Inhibition of Human Drug Transporter Activities by the Pyrethroid Pesticides Allethrin and Tetramethrin

Pyrethroids are widely-used chemical insecticides, to which humans are commonly exposed, and known to alter functional expression of drug metabolizing enzymes. Limited data have additionally suggested that drug transporters, that constitute key-actors of the drug detoxification system, may also be targeted by pyrethroids. The present study was therefore designed to analyze the potential regulatory effects of these pesticides towards activities of main ATP-binding cassette (ABC) and solute carrier (SLC) drug transporters, using transporter-overexpressing cells. The pyrethroids allethrin and tetramethrin were found to inhibit various ABC and SLC drug transporters, including multidrug resistance-associated protein (MRP) 2, breast cancer resistance protein (BCRP), organic anion transporter polypeptide (OATP) 1B1, organic anion transporter (OAT) 3, multidrug and toxin extrusion transporter (MATE) 1, organic cation transporter (OCT) 1 and OCT2, with IC50 values however ranging from 2.6 μM (OCT1 inhibition by allethrin) to 77.6 μM (OAT3 inhibition by tetramethrin) and thus much higher than pyrethroid concentrations (in the nM range) reached in environmentally pyrethroid-exposed humans. By contrast, allethrin and tetramethrin cis-stimulated OATP2B1 activity and failed to alter activities of OATP1B3, OAT1 and MATE2-K, whereas P-glycoprotein activity was additionally moderately inhibited. Twelve other pyrethoids used at 100 μM did not block activities of the various investigated transporters, or only moderately inhibited some of them (inhibition by less than 50%). In silico analysis of structure-activity relationships next revealed that molecular parameters, including molecular weight and lipophilicity, are associated with transporter inhibition by allethrin/tetramethrin and successfully predicted transporter inhibition by the pyrethroids imiprothrin and prallethrin. Taken together, these data fully demonstrated that two pyrethoids, i.e., allethrin and tetramethrin, can act as regulators of the activity of various ABC and SLC drug transporters, but only when used at high and non-relevant concentrations, making unlikely any contribution of these transporter activity alterations to pyrethroid toxicity in environmentally exposed humans.

Introduction them are insecticides used in large-scale commercial agricultural and domestic applications and have already been tested in in vitro toxicity assays [30][31][32][33]. Our data demonstrate that the pyrethroids allethrin and tetramethrin are inhibitors of various drug transporters, but only when used at relative high concentrations likely not reached in humans environmentally exposed to these insecticides.

Materials and Methods Chemicals
Pyrethroids were provided by Sigma-Aldrich (Saint-Quentin Fallavier, France) and Cluzeau Info Labo (Sainte-Foy-La-Grande, France). The chemical structures of the fourteen pyrethroids whose the potential inhibitory effects towards activity of drug transporters were extensively tested are shown in S1 Fig. It is noteworthy that pyrethroid insecticides generally have complex configurations and contain one to three chiral centers, thus resulting in two to eight stereoisomers, with only some of them displaying insecticide properties [9,34,35]. Most, if not all, of these insecticides can therefore be theoretically considered as mixtures of geometric and optical isomers, knowing however that some commercial preparations of pyrethroids available on the market may contain only one or some of possible stereoisomers [34]. Unfortunately, the exact composition and stereoisomer proportion of the pyrethroids used in the present study were not provided by the suppliers. The total number of possible stereoisomers for each of the fourteen pyrethroids extensively analyzed in the study is given in S1 Table. Pyrethroids were initially prepared as stock solutions (50 mM) in dimethyl sulfoxide. Such stock solutions were next dissolved in the transport assay medium described below, for getting working pyrethroid concentrations tested on transporter activities. Rhodamine 123, verapamil, probenecid, amitriptyline, fumitremorgin C, fluorescein, 4',6'-diamidino-2-phenylindole (DAPI), and tetra-ethylammonium bromide (TEA) were purchased by Sigma-Aldrich, whereas carboxy-2,7-dichlorofluorescein (CF) diacetate and Hoechst 33342 were from Life Technologies (Saint Aubin, France). [1-14 C]-TEA (sp. act. 3.5 mCi/mmol), [6, H(N)]-estrone-3-sulfate (E3S) (sp. act. 54 Ci/mmol) and 3,4-[Ring-2,5,6-3 H]dihydroxyphenylethylamine (dopamine) (sp. act. 46 Ci/mmol) were from Perkin-Elmer (Boston, MA, USA). All other chemicals were commercial products of the highest purity available.
