Proteomic Analysis Reveals Multiple Patterns of Response in Cells Exposed to a Toxin Mixture Gian Luca Sala, †,‡ Giuseppe Ronzitti, †,‡,§ Makoto Sasaki, | Haruhiko Fuwa, | Takeshi Yasumoto, ⊥ Albertino Bigiani, † and Gian Paolo Rossini* ,† Centro Interdisciplinare di Scienze e Tecnologie per la Qualità e Sicurezza degli Alimenti, Dipartimento di Scienze Biomediche, UniVersita ` di Modena e Reggio Emilia, I-41100 Modena, Italy, Graduate School of Life Sciences, Tohoku UniVersity, Sendai 981-8555, Japan, and Okinawa Science and Technology Promotion Center, Okinawa, 904-2234, Japan ReceiVed February 4, 2009 We have used proteomic analyses to probe the responses induced by a pair of marine biotoxins, okadaic acid (OA) and gambierol (GB), added alone or in combination to a cultured cell line and carried out a preliminary investigation into the possible interactions between toxins possessing two different molecular mechanisms of action at a cellular level. When MCF-7 cells were treated with OA, we found that cellular levels of 30 proteins were significantly affected, including several isoforms of nonphosphorylated and phosphorylated hsp 27, as well as enzymes involved in the maintenance of nucleoside triphosphate pools and the control of redox states of the cell. When we repeated our analysis using GB, nine proteins were significantly affected, including some isoforms of nonphosphorylated hsp 27, as well as semenogelin-1, myosin-7, and the ATP synthase subunit δ. The combined addition of OA and GB to MCF-7 cells, in turn, affected 14 proteins, including some isoforms of nonphosphorylated and phosphorylated hsp 27, as well as myosin-7, the ATP synthase subunit δ, and enzymes involved in the control of redox states of the cell. If components affected by either OA or GB, as well as by the combined treatment, were classified according to the detected changes, two sets of data were obtained, including the components whose levels were found affected by the combined treatment, regardless of the effect observed after addition of only one agent, and those that had been found affected in cells that had been challenged with only one toxin but not when cells had been subjected to the combined treatment. Multiple patterns of responses to the toxin mixture were recorded in the two sets, consisting of both independent and interacting actions, among which we detected synergistic, similar, and antagonistic effects. Introduction The use of biomarkers for the assessment of the exposure of biological entities to potential hazards, including toxic sub- stances, is an established practice that has gained new impetus by the introduction of high-throughput technologies, with their potential to process large amounts of analytes at the same time (1, 2). These systemic analyses provide the fingerprints of exposures to individual groups of chemicals, avoiding the risk inherent in the use of single end point bioassays (1-3). In the real world, living organisms are exposed to complex mixtures of bioactive compounds and hazardous chemicals, so that the identification of biomarkers for a single class of compounds may not be sufficient for a proper understanding of any molecular fingerprint. Furthermore, the possibility that mixtures of toxins at concentrations below their nonobservable adverse effect levels might cause detectable responses in biological systems should be considered (4). To approach this level of complexity and taking into account the limited information regarding possible interactions between different classes of marine biotoxins in biological systems, we set out a preliminary investigation to probe the general behavior of an experimental model challenged with two different toxins. The aim of our study was to carry out a system-level analysis, to ascertain whether general patterns of responses could be found, independently of specific sequences of events set in motion by the agents present in a mixture. In this study, therefore, the biological system was treated as a black box, and we have analyzed the overall outputs generated by the chemicals added to the system (5). Some methodological choices were made to develop our study. To simplify the analysis and obtain sets of data that could represent the overall responses of the system to the tested compounds at the effector level, we characterized the changes induced by biotoxins in the proteome of a cultured cell line. The use of marine biotoxins was made because contamination of seafood by algal toxins is a widespread phenomenon, posing significant threats to human and animal health (6), and it has long been recognized that contamination can involve multiple chemical classes of compounds (7-9), often present in con- taminated materials at the same time (9-12). A second choice regarded the tested compounds, and we sought to minimize the occurrence of interactions among different biotoxins. To this end, we have challenged the experimental system with okadaic acid (OA) and gambierol (GB), because they have distinct mechanisms of action. OA, in fact, is an inhibitor of serine/ threonine phosphoprotein phosphatases (13), whereas GB affects * To whom correspondence should be addressed. E-mail: gianpaolo.rossini@unimore.it. † Universita ` di Modena e Reggio Emilia. ‡ These authors have equally contributed to this paper. § Present address: Neuroscience and Brain Technologies Department, Italian Institute of Technology, Via Morego 30, 16163, Genova, Italy. | Tohoku University. ⊥ Okinawa Science and Technology Promotion Center. Chem. Res. Toxicol. 2009, 22, 1077–1085 1077 10.1021/tx900044p CCC: $40.75  2009 American Chemical Society Published on Web 04/27/2009