Heritable hazards of smoking: Applying the “clean sheet” framework to further science and policy

Abstract All the cells in our bodies are derived from the germ cells of our parents, just as our own germ cells become the bodies of our children. The integrity of the genetic information inherited from these germ cells is of paramount importance in establishing the health of each generation and perpetuating our species into the future. There is a large and growing body of evidence strongly suggesting the existence of substances that may threaten this integrity by acting as human germ cell mutagens. However, there generally are no absolute regulatory requirements to test agents for germ cell effects. In addition, the current regulatory testing paradigms do not evaluate the impacts of epigenetically mediated intergenerational effects, and there is no regulatory framework to apply new and emerging tests in regulatory decision making. At the 50th annual meeting of the Environmental Mutagenesis and Genomics Society held in Washington, DC, in September 2019, a workshop took place that examined the heritable effects of hazardous exposures to germ cells, using tobacco smoke as the example hazard. This synopsis provides a summary of areas of concern regarding heritable hazards from tobacco smoke exposures identified at the workshop and the value of the Clean Sheet framework in organizing information to address knowledge and testing gaps.

A report from a meeting of the International Workshop on Genotoxicity Testing (IWGT) detailed heritable outcomes that need to be considered when asking what risk management questions must be addressed and what testing strategies are needed for risk assessment (Yauk et al. 2015). However, most regulatory bodies across the world do not have an absolute requirement to test drugs and environmental chemicals prior to approval for their ability to cause heritable damage or to assess the ramifications that such damage might have on human health (Cimino 2006). Furthermore, regulatory attention on genotoxicity has been based on a standard, one-size-fits-all testing approach that does not incorporate risk specific to germ cells very well, as described further below.
Recently, an approach referred to as the "Clean Sheet" was proposed to encourage the examination of a substance's potential adverse effects based on its projected mode of action and anticipated human exposure versus relying exclusively on a predetermined set of standard assays (Dearfield et al. 2017;Luijten et al. 2020). This strategy emphasizes flexible and innovative exposure-based approaches that analyze endpoints most relevant to human health risks. Broader application of the Clean Sheet approach can be facilitated by using case studies to explore this novel regulatory framework in different decision-making contexts.
On September 19, 2019, a workshop entitled Heritable Hazards of Smoking: Applying the "Clean Sheet" Framework to Further Science and Policy was held in advance of the 50th Annual EMGS Meeting in Washington, DC. There is mounting evidence suggesting tobacco smoke exposure imparts heritable effects on germ cells with negative health implications for the resulting offspring .
Given the historical prevalence of tobacco smoking, it is also a substance with a legacy of exposure that is carried among us today and will continue to be passed along to the next generations. Also, the availability of alternatives to traditional tobacco products (e.g., e-cigarettes), which are increasingly being used by young people, means that exposure to tobacco and related products will likely remain a public health concern over the long term.
The workshop assembled researchers to present evidence regarding the genetic and epigenetic toxicity of tobacco smoke and related substances to germ cells, as well as the health effects identified in progeny born of those cells. In addition to exploring the utility of the Clean Sheet framework, the focus of this workshop was to heighten the awareness and urgency of heritable hazards via germ cell exposures and to encourage regulatory bodies to consider equally the health impacts to germ cells and somatic cells from exposures to environmental chemicals and pharmaceuticals.

