Chemical Safety and the Exposome

Abstract

Air pollution and rapid chemical intensification are major threats to the environment and human health. Today, we have produced over 350,000 chemicals, and current testing strategies do not meet the demands. Therefore, it is important to develop new approach methodologies (NAMs) that can help fill current information gaps. Toxicology needs to evolve from hazard and risk assessments based on morphological endpoints in animal tests towards a mechanism-driven integrated approach that better includes computational modelling as well as molecular, human, and in vitro data. The application of new science and technology such as different types of imaging and omics methods can allow faster collection of high-quality toxicological data for hazard identification and better prediction of toxicological potential using advanced in silico approaches including machine learning. A shift toward active prevention of pollution through a safe and sustainable-by-design approach based on cutting-edge science could significantly help safeguard the population and planetary health. Moreover, it is necessary to improve the understanding of how interactions among chemical mixtures, climate change, infectious agents, and other stressors that constitute the exposome, may affect biota and human health. Individual responses to current exposures and susceptibility to disease are influenced by factors such as genetics, epigenetics, physiology, and health status, which involve changes in biological pathways caused by previous exposures or even ancestral exposures. It is therefore important to better consider individual exposomes and susceptibility in future risk assessments and precision medicine. This review describes the central role of mechanistic toxicology in chemical safety and in the cross-disciplinary efforts needed to characterize the exposome and its complex interactions in detail.

Introduction

Chemical pollution is one of the largest threats to the environment and human health. A well-known example is the emissions of carbon dioxide and other greenhouse gases and their adverse impact on the oceans, atmosphere, and climate¹. However, greenhouse gas emissions constitute only approximately 20% of the annual anthropogenic chemical release, which is estimated to be 220 billion tonnes^{2, 3}. The societal benefits of manmade chemicals are significant and their release into the environment is a direct consequence of economic development². The extensive burning of fossil fuels was a driver of the Industrial Revolution, and since then a large variety of chemicals have been released globally. This includes the anthropogenic dispersal of geogenic chemicals and the continuous production of a large variety of manmade chemicals. Over the recent decades, there has been massive chemical intensification in most sectors, including agriculture, pharmaceuticals, health, mining, construction, services, plastic and rubber production, and other industries². In 2017, approximately 2.3 billion tonnes of manmade chemicals were produced globally, doubling the amount from 2000⁵. Currently, we have produced over 350,000 chemicals, and while the safety of pharmaceutical drugs has been extensively tested before they enter the market, the risk assessment of industrial chemicals is often inadequate, making environmental and human safety uncertain^{1, 2, 4}.

Polychlorinated biphenyls (PCBs) were used in industry for over 30 years before Jensen, for the first time in the sixties, discovered that they could accumulate in nature⁶. This was an eye-opener, as it showed that in addition to chemicals such as pesticides that are directly used in the environment, chemicals are unintentionally released into it, and be a potential threat to biodiversity. Since then numerous reports have confirmed that the relationship between humanity and nature has been damaged^{1, 2, 7–10}. For example, using monitoring data from over 20,000 populations of more than 4,000 vertebrate species, the Living Planet Report from 2022 revealed an average 69% decrease in population sizes between 1970 and 2018 due to unsustainable human activities such as overexploitation, habitat destruction, environmental pollution, and climate change¹¹. In line with this, there are currently over 150,300 species that have been assessed for the IUCN Red List and more than 42,100 species are threatened with extinction, including 41% amphibians, 37% sharks and rays, 34% conifers, 36% reef building corals, 27% mammals, and 13% birds¹².

Furthermore, research shows that only a small proportion of human diseases are solely caused by genetic factors. The disease risk attributable to genetics ranged from 3% to 49%, with a median below 20% when evaluating a large number of chronic diseases in monozygotic twins¹³. Degenerative disease and cancer risk attributable to genetic factors is approximately 10%^14–16. It has been estimated that pollution caused nine million premature deaths globally in 2015, which is three times more than the combined mortality from AIDS, tuberculosis and malaria, and corresponds to 16% of all deaths¹⁷. Air pollution is estimated to cause 6.5 million of the pollution-related premature deaths, which is in line with the World Health Organization number of 7 million deaths per year caused by ambient (outdoor) and household air pollution¹⁷. A recent update of this estimate using data from 2019 found that pollution remains responsible for approximately nine million deaths per year¹⁸. However, it is likely that the actual contribution of pollution to the global disease burden is underestimated because the potential adverse effects of many chemical contaminants are unknown, involve complex interactions with genes and other environmental exposures that constitute the exposome, and can potentially be transmitted over generations (Figure 1). This review illustrates the central role of mechanistic toxicology in chemical safety and the cross-disciplinary efforts needed to characterize the exposome and its complex interactions in detail.