For transport assays, cells were usually seeded in 48-multiwell Falcon TM tissue culturetreated polystyrene or Corning TM BioCoat TM poly-D-lysine plates (Corning Incorporated, NY, USA). The type of multiwell plates, the initial cell seeding and the number of culture days before performing transport assays are indicated for each cell line/clone in S2 Table. ABC and SLC transporter activity The effects of pyrethroids on activity of ABC and SLC transporters were determined through measuring cellular accumulation or retention of fluorescent or radiolabeled reference substrates for transporters, in the presence or absence of reference inhibitors, as previously described [41]. The nature of cells and reference substrates and inhibitors used for transport assays are summarized in S3 Table. For accumulation assays (performed for all transporters, excepted BCRP), transporterexpressing cells usually cultured in 48-well plates were first incubated at 37˚C with reference substrates in the absence (control) or presence of pyrethroids or reference inhibitors, in a well-defined transport assay medium [42], consisting of 136 mM NaCl, 5.3 mM KCl, 1.1mM KH 2 PO 4 , 0.8 mM MgSO 4 , 1.8 mM CaCl 2 , 10 mM HEPES, 11 mM D-glucose and adjusted to pH = 7.4 (excepted for MATE transporter assays, for which pH was set to 8.4). The nature of substrates and reference inhibitors and the incubation times with substrates, that varied according to the transporter, were selected in agreement with previous studies [41,43] and are indicated in S3 Table. The total incubation volume with substrates was 200 μL/well. After washing twice in phosphate-buffered saline (PBS) (200 μL/well/washing), cells were lysed in distilled water (155 μL/well). An aliquot of cell lysate (5 μL) was then taken for determining cellular protein content by the Bradford method [44]. The rest of lysate (150 μL/well) was used for measuring intracellular accumulation of reference substrates by scintillation counting or spectrofluorimetry using a SpectraMax Gemini SX spectrofluorometer (Molecular Devices, Sunnyvale, CA, USA) (excitation and emission wavelengths were 485 and 535 nm, respectively, for rhodamine 123 and fluorescein, and 355 and 460 nm, respectively, for DAPI).
For retention assay (used for BCRP activity), HEK-BCRP cells cultured in 48-well plates were first loaded at 37˚C with transport assay (200 μL/well) containing 16.2 μM Hoechst 33342 for 30 min. After washing twice in PBS (200 μL/well/washing), cells were reincubated in Hoechst 33342-free transport assay medium (200 μL/well) at 37˚C for 90 min in the absence or presence of pyrethroids or of the BCRP reference inhibitor fumitremorgin C used at 10 μM. Cells were next lysed in distilled water (155 μL/well) after washing twice in PBS (200 μL/well/ washing). The cell lysate was finally used for determining cellular protein content and intracellular retention of Hoechst 33342 (by spectrofluorimetry: excitation and emission wavelengths were 355 and 460 nm, respectively), as described above.
