Nina Notman explores the challenge of assessing and managing risk from the coincidental chemical mixtures to which humans and the environment are exposed

The natural environment – our air, water and soil – hosts an ever-changing cocktail of synthetic compounds originating from a wide variety of sources. Wildlife is constantly exposed to these coincidental chemical mixtures, as are humans. We are also continuously exposed to a concoction of chemicals in our indoor environments too – from chemicals leaching out of carpets, furnishings and household objects, to those coming from foods, as well as cleaning and personal care products.

This has been known for a long time. It’s also well established that exposure to chemical mixtures can cause harm – to the environment and human health – even if all the individual chemical components in the mixture are present at concentrations that would be safe if they were met alone.

This isn’t a huge surprise – mixture effects, as they’re called, are a well-known issue in pharmaceuticals. The sedative effects of some painkillers and antihistamines, for example, add together to create a larger sedative effect that each one alone. And some antibiotics render the contraceptive pill less effective.

But while the medical and pharmaceutical industries are expected to be very aware of potential issues with drug–drug interactions in patients, the chemical industry is largely not. The majority of regulative frameworks set safe exposure levels for chemicals with the assumption that the chemicals are released into a pristine location (in either the body or the environment) where no other chemicals are present. But this is not the case. Regulations may well cover intentional mixtures of chemicals, such as those in detergent or paint formulations, but not those that occur coincidentally after their intended use.

‘When you look only at a single chemical and ignore the risks associated from a mixture composed of chemicals that also produce the same effect, then clearly you are heading towards a situation where you drastically underestimate the real, existing, practically relevant, combined risk,’ explains Andreas Kortenkamp, a professor of human toxicology at Brunel University in London, UK.

We know that chemicals can have combined effects and that regulating single chemicals is not enough

‘For a long time, there was an assumption that if [concentrations of] individual chemicals were below their set [safe exposure] thresholds, then that would also protect us from mixture effects,’ says Olwenn Martin, a colleague of Kortenkamp at Brunel. ‘There is sufficient evidence now that this is not necessarily the case. We now know that chemicals co-occur, that they can have combined effects and that regulating single chemicals is not necessarily protective for health or the environment,’ she adds.

In October 2020, the European Commission (EC) proposed taking a major step forwards in addressing the risks posed by coincidental chemical mixtures in the EU, through amendments to its chemical regulation framework Reach (Registration, Evaluation, Authorisation and Restriction of Chemicals).

Potential change ahead

The Reach framework puts the obligation to examine and manage the safety of chemicals onto those at the top of the supply chain. But asking chemical manufacturers (or importers) to consider the risks posed by any coincidental mixtures their products may form out in the wild throws up some major questions. How can they predict exactly where their chemicals will wind up after their intended use and how can they know what other chemicals will be present in the same location at the same time? ‘The number of substances out there is enormous and the number of possible mixtures is endless,’ explains Urban Boije af Gennäs, a chemical strategy expert at the Swedish Chemicals Agency located near Stockholm.

The mixture assessment factor is a pragmatic way of tackling this problem

Modelling is one option, although its utility is currently hindered by a lack of data. The EC proposal takes a simpler approach: to reduce the acceptable exposure thresholds for all chemicals. If implemented, a so-called mixture assessment factor will be added into exposure threshold calculations for the registered chemicals in the Reach database. ‘An extra safety margin would be added to the single substance risk assessment, through this mixture assessment factor, to take account of the fact that all substances will always appear in reality together with other substances,’ says Boije af Gennäs, who was involved in developing this proposal while on secondment to the EC.

The risks posed by coincidental mixtures will be accommodated by reducing ‘the overall pressure on the environment and human health, because every chemical will have lower exposure levels than it would otherwise potentially have’, explains Jack de Bruijn, director of prioritisation and integration at the European Chemicals Agency in Helsinki, Finland. ‘The mixture assessment factor is a pragmatic way of tackling this problem,’ he adds.

This proposal is one of a number of changes to Reach under consideration that were outlined in the EU Chemicals Strategy for Sustainability published in October 2020. ‘The EC are working towards having a proposal to change the regulation by the end of next year [2022],’ says de Bruijn. If mixture assessment factors are adopted, decisions ahead include the extent to which safe exposure levels for chemicals will need to be reduced (the size of the mixture assessment factor) and how and when it will be implemented.

Human biomonitoring

The sweeping mixture assessment factor is being heralded as a potential major step forward in addressing the safety of coincidental chemical mixtures through Reach, but this isn’t the first time concerns about these types of mixtures have been accommodated in this regulatory framework. The safe exposure levels of a handful of chemical groups have already been reduced for this reason.

