Epigenetics: a promising paradigm for better understanding and managing pain

The regulation of gene expression through chromatin modifications and remodeling underlies the phenomenon of epigenetics. Epigenetic mechanisms confer pluripotent progenitor cells that possess identical genomic DNA with the ability to differentiate into morphologically and functionally distinct populations, such as neurons, astrocytes, and Schwann cells. This wide range of differentiation originates from modulating genome expression in different manners that are inheritable with each somatic cell division. Beyond determination of cell fate, epigenetic regulation is of special interest in the nervous system due to the longevity of neurons. Unlike other organs in the body, the regeneration rate of neurons in both central and peripheral nervous system is remarkably low, suggesting that neurons are designed to survive for decades, possibly even the perpetuity of the organism. In order to survive such a long period of time, neurons must possess the greatest capacity for adaptation to changes in the environment, including physical and psychological stress. Given the inherent stability of DNA, such flexibility is most likely the result from epigenetic regulation of genes. Easily detectable phenotypes that reflect the wide range of epigenetic mechanisms in the neurons are observed in extreme conditions such as development, where the cells experience rapid environmental changes and chemical stress. Likewise, injury and inflammation cause such environmental changes to the innervating neurons that perceive pain. So far, the focus of pain related research has been mainly on genomics and pharmacogenomics 38 , 44 , 45 . Yet, recently, increasing number of investigators have started to realize the importance of epigenetic mechanisms that link the phenotype and genotype related to pain 71 . Those studies on epigenetics of pain are reviewed in this manuscript as initial evidence showing participation of epigenetic mechanisms in pain state. Understanding the mechanism of how injury and inflammation induce changes in gene expression leading to changes in pain perception (e.g. analgesia or allodynia), would allow us to develop better therapeutic strategies. Our goal in this review is to introduce the basic concepts of epigenetics, and to review direct evidence of epigenetic mechanisms involved in pain signal. Epigenetic drugs that are already available for other diseases will be discussed briefly as well. In this review we will: 1) briefly discuss the achievement and limitation of pain genetics, 2) discuss basic aspects of molecular mechanisms that are important for understanding gene regulation, 3) examine existing evidence of epigenetic mechanisms’ contribution to pain pathways by reviewing transcriptional regulation of opioid receptors, and 4) mention some epigenetic drugs that may be used for intervention.

Pain has an impact on the patient’s life at several levels. In addition to the suffering, pain, especially chronic pain, can impair the individual’s ability, quality of life, mood, and sleep, which can result in difficulties in personal and social life. Even though objective assessment of pain has been challenging, many self report scales have been developed in order to standardize the measurements, and have been utilized successfully both in clinical practice and research 25 . The inter-individual variability in pain response and analgesic response to pharmacological agents, and variability in analgesic response patterns in different stains of mice have indicated genetic basis of pain perception 11 , 59 , 60 . Following such observation, significant efforts have been made to identify genes and polymorphisms that are involved in specific pain phenotype, and to develop tailored therapy for each patient 14 , 75 . Numerous genes and single nucleotide polymorphisms (SNPs) have been identified to affect pain perception and response to analgesic drugs. However, accumulating data have been showing some controversial results on the genetic associations with pain related phenotype 74 . For example, the most prevalent single nucleotide polymorphism (SNP) on human μ-opioid receptor ( ORPM1 A118G, asn40asp, rs1799971) was initially reported to show increased binding affinity for β-endorphin and to cause subsequent increase in G-protein coupled potassium channel activation, compared to the wild type 2 . Yet, more recent independent study found no significant difference in ligand binding affinity, including that of β-endorphin, and in the activation of intracellular signaling cascade, between cells expressing the mutant and the wild type ORPM1 1 . Similarly, catechol-O-methyltransferase (COMT) val158met SNP (rs4680) was reported to be associated with higher pain sensitivity by some groups, while others found the same SNP to have no association with pain perception 39 , 89 . Such inconsistencies from genetic studies suggest possible existence of additional, DNA sequence independent, factors that play a role in pain phenotypes. Such pathways, proteins, and mechanisms are essential topics in the field of epigenetics.

