The Ecotoxicology of Chemical Mixtures: Comprehensive Study Notes

Lecturer Background and Objectives

  • Speaker Profile: Mikael Gustavsson

    • Current focus: Modelling of chemical hazard, exposure, and risk.

    • Specific expertise:

      • AI-based predictions of chemical hazards.

      • Research spanning from consumer use to environmental concentrations.

      • Mixture risk assessment.

      • Science and policy implementation.

      • Regulatory ecotoxicology.

      • Data analysis and chemoinformatics.

    • Areas for improvement: Information evaluation for chemical substitution and biodegradation.

    • Personal interests: Board games and lifting heavy things (strength training).

  • Lecture Goals

    • Attain a deeper understanding of mixture toxicity from conceptual, practical, and regulatory perspectives.

    • Method: Utilization of critical questioning to challenge simplistic assumptions and make complex topics seemingly difficult again.

  • Lecture Structure

    • Part 1: Relevance of mixtures and empirical data (proving all exposures are mixtures).

    • Part 2: Fundamental strategies for assessing mixtures, focusing on effect estimation methods like Concentration Addition (CA) and Independent Action (IA).

The Ubiquity of Chemical Mixtures

  • The Fallacy of Single Substance Exposure

    • Technically, a pure single-substance exposure does not exist.

    • Example: Pure Standards from Sigma-Aldrich

      • Acetone (purportedly pure): 99.8%99.8\% purity, with a maximum of 0.0005%0.0005\% non-volatile matter.

      • Water: maximum of 0.0005%0.0005\% non-volatile matter.

      • The remaining percentage represents a mixture of impurities.

  • Petroleum Products

    • Crude oil supply undergoes separation (e.g., normal boiling point ranges from 30 ∘C30\,^{\circ}C to over 700 ∘C700\,^{\circ}C) to create products like gasoline, kerosene, diesel, lubricating oil, asphalt, jet fuel, and home heating oil.

    • Complexity is visualized through 2D separation (polarity/functionality vs. volatility/carbon number), showing hundreds of overlapping compounds (Paraffins, Naphthenes, Mono-Aromatics, etc.).

  • UVCBs and Polymers

    • UVCB Definition: Unknown or Variable composition, Complex reaction products, or Biological materials.

    • Globally, over 70,00070,000 substances are registered as polymers or UVCBs with ambiguous identities.

    • Under EU REACH regulation, approximately 3,7503,750 (15%15\%) of all registered substances are UVCBs.

Documented Human and environmental Exposure

  • Biocide Products (Antifouling Paints)

    • A 2006 study by I. Konstantinou analyzed 380 products:

      • 2121 products contained no biocides.

      • 6060 products contained one biocide.

      • 200200 products contained a two-biocide combination.

      • 8080 products contained a three-biocide combination.

      • 1919 products contained a four-biocide combination.

    • Common active biocides include Cu2OCu_2O, CUPTCUPT, ZnPTZnPT, TPBPTPBP, Diuron, CuSCNCuSCN, and Zineb.

  • Human Consumer Exposure: Cosmetics

    • Analysis of 8,5008,500 products revealed 318,000318,000 listed ingredients.

    • There are 2,2792,279 unique ingredients circulating in the personal care market.

  • Indoor Environments (Silicone Wristband Study)

    • A study of 243 office workers (USA, UK, China, India) used silicone wristband samplers.

    • Findings: Every participant was exposed to hormonally bioactive mixtures that mimic or block sex or thyroid hormones in human cells.

    • Caveat: Be mindful of over-interpreting; the study lacked a "living in a forest" control group.

  • Maternal and Newborn Exposure

    • Suspect Screening Study (Wang et al., 2021): Screened 3,5003,500 industrial chemicals in maternal and cord serum samples (n=60n = 60).

    • Matched 662662 suspect features in positive ionization mode and 788788 in negative mode (557557 unique formulas).

    • BodyBurden 2005 Study ("The Pollution in Newborns"): Tests showed 287287 industrial chemicals/pollutants in umbilical cord blood across 1010 babies, including banned industrial chemicals, consumer ingredients, and waste byproducts.

  • Environmental Monitoring: The Danube and Swedish Waters

    • Joint Danube Survey (2023): Sampled amounts reached up to 500 μg/sampler500\,\mu g/sampler. Quantified compounds included industrial chemicals, pesticides, and pharmaceuticals/personal care products (PPCPs).

    • Swedish Pesticide Monitoring (2002-2022): Based on 2,3802,380 samples from 6 locations (Skivarpsn, Vege, etc.).

    • Swedish Coastal Waters (Gustavsson et al., 2017): Of 172172 compounds analyzed, 6262 were detected. The number of detects per sample ranged from 3030 (Lerkil) to 4141 (Fiskebckskil).

Fundamental Concepts in Mixture Toxicity

  • Core Findings

    • Mixture effects frequently exceed the effects of individual substances.

    • Concentrations deemed "safe" individually can still lead to measurable ecotoxicological effects when combined.

  • Concentration Addition (CA)

    • Assumption: Similar Pharmacology/Toxicology; similar mode/mechanism of action (MoA).

    • Chemicals differ only by their potency.

    • Toxic Units (TU): Expresses concentration as a fraction of the effect concentration (ECxECx).

    • Formula for a single toxic unit: TU=CEC50TU = \frac{C}{EC50}

    • Formula for the Sum of Toxic Units (STUSTU): TUMix=∑i=1nTUiTU_{Mix} = \sum_{i=1}^n TU_i

    • A mixture with STU=1STU = 1 is predicted to cause a 50%50\% effect.