Figure 1. Gene-environment interactions underlie the majority of chronic human diseases The exposome can be defined as the cumulative lifetime environmental exposure and related biological responses of an individual.

The individual response to current exposure and susceptibility to disease is influenced by factors such as genetics, epigenetics, physiology, and health status, which involve changes in biological pathways caused by previous exposures, or even ancestral exposures. Currently, there are more than 350,000 manmade chemicals, and the potential adverse effects of most of these are still poorly understood. Mechanistic toxicology will therefore play a fundamental role in the large cross-disciplinary efforts needed to better understand how complex exposome interactions may affect the biota and human health.

Toxicology for Life

Because chemical contaminants can cause adverse effects on humans, biota, and ecosystems, toxicology will play an important role in the local, national, and global transdisciplinary collaborations needed for successful transformation to a safe and sustainable society. Toxicology can be defined as the study of how chemicals interact with living systems and affect biological processes and includes predictions of safe exposure levels. The Stockholm Convention on persistent organic pollutants (POPs) is an international attempt to regulate the release of toxic chemicals globally. To date, only 26 chemicals have been banned and nine more compounds are under review². At the current progress rates, the Stockholm Convention and other international, national, and regional initiatives will take more than 100,000 years to assess all currently existing manmade chemicals for human and environmental safety³⁷. Moreover, every year an additional 2000 years of assessment needs to be added to this estimate due to the massive annual production of new chemicals³⁷. While the industry is focusing on material properties to provide profitable products, regulatory authorities require reliable toxicological data to ensure safety. With the increasing amount and diversity of chemicals, current testing strategies do not meet the demands and fail to sufficiently protect humans and the environment^{2, 38}. Therefore, it is important to develop new approach methodologies (NAMs) that can help fill current information gaps. Higher throughput methods that allow rapid collection of high-quality toxicological data for hazard identification and prediction of toxicological potential should be implemented. Toxicology needs to develop from hazard and risk assessments based on morphological endpoints in animal tests towards a mechanism-driven integrated approach that also includes computational modelling as well as molecular, human, and in vitro data. The application of new science and technology, such as different types of imaging and omics methods, including epigenomics, transcriptomics, proteomics, metabolomics, and metagenomics, that permit the collection of novel high-content data directly from tissues, bulk, or single-cell samples, can accelerate the discovery of specific chemical targets and modes of action in a systems toxicology approach^{1, 39}. The adverse outcome pathway (AOP) framework is gaining popularity as a tool for describing the relationship between mechanistic data and toxicity endpoints required for regulatory assessments⁴⁰. An AOP can be used to structure current understanding of the chemical mechanism of action. This pathway begins with a molecular initiating event, such as the binding of a receptor, progresses through key events that are critical steps necessary for the toxicity to occur, and culminates in an adverse outcome, like diseases or developmental abnormalities. AOPs are useful for identifying areas where more research is necessary to comprehend the underlying mechanisms, aiding in characterizing chemical hazards, and directing the development of NAMs. Other scientific advances that have the potential to drive the development of toxicology include the generation of human-induced pluripotent stem cells, organoids, organ-on-a-chip, CRISPR/Cas9 gene editing, and artificial intelligence (AI).

