Learning from genetic responsibility

Rapid improvements in sequencing technologies and the performance of genetic testing has resulted in the controversial notion of genetic responsibility.8–11 It has been argued that the predictive nature of genetic information should be considered a valuable additional tool for empowering at-risk individuals who are now able—and possibly ought—to adopt ‘life strategies’ that are in line with their genetic susceptibilities, in order to prevent disease and optimise health.12 More specifically, narratives of responsibility have emerged in contexts such as a patient's choice whether to disclose potentially useful genetic information to family members13 or prospective parents' alleged moral responsibility to consider their genetic profiles before conceiving, or to use prenatal testing, in order to give birth to the healthiest child possible.14 Consequently, the implementation of genomic sciences into healthcare demonstrated how molecular explanations of disease—and the availability of new biomedical means to prevent or treat it—can contribute to the construction of novel imperatives for different actors in society.

In contrast with the genome, epigenetic mechanisms and variants are determined partly by living conditions and lifestyle. Modifications to the three-dimensional structure of DNA at specific genes in specific cells influence gene expression and thus health. Changing the level of accessibility of the transcriptional machinery to some genes has an effect on the subsequent production of proteins that tissues need in order to execute their normal biological functions.15

By connecting external and internal environments to genes, epigenetics has been perceived as a revolutionary field of study.16 Anthropologists, for instance, have claimed that recent findings in epigenetics are transforming the long-standing ‘nature–nurture’ debate that traditionally focused on them as two distinct sets of determinants of identity, behaviour and health. They argue that epigenetics demonstrates the fallacy of this dichotomy:In an epigenetic world, recognition of intergenerational continuities other than by the transmission of DNA brings about a crucial ontological shift; an embedded body is not the product of interactions of nature and nurture but, by definition, is situated in an entanglement of nature/nurture that transcends generations, raising profound questions about concepts of self and body as clearly bounded entities.17, ^{p. 303}

The adaptive nature of some epigenetic modifications and their transgenerational inheritance have also contributed to epigenetics being labelled as revolutionary. These features have revived the discarded Lamarckian theory, according to which an organism's acquired trait and biological adaptation to its environment can be biologically passed on to future generations through germ cells.17

So far, ethical and legal discussions regarding epigenetics have been framed mainly in terms of comparisons between the fields of epigenetics and genetics. For instance, Rothstein has argued that while epigenetic changes may to some extent be exceptional when compared with genetic variants at the scientific level (high frequency of occurrence, dose-dependent, reversible, tissue-specific and species-specific), they are not that exceptional from an ethical and legal perspective and require no ‘new ethical paradigm [or] legal regime’2 ^{(p. 734)}. Already existing legal frameworks such as environmental regulations and the U.S. Federal Genetic Information Non-Discrimination Act of 2008 (GINA) may be sufficient, he argues, for addressing the concerns that arise with epigenetics (such as environmental justice, intergenerational equity, equitable access to healthcare, privacy and confidentiality).

While we agree that epigenetics does not necessarily require a novel ethical framework, we argue that the normative accounts of epigenetics do require a heightened degree of bioethical attention, especially considering its potential impact on the political theory of the family and its relation to social justice18 ,19 as well as intergenerational justice.20 In fact, epigenetics opens a much wider range of opportunities for health intervention than genetics have in the past. In contrast with genetic mutations, detrimental epigenetic variants may be more easily preventable or curable, because they are plastic and reversible.5 While genetics are mainly determined by biological inheritance—something upon which individuals and society have little direct control—a better understanding of epigenetic programming could imply new opportunities to improve individual and public health—and thus more control and arguably new obligations. Hence, it reveals a set of moral agents (eg, individuals, parents, corporations, governmental agencies, international organisations) that could to some extent be held morally accountable for epigenetic health, and consequently significantly broadens the range of possibilities for holding these agents accountable for voluntary negligence when it results in epigenetic harm to others.

