November 15, 2016 (Vol. 36, No. 20)
Richard A. A. Stein M.D., Ph.D.
If Epigenetic Circuitry Blows a Fuse, It May Be Possible to Do a Little Rewiring
While epigenetic changes have been recognized for decades, recent years have witnessed an increasing appreciation of their role in development and disease.
This new knowledge, together with revelations about the reversibility of epigenetic modifications, catalyzed efforts to develop epigenetic therapeutics. Targetable epigenetic modifications—and the number of medical conditions with which they are associated—appear to be on the rise.
If drug developers are to identify and engage epigenetic targets, they will need biological assays capable of interrogating epigenetic alterations. There is an acute need for such assays in the field.
“We developed two bioluminescent assays that can be used in high-throughput screens for small molecule drugs,” says Hicham Zegzouti, Ph.D., senior research scientist at Promega. One is the MTase-Glo™ Methyltransferase Assay; the other, the Succinate-Glo™ JMJC Demethylase/Hydroxylase Assay.
The MTase-Glo assay monitors the formation of S-adenosyl homocysteine from S-adenosyl methionine during an N-methyltransferase reaction. This assay can be used with lysine and arginine as well as DNA methyltransferases and provides information about substrate requirements and enzyme specificity. This tool is complemented by the Succinate-Glo assay, which detects the activity of demethylases that use alpha-ketoglutarate as a substrate to produce succinate.
“The universality of these two assays is a key advantage,” asserts Dr. Zegzouti. “New targets for methylation and demethylation are constantly unveiled, and investigators need the tools to study them.”
The assays detect the products of the reactions they monitor. Accordingly, the assays are suitable for drug discovery using new targets, such as RNA and small molecules, which were recently shown to undergo reversible methylation.
Another advantage of these assays is their sensitivity. Because they can detect low-level product formation, the assays can help interrogate enzymes with low activity. “Both of these assays have very good linearity,” states Dr. Zegzouti. “And both can detect a broad range of substrate conversions.”
In investigations that vary substrate concentrations, the two assays are well positioned to identify the pocket into which a compound binds, unveiling information about the activation or inhibition mechanism.
Yet another advantage of these two assays is their ease of use. The assays’ homogenous “add and read” format allows investigators to perform them without the need for sophisticated equipment. During the early stages of drug discovery, their use promises to reduce the time required to identify promising compounds for subsequent steps, such as cell-based assays, animal studies, and clinical trials.
The assays could help “investigators identify targets or understand how the specific modifications are involved in the progression of pathology,” notes Mike Curtin, global project manager, Promega. “This would represent a great step toward advancing the field.”
Inhibiting Hedgehog Signaling
“We entered the area of epigenetic drug discovery because we realized that this is a very promising field,” says Elisabeth D. Martinez, Ph.D., assistant professor of pharmacology at UT Southwestern Medical Center in Dallas. The field, she points out, is one in which few tools are available, which suggests that new tools are bound to be in demand.
During the development of a cell-based screen that proposed to find chemical modulators of epigenetic changes, Dr. Martinez and colleagues identified JIB-04, a small molecule that inhibits Jumonji family histone demethylases and prolongs survival in animal models of cancer. “Jumonji family histone demethylases are excellent epigenetic targets,” notes Dr. Martinez. “They are upregulated in several cancers and have enzymatic activity.”
JIB-04 did not affect other alpha-ketoglutarate dependent enzymes. Also, it did not affect transcriptional programs in normal cells.
“Using our inhibitors, we are beginning to make connections with other people in the field for conditions that were not so obvious previously in terms of their connection with Jumonji demethylases,” informs Dr. Martinez.
For example, Dr. Martinez and colleagues collaborated with investigators in the laboratory of Matthias Lauth, Ph.D., a researcher at Philipps University in Marburg, Germany. This collaboration showed that JIB-04 induces GLI1 protein degradation in vitro and in vivo, and may even serve as an inhibitor of the Hedgehog signaling pathway.