| APPLYING THE CLEAN SHEET FRAMEWORK TO THE CASE OF GERM CELL TOBACCO SMOKE EXPOSURE
The purpose of the Clean Sheet framework is to provide a systematic but flexible approach to risk assessment aimed at producing the most relevant and informative data for decision making (Dearfield et al. 2017). It details a series of steps that begins with identifying what the problem is and what risk management questions are most relevant to address (planning and scoping). During this step, the exposed populations and anticipated exposure levels are projected to initiate an exposure analysis. Next, the effort assembles what data, evidence, and information are already extant regarding the problem (build the knowledge base). After the knowledge base is assembled, the flexibility of the approach is brought into play with development of hypotheses about how the specific genetic endpoint being analyzed will lead to the adverse outcome of interest (create a rational biological argument). This step provides the basis for determining what genetic testing is most relevant to perform for risk assessment purposes (select assays, perform them, and review the results steps). The resulting data are then used to quantify risk, that is, dose-response analysis and selection of a point of departure (PoD) on the dose-response curve for extrapolation to human exposure levels. Once risk is characterized by combining the relevant testing data and exposure analyses, risk managers can use this information to develop actions to prevent or mitigate the adverse outcome.
In following the Clean Sheet framework, the workshop coordinators first undertook problem formulation and agreed that the current risk assessment paradigms fail to fully address the heritable consequences of environmental exposures (Table 1). The planning and scoping involved establishing the specific endpoints of concern: germ cell and heritable mutations, intergenerational phenotypic changes, and epigenetic effects transmitted across generations. Following this, a decision was made to use tobacco smoke as the case study for the problem, given the widespread population-level exposure and extensive data. An expert group was then assembled to present the evidence and information associated with this exposure.
At the workshop, experts first reviewed current regulatory processes for germ cell mutagenicity, considered the unique features of germ cell biology and heritable effects, and reviewed evidence from recent key studies supporting the potential for heritable genetic and epigenetic effects of tobacco products. A directed discussion with audience participation focused on key questions developed during the problem formulation. This discussion involved evaluating the weight of evidence, identifying data gaps and uncertainties, and interpreting the implications for risk assessment. Finally, ethical considerations and advocacy needs were examined in the context of risk assessment and regulatory action. Ultimately, the Clean Sheet framework was applied to the heritable risk of tobacco smoking with the goal of providing regulatory agencies with information on which to base future actions.

| STATE OF REGULATION OF GERM CELL MUTAGENS
Given the far-reaching health implications of germ cell mutations and other genomic changes (Stenson et al. 2017), identifying chemical agents that may produce such effects should be of high importance to regulatory agencies. However, standard toxicity testing paradigms are generally poorly designed to fully capture perturbations to germ cells that could result in heritable effects (Marchetti et al. 2020). This is in large part due to the limited window of development to which germ cells are typically exposed during testing, in addition to the fact that many assays expose only male germ cells, thus ignoring female germ 1. Somatic adverse effects from smoking are well established (e.g., cancer), but germ cell effects and effects on heritability awareness is very low 2. Reason for workshop-Examine the evidence that exposure to tobacco smoke is a heritable concern, that is, to focus on heritable effects of germ cell exposures 3. Brought together a diverse group of stakeholders (researchers, regulators, public interest groups, bioethicists) to share information 4. Goal is to provide quantitative data to model the potential risk levels of substances that induce genomic damage and contribute to human adverse health outcomes 5. If such data sets are not available or amenable to quantitation, then identify what needs to be done (e.g., fill data gaps) and whether there are alternative actions risk managers can take in lieu of setting regulatory limits 6. Describe what exposed populations are at risk (e.g., young adults, fetuses) and discuss the potentially vulnerable periods of exposure to focus concern(s) 7. Major risk management question -what could assist regulatory agencies to act on the risk posed to germ cells?
Build knowledge base 1. Review current regulatory processes for germ cell mutagenicity 2. Detail the unique features of germ cell biology and heritable effects (to assist targeting of appropriate studies to assay potential adverse effects) 3. Review available evidence from recent key studies supporting the potential for heritable genetic and epigenetic effects of tobacco products Create rational biological argument 1. Examine which studies can help to characterize the mode of action(s) of potential germ cell mutagens and thus provide insights on how to prevent damaging exposures 2. Identify the most relevant pieces of available evidence that demonstrate that there is indeed a heritable risk due to tobacco smoke exposure 3. What studies provide dose-response data for quantitative analysis and if not available, identify data gaps to fill Select assays and perform them 1. Many assays provide qualitative information, but none identified for use in quantitating heritable risk 2. Recommend assays that can be performed to address data gaps and to obtain useful data for quantitative analysis 3. Further studies identifying the specific cellular and molecular mechanisms by which heritable effects are manifested to bolster the evidence and aid in creating a rational biological argument for heritable risk