One important barrier to being able to deal with mixtures is the availability of data

In 2019, for example, acceptable exposure thresholds were lowered for four phthalates: butyl benzyl phthalate, di(2-ethylhexyl) phthalate, dibutyl phthalate or diisobutyl phthalate. These plasticisers are found in products ranging from floorings, mattresses and footwear. Exposure to them is known to impact male sexual development and can lead to infertility.

Action was taken after an assessment of human biomonitoring data found additive biological effects from coincidental mixtures of these phthalates. Biomonitoring is a popular tool for scientists working to identify risks from coincidental chemical mixtures: the levels of hundreds of toxic compounds, elements and their metabolites are monitored in volunteers’ bodies, by analysing their blood, urine and hair. These tests identify what chemicals the volunteers are each exposed to over time. But to grasp exposure levels on a population scale, huge numbers of people need to be monitored. ‘One important barrier to being able to deal with mixtures is the availability of data,’ explains Martin.

A few countries have long had human biomonitoring programmes in place to supply this information. The US, for example, has run programmes since the late 1960s. The current incarnation is the National Biomonitoring Program run by the Centers for Disease Control and Prevention. Germany also started a programme in the mid-1980s and, since 2017, it has co-ordinated a human biomonitoring effort involving 30 European countries: the European Human Biomonitoring Initiative or HBM4EU.

Martin and Kortenkamp are both participating in the chemical mixtures arm of HBM4EU. This programme is providing ‘more data on what people are actually exposed to, what we can measure in their blood, urine, hair etc, to [give us] a better handle of what the mixtures of concern are’, Martin says.

Analytical tools

An alternative way to assess the chemical mixtures to which humans and the environment are being exposed is to identify all the chemicals present in the particular locations where exposure is suspected. Juliane Hollender and her team at Eawag – the Swiss Federal Institute of Aquatic Science and Technology in Dübendorf – hunt for organic contaminants in water samples. ‘What we are interested in is getting a picture of the whole [organic] chemical space present in water,’ says Hollender.

The team’s hunting grounds include surface water, raw and treated wastewater, ground water and sediment. Samples are analysed using liquid chromatography coupled to high-resolution mass spectrometry. ‘With this technique we can detect, very sensitively and selectively, a broad range of compounds and get a picture on their concentrations,’ Hollender explains, and by linking the concentrations to published biological effects ‘we can also calculate the risk’.

Trying to match tens of thousands of peaks in a spectrum to these hundreds of millions of chemicals is a huge data challenge

‘When using high resolution mass spectrometry, we’re usually confronted with thousands to tens-of-thousands of peaks in a typical sample [containing a coincidental chemical mixture] and, generally, the more complex the samples, the more peaks we get,’ explains Emma Schymanski, an environmental cheminformatics expert at the University of Luxembourg. To identify the chemicals present, the first step is to match as many peaks as possible to those of suspected organic contaminants, whose mass spectral ‘fingerprints’ are held in specialised databases and libraries. The database Hollender uses, for example, contains around 2000 organic compounds.

Unidentified peaks in spectra, that don’t match with compounds in specialised databases, can then be identified by matching them to information in much larger chemical databases. This can be like searching for a needle in a haystack, however: the CAS Registry contains data for more than 180 million chemicals and PubChem over 110 million. ‘Trying to match tens of thousands of peaks in a spectrum to these hundreds of millions of chemicals is a huge data challenge,’ says Schymanski. ‘This is where the cheminformatics comes in; we can help people find the chemicals that are relevant for them without looking at the other 109,999,999 chemicals that are not so relevant.’

Chemical identifiers were never designed to deal with mixtures

Data scientists, such as Schymanski, are working to improve the searchability of these huge databases. Their efforts include converting chemical data into formats that Google can index. It’s not possible to just post a picture of a spectra or chemical structure into Google, explains Schymanski. ‘We do a lot of formatting that makes the searching easier.’ For structures, for example, InChIs (international chemical identifiers) are used. These textual identifiers contain layers of information including atoms present, how they bond together, and isomer, isotope and stereochemical information.

But while concepts such as InChls have revolutionised the searchability of huge databases, ‘they were never designed to deal with mixtures’, admits Schymanski. ‘One of the challenges we have now is trying to upscale this single chemical approach to work when there are going to be lots of chemicals occurring at the same time.’ Approaches aimed at mixtures include using standardised keywords to allow for easy grouping of similar chemicals, so you only search the databases for endocrine disruptors if you are looking for an endocrine disruptor, for example, or to allow you to eliminate chemicals from a search that are probably not going to be produced in the environment if you are looking at mass spectra from chemicals found in the environment.