Initiatives such as Tox21 and ToxCast have spurred the use of in vitro and computational toxicology models. The application of advanced in silico approaches, including AI, can aid in defining the modes of action and improving the ability to predict whether the effects caused by a chemical in experimental models can also occur in humans and animals. With the increasing amounts and types of data available, and new advanced computing technologies, machine learning will likely play a crucial role in toxicology^{41, 42}. Machine learning (subfield of AI), utilizes data and algorithms to continually improve predictions without being explicitly programmed. Its iterative nature allows the models to adjust on their own when they are exposed to new information. Through learning from past computations, machine learning models can produce consistent and reliable decisions and results. However, good data quality is an essential part of the overall data science and machine learning process. Low quality data can greatly affect the outcome and have negative consequences when decisions are based on those results. Mechanistic in vitro data can also be used to strengthen the links between chemical exposure and health outcomes found in epidemiological studies. For example, human data have revealed an association between per- and polyfluoroalkyl substances (PFAS), such as perfluoroctane sulfonate (PFOS), perfluorooctanoic acid (PFOA), and perfluorohexane sulfonate (PFHxS), and breast cancer. Our in vitro studies confirmed that these PFAS can transform normal human breast epithelial cells to a malignant phenotype via mechanisms that include epigenetic alterations, providing potential pathways involved in the association with breast cancer^43–45. Interestingly, very low concentrations (500 pM) of the binary PFOS and PFOA mixture were found to induce synergistic effects on breast cell proliferation⁴⁶. The same cell model was then used to assess the potential tumorigenic activity of several alternative PFAS, which had no effects and were clearly less harmful in our experimental model⁴⁷.

In addition to the benefits of improved possibilities to generate and use mechanistic information at higher throughput, which allows screening of more compounds, an increased use of NAMs is important considering the 3R concept. The principles of replacement, reduction, and refinement are vital for safeguarding animal welfare during research. However, the development and validation of NAMs for regulatory testing is a complex process. Before new methods can replace animals, they must be verified to provide equivalent or improved protection as the animal test that they would replace⁴⁸. Currently, several NAMs have been approved by regulatory authorities; however, most of these methods have not yet been widely implemented⁴⁸. Although scientifically valid NAMs are necessary to screen the large number of chemicals produced and contribute to improved animal welfare, the complete replacement of animals in toxicological research and testing remains a complex and challenging task. The processes in any organism involve a large number of intricate interactions and are more complicated than the sum of all parts. When a chemical enters the body, it can simultaneously interact with several target tissues, affecting numerous biological processes and physiological functions, which may cause unexpected consequences for other pathways. This complexity can in certain cases be difficult to fully study and predict using test approaches that do not include animal models at all. For example, compound toxicity depends on its kinetics and metabolism (absorption, distribution, metabolism, and excretion; ADME). This can include site-specific accumulation and/or bioactivation of the compound, which causes unpredicted toxicity in certain tissues or even particular cells. In addition to dose and ADME properties, the adverse effects of a chemical depend on the timing of exposure. During vulnerable periods in utero and in early infancy, exposure to chemical concentrations regarded as safe for adults may impair child development and result in disease later in life^{49, 50}. Epigenetic mechanisms such as histone modifications, DNA methylation, and regulation by non-coding RNAs are central pathways for the long-term effects of developmental exposure^{49, 51}. They may even mediate multi- and transgenerational effects of environmental exposure through epigenetic inheritance^52–54, which is very challenging to study in detail without any animal model. Our recent study demonstrated that male frogs (F0) exposed to environmental concentrations of the pesticide linuron fathered male offspring (F1) with reduced body size, altered spermatogenesis, decreased fertility, and other signs of endocrine system disruption. The impacts were further propagated to the F2 generation, providing evidence of transgenerational effects in amphibians. F2 males demonstrated a decreased number of germ cell nests, increased weight and fat body palmitoleic-to-palmitic acid ratio, and decreased plasma glucose levels. The findings provide important cross-species evidence of paternal epigenetic inheritance and pollutant-induced transgenerational toxicity, supporting a causal and complex role of environmental contamination in ongoing amphibian extinction. This also exemplifies that animal research not only is important for human safety but also for the benefit of animals⁵⁵.

Even if animal experiments may be difficult to fully replace in the near future, continued investments in the development of NAMs are essential. The first stage of toxicity testing should be performed in vitro and in silico. Better incorporation of mechanistic knowledge gained from NAMs will be important for chemical safety as well as reduce the number of animals used for toxicological testing and generate biomarkers that contribute to reduced distress in animals. The discovery and application of new science and technology to develop and validate NAMs and earlier humane endpoints will further refine, reduce, and replace animal use while ensuring and advancing the health of people, animals, and the environment.