Intracellular concentrations of reference substrates were initially expressed as fluorescence arbitrary units (for fluorescent dye substrates) or as moles (for radiolabeled substrates) and normalized to cellular protein content, determined by the Bradford method [44]. They were finally used for calculating percentage of transporter activity, according to the following equations [41,45] ] = cellular concentration of reference substrate in control cells not exposed to pyrethroid or reference inhibitor. For OCT1-mediated transport inhibition by pyrethroids, data were also expressed as percentages of inhibition of OCT1 activity using the following equation: % OCT1 activity inhibition ¼ 100% À % OCT1 activity ðin the presence of pyrethroidÞ ð3Þ The initial tested concentration of pyrethroids was 100 μM. When a pyrethroid used a this concentration modulated transporter activity by more than 50%, the effects of various additional concentrations of the pyrethroid (from 0.1 to 300 μM) on transporter activity were analyzed, in order to determine half maximal inhibitory concentration (IC 50 ) (for inhibitory effects of pyrethroids) or half maximal effective concentration (EC 50 ) (for stimulatory effects of pyrethroids). Such concentrations were calculated using GraphPad Prism software (Graph-Pad Software, La Jolla, CA), through nonlinear regression based on the four parameter logistic function using the following equations: For IC 50 : A ¼ 100 1 þ 10 ðð½I À LogIC 50 Þx Hill slopeÞ ð4Þ For EC 50 : where A is the percentage of transporter activity for a given concentration of pyrethroid determined as described in Eq (1) for ABC transporters and in Eq (2) for SLC transporters, [I] is the pyrethroid concentration in the medium, and Hill slope is a coefficient describing the steepness of the curve.

Dopamine accumulation in OCT1-transfected HEK293 cells
HEK-MOCK and HEK-OCT1 cells were incubated with transport assay medium containing 11 nM 3,4-[Ring-2,5,6-3 H]-dopamine, in the absence or presence of the reference OCT1 inhibitor verapamil or of pyrethroids for 5 min at 37˚C in the transport assay medium described above. After washing twice with PBS, cells were lysed in distilled water and intracellular accumulation of dopamine was finally measured by scintillation counting.

Trans-stimulation assays in OCT1-transfected HEK293 cells
Trans-stimulation assays were performed as previously described [46]. Briefly, OCT1-transfected HEK293 cells were first preloaded with 2 mM unlabeled TEA or 100 μM pyrethroid for 60 min at 37˚c in the transport assay medium described above. After washing twice with PBS, cells were next re-incubated in transport assay medium containing 29 μM [1-14 C]-TEA for 5 min at 37˚C. After washing twice with PBS, cells were lysed in distilled water and intracellular accumulation of [1-14 C]-TEA was then determined by scintillation counting.

Molecular descriptor generation
Molecular descriptors were evaluated using the Dragon 6 software (Talete, Milano, Italy), that provides 4885 molecular descriptors divided into 29 blocks (See http://www.talete.mi.it/ products/dragon_molecular_descriptor_list.pdf for a complete list of these descriptors). Pyrethroids, initially expressed in SMILES format, were converted to 3D format using the Marvin-View software (ChemAxon, Budapest, Hungary) before processing by Dragon 6 to obtain molecular descriptors.

Statistical analysis
Experimental data were expressed as means ± SEM of at least three independent experiments, each being usually performed in triplicate (for transport assays, one typical independent experiment therefore corresponded to 3 independent wells on the same multiwell plate). They were statistically analysis through the Student's t-test or analysis of variance (ANOVA) followed by a Dunnett's or a Newman-Keuls post-hoc test; the criterion of significance was p < 0.05. Correlation between molecular descriptor indexes and percentages of OCT1 activity inhibition by pyrethroids was initially done with Dragon 6 software through Pearson correlation, using a cut-off for Pearson's correlation coefficient of r > 0.8 (positive correlation) or r < -0.8 (negative correlation), as recommended by the software user's manual and as previously described [47]. P-values for Pearson correlations, as well as linear regressions for molecular descriptors exhibiting a high level of correlation with OCT1 activity inhibition (|r| > 0.92), were next determined using GraphPad Prism software after confirmation of normality of data distribution by D'Agostino and Pearson omnibus normality test. The statistical analysis of molecular descriptors belonging to the block of constitutional indices or to that of molecular properties and allowing to discriminate allethrin/tetramethrin from the other pyrethroids was performed using the unpaired Student's t-test, which is applicable to very small sample sizes [48,49].