Moral epigenetic responsibilities

Epigenetics is already the object of bioethical attention, which has shed light on ‘individual and societal responsibilities to prevent hazardous exposures, monitor health status and provide treatment’1, ^{p. 224}. As observed by Landecker, ‘[t]he normative implications of a science that draws direct causal links, at the molecular level, from such culturally tender and politically contested issues as parenting and air quality regulation to human health are a subject of much speculation in this field’21^{,p. 2}. At the policy level, epigenetics was suggested, for instance, to feed a ‘social work imperative’ meaning that it should justify and encourage ‘a mandate for social workers to intervene at the policy level, both for today's children and for those in future generations’4, ^{p. 23}. At the clinical level, the malleability and reversibility of some detrimental (or risky) epigenetic variants arguably call for further techno-scientific innovation as well as the creation of new types of biomedical interventions to ‘cure’ the epigenome of at-risk individuals.5

Political scientist Maria Hedlund7 explores the concept of epigenetic responsibility, building on four important conditions for ascribing moral responsibility. She claims that in order to be morally responsible for adverse consequences related to epigenetic health, one must be causally responsible for a certain epigenetic programming or modification (the causation criterion) as well as aware of the causal link between this programming and its consequences, making free and informed choice possible (the cognisance criterion). There must also be a recognised individual or collective obligation to strive for a specific ideal of epigenetic health (the obligation criterion). Finally, one must be able to do what is morally required in order to achieve that ideal (the capacity criterion). Building on these criteria, Hedlund argues that epigenetic responsibility should be primarily prospective, rather than retrospective—that is, oriented towards guiding future actions rather than assigning blame for past actions—and be ascribed primarily to the State, rather than the individual.7

Unfortunately, Hedlund's account of epigenetic responsibility has a number of flaws limiting its utility as a framework for identifying and assigning various types of moral epigenetic responsibility among different actors of society. While we are aware of—and concerned with—the ethical pitfalls of ascribing epigenetic responsibility to individuals, we argue that defining epigenetic responsibility as a priori prospective and belonging mainly to the State is misleading because it is simplistic, ineffective and ethically problematic. It is simplistic because it reduces discussion to the tension between the citizen and the State, whereas other actors may have a role to play (eg, corporations, international organisations). It is ineffective because attributing mere prospective responsibility without the possibility of holding actors responsible for past negligence (through health policies or laws) may result in a very limited upholding of the suggested prospective responsibility. It can be ethically problematic in the sense that ascribing epigenetic responsibilities entirely to the State fails to recognise that overpaternalistic health policies run the risk of turning into coercive measures threatening individual autonomy.

It is oversimplistic to understand epigenetic responsibility as a monolithic concept emerging solely from comparisons between epigenetics and genetics. Such understanding fails to take into account important scientific nuances within the field of epigenetics itself that should be incorporated into the debate. For this reason, we offer an ethical perspective that addresses the diversity of types of epigenetic responsibility. We argue that this diversity is likely to emerge from future developments in epigenetics. Using Hedlund's criteria for assigning moral responsibility (causation, cognisance, obligation and capacity), we make the assumption that in the near future, the two first criteria will be fulfilled, that is, there will be some agreement in the scientific community that specific types of epigenetic programming can be causally responsible for the development of disease (causation) and that this knowledge will be successfully disseminated to the larger public (cognisance). Our framing thus focuses on the last two criteria and underscores two important ‘biological ambiguities’ inherent to epigenetic mechanisms themselves: ‘epigenetic normality’ (in relation to the criterion of obligation) and ‘epigenetic plasticity’ (in relation to the criterion of capacity). These ambiguities have already been reported and discussed by social scientists,17 ,21 ,22 but their important normative relevance has been largely overlooked.

Defining epigenetic normality

The above-mentioned studies are expected to generate a list of detrimental epigenetic variants that may provide causal links between living conditions and health/disease. By mapping these ‘harmful’ variants, these studies determine what a ‘reference’ or ‘normal’ epigenome is, that is, what epigenome is associated with health or at least not associated with specific diseases. Notwithstanding the promising preventive and therapeutic opportunities that this endeavour may provide—and accordingly its potential implications for assigning novel responsibilities for epigenetic health—defining epigenetic normality (and by extension abnormality) is scientifically and ethically challenging for many reasons.