This finding provides opportunities to control malignancies that are driven by GLI1. “This is something we would have never predicted directly from what we know about the enzymes and the genes they regulate,” says Dr. Martinez.
In drug discovery, the initial identification of targets is often a rate-limiting and critical step. Typically, subsequent studies examine the ability of specific compounds to penetrate cells, assay their biological activities, and optimize their structures.
“In our approach, we usually start with a complex system such as a cell and then work backwards to identify the target of a compound,” explains Dr. Martinez. While technologies that help identify drug targets faster and more accurately are critical, their development strongly depends on approaches such as mass spectrometry that can be coupled with other techniques in a high-throughput fashion.
“Understanding protein complexes and the enzymes they associate with in different cell types, in different parts of the genome, and in different developmental stages is critical to be able to characterize these large protein complexes functionally,” insists Dr. Martinez.
A topic of interest in epigenetic drug discovery is that in addition to histones, other proteins that have for a long time been functionally implicated in chromatin dynamics are substrates for therapeutic modulation.
“The field so far has focused on histones as substrates for these enzymes, but we have suspected all along that certain non-histone proteins may also be regulated by methylation and demethylation,” comments Dr. Martinez. “Non-histone proteins likely represent a whole new area for us to explore.”
Determining Substrate Specificity
For some of the demethylases, wrong or incomplete in vitro functional assignments have been made based on inadequate assays, notes Udo Oppermann, Ph.D., professor of molecular biology at the University of Oxford. This problem stems, in part, from the unknown substrate specificity for many of the enzymes that are discovered.
“We know of two distinct types of histone demethylases, the flavin-dependent monoamine oxidases and the 2-oxoglutarate-dependent oxygenases,” explains Dr. Oppermann.
Inhibitor development for the lysine-specific histone demethylase LSD1 has progressed rapidly, with several compounds currently in clinical trials, and recent years have also witnessed the identification of selective inhibitors of 2-oxoglutarate demethylases.
“Very good progress has been made—especially in the private sector—in generating apparently good inhibitors against selected targets of clinical relevance,” maintains Dr. Oppermann. “This progress demonstrates that these enzymes constitute a druggable class.”
Dr. Oppermann recently participated in a study led by Frédérick Mallette, Ph.D., an assistant professor at University of Montreal. The study confirmed that specific mutations in isocitrate dehydrogenases 1 and 2, enzymes that participate in the Krebs cycle, lead to the formation of 2-hydroxyglutarate, which can impair histone demethylation and affect cellular differentiation.
As part of this work, for the first time, the investigators found that formation of 2-hydroxyglutarate in this pathway activates mTOR signaling. Also, the investigators used a targeted siRNA screen to determine that KDM4A has an mTOR regulatory function it was not previously known to possess. KDM4A copy number amplification, genomic loss, and overexpression have been reported in several human malignancies. These findings underscore the contribution of metabolic dysregulation to malignant transformation, and provide new opportunities for therapeutic strategies in cancer.
“It is often difficult for academic laboratories to develop and systematically optimize molecules that are good “tool” compounds, and to bring them to a stage where they have suitable pharmacokinetic properties for in vivo experiments,” observes Dr. Oppermann. “But this is less of an issue for dedicated medicinal chemistry laboratories, particularly in industry.”
Estimating the Cancer Fraction
“The marker that we developed for gastric cancer has diverse applications,” says Toshikazu Ushijima, M.D., Ph.D., chief, Carcinogenesis Division, National Cancer Center Research Institute, Tokyo. “A similar strategy can be used for other cancers.”
One of the challenges in examining cancer samples is that specimens from primary tumors are always contaminated with co-existing nontransformed cells, making it difficult to detect genetic and epigenetic modifications in samples with low fractions of cancer cells.