Review results
Evidence strongly suggests that exposure to tobacco smoke can result in genomic and epigenomic changes to germ cells that are inherited and produce negative health effects in the resulting offspring Select appropriate point of departure (for quantitative analysis) 1. The workshop did not identify studies that could provide a point of departure for performing quantitative analysis 2. Recommendations for dose-response data to assist in identifying an appropriate point of departure Determine expected exposure The workshop focus did not include an exposure assessment, so quantitation of exposure was not discussed Estimate acceptable levels for endpoints of human relevance As there were no quantitative data and an exposure assessment, estimated levels could not be derived Risk characterization 1. Recognize that over a billion people smoke cigarettes worldwide; current human exposure to tobacco smoke is extensive 2. There are several lines of evidence that support tobacco smoke as a genotoxicant in germ cells 3. Evidence indicates a combination of germ cell DNA/chromosomal damage, DNA mutations, and epigenetic alterations contribute to the observed effects 4. Though quantitation of risk not possible at this time, evidence heavily suggests regulatory agencies need to focus on reducing harm to germ cells and mitigating heritable risk, particularly to young people and other vulnerable populations 5. Alternate actions discussed in lieu of quantitative levels for regulatory consideration 6. Effective communication strategies to better inform the public about heritable germ cell effects on health and consistent messaging are needed result in fertilization failures or embryonic death (Marchetti and Wyrobek 2005), it would not identify toxicants that cause non-lethal mutations or epimutations. Furthermore, because the males are exposed only as adults, the test would not identify effects that result specifically from exposure to primordial germ cells (PGCs) or other early stages of germ cell development.
The mammalian spermatogonial chromosomal aberration test (OECD 2016b) likewise examines effects only during the mitotic proliferation of spermatogonia in adult animals, and there is no equivalent standard test for female germ cells. Similarly, the transgenic rodent gene mutation assay (OECD 2013) can be applied to identify chemicals that are mutagenic in male germ cells (Marchetti et al. 2018), but it is not applicable to female germ cells because not enough of them can be collected to conduct the assay. The lack of practical methods to assess mutagenicity in female germ cells was acknowledged as a critical research gap in the IWGT report (Yauk et al. 2015).
Under the global harmonized system, a substance may be classified as a Category 1A germ cell mutagen only if there is positive evidence from human epidemiological studies (UN 2017). However, these data are inherently difficult to produce given the implausibility of obtaining human populations exposed to a single toxicant, the invasive procedures required to obtain female germ cells or developing embryonic germ cells, and the technical limitations in identifying small genomic changes out of the 3 billion base pairs of the human genome.
Consequently, no substances to date have been designated as Category 1A germ cell mutagens. However, this result is more likely a reflection of regulatory testing shortcomings than of biological reality.
At present, there is no standardized test guideline adequately designed to identify heritable effects from germ cell exposures, particularly those mediated through epigenetic changes. This omission means that there is essentially no place in the risk assessment process to account for such effects. Therefore, it is necessary for risk assessment paradigms to shift in ways that are better able to protect the health of present and future generations. This is where the Clean Sheet approach is most applicable because the hazards specific to germ cell biology necessitate different testing design from the standard genetic toxicity testing battery.