At least three biological features of the human epigenome contribute to making its characterisation an even more daunting task than the Human Genome Project represented 20 years ago. First, in comparison to genomic information which is unique and ubiquitous in all the cells of an organism, each cell line—thus each tissue and biological system—has its own epigenome. Second, as will be discussed more extensively in the next section, the multiple epigenomes of an individual are relatively plastic and subject to multiple modifications that depend on various factors (eg, prenatal environment, lifestyle, age). Third, epigenetic variants interact with each other and are highly influenced by the microenvironment in which these interactions take place, that is, on variations in the microstructure of chromatin at specific loci. Considering these biological features, it becomes a very complex endeavour to define the reference epigenome, and before judging the level of normality of someone's epigenetic programming, the larger context of its occurrence (eg, cell type, age of the person, microenvironment surrounding the variant) must be understood with precision.

The mismatch model of disease development

Another—possibly greater—challenge of characterising epigenetic normality is caused by the ‘mismatch model’ of epigenetic disease development. According to this model, an adverse phenotype does not depend merely on the presence or absence of a specific epigenetic variant, but rather on the mismatch between the previously programmed variant and the individual's lifestyle or living conditions.27 ,28 In other words, the early programming of some distinct epigenetic patterns is sensitive to the developmental environment and aims to better prepare the organism for the environment it will most likely be surrounded by in the future. The mismatch model of disease has been suggested to account for the fact that a specific lifestyle or living condition can be more or less detrimental to different individuals and populations, depending on the environment in which the early epigenetic programming occurred.29 ,30

The mismatch model was corroborated mostly by experiments on obesity. It was suggested that the epigenetic programming of obesity-related genes during fetal development was influenced by the availability of nutrients in the womb during pregnancy: that is, low nutrient availability would result in a programmed increased efficiency of the body for storing calories in the lipid tissues later in life, in order to optimise their uptake and use.31 ,32 Subsequently, if the environment changes during the life course and nutrients become more available and diets become richer in calories, this epigenetic programming may become disadvantageous and lead to a higher risk of obesity and its associated metabolic syndromes. In other words, epigenetic programming can be perceived as an active biological mechanism that aims to increase an individual's ‘fit’ to her surrounding environment. At the same time, this programming can become maladaptive when the environment changes and thus become detrimental to health.

Consequently, it appears that when taken individually without consideration for their environment, epigenetic variants are unreliable biological markers for disease susceptibility and epigenetic normality cannot be regarded as universal. Rather, epigenetic normality has to be rooted in geographical and temporal references that are specific to individuals and populations.33 For this reason, it is often unclear whether epigenetic differences—potentially perceived as epigenetic abnormalities—should be treated as impairments or rather functional adaptations conferring advantages in specific environments.34

Personal and public health ideals

The biological features of epigenetic mechanisms discussed above and the mismatch model of disease development introduces significant complexity to the task of defining and ascribing epigenetic responsibilities. Central to this difficulty are important questions that need to be addressed in order to determine what kind of epigenetic ideal we ought to pursue, individually and collectively, using preventive or therapeutic means, and whether we should pursue such an ideal at all. Such questions may impact the construction of different types of novel personal and collective obligations.

For instance, should public health policies aim at favouring the programming of a specific epigenome that best corresponds to that of the majority or should physicians cure deviations from it? Should we consider the idea of a normal epigenome to necessarily be bound to that of a healthy epigenome? And what could the normative implications of such simplistic associations be? Should we focus on the best environment for the healthy programming and maintenance of epigenetic health or rather promote programming the epigenome so that it best matches the living environment or lifestyle of the individual? These questions should be debated, since each may lead to distinct ethical, legal and social implications that we must anticipate before defining and ascribing moral epigenetic responsibilities. We argue that the three examples presented below require ethical attention.