In a study that compared genome-wide DNA methylation in gastric cancer cell lines, gastric cancer samples, and normal gastric mucosa, Dr. Ushijima and colleagues identified a panel of three genes that were highly methylated in cancer cells but showed little or no methylation in the non-transformed gastric mucosa. Using data generated with Illumina’s BeadArray technology, Dr. Ushijima and colleagues validated the strength of this marker in estimating the cancer fraction and in improving the accuracy of molecular analysis.
Understanding the physiological and pathological significance of epigenetic changes, and how they relate to the development of biomarkers, is intimately dependent on technological advances, some of which promise to bring conceptual changes to the field. “Single-cell level studies are ideally positioned to collect information about epigenetic changes that occur as part of cellular differentiation, cell-cell communication, and chronic inflammation,” explains Dr. Ushijima. “Conducting such studies will help us understand epigenome diversity.”
A major effort in Dr. Ushijima’s lab focuses on dissecting the mechanisms by which chronic inflammation induces epigenetic changes such as aberrant DNA methylation. These mechanisms may be pertinent, for example, in the inflammation that accompanies bacterial or viral infections.
Although epigenetic changes are also known to take place physiologically, the ones that occur as part of chronic inflammation have been found to lead to cancer development and progression. Once a tumor develops, a critical consideration in epigenetic therapies is the need to selectively modulate aberrant epigenetic changes and minimally affect the ones that occur physiologically. Several studies have found that aberrant epigenetic modifications are more easily corrected than physiological epigenetic changes.
“One model to explain this proposes that cancer cells have aberrant combinations of epigenetic modifications, and these can be recognized and more easily corrected,” states Dr. Ushijima.
A second hypothesis proposes that when aberrantly methylated genes are demethylated and transcribed, they become resistant to remethylation.
“On the other hand, physiologically methylated genes have low transcription levels,” Dr. Ushijima continues. “Therefore, even if they are demethylated, they return to their methylated states.”
Instituting Compound Triage
In collaboration with investigators from Genentech, researchers at Constellation Pharmaceuticals made significant progress on the druggability of Jumonjis, according to Patrick Trojer, Ph.D., vp for research at Constellation. In a recent study, Dr. Trojer and collaborators from Constellation and Genentech described the crystal structure of the specific KDM5 inhibitor CPI-455 and unveiled mechanistic details about the binding. The KDM5 family of histone demethylases includes five enzymes that catalyze the demethylation of trimethylated lysine 4 on histone H3. Several of these family members are amplified or overexpressed in human malignancies.
“We were surprised to see that CPI-455 was not a typical iron chelator, but that it establishes a monovalent interaction with the iron ion in the active site,” recalls Dr. Trojer.
However, instead of simply chelating the iron ion, the small molecule bound via an induced fit mechanism, a slightly different binding mode from that of other Jumonji inhibitors known today. “With this different type of binding type came a fair amount of selectivity,” notes Dr. Trojer.
Prior to the discovery of CPI-455, one of the challenges with existing Jumonji inhibitors has been their relative lack of selectivity, translating in their ability to change gene expression programs in a pleiotropic manner.
“We kept advancing our inhibitor series beyond CPI-455,” informs Dr. Trojer. “In our push to improve the compound series’ properties, we developed a compound with significantly better cellular potency and much improved pharmacokinetic profile in mice.”
Concomitantly, Dr. Trojer and colleagues also generated 1,7-naphthyridones, a new series of competitive KDM5 inhibitors with nanomolar potencies. These inhibitors exhibited increased selectivity over several other KDM subfamilies. The X-ray co-crystal structure of one of the representative molecules from this series with KDM5A showed that it competitively binds to the 2-oxoglutarate substrate.
“One experimental aspect that is very important to us is the biophysical validation of small molecules that are identified in high-throughput screens,” maintains Dr. Trojer. For many targets identified in small molecule screens, the initial hits are generally not very potent. “This is why it is very important to distinguish between false positives and valid hits, and to explore a plethora of biophysical methods to triage the compounds.”
Several new biophysical methods have become available over the last few years. “We tested many of these methods, and an important consideration is that there is no one-size-fits-all approach,” explains Dr. Trojer. “In our experience, different biophysical methods play critical roles for different targets.”