| UNIQUE GERM CELL BIOLOGY
To design relevant testing for germ cell damage, it is critical to understand the underlying biology of germ cell development (crucial for building the knowledge base). Using the knowledge of spermatogenesis and oogenesis can help researchers make better decisions regarding what testing assays to apply in determining when and how genomic damage to germ cells can occur ( Figure 1). It also furthers the understanding of what chemical insults to germ cells mean in terms of adverse outcomes in exposed individuals. These are the types of data that are most useful for regulatory bodies to consider because they are most relevant to human heritable risk.
In humans, the earliest germ cells, PGCs, are first formed in the embryo approximately 2-3 weeks after fertilization (De Felici 2013).
In both XX and XY embryos, the initial PGCs migrate to the genital ridge by the sixth week and continue mitotic proliferation through the 10th week post fertilization (Tang et al. 2016). After this time, germ cell development progresses in a sex-specific manner.
In females, immature oogonia first enter meiosis at Embryonic Week 10  The ability of germ cells to cope with DNA damage is dependent upon both sex and developmental stage. For instance, fetal primary oocytes likely possess efficient DNA repair capacity but also appear to have a highly sensitive apoptotic response that correlates with the high rate of germ cell apoptosis that occurs prior to birth (Winship et al. 2018). During the protracted postnatal period of meiotic arrest, primary oocytes may be particularly sensitive to DNA damage but exhibit a decreasing ability to repair the damage with age (Myers and Hutt 2013). Oocytes also exhibit a relatively high incidence of aneuploidy (Jones et al. 2013) that substantially increases with increased maternal age at meiotic completion (Stuppia 2013 proper embryonic development (Marchetti and Wyrobek 2005).
In addition to DNA damage, the processes involved in germ cell specification and differentiation introduce periods of susceptibility to aberrant reprogramming of the epigenome. PGC specification is driven by expression of BLIMP1 and SOX17, which suppress somatic gene expression and promote germ cell gene expression, respectively (Tang et al. 2015). These transcription factors further lead to repression of the maintenance DNA methyltransferase DNMT1 and the de novo DNA methyltransferases DNMT3A and DNMT3B, as well as upregulation of TET1, which converts 5-methylcytosine to 5-hydroxymethylcytosine. As a result, global CpG methylation levels fall from approximately 80% at the time of PGC specification to approximately 5% in week 9 fetal PGCs (Tang et al. 2015). This demethylation facilitates parental imprint erasure as well as X chromosome reactivation.
Despite global demethylation, certain regions of the PGC genome are more resistant and retain their DNA methylation (Tang et al. 2015).
In particular, evolutionarily young retrotransposons retain higher levels of methylation, which maintain their repression. In addition, approximately 6% of regions escaping DNA demethylation are depleted of repetitive elements and occur in genes with enhanced expression in the brain (Tang et al. 2015). These demethylationresistant genes, which are associated with diseases such as metabolic disorders, schizophrenia, and multiple sclerosis, may be important for epigenetic inheritance.
Although hypomethylation of gene promoter regions is generally associated with derepression of the gene, promoter methylation is F I G U R E 1 Overview of select developmental stages and genomic processes subject to environmental insults in (a) female and (b) male germ cells decoupled from gene expression in PGCs, with only about 12% of hypomethylated genes exhibiting upregulation (Tang et al. 2015).
These upregulated genes include those in the piRNA pathway as well as Krüppel-associated box zinc finger genes that recruit repressive complexes to specific sequences, resulting in heterochromatin formation. Changes to posttranslational histone modifications during PGC development also serve to alter chromatin configuration (Tang et al. 2015). Therefore, chromatin reorganization adds an additional layer of transcriptional control during germ cell reprogramming.
Male fetal germ cells begin to regain DNA methylation in the second trimester after entering mitotic quiescence (Tang et al. 2015). De novo methylation appears to be largely completed within a few days after birth and, thus, prior to any meiotic activity (Wermann

| EVIDENCE PRESENTED AT THE WORKSHOP
The biological knowledge described above set the stage for the interpretation of results from different studies, and the Clean Sheet provided a framework for considering them regardless of whether they were conducted following established test guidelines or not. The data from such studies can help to characterize the mode of action(s) of potential germ cell mutagens and, thus, provide insights on how to prevent damaging exposures Sasaki et al. 2020). It should be noted that the evidence summarized below should not be construed as a comprehensive review of the existing literature.
Rather, presenters described empirical data from key studies in both animal models and humans (epidemiological studies) on the effects of tobacco smoking directly on germ cells or on the offspring of smokers.
In pregnant females who smoke, both the somatic and germ cells of the F 1 generation are exposed, with potential heritable effects manifesting in the F 2 generation. In adult female smokers who cease smoking prior to conception, impacts on their own oocytes manifest in the F 1 generation. In the case of paternal smoking, germlinemediated effects are evident in the F 1 generation. The presenters cautioned that these studies presume germ cell exposure through the smoker, not secondhand or sidestream smoke.

| Effects of smoking on the germ cell genome
The Less is known about direct effects of tobacco smoke on female germ cells. However, according to a review by Zenzes (2000), smoking may decrease the quality and quantity of oocytes. For example, smoking was associated with an 8-17% reduction in oocytes retrieved during in vitro fertilization, with higher smoking rates corresponding with a greater reduction. The effect of smoking upon oocyte number was further compounded by increasing age. Conversely, a dosedependent relationship was found between increased smoking and earlier age of menopause, which is caused in part by oocyte depletion.
A dose-dependent relationship has also been found between maternal smoking and incidence of spontaneous abortion, further suggesting smoking may cause genomic aberrations in female germ cells.