Combining scientific development and knowledge within toxicology and other related research fields, with far-sighted political leadership and carefully defined environmental policies, will be central to driving the changes needed to better protect environmental and human health. The European Commission’s Chemicals Strategy for Sustainability (CSS), released in 2020, is an ambitious political action plan to restructure the chemical regulatory framework in the EU. Safe and sustainable-by-design is an important concept in this initiative, which could become one of the pillars of chemical regulation in the EU. Based on cutting-edge science, this holistic approach aims to promote the development of chemicals that are safe for humans and the environment from production throughout the lifecycle, offer an improved environmental footprint, and deliver the expected performance and societal value. This requires close collaborations between toxicologists and chemists, and innovative green chemistry approaches.

The Exposome: Multiple Exposures and Individual Susceptibility

Another challenge in toxicology is the interaction of chemicals in mixtures, which can occur via several mechanisms. Chemicals are almost always considered individually in current risk assessments, possibly leading to a systematic underestimation of risks. Toxicological research must begin to fully assess the combinatorial effects of chemicals in mixtures. By examining the toxicological impact of complex real-world chemical mixtures instead of one chemical at a time, it may be possible to more accurately identify harmful combinations and underlying mechanisms. This is critical because the combined effect of chemicals in mixtures may be greater or lesser than the effect of individual chemicals on biological processes^{46, 56, 57}.

Moreover, the adverse effects of chemicals also depend on their complex interactions with genes and other environmental factors. The exposome can be defined as the cumulative lifetime environmental exposure and related biological responses of an individual (Figure 1)²⁷. In the exposome concept, all non-genetic factors contributing to disease are considered environmental, including for example air pollutants, chemicals, drugs, radiation, microbiota, climate, and psychosocial stress^28–30. It is therefore central to study how interactions among chemical exposures and other environmental factors that make up the exposome, may affect our health. For example, the importance of the gut microbiome in human health has become increasingly evident. Pharmaceuticals and other chemicals have also been shown to affect the gut microbiome and vice versa; however, the complex relationship and its possible role in the etiology of disease and response to medical treatment warrant further research^{58, 59}.

Moreover, closely linked factors, such as diet and environmental contaminant exposure, can strongly influence gut flora^{60, 61}. People immigrating from Thailand to the US demonstrated a clear westernization of their gut microbiome and an increased risk of obesity, illustrating how different parts of the exposome may interact and affect health⁶⁰. Pathogens and environmental pollution can also interact and worsen adverse health effects. For example, people living in regions with high air pollution are more likely to have compromised respiratory and cardiac systems and may therefore be more vulnerable to infectious agents targeting the respiratory system^{62,63, 64}.

While it is generally accepted that human activities, including chemical pollution, drive the severe ongoing decline of biodiversity, the close link between the health of our planet and human health is less often considered^21–23. For example, biodiversity loss alters ecosystem functions and reduces the ability to provide society with necessary goods and services, such as food^21–23. Furthermore, one consequence of climate change that was recently highlighted in a review is its potential to alter the distribution and toxicological effects of environmental pollutants¹. Global warming together with eutrophication of aquatic and terrestrial environments by widespread agricultural use of fertilizers and manure, can also contribute to ecosystem dysfunction and an increase in cyanobacterial and algal blooms, which are a threat to human and animal health as they can produce a wide range of potent toxins^{9, 10, 33–36}.

