April 1, 2013 (Vol. 33, No. 7)
Richard A. A. Stein M.D., Ph.D.
A remarkable scientific feat of the past century, the discovery of the DNA double helix, set in motion advances that reshaped the biomedical, physical, social, and behavioral sciences more powerfully than any other event in history.
The ensuing decades marked a vibrant period, one in which existing research areas were redefined and reshaped, new disciplines were born, and fields that had previously been perceived as unrelated converged to forge inter- and multidisciplinary endeavors.
Our progress in characterizing genes, followed by the interest and the need to learn about their organization into genomes, was assisted and paralleled by key developments in biotechnology. In 1977, when the first viral genome, that of bacteriophage Φ174, was reported, approximately 1,000 base pairs could be sequenced annually. Later estimates projected that over a millennium would be needed to sequence the Escherichia coli chromosome, and over a million years to sequence the human genome, using similar approaches.
Yet, 1995 witnessed the first fully sequenced genome of a free-living organism, that of the bacterium Haemophilus influenzae, and at the end of 2009, when sequencing and annotating a bacterial genome took less than 24 hours, the 1,000th bacterial genome sequence was published.
While approximately 13 years and $3 billion were required to sequence the first human genome, most recently this task has become possible within a day or less, for $1,000.
Genome-Wide Association Studies
As the Human Genome Project revealed that people are 99.9% identical at the DNA level, the remaining 0.1% emerged as one of the most exciting components for further study. Many of the estimated 3–10 million single nucleotide polymorphisms, and the more recently unveiled copy number variants, started to shed light on inter-individual differences in traits, disease susceptibilities, and therapeutic responses.
Subsequently, genome-wide association studies established increasing numbers of links between genes and phenotypes. However, despite the strength of this approach, many of the links were not causal, others were not statistically significant, and the ones that conferred an enhanced risk often explained only a small proportion of the heritability for specific complex diseases.
For example, 32 loci associated with Crohn’s disease explain approximately 20% of the heritability for this condition, and about 47 loci linked to type 2 diabetes and glycemic traits account, collectively, for only approximately 10% of the heritability.
This phenomenon is known as “missing heritability.” The phenotypic discordance between monozygotic twin pairs that is often apparent for many medical conditions is also supported by studies showing increasingly divergent DNA methylation, histone acetylation, and gene expression patterns between monozygotic twin pairs as they are advancing with age. Collectively, these findings point toward the importance of additional, nongenetic factors, in shaping phenotypes.
Epigenetics During Differentiation
That notion that DNA sequence changes are not the only factor shaping gene expression and phenotypes is neither new, nor unexpected. It has, in fact, been apparent for decades. In 1957, Conrad Hal Waddington introduced the term “epigenetic landscape” to refer to the causal interaction between genes and their products, as phenotypes are being shaped in a differentiating embryo.
To illustrate the multitude of interconnected choices that a differentiating cell can make, Waddington depicted it as a ball rolling down a landscape of ridges and valleys that branch at different points, representing the alternative fates that it could assume.
As the ball moves downhill, its options are progressively narrowed, and by the time it reaches the valley, it becomes a differentiated cell. It has become increasingly apparent that epigenetic modifications can explain the ability of totipotent cells to generate the over 220 cell types of an adult organism that, with small exceptions, share the same DNA, but nevertheless exhibit significantly different gene expression patterns and perform widely distinct functions. Thus, development provides an ideal system to visualize gene expression changes that are epigenetically shaped.
Over the years, the definition of epigenetics as a field has shifted and, most recently, the term has been used to describe potentially heritable gene expression changes that occur without alterations in the DNA nucleotide sequence.
The groundbreaking discovery that overexpressing four embryonic transcription factors is sufficient to reprogram terminally differentiated fibroblasts into induced pluripotent stem (iPS) cells, which resemble embryonic stem cells, demonstrated the possibility to reverse cellular differentiation, a process known as reprogramming. This was first reported in 2006 for mouse fibroblasts and in 2007 for human fibroblasts, and subsequently for additional cell types.
The iPS cells exhibit the ability to differentiate into many cell types and the capacity for infinite self-renewal. In addition to unveiling details about the molecular basis of differentiation and about mechanisms that allow cells to maintain their identity, this finding opened novel therapeutic opportunities, including the possibility to generate patient-specific embryonic stem cells for use in regenerative medicine and to treat various conditions.
The ability of physical, chemical, biological, and socio-emotional factors to change gene expression by epigenetic modifications has opened a fascinating chapter in biology. Ultraviolet B exposure, previously known to induce mutations, was shown to cause DNA hypermethylation and to transcriptionally downregulate tumor-suppressor genes.
In many instances, these links provide the mechanistic basis for epidemiological observations made decades ago. For example, divalent nickel compounds have long been implicated in carcinogenesis based on animal and human epidemiological studies, even though they do not appear to be strong mutagens in vitro, suggesting that carcinogenesis may also occur by nongenotoxic pathways.
Divalent nickel salts were shown to cause epigenetic changes that include aberrant DNA methylation and post-translational histone modifications, causing changes in chromosomal condensation and gene silencing. These findings also illustrate that, for a long time, we erroneously visualized carcinogens as being mutagens, and neglected to consider the possibility that gene expression changes may occur by mechanisms that do not involve mutagenesis.
This fallacious strategy, similar to searching for the lost keys only under the lamp, because that is where the light reaches, was relevantly called, by Trosko and Upham, the “lamp post effect.”
In 2012, it was reported that slightly over 20% of all human cancers are causally linked to infectious diseases, and epigenetics played a pivotal role in rekindling and providing a mechanistic understanding of this link, which was first reported over a century ago but subsequently fell into oblivion for decades. These advances helped characterize the “epigenetic field for cancerization” or “epigenetic field defect,” a region with aberrant CpG methylation that has a higher likelihood of undergoing malignant transformation.