| Effects of smoking on the germ cell epigenome
The effects of smoking on the germ cell epigenome have been evalu- however, it did appear to be associated with an increase in the variability of methylation levels. Differential methylation appeared to be enriched at sites reported to escape protamine replacement during spermatogenesis, suggesting that DNA with retained histones may be more sensitive to environmental insults. This study also demonstrated that CpG sites that are resistant to demethylation during germ cell reprogramming appeared to be overrepresented in the differentially methylated sites.
In laboratory studies on mice, nicotine alone has also been found to induce DNA methylation changes in sperm. According to a study by McCarthy et al. (2018), the sperm of mice exposed to nicotine via drinking water exhibited a global increase in DNA methylation levels.
However, DNA methylation levels specifically in the promoter region of the D2 dopamine receptor were decreased. This finding corresponded with a significant decrease in D2 receptor mRNA expression in the striatum of male F 1 mice descended from nicotine-exposed fathers. Although there were tissue-and sex-specific differences in the effects observed in the F 1 generation, these data suggest that nicotine may be capable of inducing epimutations in sperm that affect the phenotype of the resulting offspring.

| Germ cell-mediated effects of smoking on offspring genome
A few human studies provide evidence supporting an impact of paternal tobacco smoking on the offspring genome. One small human study by Linschooten et al. (2013) indicated that paternal smoking within 6 months of conception correlated with a significant increase in tandem repeat minisatellite mutations in their offspring. Furthermore, the effect appeared to be dose dependent, with 5.3, 19, and 33% of children from nonsmoker, irregular smoker, and daily smoker fathers, respectively, exhibiting mutations. An additional human study by Laubenthal et al. (2012) found that paternal preconception smoking was positively associated with single-and double-strand DNA breaks in the cord blood of their offspring. These human findings are complemented by a recent mouse study by Beal et al. (2019) showing that paternal exposure to B[a]P significantly increased copy number duplications and de novo mutations in the offspring. Thus, there is growing evidence that paternal tobacco smoking and exposure to the components of tobacco smoke lead to an increase in the number of mutations that offspring inherit.

| Germ cell-mediated effects of smoking on offspring epigenome
The impact of tobacco smoking on the offspring epigenome is less established. However, recent studies provide support for epigenomic alterations. In a human study by Knudsen et al. (2019), whole blood samples were collected from adolescent and adult F 1 offspring to examine the effects of paternal (F 0 ) smoking on DNA methylation. This study identified differentially methylated regions associated with paternal smoking in genes related to innate immune system pathways as well as lipid metabolism and fatty acid biosynthesis. The study was unable to conclusively determine if the observed effects were due to altered methylation in the fathers' sperm because it is possible that secondhand smoke during gestation and/or childhood could have contributed to these changes. However, it was notable that persistent DNA methylation changes were discernible in adult F 1 offspring up to 54 years of age.
Based upon a separate study by Joubert et al. (2016) of cord blood collected from 6,685 newborns, maternal smoking was associated with significant differences in DNA methylation of the aryl hydrocarbon receptor repressor gene. However, because this study included women who continued to smoke during pregnancy and, thus, during somatic development of the newborns, it cannot be determined how many of these methylation changes, if any, were attributable specifically to changes induced in the maternal germ cells.

