Dr Willard ‘Bill’ Freeman leads a research group in the Department of Physiology at OU Medicine and is a member of the Reynolds Oklahoma Center on Aging, Harold Hamm Diabetes Center, and Stephenson Cancer Center. His research focuses on systems neurobiology related to brain aging, diabetic retinopathy and substance abuse.
We spoke to Bill to find out more about his work in neuroepigenetics and how understanding epigenetic DNA modifications, particularly hmC, is important in this field
What are your current research interests?
Our research is primarily in neuroepigenetics, specifically the DNA modification changes associated with aging, diabetes or drug abuse in the central nervous system. We initially went down this road to understand how stimuli, be they stressors, endocrine changes, hyperglycemia, or drug use can continue to have effects on the brain or retina long after that stimuli has subsided. How this signal is propagated for years after the stimuli is gone, potentially through epigenetic changes that occur during the period of hyperglycemia, is a focus of our research. We use a variety of animal models and next generation sequencing (NGS) approaches to map and understand altered cytosine methylation (5mC) and hydroxymethylation (5hmC).
How is epigenetics important to your research?
Epigenomic changes seem to be a hallmark of the aging process. However, we still know very little detail on how the epigenome changes with aging in any given organ or cell. At this stage, much of our work is discovery analyses to define the changes to the epigenome so that we can understand how they may alter genome organization and gene expression. The process is full of surprises. For example, once we started including males and females in our aging studies, we determined that the vast majority of changes in 5mC were sexually divergent! So, if the neuroepigenomic response to aging is different between males and females, what is the functional impact and what are the mechanisms causing this sexual divergence? With this base-level analysis of DNA modifications we can then begin to work both forward to determine function, and backward to understand regulation.
What first ignited your interest in epigenetics?
As a graduate student, I worked primarily in the neurobiology of drug abuse with Dr. Kent Vrana and Dr Dave Roberts at Wake Forest University. In rats, after a period of drug self-administration we could withdraw access to drugs for 10, 30, or 100 days and then reintroduce access to the drug. When the rats again had access to drugs they went right back to where they were in the addiction process, just like humans, despite long periods of abstinence. This was accompanied by gene expression changes that persisted through drug abstinence. We identified epigenetic changes during that period of initial drug abuse as a potential cause for this persistent phenotype.
At Cambridge Epigenetix we’ve a long-standing interest in DNA modifications. Is understanding DNA modifications important for your research?
While we have studied histone modifications, we are primarily focused on the long-lasting epigenetic changes mediated by DNA modifications, which have the potential to persist without any additional signal – especially in long-lived cells like neurons. Eventually histone proteins turn over, transcription factors are degraded and replaced, but in a post-mitotic/non-dividing cell, a DNA modification has the potential, without any new signal or energy, to persist for the rest of the lifetime of the cell.
What do you think are the main outstanding challenges associated with studying DNA modifications and epigenetics, and how are you approaching them?
The primary challenges for studying DNA methylation are data analysis and DNA input. We have scaled up our bioinformatic capabilities dramatically. While we were accustomed to sequencing analysis for genomics and transcriptomics, the scale of data needed for genome-wide bisulfite and oxidative bisulfite (oxBS) data analysis is much larger. We are now looking at tens of millions of cytosines, both CpG and non-CpG, rather than tens of thousands of genes. On the DNA input front, as neuroscientists, we are dealing with a very heterogenous tissue that is difficult to meaningfully replicate in culture systems. A critical next step is to be able to examine specific cell types in the central nervous system (CNS) which provides very little starting material.
A further challenge for the field is addressing how epigenetics research expands out of specialized labs to more general molecular biology or disease-focused labs given the very specific skills and instrumentation required to perform these studies. Getting these techniques out into the broader biomedical research community is a critical challenge that needs to be addressed.
How important do you think it is to study 5hmc as well as 5mc?
Our research is primarily in the CNS, the organ with probably the highest 5hmC levels in the body. We have already found through oxidative bisulfite sequencing (oxBS-Seq) that there are specific changes in 5hmC associated with aging, diabetes and drug abuse. These changes would have been mis-identified as changes in methylcytosine if we had not been using oxBS-Seq. Looking to the future, we would like to routinely examine both 5hmC and 5mC in all our studies.
Do you consider bisulfite and oxBS to be the gold-standard for 5mC and 5hmC analysis?
Absolutely. In our research, especially in brain aging, we have found that published results with antibody-based methods (ELISAs, meDIP) in many cases do not replicate and the relative quantitation data is unfortunately often misleading. Additionally, base-specific analysis gives us a resolution, with regards to genomic features like transcription factor sites, that cannot be achieved using other methods.
Do you validate your bisulfite and oxBS results?
We focus most of our efforts on performing genome-wide analyses with increasing sample numbers in order to gain a broad understanding and to recapitulate previous findings. Where a locus is of special interest, we also perform focused confirmations of our genome-wide analyses through bisulfite amplicon sequencing.
Is data analysis challenging? How do you handle data analysis?
Yes, data analysis is challenging. We perform all of our own data analyses and have devoted a great deal of time and resources to being able to do this part of our studies in-house.
As a university laboratory, in addition to our research goals, we have a duty to train the next generation of scientists. For graduate students and other trainees, developing bioinformatic skills is critical to their future careers. Ultimately, my hope is having experimentalists who also can perform advanced bioinformatic analyses not only improves our research but sets trainees on a path to success in their careers.
What breakthroughs in methodologies do you think are needed to move the field of epigenetics forward?
In addition to anything that brings the cost of sequencing down, which is the greatest limiting factor in the number of studies we can perform, being able to analyse single cells or very small numbers of cells would tremendously advance the field.
What excites you most about the future of epigenetics?
In the near future, expanded abilities to ask new questions of the data generated from sequencing looks very exciting. For example, we are testing using wavelet analysis to understand the inter-relatedness of 5mC and 5hmC across sites; however; this type of deeper understanding is still in its infancy. Also, integration of DNA modification data with other forms of epigenetic and transcriptomic data, and manipulation of modifications at specific sites in the genome are areas of great interest for us. As the field begins to understand the patterns of DNA modifications, being able to manipulate specific sites or sets of sites through approaches like CRISPR will allow hypothesis testing.
Are you presenting your work at any conferences this year?
The paper from Rudolph Jaenisch’s lab last year (Editing DNA Methylation in the Mammalian Genome. Liu, X. Shawn et al. Cell, Volume 167 , Issue 1 , 233-247) on site-specific modification of methylation in vivo is a great example of how data will move to testable hypotheses.
What advice would you give any young researchers entering the field of epigenetics today?
In what promises to continue being a very competitive funding environment, having a good experimental question for epigenetic analyses is critical. As epigenome analysis joins all the other ‘omes (transcriptomics, proteomics, lipidomics) having a well-thought-out rationale for your experimental question is critical to rise to the top of the funding list.