Norman network databases, that Schymanski helps co-ordinate, hold information about emerging substances of concern found in the environment, and are trailblazers in addressing this grouping approach. Data in the Norman databases is uploaded by a network of over 80 research laboratories from 20 countries, and Schymanski and her collaborators are currently annotating these to make the databases easier to search. ‘This metadata is really important for finding the chemicals again,’ she explains. The data and associated metadata are also being integrated into larger databases such as PubChem.

Bioanalytical tool

Beate Escher and her team of environmental toxicologists at Helmholtz Centre for Environmental Research UFZ in Leipzig, Germany, use a different approach for assessing risks posed by coincidental chemical mixtures – looking at the biological effects they trigger in cell-based bioassays. ‘We use cell lines, modified to enhance one specific mode of action, to probe [the biological effect of coincidental] mixtures,’ she says.

Bioassay results complement those obtained using mass spectrometry. ‘Chemical analysis is never comprehensive – you can never analyse all the peaks from a chromatograph. Bioassays let you quantify the [biological] effect of the entire mixture, but you do not know which chemicals have caused it. But by throwing the ball back and forth, [and using some single chemical in vitro bioassay data in databases like ToxCast and Tox21], we can make connections between the chemical analysis and the bioassays,’ says Escher. ToxCast contains toxicity data for around 4500 chemicals while Tox21 holds data for around 10,000. Cheminformatics is helping to make these databases more easily searchable too.

Escher and Hollender collaborate on back-and-forth investigations with water samples. Using analytical data and published biological effect data ‘we can calculate what kind of effect we would expect for a certain assay and then we can compare this with the outcome of the actual assay’, explains Hollender. If the observed biological effect is different to the predicted one, Hollender’s analytical chemistry team know to go hunting for other chemicals in the water sample.

The use of this back-and-forth approach to improve understanding of the composition and risk from chemical mixtures is ‘really dramatically accelerating’, says Escher. It is now possible to identify up to 1000 chemicals in complex mixtures and to pinpoint which of these are causing the biological effect, she adds. Her team is currently developing sample preparation approaches to enable the same bioassay approach developed for water samples to be used on sediment and human and animal blood and tissues. It’s really tricky to not extract anything natural and bioactive during the sample clean-up but ‘we are getting better and better’ at this, Escher explains.

Dust buster

Heather Stapleton, an environmental chemist at Duke University in the US state of North Carolina is concentrating her efforts on house dust. Young children are exposed to dust more often than adults, due to crawling and playing on the floor, where dust tends to accumulate. Exposure risks from dust were first identified by scientists looking at lead exposure. ‘When lead used to be used in certain paints, it would accumulate in dust and become a significant exposure pathway for children which, unfortunately, was then associated with neurodevelopmental deficits in children,’ says Stapleton.

It’s really surprising just how many chemicals are commonly found in dust

Her lab began research in the area by focusing on brominated flame retardant accumulations in dust. ‘We visit people’s homes, vacuum their house and bring the dust back to the laboratory for analysis,’ says Stapleton. It was during these mass spectrometry analyses that she started realising the scale of the challenge. ‘It’s really surprising just how many chemicals are commonly found in dust,’ she says.

Many of these chemicals are endocrine disrupters that impact thyroid hormone regulation, and Stapleton’s group is examining the way mixtures in dust impact this system. The team has started testing dust samples in cell-based bioassays. ‘We often see thyroid disrupting activity in dust samples at a concentration that I would consider to be relevant for human exposure. As opposed to the very high concentrations needed to elicit a response when you test an individual chemical,’ she says. The team is now starting to combine the mass spectrometry and cell-based bioassay findings in an attempt to identify the active players in the chemical mixtures present in dust.

The overarching purpose of the scientific analysis into the composition and risks posed by coincidental chemical mixtures to humans and the environment, is to drive chemical policy to better protect against the harm from coincidental chemical mixtures. Today, the regulation of coincidental chemical mixtures is extremely piecemeal. But in addition to the mixture assessment factors, the EC – in the same chemical strategy document – has indicated that it will introduce or reinforce provisions to take account of combination effects in other legislation, such as those for water quality, food additives, cosmetics and toys.

There are signs of movement outside the EU too. The UK, for example, hasn’t ruled out allowing for risks posed for coincidental mixtures in its newly launched chemical regulation framework UK Reach. ‘Together with other regulators [we] are undertaking work to further our scientific understanding and develop our management approach. This will help us assess the potential risks from chemical mixtures to inform future policy and regulatory approaches,’ says a spokesperson for the Department for Environment, Food and Rural Affairs in response to Chemistry World’s questions.

With the UK Reach only recently launched, any additional regulations on mixtures will probably be a few years away. ‘The EU is paving the way for the rest of the world in terms of how they’re going to approach this,’ concludes Martin.

Nina Notman is a science writer based in Salisbury, UK