Various approaches for evaluating chemical mixtures have been proposed by the scientific community [148
]; however, there is no internationally agreed-upon procedure. The proposed approaches fall within two general categories, the whole-mixture (evaluation as though mixtures are single entities) and the component-based (evaluating individual chemicals in a mixture to estimate response) approaches [72
]. Whole-mixture approaches can be impractical due to the multiple interactions that can potentially occur in real-world mixtures, some of which do not necessarily occur via a common mode of action by structurally similar compounds. Furthermore, this approach does not identify which types of chemical interactions are responsible for additive, synergistic, or antagonistic effects. Most studies have thus resorted to using the component-based approach, which requires information on each individual component within the mixture [150
]. The component-based approach operates on the calculated sum from either of the following methods: (1) concentration or dose addition method
, which assumes that mixture components act on a similar target and therefore elicit a common response; and (2) the response addition method
, which assumes that components act on different targets (the overall response is calculated from individual components) [148
]. The latter method is not commonly used for XE mixtures.
The overall goal of every method is to establish principles of how chemicals behave based on structure and mode of action. Once enough examples are processed and the model optimized, then predictions for unknown mixtures should be possible. In a review, Kortenkamp extensively discussed mixture effects of several classes of EDCs (i.e.
, estrogenic, anti-androgenic, and thyroid-disrupting agents) [28
], focusing primarily on their genomic and functional responses. Below, we will briefly focus on the few existing non-genomic studies of synthesized estrogen mixtures.
Jeng and Watson studied the phospho-activation of MAPK (p
ERK) upon exposure to binary mixtures of endogenous estrogens (E1
, and E3
) at single physiologic (nM) concentrations with increasing (10−15
M) concentrations of alkylphenol compounds in the GH3/B6/F10 pituitary cell line [25
]. Individual compounds caused non-monotonic dose-responses, but with varying weak, moderate or strong response levels compared to E2
. The composite responses were not additive, and often showed attenuation at the higher concentrations. The degree of attenuation was based on the response magnitude and potency of the paired xenoestrogen. The stronger the XE’s activating response, the more it was able to attenuate the physiologic estrogenic response.
When assessing the effects of XEs on dopamine efflux through its transporter in PC12 cells, Alyea and Watson also found that 10−14
M DDE caused a weak efflux as a lone compound, but in a binary mixture with 10−9
it additively enhanced dopamine efflux. BPA in contrast evoked a strong efflux response on its own, but when mixed with 10−9
it inhibited efflux [64
The overall pattern observed in these two studies was that when a compound with a weak response is paired with a physiologic estrogen, the response is enhanced. But, when a compound elicits a potent estrogenic effect, then it inhibits the paired physiologic estrogen’s response. This progression is summarized graphically in . In a very recent tertiary mixture study, we have observed further inhibition of responses to the physiologic estrogen E2 by two added XEs However, the same mixture resulted in a synergistic positive response for pJNK (Viñas and Watson, unpublished); hence, when assessing non-genomic pathways one has to take into consideration the variety of signaling responses, and probably the interactive nature of signaling “webs”. Response inhibitions by combinations of estrogens may be governed by cellular protective mechanisms against combined hormone overstimulation. Overstimulation can be wasteful and even dangerous when the enhanced function (such as peptide release or cell proliferation, for example) can lead to diseases like cancer.
Figure 1 Working model of xenoestrogen (XE) alteration of physiologic estrogen non-genomic response effects. XEs of increasing dose were used to challenge the responses of the physiologic estrogen estradiol (10−9 M E2). These combinations examples are (more ...)
Interestingly, not all signaling pathways culminating in different functional responses may behave in the same fashion. When another type of response (PRL release) was monitored under the same (binary mixture) circumstances, BPA’s strong p
ERK response when present by itself did not correlate with a strong PRL secretion. However, PRL release did decrease when BPA was paired with either E2
(but not the weaker E3
) in pituitary cells [24
]. This means that one has to study sufficient examples of compounds over a wide range of times points and concentrations, assessed for a spectrum of different signals and functional endpoints. It will take an adequately representative set of such data to finally hone our predictive principles. In addition, for some complex responses such as cell proliferation, there will undoubtedly be both genomic and non-genomic contributory components to consider, as well as cross-talk between signaling pathways.