| Germ cell-mediated effects of smoking on offspring phenotype
Evidence is accumulating that preconception tobacco smoking can Two types of childhood cancers appear to exhibit elevated risk from germ cell exposures to tobacco smoke. Pang et al. (2003) reported that leukemia risk was elevated in children of fathers who smoked prior to conception. They also found that hepatoblastoma risk was elevated with both paternal and maternal preconception smoking, and this risk was compounded further when both parents smoked prior to conception. These and other supporting data were considered sufficient by the International Agency for Research on Cancer to conclude that paternal tobacco smoking is a causative factor for an increased risk of childhood cancer (IARC 2012).
Asthma is another disease linked to germline exposure to tobacco smoke. One large cohort study by Accordini et al. (2018) utilized data collected from 2,233 mothers and 1,964 fathers to examine the associations between parental and grandparental smoking and asthma risk in offspring. In the case of offspring (F 2 ) whose mothers (F 1 ) were exposed to maternal (F 0 ) smoking in utero, there was a nonsignificant relative risk ratio (RRR) of 1.3 for asthma. The RRR was lower for offspring (F 2 ) of fathers (F 1 ) exposed to maternal (F 0 ) smoking in utero.
However, a significant increase in asthma was found for offspring (F 2 ) of fathers (F 1 ) who started smoking prior to the age of 15 years, with an RRR of 1.43. This elevated risk was eliminated if the father began smoking after 15 years of age. A previous study by Svanes et al. (2017) had found similar results, with significant increases in risk for childhood asthma associated with grandmaternal smoking as well as with paternal smoking starting before the age of 15 years.
An earlier, smaller cohort study by Li et al. (2005) found a higher risk of asthma associated with grandmaternal smoking compared to the Accordini et al. study. In offspring (F 2 ) whose grandmothers (F 0 ) smoked while pregnant with their mothers (F 1 ), there was an odds ratio (OR) of 2.1 for developing childhood asthma. No elevated risk of asthma was found for offspring (F 2 ) whose mothers (F 1 ) quit smoking prior to pregnancy.
Cohort studies have indicated that grandparental smoking also elevates the risk of autism, autism traits, and attention-deficit/hyperactivity disorder (ADHD). A study by Golding et al. (2017) using data from the Avon Longitudinal Study of Parents and Children cohort showed significantly elevated odds for autism traits and diagnosed autism in grandoffspring (F 2 ) of pregnant smokers (F 0 ). A separate study by Yim (2019) using data from the Nurses' Health Study II found that grandmaternal (F 0 ) smoking was associated with an adjusted OR of 1.18 (95% CI, 1.11, 1.25) for ADHD in the F 2 generation.
Animal studies indicate an elevated risk for asthma from germ cell exposure to nicotine alone. In a multigenerational study by Rehan et al. (2012), the F 2 offspring of F 1 rats exposed to nicotine both in utero and for 21 days postnatally exhibited significant decreases in pulmonary function, indicative of an asthma-like phenotype. Because the F 2 generation was exposed to nicotine via the germ cells developing in the F 1 generation, the phenotypic response in the F 2 generation was mediated likely by germ cell changes. In a follow-up study (Rehan et al. 2012), this asthma-like phenotype was found to persist into the F 3 generation, further supporting the possibility of germline inheritance.
Mouse studies indicate that both grandmaternal and paternal exposure to nicotine alone can also increase ADHD-like phenotypes in offspring (Zhu et al. 2014;McCarthy et al. 2018). In the former case, Zhu et al. (2014) found that the F 1 offspring of female (F 0 ) mice exposed to nicotine for 3 weeks prior to mating and for the duration of pregnancy exhibited ADHD-like behaviors in both sexes. However, only F 2 mice descended from F 1 females continued to exhibit these behaviors. The behaviors were also found in F 3 mice descended from F 2 females. Similarly, a study by McCarthy et al. (2018) from the same lab found that when adult male (F 0 ) mice were exposed to nicotine, both male and female F 1 mice exhibited ADHD-like behaviors.
Overall, the evidence presented at the workshop strongly suggests that exposure to tobacco smoke can result in genomic and epigenomic changes to germ cells that are inherited and produce negative health effects in the resulting offspring. However, despite the breadth of information provided, the workshop attendees agreed that additional pieces of information are required to further the risk assessment of tobacco smoke exposure to germ cells. Similarly, further studies identifying the specific cellular and molecular mechanisms by which heritable effects are manifested would bolster the evidence and aid in creating a rational biological argument for heritable risk. However, a major complicating factor with evaluating risk from tobacco smoke is that it is a complex and variable mixture with over 4,000 individual compounds. Any number of these compounds could lead to various molecular initiating events that could synergize, antagonize, or otherwise interact with one another.
Furthermore, products such as e-cigarettes are increasingly being used as alternatives to tobacco cigarettes, particularly among young people (Jaspers 2019). These products present their own unique combinations of active and inactive ingredients that require specific attention in the risk assessment process.
Beyond mixture toxicity, more concerted study of mechanisms based upon germ cell biology could better inform which stages of germ cell development are most sensitive to tobacco smoke exposure.
Studies specifically examining the link between genomic and/or epigenomic changes in parental germ cells to phenotypes observed in offspring and grandoffspring are lacking. Although many of the phenotypic outcomes reported in offspring are likely due to epigenomic changes in parental germ cells, this link has not been explicitly examined in a meaningful way. Information regarding epigenomic changes in germ cells that may be inherited in offspring are especially needed.
The question of how many generations through which such epigenetic changes may persist also needs to be addressed systematically.