Chromatin is the structural conformation of DNA in association with assembly proteins. The nucleosome is the fundamental repeating unit of chromatin, consisting of about 140 base pairs of DNA wrapped around a histone octamer (consisting two copies of H2A, H2B, H3 and H4). Chromatin is dynamic, often switching between two higher order chromatin structures: euchromatin and heterochromatin. Euchromatin is de-condensed chromatin, consists largely of coding sequences that are “poised” for transcription 63 . The transcriptional activity of the genes in euchromatin is determined by recruitment of transcription factors and other nucleosome-remodeling complexes, which may be either repressors or activators. On the other hand, heterochromatin is a highly compacted chromatin structure where genes are silenced. Heterochromatin is known to define centromeres, facilitating chromosome segregation during mitosis and meiosis. Also, heterochromatin is formed at telomeres, ensuring stability of genome. Histone tail deacetylation and methylation of specific lysine residues recruit heterochromatin-associated proteins (e.g. HP1) which establish DNA methylation and the formation of heterochromatin 61 , 66 . The nucleosome is also dynamic with post-translational histone modifications, subunit variation, DNA methylation, and small non-coding RNA interaction. All these modifications on nucleosome are interrelated to each other and regulate euchromatin and heterochromatin formation in an orchestrated manner 32 . In other words, the complexity achieved by combinations of all the different modifications mentioned above provides multidimensional layers to the readout of DNA and results in different patterns of gene expression or silencing from the same genome. Therefore, in order to gain a complete understanding of the epigenetic regulation of gene networks, all the basic paradigms must be incorporated into an integrated picture. Histones are small, basic proteins that are extremely conserved throughout evolution. The first 20 or so amino acids of histones, known as the histone tail, are very highly conserved, and they structurally extend from the nucleosomal disk, which allows access of post translational modification enzymes 54 . In the histone tail, serine (S), threonine (T), and tyrosine (Y) residues can undergo phosphorylation, and other residues, such as lysine (K) and arginine (R), can be methylated, acetylated, ubiquitinated, and sumoylated 76 . Furthermore, lysine residues have potential to be mono-, di-, or tri-methylated 42 . These histone modifications are known as “marks” and provide specific docking sites for many chromatin-associated proteins. The Histone Code hypothesis predicts that the type, location, and combination of histone marks determine recruitment of a specific chromatin-associated proteins or transcription factors, and subsequently, determine whether the gene would be expressed or silenced under a particular set of circumstances 32 . The histone marks can be created, recognized, and reversed by different proteins: writers, readers, and erasers respectively 84 . The writers include HAT (histone acetyltransferases), histone kinases, and HMT (histone methyltransferases) 56 , 73 . Marks imprinted by writers are recognized by the reader proteins containing domains engineered to recognize particular mark, i.e. chromodomain for methylation and bromodomain for acetylation 34 , 43 . The modifications can also be reversed by eraser proteins, such as HDAC (histone deacetylases), phosphatases, LSD (lysine specific demethylases), and Jumonji domain-containing proteins. The histone subcode hypothesis predicts that some of these chromatin-associated proteins may also be post-translationally modified to create an additional layer of the chromatin code, a subcode, that determines transcriptional activity of the target genes 51 . This hypothesis was supported by changes in localization of pan-nuclear HP1γ, which becomes exclusively euchromatic with phosphorylation of Ser 83 residue, suggesting that p-Ser 83-HP1γ may adopt different modes for silencing the target gene compared to non-phosphorylated HP1γ 51 . Together, the histone code and histone subcode hypotheses show how post-translational modifications of different nuclear proteins precipitate differential protein complex recruitment, gene expression, and chromatin dynamics. In addition to histone modifications, chromatin structures can be altered by recruitment of nucleosome remodeling complexes. Nucleosome remodeling machines, such as SWI/SNF, NuRD (nucleosome remodeling and deacetylation) and CHRAC (chromatin accessibility complex), can move nucleosomes along DNA in ATP-dependent manner 47 , 77 , 88 . Several of these molecular machines have been found to be conserved from yeast to human 49 , 82 . The SWI/SNF family of nucleosome remodeling complexes transiently alters the structure of nucleosome, exposing DNA 67 . Similarly, FACT (facilitates chromatin transcription) initiates nucleosome release from DNA, and allows transcriptional elongation by RNA Pol II 64 . These nucleosome remodeling complexes work in concert with histone modification complexes such as SAGA (Spt-Ada-Gcn5-acetyltransferase) to regulate gene transcription 19 . Other nucleosome remodeling complexes, such as the SWR1 complex, are involved in exchange of histones and histone variants 57 . There are a number of histone variants that may occupy a nucleosome to yield an effect on transcriptional activity 22 . For instance, histone H3 has three variants in mammals, H3.1, H3.2, and H3.3, which differ from each other by