An epigenetic field for cancerization was visualized after exposure to various carcinogens, and was reported for multiple malignant tumors, including stomach, breast, liver, and colon cancer.
Social Environment Impact on Gene Expression
Studies showing that the social environment influences gene expression by epigenetic modifications are unveiling a fascinating link between social adversity and pathology in later life. As part of these studies, it is relevant to remember that, in an animal model, licking and grooming were linked to CpG methylation changes in the hypothalamic glucocorticoid receptor of newborn pups, establishing causality between maternal care and the stress response in the offspring later in life.
Moreover, adversity during childhood was shown to leave an epigenetic imprint that changes gene expression in very specific regions of the brain. The involvement of epigenetic changes in synaptic plasticity, cognitive processes, and the formation of long-term memories, and the epigenetic modulation of post-traumatic stress disorder, are emerging topics that promise to fathom some of the most challenging areas in biology.
Fetal Origins of Adult Disease Hypothesis
While epigenetic changes may occur throughout life, embryogenesis is thought to be the period with the highest vulnerability. Epigenetics is providing a missing link to understand the ability of adverse conditions, acting during intrauterine development, to shape the risk of adult-onset diseases that clinically start decades later.
This relationship, known as the fetal origins of adult disease hypothesis, or the Barker hypothesis, was first published in 1992 and subsequently confirmed epidemiologically for several health conditions. Studies on women prenatally exposed to famine during the Dutch Hunger Winter at the end of World War II revealed that nutritional deprivation, when occurring at critical intrauterine developmental stages, may cause epigenetic changes that shape the risk of specific adult-onset diseases in the fetus.
Timing appears to be crucial, because exposure during critical windows of development was associated with specific disorders later in life. Recent years revealed that other types of exposures during intrauterine development, including nutritional compounds, maternal cigarette smoking, environmental polycyclic aromatic hydrocarbons, and endocrine-disrupting chemicals, may cause epigenetic changes in the fetus.
These changes include aberrant DNA methylation, histone post-translational modifications, and microRNA dysregulation, and shape disease risk later in life.
The possibility to transmit epigenetic modifications across several generations, a phenomenon known as transgenerational epigenetic inheritance, is emerging from animal and human studies. Vinclozolin, a fungicide used in the wine industry, whose two major metabolites are anti-androgenic compounds, was shown, in an animal model, to act at the time of embryonic sex determination and cause epigenetic modifications in the male germ cell line.
These inherited modifications, occurring at multiple locations throughout the sperm epigenome, increased the risk of diseases for at least three generations after the initial exposure. Evidence is also accumulating for the ability of diethylstilbestrol, a nongenotoxic synthetic estrogen that decades ago was prescribed to prevent miscarriages, to cause transgenerational epigenetic effects.
Animal studies revealed that this compound alters DNA methylation, and the increased susceptibility for adverse health effects appears to be transmitted across several generations though the maternal and, as some studies reveal, paternal germ cell lineage, and human studies showing transgenerational inheritance are attracting considerable attention.
The potential reversibility of epigenetic modifications, along with additional aspects, such as the recent finding that oncogene addiction extends to certain microRNA molecules, are opening new therapeutic perspectives. Four epigenetic targets, two DNA methyltransferase inhibitors and two histone deacetylase inhibitors, were approved in recent years by the FDA, and several other compounds are at various stages of preclinical and clinical development.
One of the challenges that these therapeutic targets present is that they may change epigenetic marks on multiple genes, not all of them involved in pathogenesis. More specific drug targets, known as second-generation epigenetic compounds, are currently at various stages of preclinical and clinical development.
Some of these, such as the small molecule inhibitors of the BET family bromodomains, or long noncoding RNA molecules that modulate gene expression, promise to much more selectively target the epigenetic dysregulation of specific genes.
In addition, epigenetic biomarkers that visualize and monitor epigenetic changes at the molecular level, sometimes years before histological changes become apparent, emerge as a major practical application. These changes, which can be used for early detection and in prognostic, risk stratification, and therapeutic decisions, found an exciting place in the drug discovery pipeline.
The Road Ahead
As a vibrant field, epigenetics promises much excitement and many surprises for years to come. While most research has focused on epigenetic changes within promoters and repetitive chromosomal regions, their occurrence within genes and intergenic regions is gaining considerable attention.
Epigenome-wide association studies recently emerged as a new approach to study complex human diseases. Particularly given the dynamic nature of epigenetic marks, which change during development and differentiation, vary in a tissue-specific manner, and are reshaped in differentiated tissues over time, this approach presents multiple layers of challenges, but promises a wealth of information.
Our knowledge about epigenetic crosstalk, the interplay among different layers of epigenetic modifications, is still in its infancy. The approximately 30 million CpG sites, the tens or hundreds of millions of histone tails from a haploid human genome that each can undergo various types of post-translational modifications, and the over 2,000 microRNAs discovered to date, forecast an epigenetic network with a complexity whose spectrum is far from being fathomable, even in the foreseeable future.
As certain microRNAs may target tens to hundreds of gene products, and the same mRNA can be regulated by multiple micro-RNAs, and with microRNAs being estimated to regulate over 30% of the protein-coding genes, epigenetics develops into a more thought-provoking and intriguing area than we could ever have envisioned.
Additionally, as recently pointed out, the bidirectional complex crosstalk between epigenetic and genetic changes has numerous far-reaching implications and should be more carefully scrutinized.
Richard A. Stein, M.D., Ph.D., is a research scientist at the New York University School of Medicine.