| IMPLICATIONS FOR RISK ASSESSMENT
Risk assessment should be focused on addressing the most appropriate risk management questions relevant to human health outcomes (Dearfield et al. 2017). The subject workshop sought to address the question of the heritable risk posed to germ cells by exposure to tobacco products (see Table 1), which is of high importance for population health but has not been accounted for specifically in standard risk assessment processes. Given that 1.1 billion people smoke cigarettes worldwide (WHO 2019a), current human exposure to tobacco smoke is extensive. Furthermore, this number does not include individuals who may be exposed currently to secondhand/sidestream smoke or those who were exposed ancestrally, either from parental or grandparental smoking going back several decades. Therefore, the vast scale of human exposure, including germ cell exposure during various potentially susceptible windows of development, supports the need for further assessment.
As summarized in Table 1, the workshop made substantial progress in assessing the heritable risk of tobacco product exposure to germ cells using the Clean Sheet framework. As was shown during the workshop, there is a wealth of qualitative information supporting the biological argument that tobacco smoke can alter the genome and epigenome of both nascent and adult germ cells. However, the available data for tobacco smoke exposure are not sufficient for quantitating the risk of adverse health effects in the offspring of smokers, although certain components of tobacco smoke may provide some useful information.
Based on the evidence presented and the data gaps identified above, the key information needed now is ( The Clean Sheet framework was employed as a tool to better articulate the problem and provide direction for future actions in the interest of improving public health.
During the workshop, a wide range of evidence was presented indicating that tobacco smoke can cause genomic and epigenomic alterations to germ cells, resulting in impaired fertility as well as negative health effects in subsequent generations. However, it was widely acknowledged that additional studies that demonstrate a dose response and provide more information on molecular mechanisms are necessary to aid in the development of an AOP and provide a more compelling argument for regulatory action. Therefore, future research should prioritize addressing these gaps through use of large cohort data, modern omics technologies, and targeted animal studies.
In addition to advancing the science, researchers should improve their communication of the science to the public. Raising awareness about concerns for heritable germ cell effects among those beyond the research community would help to both inform individual decision making and build a larger base of support for regulatory change.
Although the data are insufficient for quantitative risk purposes at the present time, based on the evidence presented, there are many other actions regulatory agencies can take to protect individuals from tobacco-related harm and subsequent health risks; see Codex Alimentarius (2015) for information on risk management options from risk assessment outcomes. An important action already taken is the labeling of tobacco products that warn of health risks, although the messaging specific to heritable effects should be more prominent.
Guidance to particularly susceptible populations to avoid exposure would also be extremely useful. Young people as a susceptible population were discussed at length during the workshop. Based on the biology of oogenesis and spermatogenesis, there are a multitude of opportunities for deleterious genetic and epigenetic effects on the germ cells of exposed humans. Exposures to young people, while their germ cells are still developing, can be particularly deleterious. Effective measures to minimize the risk posed to their genetic material are needed from regulatory agencies.
Although tobacco smoke was the primary focus of the workshop, the workshop acknowledged that there are emerging concerns about germ cell harms from exposures to related substances such as e-cigarettes, smokeless tobacco, and cannabis. Although the general health concerns surrounding cigarette smoking have been promoted widely and internalized by the public, these alternative products may be seen as relatively harmless, particularly among young people (Jaspers 2019).
Thus, as the public health ramifications of cigarette smoking are abating with reduced usage (WHO 2019b), a new wave of current and future health impacts may be rising with increased usage of these alternative products. In response, regulatory agencies should take a more proactive and comprehensive approach to mitigating the potential risk to avoid a new health crisis.
Regardless of the substance, hazards to germ cells and the progeny to which they give rise must be accounted for explicitly in chemical risk assessment and regulatory decision making. Insults to the genomic integrity of germ cells introduce a significant risk to the welfare and health of people who inherit these genomic changes. If we continue to ignore these risks, we may be jeopardizing irreversibly human health and wellbeing for generations.