Effects of pyrethroids on ABC transporter activity
For investigating the effects of pyrethroids on ABC transporter activity, we used MCF7R, HuH-7 and HEK-BCRP cells. Such cells exhibited increased accumulation (MCF7R and HuH-7 cells) or retention (HEK-BCRP cells) of dyes substrates for P-gp (rhodamine 123), MRP2 and MRP2-like transporters (CF) or BCRP (Hoechst 33342) in response to reference inhibitors of the pumps (verapamil for P-gp, probenecid for MRP2 and fumitremorgin C for BCRP) (S2 Fig), thus demonstrating that P-gp, MRP2 and BCRP are fully functional in these cells, as previously described [45]. By contrast, verapamil failed to enhance rhodamine 123 accumulation in parental MCF7 cells (S2 Fig), which is fully consistent with the absence of detectable P-gp activity in these cells; in the same way, fumitremorgin C did not augment Hoechst 33342 retention in HEK-MOCK cells (data not shown). Pyrethroids were routinely used at 100 μM for screening their potential effects towards drug transporter activity. This 100 μM concentration was retained because it was in the range of those previously used to interact with the pharmacological targets of pyrethroids, i.e., voltage-gated sodium channels, in cultured cells [50,51], thus underlining its potential relevance for in vitro studying interactions of pyrethroids with membrane proteins. P-gp activity was found to be moderately, but significantly, inhibited (32.7% inhibition) by 100 μM allethrin (Fig 1), whereas MRP2 and BCRP activity were more potently inhibited (83.8% and 83.9% inhibition for MRP2 and BCRP, respectively) by the pyrethroid (Fig 1). For MRP2 inhibition, allethrin IC 50   For P-gp, because 100 μM allethrin and 100 μM tetramethrin inhibited activity by less than 50%, IC 50 values may be predicted to be higher than 100 μM. Other pyrethroids used at 100 μM did not significantly inhibit P-gp, MRP2 or BCRP activity (Fig 1).

Effects of pyrethroids on organic anion SLC transporter activity
For investigating effects of pyrethroids on activity of organic anion SLC transporters, i.e., OATPs and OATs, we used OATP1B1-, OATP1B3-, OATP2B1-, OAT1-and OAT3-transfected cells. As shown in S4 Fig, these cells exhibited increased accumulation of reference substrates, i.e., E3S for OATP1B1 and OATP2B1 and fluorescein for OATP1B3, OAT1 and OAT3, when compared to OATP-or OAT-untransfected counterparts; moreover, the OATP inhibitor probenecid was able to reduce accumulation of reference substrates only in OATPand OAT-transfected cells (S4 Fig). Taken together, these data indicated that OATPs and OATs were fully functional in OATP-and OAT-transfected cells.

Effects of pyrethroids on organic cation SLC transporter activity
For analyzing effects of pyrethroids on activity of organic cation SLC transporters, i.e., MATEs and OCTs, we used MATE1-, MATE2-K-, OCT1-and OCT2-transfected HEK293 cells. As shown in S6 Fig, these cells exhibited increased accumulation of reference substrates, i.e., TEA for MATE1 and MATE2-K, DAPI for OCT1 and rhodamine 123 for OCT2 when compared to HEK-MOCK cells; moreover, reference inhibitors (verapamil for MATEs and OCT1 and amitriptyline for OCT2) reduced accumulation of reference substrates in MATE-and OCT-transfected cells, but not in HEK-MOCK cells (S6 Fig). Taken together, these data indicated that MATEs and OCTs were fully functional in MATE-and OCT-transfected cells.