    • Formula for CA: 1=∑i=1nciECxi1 = \sum_{i=1}^n \frac{c_i}{ECx_i}

  • Independent Action (IA)

    • Assumption: Dissimilar mode/mechanism of action. Chemicals act on different biological targets but affect the same broad endpoint (e.g., organism death).

    • Toxicity of one component is not influenced by the presence of others.

    • Formula for IA: E(cmix)=1−∏i=1n(1−E(ci))E(c_{mix}) = 1 - \prod_{i=1}^n (1 - E(c_i))

    • Example (Binary): If compound 1 kills 10%10\% and compound 2 kills 10%10\%, the mixture survivability is 0.9×0.9=0.810.9 \times 0.9 = 0.81 (total effect = 19%19\%, not 20%20\%).

  • Comparison of Models

    • Differences between CA and IA are often small in environmentally occurring mixtures.

    • CA is generally more conservative (predicts higher toxicity than IA).

    • CA usually performs within a factor of two compared to empirical observations.

Deviations from Additivity: Synergy and Antagonism

  • Definitions

    • Synergism: The mixture effect is greater than predicted by additivity (CA or IA).

    • Antagonism: The mixture effect is less than predicted by additivity.

  • Significance in Environment

    • Synergistic interactions generally require high chemical concentrations, often higher than those typically found in the environment.

    • Additive effects of many co-occurring pollutants usually carry a larger hazard than a few synergists.

    • Case Study (Laetz et al., 2009): Combined exposure of Organophosphates (OP) and Carbamates (CB) showed synergism in inhibiting Acetylcholinesterase (AChE) activity.

    • Case Study (PAH + CLO): Benz[a]pyrene (BaPBaP) + Clotrimazole (CLOCLO) can trigger synergistic CYP1A activity responses.

  • The Funnel Hypothesis (Warne, 1995)

    • As the number of components in a mixture increases, the likelihood of synergistic or antagonistic effects dominating decreases, and the mixture tends toward additivity.

Advanced Analytical and Bioassay Methods

  • Problems with Traditional Approaches

    • We often fail to detect compounds below the Limit of Detection (LOD).

    • Many chemicals present in environmental samples are not even included in the measurement targeted list.

  • Chemical Screening Techniques

    • Target Screening: Known analytes, reference substances available, quantitative data, high sensitivity.

    • Suspect Screening: Large list of potential analytes, reference substances not available, semi-quantitative data.

    • Non-Target Screening: Generic goal (identifying anything present), complex, often non-conclusive.

  • Effect-Based Methods (EBM) / Bioassays

    • Approach: Test the mixture itself using biological systems (cells, organisms) to capture unknown compounds and mixture effects.

    • Iceberg Modeling: Comparing expected effects (from detected compounds) vs. measured effects (from bioassays).

    • Often, detected chemicals explain only a tiny fraction of the effect (e.g., in a bacteria study, 269269 chemicals explained only 0.1%0.1\% of the effect).

  • Effect-Directed Analysis (EDA)

    • A "forensic" approach combining fractionation (e.g., via Solid Phase Extraction) with bioassays to identify the specific chemicals responsible for an observed effect.

    • Resource-intensive and carries the risk of "fractionating away" the effect if required components are separated.

Regulatory implementation and Risk Mitigation

  • Mixture Assessment Factor (MAF)

    • A proposed safety factor applied to individual Predicted No Effect Concentrations (PNECs).

    • PNECmixture−adjusted=min(EC50Algae,EC50Crustaceans,EC50Fish)AF×MAFPNEC_{mixture-adjusted} = \frac{min(EC50_{Algae}, EC50_{Crustaceans}, EC50_{Fish})}{AF \times MAF}

    • Data from Swedish pesticide monitoring (1,513 samples) suggests a MAF of 10 would cover 95%95\% of samples where the sum of Risk Quotients (RQSumRQ_{Sum}) exceeds 11.

  • Recent EU Legislation

    • Urban Wastewater Directive (EU 2024/3019): Encourages identification of risks through broad chemical screening and/or biological EBM.

    • Council of the EU (September 2025): Mandatory use of EBM for estrogenic substances in surface waters for a 2-year period to detect harmful mixtures.

  • Summary of Priorities

    • A small number of compounds (often 1-10) typically dominate the total toxicity of a sample (average contribution of top compound = 61.5%61.5\%; top 10 compounds = 99.8%99.8\%).

    • The dominating compound shifts over time, meaning risk reduction must address the whole mixture, not just individual substances.

Questions & Discussion

  • Question Check: How do we assess compounds not detected (<LOD)?

    • Conservative approach: set concentration = LOD.

    • Best case: set concentration = 0.

    • Analytical standard: set concentration = LOD2\frac{LOD}{\sqrt{2}}.

    • Statistics solution: Impute data based on correlations with other compounds.

  • Discussion Point: Does Ivermectin B1a (90%90\%) vs B1b (10%10\%) matter for TU? Yes, because if B1a mean TU is 24332433, even the unmeasured B1b could add Significant toxicity (est. 270270).

  • Discussion Point: Pro vs Cons of EBM.

    • Pros: Measures all compounds causing a specific effect; includes unknowns.

    • Cons: No established threshold values; sensitive methods can produce signals that don't necessarily correspond to actionable environmental risk.