First, aiming for the ‘typical epigenome’—the one that is found in most people in a given population—may be misleading. Indeed, we should be careful not to conflate the atypical epigenome with the detrimental one, or the one that requires intervention. For instance, it was suggested by Rice et al35 that homosexuality could in part be explained by epigenetic mechanisms during development and ‘atypical’ epigenetic variants associated with distorted prenatal androgen exposure. Needless to say, such findings are very sensitive, since they frame sexual orientation in medical terms. Using the language of biology to discuss homosexuality, we are at great risk of falling into a ‘normal versus abnormal’ debate that could undermine the human rights framing that focuses on protecting individual autonomy and promoting the standing of sexual minorities. In order to avoid simplistic conclusions with regard to the typical epigenome, we should keep in mind David Hume's naturalistic fallacy and be careful not to derive too hastily an ‘ought’ from an ‘is’.

Second, in a context of resource scarcity, the development, implementation and promotion of public health measures that would target particular epigenomes, perceived as an ideal public health goal, might become grounds for discriminating (and potentially eugenic) policies and behaviours. For instance, it was argued that the potential for transgenerational inheritance of some epigenetic variants may be used as an additional argument for the development of sustainable obesity prevention policies.36 Hence, those desiring to have children, but carrying an epigenetic variant (acquired or inherited) recognised to increase one's risk for obesity and transferable to future generations, could be discouraged from—or in the worst case prohibited from—procreating, in order to improve future public health. While we agree that epigenetics may and eventually should contribute importantly to the implementation of more effective and ethically sound health policies, we also need to remain aware of the potential ethical pitfalls of such policies. To some extent, health promotion aiming for the healthy epigenome risks exacerbating social pressure on individuals to optimise, for instance, their offspring's epigenetic programming so that they meet the social norm. Ultimately, this may lead to stigmatisation and discrimination of those whose epigenomes differ significantly from the socially desired or expected ones.

Third, while it may seem wise to consider both the epigenome and its environment when defining social obligations and orienting health interventions, the mismatch model may undermine the implementation of health policies aiming to facilitate the adoption of healthy lifestyles and favouring fundamentally healthy living conditions widely recognised as contributing to the programming of healthy epigenetics (eg, enhancing food availability or socio-economic status in general). In some way, the mismatch model defines ‘normal’ epigenetic programming as functional adaptation to the environment. Speculating regarding such a capacity, how important is it to improve people's living environments through public health policies? The mismatch model may lead to the troubling implication that by trying to improve the social circumstances of disadvantaged groups, we risk making their health even worse. For instance, populations moving from low to high caloric intake may be more susceptible to obesity and its associated diseases, because their metabolism was ‘epigenetically programmed’ to stock calories in an extremely efficient way. Consequently, the mismatch model may devaluate requirements for health policies and could be exploited to justify existing environmental and social injustice. One might argue, for instance, that governments and public health agencies should not intervene to improve the epigenetic programming of disadvantaged populations, because this biological mechanism ‘naturally’ prepares them for facing adverse living environments later in life.

In sum, when trying to identify moral epigenetic responsibilities, a few important difficulties arise in relation to the complex concepts of ‘reference epigenome’, ‘epigenetic normality’ and ‘epigenetic ideals’. What healthy behaviour should individuals adopt if they do not even know what epigenome to strive for? In this context, on what grounds should epigenetics drive the creation and implementation of health policies? Most importantly, we argue, none of the different perspectives presented above should be taken for granted or be privileged prima facie. It is also necessary to recognise that epigenetic normality may be both scientifically and socially constructed, and that our definition and ascription of epigenetic responsibilities may vary according to the preferred perspective. For this reason, a simplistic view of epigenetic normality and associated moral epigenetic responsibility should be avoided.