With respect to OCT1 activity, it was markedly inhibited by 100 μM allethrin (87.9% inhibition) and 100 μM tetramethrin (73.8% inhibition); corresponding IC 50 values, i.e., 2.6 μM (allethrin) and 4.9 μM (tetramethrin), were rather low (S7B Fig). Pyrethroids such as β-cyfluthrin, λ-cyhalothrin, β-cypermethrin, deltamethrin, fenpropathrin, trans-permethrin, resmethrin and tefluthrin also significantly inhibited OCT1 activity, but in a weaker manner; OCT1 inhibition thus ranged from 30.1% (β-cypermethrin) to 48.8% (resmethrin) (Fig 4). By contrast, bifenthrin, cis-permethrin, esfenvalerate and τ-fluvalinate failed to significantly alter OCT1 activity (Fig 4). Allethrin and tetramethrin were finally also shown to markedly inhibit OCT2 activity (Fig 4); IC 50 values were 42.6 μM (allethrin) and 11.2 μM (tetramethrin) (S7C Fig). OCT2 activity was Data are expressed as percentages of activities found in untreated control cells, arbitrarily set at 100% and indicated by dotted lines on graphs; they are the means ± SEM of three independent assays, each being performed in triplicate. *, p<0.05 when compared to untreated control cells.  Characterization of pyrethroids-mediated OCT1 inhibition We next focused on pyrethroids-mediated OCT1 inhibition. Indeed, OCT1 can be considered as the drug transporter most impacted by pyrethroids because it was the transporter most sensitive to allethrin and tetramethrin according to IC 50 values (S7B Fig) and it was also the transporter significantly inhibited by the largest number of pyrethroids, as 10/14 pyrethroids significantly impaired its activity (Fig 4). We first determined whether allethrin and tetramethrin can inhibit OCT1-mediated transport of endogenous substrates such as dopamine. HEK-OCT1 cells exhibited enhanced verapamil-inhibitable accumulation of dopamine comparatively to HEK-MOCK cells (Fig 5A), thus fully confirming that OCT1 handles dopamine [52]. Allethrin used at 10, 100 or 300 μM was found to significantly inhibit this OCT1-mediated transport of dopamine ( Fig 5B); tetramethrin also decreased it, but only when used at 100 or 300 μM (Fig 5B). Allethrin and tetramethrin used at 100 μM were next shown to markedly cis-inhibit accumulation of the reference OCT1 substrate TEA in HEK-OCT1 cells (Fig 6A). Finally, we investigated whether allethrin and tetramethrin can trans-stimulate TEA uptake, which may constitute an argument in favor of the transport of the two pyrethroids by OCT1 [46]. Pre-loading with allethrin and tetramethrin however resulted in trans-inhibition, and not trans-stimulation, of radiolabeled TEA uptake ( Fig 6B). By contrast, pre-loading with unlabeled TEA led to a trans-stimulation of radiolabeled TEA uptake (Fig 6B), as expected for an OCT1 substrate like TEA [46].
In order to identify the specific physico-chemical properties associated to OCT1 inhibition by pyrethroids, molecular descriptors, including 0D-constitutional, 1D-structural, 2D-topological and 3D-geometrical descriptors, were determined using Dragon 6 software. Putative correlations with OCT1 activity inhibition were next analyzed using Pearson's correlation test. Using a cutoff of |r| > 0.8, 602 molecular descriptors were found to be correlated with inhibition of OCT1 activity; the correlation was positive (r > 0.8) or negative (r < -0.8) for 93 and 509 descriptors, respectively (Table 1). These molecular descriptors associated with OCT1 inhibition belong to different blocks of descriptors, especially those of 2D-matrix based descriptors (n = 350), of edge adjacency indices (n = 85), of topological indices (n = 26), of 3D-MoRSE descriptors (n = 24) and of walk and path counts (n = 20) (Table 1). A complete list of these descriptors is given in S1 File. Notably, lipophilicity, evaluated through octanolwater partition coefficients (LogP) calculated using Moriguchi LogP model (MLOGP), was negatively correlated to OCT1 inhibition, whereas no correlation was found for molecular weight (MW) or number of hydrogen bond acceptors (nHAcc) and donors (nHDon) (S1 File). When a more stringent cut-off value of |r| > 0.92 was applied, 9 molecular descriptors were found to be correlated with OCT1 activity inhibition (S4 Table); linear regression analysis next confirmed a highly significant linear relation between the index values of these descriptors and the percentages of OCT1 activity inhibition (S8 Fig). Such molecular descriptors exclusively belong to the category of 2D topological indexes, notably to walk and path counts (S4 Table).