Efforts to generate systematic information on the human exposome and its complex interactions would significantly advance our understanding of human diseases^{1, 75}. First, exposure assessments need to be further improved. Traditionally, environmental exposure is measured through questionnaires, geographical information, or targeted analytical methods⁷⁶. The development of high-resolution mass spectrometry and approaches that combine targeted chemical analysis with suspect screening and nontarget analysis can lead to better characterization of current chemical exposures^{28, 29, 77}. Methods for collecting representative information on air pollution exposure have also been developed. While the nearest air pollution monitor may provide a suboptimal estimate of the actual exposure, portable air pollution sensors, dispersion modelling, and land use regression modelling may offer a better and more personalized estimate⁷⁸. It is also important to leverage new omics methods and other technological developments to characterize the composition and effects of the exposome. Epidemiologists and biostatisticians will play an important role in this challenging task, as further implementation and development of suitable data analysis methods is necessary⁷⁹. The increased access to high-throughput technologies has led to a massive growth in omics data, which aim to depict various biological aspects such as genomics, epigenomics, transcriptomics, proteomics, metabolomics and metagenomics. Machine learning techniques have been applied to these data to develop diagnostic and classification biomarkers. However, these biomarkers are often limited as they only utilize data from a single omics measurement, not fully utilizing the comprehensive nature of recent multi-omics studies⁸⁰. To make the most of these multi-omics experiments, a strategy that integrates data from all omics layers is required^80–82. The field of environmental epidemiology has also begun to move towards using an exposome approach that include multiple environmental exposures. Various statistical methods have been proposed to evaluate the individual and combined effects of chemical combinations and other environmental factors^{79, 83–87}. Studies that have compared various machine learning approaches for examining the impact of the exposome suggest that there is no one-size-fits-all model^88–90. It is therefore recommended to explore various machine learning techniques, such as and random forest models and artificial neural networks (deep learning), to characterize the impact of the exposome in the best possible way^{82, 88, 91–96}.

Mechanistic toxicology will play a central role in the large cross-disciplinary efforts needed to better understand complex exposome interactions and their impact on biota and human health. Human health risk assessment aims to protect broad populations, which may include a wide range of susceptibility to the adverse effects of chemical exposure. The exposome can alter important biological processes through epigenetic mechanisms such as DNA methylation, histone modifications, and regulation by non-coding RNAs that alter the expression of key genes^97–100. Individual responses to current exposures and susceptibility to disease are influenced by genetics, epigenetics, physiology, and health status, which involve changes in biological pathways caused by previous exposures and potentially parental or even grandparental exposures via epigenetic inheritance (Figure 1). People living in socially or economically disadvantaged areas or in developing countries where pollution is generally high and health status is low may therefore be particularly sensitive to chemical exposure. Life stage is also an important factor; both the elderly and young can be highly vulnerable to chemical exposure. Chemical exposure in utero or in early infancy may adversely affect vital developmental processes at concentrations regarded safe for adults^{49, 50}. Currently, uncertainty regarding human variability and susceptibility is considered by applying the conventional safety assessment factors. However, more research is needed to better understand gene-environment interactions, characterize the exposome and its complex interaction at the individual level, and ultimately include this information in the assessment of population-level variability and susceptibility. NAMs can also provide the possibility to experimentally investigate human population variability and susceptibility factors by, for example, using human cell models derived from diverse populations or in silico approaches integrating information from different sources and subpopulations. Chemical exposures are likely to be underestimated contributors to human disease, yet gene-environment interaction studies rarely assess their modifying effects. We recently published a computational approach to profile the human genome and identify the genetic regions associated with differential population susceptibility to chemical exposures⁷⁵. By integrating high-throughput screening data from sources such as the ToxCast program, and literature-curated chemical-gene data with gene-disease data, this method predicts high-risk genetic regions for differential population susceptibility to chemical exposures and characterize chemical modifying factors in specific diseases⁷⁵. Another recent study has used an integrative approach, combining an in vitro CRISPR screen with molecular epidemiology to discover individual susceptibility to adverse effects of environmental contaminants¹⁰¹.

Precision medicine, also called personalized medicine, is an emerging approach for disease prevention and treatment that considers differences in the genes, environments, and lifestyles of people. The role of precision medicine in daily healthcare is still relatively limited and highly focused on genetics. It is important to further expand this approach to more areas of health and healthcare and make efforts to better include individual environments and lifestyles by integrating the exposome concept. For example, although air pollution is a global public health concern, there is a lack of individual-level preventative options. In a controlled human exposure study led by Baccarelli, we demonstrated the epigenetic effects of concentrated ambient PM2.5 (fine particles with a diameter of 2.5 micrometers or less) on peripheral T helper cells, the most prolific cytokine producer mediating PM2.5-induced inflammatory responses. Interestingly, B-vitamin supplementation prevented these changes¹⁰⁰. This placebo-controlled intervention trial indicated that individual-level prevention may be used to complement regulations and control potential mechanistic pathways underlying the adverse effects of PM2.5, with possible significant public health benefits in areas with frequent PM2.5 peaks. Because of the central role of epigenetic modifications in mediating environmental effects, these findings may be extended to other toxicants, and illustrate the importance of understanding the adverse effects of environmental exposures on a molecular level.

