulate the targets. Furthermore, the very concept of targeting such fundamental cellular mechanisms leads many in the field to worry about the safety of drugs against epigenetic targets.
“Even though we have a handful of targets that we’re very excited about, these are all unprecedented targets, all targets for which there has never been an inhibitor developed and placed inside a human to see what happens,” Constellation’s Goldsmith cautions.
As drug discovery programs get under way, companies are struggling to develop compounds that are exquisitely selective for their epigenetic target. This is no small feat, considering that a specific enzyme adds a chemical group to a specific spot on a histone. “Just think about what a lysine side chain looks like: It’s a big alkyl chain with an amino group at the end, and those enzymes can somehow put it on the right spot,” Varney says. “Now you think about, ‘How am I going to find a small molecule that puts it on this lysine versus that kind of lysine?’ On the face it looks like a difficult problem.”
Goldsmith agrees that the challenge is steep. “I think that many of us, four or five years ago, would have doubted that an inhibitor of a major chromatin regulator could have a selective biological effect within a cell,” he says. The skepticism stems from industry’s experience with histone deacetylase inhibitors like the Merck and Celgene drugs, which have found limited use because they are relatively broad in the number of enzymes they hit.
Scientists compare the state of the field to the early days of drug discovery around protein kinases, when many insisted that selective kinase inhibitors would never come to fruition. “In the early days of kinases, people spun their wheels with fairly intractable small molecules and spent a lot of time with useless debates of whether selectivity could or could not be achieved,” says Stephen V. Frye, director of the Center for Integrative Chemical Biology & Drug Discovery at the University of North Carolina, Chapel Hill. Frye, who devoted many years to kinase inhibitors while working at GSK, is now developing HMT inhibitors.
Drug design has come a long way in the past 15 years, Frye points out, and scientists working on epigenetic targets have learned many lessons from the kinase field. “There’s still lots of risk and uncertainty, but I think the scientific foundation people are building on is perhaps a little more sophisticated.”
Researchers point to several recent examples of compounds in the literature and under development suggesting that highly selective drugs are not only possible but promising. Constellation has shown that its BET inhibitors are highly specific for their target, “resulting in a surprisingly small number of changes in gene expression within the cell,” Goldsmith says. He says the number of genes that are affected by a BET inhibitor is far smaller than those affected by a histone deacetylase inhibitor.
Although no one will know the value of the new epigenetic compounds until they are tested in humans, scientists are confident that the field is moving forward with the right balance of caution and enthusiasm.
“There’s a tremendous amount to learn, and so I believe that we’re still at that crucial definitional stage,” Goldsmith says. “There’s a huge genetic component to epigenetic therapeutics and a larger epigenetic component to diseases we thought were genetically driven. The nuance and the subtleties of the field are just beginning to reveal themselves. I anticipate there will be further surprises.” ◾