Transgenerational Epigenetic DNA Methylation Editing and Human Disease

Abstract

During gestation, maternal (F0), embryonic (F1), and migrating primordial germ cell (F2) genomes can be simultaneously exposed to environmental influences. Accumulating evidence suggests that operating epi- or above the genetic DNA sequence, covalent DNA methylation (DNAme) can be recorded onto DNA in response to environmental insults, some sites which escape normal germline erasure. These appear to intrinsically regulate future disease propensity, even transgenerationally. Thus, an organism's genome can undergo epigenetic adjustment based on environmental influences experienced by prior generations. During the earliest stages of mammalian development, the three-dimensional presentation of the genome is dramatically changed, and DNAme is removed genome wide. Why, then, do some pathological DNAme patterns appear to be heritable? Are these correctable? In the following sections, I review concepts of transgenerational epigenetics and recent work towards programming transgenerational DNAme. A framework for editing heritable DNAme and challenges are discussed, and ethics in human research is introduced.

Keywords: DNA methylation; cytosine; dCas; development; epigenetic editing; epigenetics; epimutation; germline; heritable; transgenerational.

Figure 1

An overview of DNAme through developmental time. Massive demethylation waves occur in primordial germ cells and pre-implantation stages of development, followed by stage-specific DNAme reacquisition.

Figure 2

An epigenetic landscape of developmental potential generated with the help of AI DALL-E3 (OpenAI, San Francisco, CA, USA). The ball of pluripotent stem cells represents early developmental potential, where fate is intrinsically regulated by the epigenetic landscape that awaits as it rolls forward in developmental time. DNAme can be visualized as a looming fog, which accumulates over developmental time and clouds cellular potential, as cells transverse narrowing valleys. Auxiliary forces visualized by a growing storm of clouds may erode and re-contour this potential. Migrating PGCs are illustrated moving along a rare path to a distal peak and escaping the downward forces of somatic cell development, transversing upward onto the next generation’s highest developmental potential. Next-generation ESCs are displayed as well as another developing blastocyst. This feat becomes more difficult when cells have accumulated toxic transgenerational DNAme information.

Figure 3

An overview of cytosine DNAme and demethylation is provided. Briefly, C becomes methylated at C5 with DNMT activity and co-factor S-adenyl-methionine (SAM). 5mC can be actively oxidized to 5-hydroxymethylcytosine (5hmc), then to 5-formylcytosine (5fC), and 5-carboxylcytosine (5caC) intermediates, which can be removed either actively by Thymine DNA glycosylase (TDG) and base excision repair (BER), or passively through DNA replication.

Figure 4

An overview of transgenerational inheritance. Potential sources of environmental exposure that may directly affect the gestating mother, developing F1 offspring, or migrating F2 germ cells within F1 offspring are displayed. The effects may also be indirect, such as F0 extreme trauma, which might alter hormone signaling and augment developmental signaling within the fetus and F2 PGCs. Phenotypic influence and epigenetic inheritance to the >F3 generation, without continued exposure to the original F0 event, is considered transgenerational.

Figure 5

Overview of dCas-based DNAme editing. Briefly, enzymatically “dead” Cas enzymes are positioned adjacent to target CGs or CGIs for effector domain recruitment. Primary concerns for a typical experiment are highlighted in blue text, with additional potential modifications that may enhance editing capacity. Transfection and genome homing efficiency are critical factors for successful editing, as are effector domain selection, the timing of dCas, orthogonal system deployment, and the avoidance of immune responses for repetitive editing strategies.

Figure 6

An overview of CpG Island Methylation Responses (CIMRs) that induce CGI-wide DNAme. Briefly, CG-free DNA is inserted into CG-dense CGI sequences (either double- or single-stranded DNA may be integrated). Uniquely, in primed pluripotency, the flanking CGI CpGs become spontaneously methylated. For most instances of promoter DNAme, this results in stable long-term gene silencing. DNAme is retained after CG-free insert removal and enables transgenerational testing of a single-engineered CGI by tracking DNAme in subsequent generations of offspring when tested in vivo.

Figure 7

A basic framework for determining whether programmed DNAme is transgenerational has been provided. Major considerations for editing are provided as follows. (I) Characterize the source of transgenerational detriment, epigenetic penetrance, and the expected generational durability. Which DNAme sites are changed relative to normal, age, and sex-matched controls without exposure? What proportion of individuals with a particular heritable DNAme exhibit the negative epiphenotype? How many generations of transgenerational inheritance may be expected? For strongly modeled candidates or those supported by long-term epidemiological studies on transgenerational inheritance, epigenetic editing for mimicry can be used to test the suspected site for programmed transgenerational effects. Closed circles represent methylated candidate CG sites, which can be found in clusters or individually. (II) Determine the targeting strategy to mimic transgenerational DNAme. Targeting of PGCs, gametes, or early embryonic windows is preferred for inheritance into entire organisms, and for targeting developmental windows, which precede next-generation PGC migration and development for inheritance testing. Editing strategies are generally protein-homing-based and may also include certain adaptations to the enzymatically dead Cas systems, such as crRNA modifications and effector domain scaffolds. Specific domains or small RNA co-factors active during key developmental windows may enhance these activities towards transgenerational outcomes, especially those building histone marks, such as H3K9me3, which promote eventual DNAme deposition during later stages of development. Demethylation strategies generally rely on the recruitment of active demethylating enzymes or activating domains, such as VP64, to override DNAme. Engineered DNAme sites may also be protected by binding certain factors, such as STELLA, to prevent first-generation removal and allow for durability testing in subsequent generations. Insertions of CpG-free DNA or mutations that drive changes in transcription, for example, may induce local CGI DNAme spreading epimutations. Retained after the original insult, these naturally aid in transgenerational testing. (III) Monitor DNAme for transgenerational inheritance. Programmable Transgenerational Epigenetic Reacquisition (PTER) is the intentional, locus-, or region-specific induction of specific epigenetic configurations, which fail to escape germline or pre-implantation erasure but that are retriggered with each ensuing generation without the continued presence of epigenetic editing systems. This is unique from Programmable Transgenerational Epigenetic Transmission (PTET), which involves protection from germline and pre-implantation stages of epigenetic erasure. Either form of inheritance is of interest to human biology, but these remain important mechanistic clarifications, which can be further defined by stage-specific isolation of developing PGCs and gametes, developing blastocysts, and the examination of multigenerational inheritance in somatic tissues of offspring. Pedigree analyses can aid in establishing transgenerational transmission, penetrance, and durability. For in vitro modeling, which may include human pluripotent stem cells, the reversible cycling of cells between primed and naïve states enables reversible global DNAme switching and the testing of DNAme between additional generations of development. Sites or regions that wane in DNAme over repeat cycling, especially with DNAme-inducing DNA or mutations removed, are less likely to maintain transgenerational activity in vivo. For human early embryo studies modeled in vitro, the 14-day rule would be sufficient for understanding DNAme inheritance through primed stages of pluripotency and up to gastrulation. Beyond this, in vitro mimics of specific human somatic lineages remain state-of-the-art. Given the difference between humans and mice in repetitive element and imprinting regulation, human studies are preferred for human biology; however, preclinical modeling remains essential to understanding and testing transgenerational DNAme correction.