Harnessing Histone Antibodies to Fuel Discovery
It can be said that the end product is only as good as the tools used to build it. This is certainly true in science, where the quality of reagents, samples and other vital components can significantly impact results.
Biomedical research, for example, relies heavily on antibodies, which allow scientists to track proteins of interest in healthy and diseased cells. However, despite their indisputable value, selecting the correct antibody for a given experiment can be challenging. Specificity and quality can vary from lot-to-lot and between vendors and, even when selected carefully, the chosen antibody must be validated to ensure it performs consistently with its intended specifications.
To help combat these issues, several databases have been developed in recent years that provide scientists with additional resources to characterize their antibodies. An example is the Histone Antibody Specificity Database (www.histoneantibodies.com), an online portal that helps scientists make informed choices about the use of histone antibodies—a type of antibody that is of particular importance to epigenetics—in their research. It is populated with validated test results, allowing scientists to access and compare real-world data, and pick the most reliable antibody for each experiment with a much higher degree of confidence than before.
The database includes profiles of more than 100 commercially available histone antibodies, which were analyzed using a specially designed peptide microarray that presents a library of many of the major chemical modifications to histone proteins. Similar microarrays are now commercially available, allowing any scientist to run their own validation experiments and submit their results to the growing database.
“Providing the research community with a better understanding of what these antibodies recognize will help us define more precisely how epigenetic mechanisms work,” said Scott Rothbart, Ph.D., an assistant professor at Van Andel Research Institute. He developed the database with Brian Strahl, Ph.D., a professor at the University of North Carolina Chapel Hill.
“As our appreciation for the role of epigenetic regulation in human health and disease is rapidly growing, it is of utmost importance that we make sound conclusions from our research utilizing antibody reagents and translate this knowledge into clinical benefit through epigenetic therapies,” added Dr. Rothbart.
Targeting an Epigenetic Enzyme Associated with Cancer
One of the areas with great potential for cancer treatment is the intersection between targeted therapy and immunotherapy, notes Robert A. Baiocchi, M.D., Ph.D., associate professor of hematology and internal medicine, Ohio State University. “This is the direction that our group is heading in.”
In recent years, several epigenetic modulators were found to also influence the tumor microenvironment and to cooperate with immune targeting agents toward driving anti-tumor responses. In parallel with the development of more selective epigenetic compounds, this finding is catalyzing a shift toward designing combination therapies.
A key effort in Dr. Baiocchi’s group is focusing on the arginine methyltransferase enzymes as therapeutic candidates in multiple malignancies including acute myeloid leukemia, aggressive lymphomas, and glioblastoma multiforme.
“If we look at how the field started, we began targeting the epigenome with broad spectrum modulators, such as HDAC inhibitors, and as the field evolved we learned about various classes of HDAC enzymes and began developing small selective inhibitors for specific classes of HDAC enzymes,” says Dr. Baiocchi.
He and colleagues approached epigenetic drug discovery from the opposite direction.
“We first identified the biology of a select class of PRMT enzymes and targeted only one enzyme, PRMT5,” explains Dr. Baiocchi.
By designing therapeutic compounds selective for PRMT5, which catalyzes the symmetric di-methylation of arginine residues on histone tails, Dr. Baiocchi and colleagues ensured that the inhibitors would not affect type 1 or 3 enzymes.
“This allowed us to focus on the role of this enzyme in cancer (mature cancer cells as well as cancer stem cells) while simultaneously examining how these drugs affect the tumor microenvironment and immunity,” says Dr. Baiocchi.
In a recent study, he and his team reported that PRMT5 is critical in supporting glioma stem cell renewal and unveiled its relevance for both the stem and the mature components of glioma tumors.
“Hopefully we are getting close to understand how to target this enzyme, perhaps in combination with other enzymes, or immune modulators to achieve a much more durable response to therapy”, says Dr. Baiocchi.