one to four amino acids in the tail 52 . These small differences allow the variants to have unique patterns of modifications and different affinities to the binding factors. Together, the combinatorial effect between histone variants and their participation in the Histone Code, known as the histone “barcode,” introduces another level of complexity to transcriptional regulation 22 . The mechanisms and timing of histone variant exchange and synthesis, and interrelationship between the chromatin dynamics and nucleosome remodeling is an active area of research. In order for the nucleosome remodeling complexes and histone modification complexes to function properly in regulating gene transcription, they must be recruited to the correct location at the correct time. In this regard, DNA sequences, especially the promoter sequences, play a significant role. Promoter footprinting and Electrophoretic Mobility Shift Assays (EMSAs) were utilized to identify sequence specific transcription factors 31 . Subsequent investigations on these sequence-specific transcription factors revealed DNA binding domains and transcriptional regulatory domains, indicating that these proteins serve as adaptor proteins between DNA and chromatin-associated protein complexes 50 . For example, the Sp1/KLF (Kruppel-Like Factor) family was established based on their similarities in the DNA binding domain, which is comprised of three zinc fingers. Some of the members of this transcription factor family were able to compete with each other for the same promoter, and due to the variations in their regulatory domain, promoter occupancy of a specific family member resulted in distinct levels of transcription 50 . Proteins that interact with transcription factors to promote transcription are called co-activators, while proteins that repress transcription are called co-repressors. These non-histone chromatin proteins function either as co-activators or co-repressors via distinct mechanisms, as mediators of histone modifications, such as acetylation, methylation, and ubiquitination. In addition to the basis of sequence specific recruitment of transcription factors, DNA can also function as a template for modifications. Unlike histone modifications which can occur in great variability, from methylation to ubiquitination, DNA modification is limited to methylation of cytosine residues. DNA methylation was the first type of epigenetic changes to be studied as a mechanism for the inactivation of tumor suppressors. DNA methylation occurs on di-nucleotide CpGs by DNMTs (DNA methyltransferases), and the methylation is enriched at non-coding regions 12 . DNA methylation is closely related to heterochromatin formation and chromatin silencing. DNMTs interact with HP1 and the SUV39H1 histone methyltransferase, which functions to establish tri-methylation of histone H3-K9, and induce heterochromatin formation and gene silencing 16 , 72 . DNA methylation-mediated gene silencing has physiological significance. First, hypermethylation of one of the two parental alleles of a gene is a process of genomic imprinting that ensures monoallelic expression of genes 5 . Second, hypermethylation of repetitive genomic sequences prevent chromosomal instability and translocations 35 . DNA methylation has been classified into two types, namely de-novo and maintenance methylation 12 . Human cells utilize three DNA methylases, DNMT1, DNMT3a, and DNMT3b. DNMT3a and 3b function to establish embryonic methylation patterns, while DNMT1 works at the replication forks during DNA replication to maintain the methylation patterns in semi-conservative way. However, DNMT1 appears to be inefficient at maintaining the methylation of many CpG dense regions. Therefore, the de-novo activities of DNMT3a and DNMT3b are also necessary in somatic cells in order to reestablish the methylation patterns 6 . The mechanism of DNA methylation inheritance is more obvious than many other epigenetic regulators and the mark appears to be very stable. However, the presence of de-novo DNMTs suggests that certain amount of flexibility to this mechanism may also exist. Beside DNA and histone modifications, RNA interference (RNAi) machinery has been linked to post-transcriptional gene silencing and heterochromatin formation. RNAi machinery, including proteins like Dicer, Argonaute, and RdRP (RNA dependent RNA polymerase), breaks down double stranded RNA (dsRNA) into small RNA molecules, named microRNA (miRNA), to interact and degrade the mRNA of homologous genes and inhibit translation 13 . Studies in S. Pombe have shown that mutations of any proteins in the RNAi machinery lead to defects in chromosome segregation due to unstable centromeric heterochromatin 23 , 83 . This evidence suggests the role of miRNA in heterochromatin formation, in addition to post-transcriptional gene silencing. Similar evidence has been found in mammals. For example, mouse embryonic stem cells with depletion in a component of RNAi machinary, Dicer, showed significant decreases in the formation of centromeric heterochromatin and failed to differentiate 33 . In human-chicken hybrid cell line, loss of Dicer resulted in abnormal localization of heterochromatin proteins, depletion of H3-K9 methylation, accumulation of abnormal mitotic cells with premature sister chromatid separation, and ultimately cell death 15 . Therefore, these studies indicate an active participation of RNAs in chromatin remodeling and gene silencing. In summary, there are multiple layers of regulatory mechanisms that are functionally interrelated for target gene silencing or activation.