Determination of basic molecular descriptors discriminating multitransporter-interacting pyrethroids (allethrin/tetramethrin) from other pyrethroids
Allethrin and tetramethrin exhibit a rather large, but specific, profile of transporter inhibition, because they markedly decreased activity of various transporters, including MRP2, BCRP, OATP1B1, OAT3, MATE1, OCT1 and OCT2, whereas other pyrethroids concomitantly exerted no, or only moderate or marginal, inhibitory effects. In addition, allethrin and tetramethrin, unlike other pyrethroids, stimulated OATP2B1 activity (Fig 2). Molecular features restricted to allethrin and tetramethrin may therefore be hypothesized to be involved in the regulations of drug transporter activities caused by the two pyrethroids. In order to identify such physical-chemical parameters, we compared constitutional indices and molecular properties from allethrin and tetramethrin with those of other pyrethroids, because important   descriptors for transporter inhibition are usually comprised among these basic descriptors [53]. As shown in Table 2, values for 19 descriptors, i.e., 15 constitutional indices and 4 molecular properties, were found to significantly differ between allethrin/tetramethrin and other pyrethroids. These discriminating molecular descriptors notably correspond to molecular weight (MW), van der Waals volume (Mv), polarizability (Mp), rotatable bond fraction (RBF), percentage of C atoms (C%) and lipophilicity/LogP, calculated using either Moriguchi (MLOGP) or Ghose-Crippen-Viswanadhan (ALOGP) LogP models. Allethrin/tetramethrin thus exhibit reduced molecular weight, van der Waals volume, polarizability, percentage of C atoms and lipophilicity when compared to other pyrethroids (Table 2). Interestingly, when some of these descriptors were combined pairwise, they allowed to easily graphically discriminate allethrin/tetramethrin from other pyrethroids (Fig 7). Applying this graphically discrimination based on MW versus ALOGP or on C% versus Mv to three type I pyrethroids not previously analyzed in the present study, i.e., imiprothrin, prallethrin and phenothrin, allowed to predict that imiprothrin and prallethrin, unlike phenothrin, may interact with drug transporter activities (Fig 8A). This prediction was next validated through demonstrating that imiprothrin and prallethrin, but not phenothrin, inhibited OCT1, OCT2 and OAT3 activities and cis-stimulated that of OATP2B1 when used at 100 μM ( Fig 8B); imiprothrin and prallethrin, like allethrin and tetramethrin (Fig 3), however failed to alter OAT1 activity (Fig 8B).