Spectrum of plasticity

Epigenetic mechanisms rely on a diversity of enzymes, such as histone acetyltransferases, histone deacetylases and DNA methyltransferases, which help create and break covalent chemical bondsⁱⁱ between a plurality of chemical groups (acetyl, methyl, phosphoryl, ubiquitinyl, sumoyl and ADP-ribosyl) and substrates (histones or DNA). Depending on the chemical group and substrate, bonds of variable strengths are formed or broken, thus creating a wide spectrum of plasticity among epigenetic variants. Histone acetylation may last only a few hours, while DNA methylation may be persistent over a lifetime and may even be transmitted across generations. In addition, non-covalent epigenetic mechanisms, such as RNA interference and histone addition, deletion or substitution for a different histone type, contribute to even greater diversity in levels of plasticity.44

The microenvironment in which these reactions are taking place can also significantly influence an epigenetic variant's level of plasticity. A very dense, and thus closed, tridimensional space (sterically hindered microenvironmentⁱⁱⁱ) or surrounding electromagnetic forces may see their stability increase or decrease over time. Thus, the proximity of other epigenetic variants can modify the local microenvironment in a way that favours or impedes other epigenetic modifications.45 ,46 Consequently, the plasticity of epigenetic variants is known to vary depending on location.47

The spectrum of plasticity is space-dependent and also time-dependent. At different moments in cell growth and differentiation, some specific epigenetic modifications are more (or less) prone to occur. For instance, during embryogenesis and fetal development, epigenetic modifications occur at a very high rate and are responsible for the specification of tissues and biological systems. From pluripotent embryonic stem cells to somatic cell lines, specific sets of genes are being silenced by DNA methylation, thus assigning tissues their specific functions within the organism.48

Windows of sensitivity, opportunity and sustainability

Epigenetics' spectrum of plasticity adds further complexity to the already challenging task of assigning moral epigenetic responsibilities. To implement efficient preventive and/or curative interventions based on evidence from epigenetics, it seems that we must consider some precise ‘windows of sensitivity’34^{, p. 222} that vary depending on multiple factors. Moral epigenetic responsibilities should be recognised as necessarily context-dependent and relying on who has a capacity to act and also on the type of targeted chemical bond, its exact location in a specific tissue and its stage of development. Only then can we better identify windows of opportunity and the actors responsible for health interventions, thus satisfying Hedlund's capacity criterion.

Once distinct biological systems are fully developed, their epigenomes' plasticity varies, depending largely on their function. For instance, the immune and nervous systems, which are highly adaptive, also have very flexible epigenomes. Periodically, they require a relatively fast process for the activation or silencing of some genes, in order to better respond to potentially recurring environmental stimuli in the future (eg, immune memory, cognitive memory). In such tissues, even the most stable epigenetic modifications—consider, for instance, DNA methylation that is potentially inheritable—may undergo intraindividual variation during the life course.49 Thus, we should remember that different levels of plasticity of specific epigenetic variants correspond to different tissues and their stages of development, thus providing different opportunities for preventive and therapeutic interventions.

As windows of opportunity become widely recognised in the public arena, it may become increasingly unrealistic to advocate for the strictly prospective nature of epigenetic responsibility. Once opportunities for intervention become commonly known and prospective responsibilities are assigned to specific actors, will it still be reasonable not to hold them accountable retrospectively? For instance, if some chemical substance released by industry is proven to be epigenetically harmful, and regulations are implemented accordingly, would it not be justified to hold accountable a specific company? These questions are of importance for epigenetic responsibility, especially when the resulting deleterious epigenetic variant is programmed early in life, stable over time and hardly reversible by modifying lifestyle or living conditions.