Important advantage for brining epigenetic drugs into therapeutics is in the reversible nature of epigenetic pathways 78 . Unlike genetic mutations, if a pain phenotype is caused by epigenetic phenomena, such as DNA methylation or histone modifications, they can be chemically reversed. Such involvements of epigenetic mechanisms in pain have been already implicated and accordingly some therapeutic strategies for pain management have been suggested 4 , 7 , 18 . However, as implicated above with the opioid receptor, transcription is regulated by complex multi-layered pathways which involve proteins with multiple target genes and substrate. Therefore, understanding the epigenetic pathway is critical to apply epigenetic drugs for intervention. In this section, we will briefly review some of the available epigenetic drugs and how they may be applicable for pain management. Histone deacetylase inhibitors (HDACis) have been evaluated as therapeutic agents in several conditions including cancer, autism, neuropathic and inflammatory pain 8 , 68 , 70 . HDACis induce increase in acetylated histones by blocking histone deacetylase activity of HDAC proteins. In general, histone acetylation has been associated with transcriptional activation. Recent study has investigated the potential of HDACis as analgesic drugs. Based on studies reporting the upregulation of mGlu2 in DRG can induce analgesia and transcription of of mGlu2 is regulated by histone acetylation, the authors tested two HDACis, MS-275 (benzamine derivative) and SAHA (suberoylanilide hydroamic acid) in a mouse model of persistent inflammatory pain 7 . With 5-day treatment, both HDACis showed analgesic effect on the mice and increased mGlu2 expression in DRG without changing the expression of mGlu1a, mGlu4, or mGlu5 receptors. Another study showed that treatment with a HDACi, Trichostatin A, in a mouse model of endometriosis induced transient decrease in the capsaicin activated type-1 cation channel (TRPV1) and significant improvement in response to a noxious thermal stimulus, suggesting its role as an analgesic drug 53 . While these evidences appear promising for practical applications, with current knowledge, clinical use of these agents is still limited. As we have previously discussed, HDAC proteins have multiple target genes. Therefore, it is evident that HDACis will not only inhibit deacetylation of desired chromatins, but also that of other chromatins which may result in severe side effect. Drugs that affect DNA methyltransferase activities, such as glucosamine and valproic acid, would present similar specificity problem. Thus, in order to understand the mechanism and regard possible application of these drugs, changes in transcription of all genes, rather than a single target gene, need to be considered. Development of more specific HDACis for each HDAC family members, and perhaps, combined with gene therapy, development of a target gene specific HDACis will be extremely beneficial for therapeutics. The siRNAs, as possible epigenetic drugs, stand on a slightly different ground from other epigenetic interventions. Unlike HDACis, siRNA sequence can be designed for a specific target gene. A number of siRNA treatments, targeting NR1 and NR2 subunits of NMDA receptors, RTPV1 channels, BDNF, and other pain related genes, in animal models have been successful for alleviating pain 17 , 21 , 36 , 79 . Some of these may become one of the first epigenetic drugs that become applicable for clinical trials in pain management.

Epigenetic regulation of gene expression is a rapidly growing area of research. As epigenetic mechanisms, including DNA methylation, histone modifications, RNAi, and chromatin dynamics, are theoretically reversible, there exists considerable confidence as to therapeutic possibilities. The fundamental concepts and general principles described in this review are only a brief summary of what is known today. Our goal is to expose the readers to these concepts so that pain-related phenotypes can be investigated from the epigenetic point of view. Many fundamental questions connecting the concepts of epigenetics and pain remain to be answered. For example, in the innervating nociceptive neurons, which epigenetic pathways are activated or repressed during injury and inflammation? What is the epigenetic basis of peripheral sensitizacion and pain centralization (e.g. hyperalgesia and allodynia)? How does inflammation at the end of a periphery neuron affect gene expression in the connected neurons of the central nervous system? What is the epigenetic basis of neuropathic pain? How do the experience, stress, and memory of pain affect patterns of gene expression in the brain, and how can they be reversed? Do emotional pain and physical pain differ or coincide in terms of changes in gene expression? Does pain relief follow the same epigenetic pathways that generate pain? The questions regarding the epigenetics of pain range from basic molecular biology to psychology and even to philosophy. We anticipate that investigations in this field of research will reveal new paradigms for understanding the pathology of pain, and for developing better pain management strategies.