Concluding Remarks and Recommendations

Scientific evidence shows that we live in a time, often referred to as the Anthropocene, when unsustainable human activity shapes the biosphere. This includes greenhouse gas emissions, which are the main cause of global climate change. In addition, air pollution and the rapid chemical intensification, pose a direct threat to biota and human health. However, risk assessment of industrial chemicals and anthropogenic emissions is often inadequate, and the adverse effects of many environmental contaminants are poorly understood.

Combining scientific knowledge and development within toxicology and other related research fields, with far-sighted political leadership and carefully defined environmental policies that include stricter requirements on the industry to conduct safety evaluations of their products, will be central to driving the changes needed to better protect environmental and human health. This strategy includes taking advantage of modern technology and methods to continue the development of NAMs. Modernizing toxicology towards a mechanism-driven integrated approach will allow faster collection of high-quality toxicological data for hazard identification and better prediction of toxicological potential. To further improve chemical safety, it is also important that we actually include the mechanistic knowledge gained from NAMs into chemical regulation and management work.

Mechanistic toxicology will play a fundamental role in the large cross-disciplinary efforts needed to better understand how complex exposome interactions may affect the biota and human health. Gene-environment interactions underlie the majority of chronic human diseases, and a relatively small proportion of chronic human diseases is caused solely by genetic factors. It is therefore essential to conduct more research and develop better tools to determine environmental contributions to diseases. The individual response to current exposure and susceptibility to disease is influenced by genetics, epigenetics, physiology, and health status, which involve changes in biological pathways caused by previous exposure or even ancestral exposures. In the future, it is therefore important to better consider individual genomes, exposomes, and susceptibility in risk assessment and precision medicine. Improved understanding of the molecular mechanisms underlying the health effects of contaminants will aid risk assessment as well as help develop individual-level preventative options as a complement to regulations.

To solve the complex challenges of humanity in the Anthropocene, we must understand that global work and commitment are needed. Given the major role of economic activity in ongoing anthropogenic changes, economics must be included in the core of sustainable development and vice versa^102–104. The concept of a circular economy aims to achieve equilibrium between economic growth, the preservation of resources, and the safeguarding of the environment¹⁰³. Currently, the economic system largely builds on linear principles focused on throughput, optimization, and cost-benefit analysis. Therefore, global sourcing has been a major industrial practice for multinational companies during the past decades^{105, 106}. This is an established strategy for lowering costs and improving competitiveness by sourcing materials/products from countries with lower labor and production costs¹⁰⁷. Companies sourcing globally are also accused of not handling social and environmental concerns sufficiently, including disregard of environmental regulations and working conditions^{106, 108}. This includes industries that do not use any type of wastewater treatment before discharge into rivers or appropriate safety equipment for their employees. However, because of increased consumer pressure, some large multinational companies have started to integrate social and environmental sustainability into their overall strategies¹⁰⁶. This reflects a growing perception that a company is not more sustainable than its supply chain¹⁰⁵. Hence, increased understanding and awareness of chemical contamination could contribute to enforcement or changes in existing environmental regulations, influence global and national companies to increase their sustainability commitments, and minimize environmental contamination in developing countries. It is necessary to consider environmental aspects and ensure that they remain within defined planetary boundaries for sustainable global economic development. This requires competent political leadership and ambitious environmental policies. The vision that EU policymakers have in the CSS will have significant consequences for the industry. Based on cutting-edge science, the concept of safe and sustainable-by-design aims to promote the development of chemicals that are safe for humans and the environment from production throughout the lifecycle, offer an improved environmental footprint, and deliver the expected performance and value to stakeholders throughout the value chain. If this ambitious political initiative to restructure the chemical regulatory framework in the EU becomes a benchmark worldwide, it will significantly help safeguard the population and planetary health, and modern toxicology will play a key role in this work.