In addition to considering when in the course of a lifetime a window of opportunity occurs, we should consider the length of exposure necessary for having a substantial effect on the epigenome. The precise biological process by which the environment gets progressively ‘imprinted’ in our genes is still unclear. It is important to consider the duration of the ‘exposure’ to newly changed conditions required to engage epigenetic mechanisms and substantially influence epigenetic memory.34 ,50

For now, we can at least anticipate that the more stable an epigenetic variant, the more challenging it will be to reverse it—that is, it may require longer sustained exposure to a modified environment. On the one hand, histone acetylation and phosphorylation are mostly dynamic chemical reactions. Thus, they are easily reversible epigenetic modifications. Therefore, short-term variations in environmental stimuli may be sufficient to reverse them. Simultaneously, such induced modifications' high instability translates into potentially very brief duration. On the other hand, modifying DNA methylation patterns may require increased and sustained efforts, but their variation would normally last longer.45 ,46 ,48 Put simply, short and sparse events may often be insufficient to impact long-term epigenetic health. Timeframes of exposure should be taken into consideration when implementing preventive and therapeutic solutions to epigenetic disorders.

Yet again, the tension between the plasticity and the stability of epigenetic variants informs us vis-à-vis important aspects of epigenetic mechanisms implicated in defining and assigning epigenetic responsibility. Depending on the position of a targeted epigenetic variant on the spectrum of plasticity, the required sustainability of the health intervention to be implemented may also vary. To avoid only temporary beneficial effects of future preventive health policies and medical treatments, discussions of epigenetic responsibility should include considerations of plasticity and sustainability.

Perspectives of distributive justice

Rothstein et al51 describe three important justice-related concerns that are highlighted by research in epigenetics: environmental justice, intergenerational equity and equitable access to healthcare. These areas differ importantly in their focus, regarding what should be distributed equitably, among whom, by whom and in what situation. Environmental justice focuses on geographically/spatially based inequalities, whereas intergenerational justice focuses on temporally based inequalities. Environmental justice focuses on the fair distribution of preventive/public health interventions, whereas equity in access to healthcare focuses on the fair distribution of curative and palliative medical treatments. Thus, proper accounts of epigenetic responsibility should acknowledge possible tensions between these different aspects of distributive justice and the ethical consequences of their prioritisation.

It also becomes increasingly difficult to distinguish the types of epigenetic risk or disease that should be perceived as ‘fixed in nature’, from those that should be perceived as developing during the course of life and depending on other ‘non-biological factors’ such as personal motivation, familial and social support, available resources and infrastructures, or favourable institutions. As discussed by Loi et al52, the blurring of the boundary between ‘nature’ and ‘nurture’ may impact the core foundations of some theories of distributive justice, and to some extent, lead to the overlapping of the Rawlsian (social structural) and luck-egalitarian views of equality of opportunity. Ultimately, this may add complexity to the already challenging task of defining interventions as prevention, treatment or enhancement strategies. It may also play a role in decisions as to which interventions should be publicly funded.

The recurring debate regarding whether health inequalities are in the domain of chance or choice53 is, we argue, central to the assignment of moral epigenetic responsibilities. For instance, in cases where the detected detrimental epigenetic variant was inherited due to parents' lifestyle or living conditions—that is, when the epigenome is perceived as an ‘archive of prenatal environment’54^{, p. 526} –, it would arguably seem wrong to assign retrospective moral epigenetic responsibility to the affected individual. In contrast, if an epigenetic modification was known to have occurred due to unhealthy lifestyle choices (causation criterion), some may be convinced to assign retrospective moral epigenetic responsibility to the person, given she was informed of the risks (cognisance criterion) and of public health prescriptions against this behaviour (obligation criterion), and had control over these causes (capacity criterion). Hence, the public may be tempted to interpret epigenetic responsibility as a form of individual accountability for health.

However, such an interpretation would not stand up to ethical scrutiny from the point of view of the ‘option luck/brute luck’ distinction (Dworkin, luck-egalitarianism). In an unequal society, even those individuals who are informed of the risk and have control over the causes may differ in relation to how informed they are and how much control they have. In some cases, people may choose to be less informed about and to have less control over epigenetic programming. Often, however, and especially in relation to social or natural disadvantages, how informed one is and how much control one has depends on circumstances outside of one's control. In other words, fulfilling the cognisance and capacity criteria for being morally accountable for epigenetic harm to oneself or others would be a matter of brute luck.^iv