Additive Mixture Effects of Estrogenic Chemicals in Human Cell-Based Assays Can Be Influenced by Inclusion of Chemicals with Differing Effect Profiles Richard Mark Evans*, Martin Scholze, Andreas Kortenkamp Institute for the Environment, Brunel University, Uxbridge, Middlesex, United Kingdom Abstract A growing body of experimental evidence indicates that the in vitro effects of mixtures of estrogenic chemicals can be well predicted from the estrogenicity of their components by the concentration addition (CA) concept. However, some studies have observed small deviations from CA. Factors affecting the presence or observation of deviations could include: the type of chemical tested; number of mixture components; mixture design; and assay choice. We designed mixture experiments that address these factors, using mixtures with high numbers of components, chemicals from diverse chemical groups, assays with different in vitro endpoints and different mixture designs and ratios. Firstly, the effects of mixtures composed of up to 17 estrogenic chemicals were examined using estrogenicity assays with reporter-gene (ERLUX) and cell proliferation (ESCREEN) endpoints. Two mixture designs were used: 1) a ‘balanced’ design with components present in proportion to a common effect concentration (e.g. an EC 10 ) and 2) a ‘non-balanced’ design with components in proportion to potential human tissue concentrations. Secondly, the individual and simultaneous ability of 16 potential modulator chemicals (each with minimal estrogenicity) to influence the assay outcome produced by a reference mixture of estrogenic chemicals was examined. Test chemicals included plasticizers, phthalates, metals, PCBs, phytoestrogens, PAHs, heterocyclic amines, antioxidants, UV filters, musks, PBDEs and parabens. In all the scenarios tested, the CA concept provided a good prediction of mixture effects. Modulation studies revealed that chemicals possessing minimal estrogenicity themselves could reduce (negatively modulate) the effect of a mixture of estrogenic chemicals. Whether the type of modulation we observed occurs in practice most likely depends on the chemical concentrations involved, and better information is required on likely human tissue concentrations of estrogens and of potential modulators. Successful prediction of the effects of diverse chemical combinations might be more likely if chemical profiling included consideration of effect modulation. Citation: Evans RM, Scholze M, Kortenkamp A (2012) Additive Mixture Effects of Estrogenic Chemicals in Human Cell-Based Assays Can Be Influenced by Inclusion of Chemicals with Differing Effect Profiles. PLoS ONE 7(8): e43606. doi:10.1371/journal.pone.0043606 Editor: Aditya Bhushan Pant, Indian Institute of Toxicology Reserach, India Received June 14, 2012; Accepted July 26, 2012; Published August 17, 2012 Copyright: ß 2012 Evans et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: Financial support from the United Kingdom Food Standards Agency (Contract Number T01045) is gratefully acknowledged. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: richard.evans@brunel.ac.uk Introduction Humans are typically exposed to multiple chemicals with diverse effects [1]. Despite this, experimental studies usually examine binary or ternary combinations. We recently reviewed 173 experimental mixture studies and found that the majority had tested binary combinations and that fewer than one in four studies had examined mixtures with seven or more components [2]. Of the available mathematical concepts for the prediction of mixture effects, concentration addition (CA) has proven the most useful and has been shown to have good predictive power, see reviews by [1,3]. Ermler et al. recently showed that mixtures of 17 anti-androgenic chemicals with varied structural features produced effects in vitro that were predictable by CA [4]. However, a study of a similar number of estrogenic chemicals observed small deviations from the predictions made using CA [5]. Silva et al. studied five mixtures with from 3 to 16 components in the ESCREEN assay. The effects of two mixtures were accurately predicted by CA, whilst three showed slight overestimation by CA. It was hypothesised that the deviation was due to increased metabolism of steroidal estrogens and it was suggested that CYP1B1 activation and reduction in steroidal estrogen concentrations could ‘‘contribute to the shortfall from [CA]’’ [5]. This hypothesis was tested by predicting the mixture effect if the steroidal estrogens had been removed by metabolism (i.e. they make no contribution to the overall effect) and comparing this scenario with the observed result. Few other studies have examined mixtures with a similar number of components, and factors that remain to be addressed include the assay used, the number and type of chemical studied and the mixture design. Further studies in this area are required in order to resolve whether predictability by CA should be the default expectation for multicomponent mixtures of estrogens. If so, CA could be routinely applied for the assessment of such mixtures. In this paper we aimed to address the impact of choice of assay system, the number and nature of included chemicals, the mixture design, and the possibility of effect modulation. Our approach to each of these issues is now described in turn. Assay System To evaluate the possible contribution of the model system, we have compared the predictability of mixture effects in two in vitro PLOS ONE | www.plosone.org 1 August 2012 | Volume 7 | Issue 8 | e43606