Discussion
Previous studies have shown that some pesticides, including pyrethroids, may interact with drug transporters [26,27,54,55]. The data reported in the present study fully support this hypothesis through demonstrating that the type I pyrethroids allethrin and tetramethrin inhibited various ABC and SLC drug transporters (See Table 3 for a summary of drug transporter activity regulation by these two pyrethroids). The type I pyrethroids imiprothrin and prallethrin were additionally shown to interact with OCT1, OCT2, OAT3 and OATP2B1 activities. By contrast, other type I or type II pyrethroids, did not, or only rather very moderately or marginally, impair drug transporter activities. Allethrin and tetramethrin significantly inhibited activities of the drug efflux pumps P-gp, MRP2 and BCRP and those of the SLC transporters OATP1B1, OAT3, MATE1, OCT1 and OCT2. The potency of inhibition nevertheless depends on transporters (Table 3). According to IC 50 values, that constitute the main parameter to be considered for drug transporter inhibition, notably under a regulatory point of view [22], transporter ranking (from the most inhibited to the less inhibited transporter) was OCT1>OATP1B1>BCRP>OCT2>MRP2> MATE1>OAT3>P-gp for allethrin and OCT1>OATP1B1>OCT2>MATE1>MRP2> BCRP>OAT3>P-gp for tetramethrin. The two pyrethroids also interacted with OATP2B1 activity, which was cis-stimulated. Only OATP1B3, OAT1 and MATE2-K activities were not altered by the two pyrethroids. The fact that allethrin and tetramethrin failed to interact with these drug transporters, associated to the stimulation of OATP2B1 and to the differential sensitivities of inhibited transporters, discards any general and non-discriminating inhibitory effect of allethrin and tetramethrin towards membrane transporter activities. In addition, IC 50 for transporter inhibition by allethrin and tetramethrin range from 2.6 μM to 77.6 μM, and such values are generally much higher than pyrethroid concentrations (around 0.5-2 μM) usually required for interacting with sodium channels in primary cultured mammalian cells [56], making unlikely that drug transporter inhibition by pyrethroids may be due to alteration of sodium channel activity. This conclusion is also fully supported by the fact that the majority of pyrethroids, excepted allethrin, tetramethrin, imiprothrin and prallethrin, failed to inhibit transporters in a major and notable way, even if they are all presumed to efficiently alter sodium channel activity [35]. Interactions of some pyrethroids such as allethrin and tetramethrin with various drug transporters can rather be considered as specific, most likely reflecting direct and transporter-dependent interactions of these two pyrethroids with substrate and/or regulatory binding sites on drug transporters, as classically thought for drug transporter inhibitors [57]. In this context, it can be speculated that binding of allethrin and tetramethrin to targeted transporters reflects physico-chemical properties specific to these two pyrethroids. Such properties are likely unrelated to those underlying activity of allethrin and tetramethrin towards sodium channels. Indeed, the sodium channels-based insecticidal activity of allethrin and tetramethrin is strictly dependent upon the entire stereospecific structure of these insecticides [9,34], as for other pyrethroids inactive or poorly active on transporters. It is consequently Inhibition of Drug Transporter Activities by Pyrethroids restricted to some stereoisomers and did not rely on a specific substructure reactive entity or molecular moiety that could be identified as the toxophore conferring pyrethroid-like insecticidal activity [35]. The inhibitory effects against transporter activities of allethrin and tetramethrin can therefore not been specifically ascribed to the pyrethroid isomers within these mixtures that are insecticides and mammalian neurotoxicants. Further studies are therefore likely required to investigate the exact contribution of the various tetramethrin and allethrin stereoisomers not implicated in primary insecticidal and neurotoxic effects to drug transporter Inhibition of Drug Transporter Activities by Pyrethroids inhibition, knowing that the effects of drug transporter inhibitors may exhibit stereoselectivity [58]. Physicochemical parameters involved in drug transporter inhibitory effects exerted by allethrin and tetramethrin likely include constitutional indices, such as molecular weight and polarizability, and molecular properties, such as hydrophobicity (LogP), which were found to significantly discriminate allethrin/tetramethrin from other pyrethroids. Importantly, these basic descriptors have already been shown to be crucial for inhibition of the transporters targeted by the two pyrethroids. Lipophilicity/LogP, molecular weight and/or molecular polarizability thus constitute important parameters to consider for inhibiting MRP2, BCRP, OATP1B1, OAT3, MATE1, OCT1 or OCT2 [59][60][61][62][63][64][65]. Interestingly, pairwise association of some of these basic descriptors allow to easily and graphically differentiate allethrin/tetramethrin from other pyrethroids not, or only poorly interacting with drug transporters. Such a graphical discrimination was moreover used to successfully predict whether pyrethroids such as imiprothrin, phenothrin and prallethrin may interact with activity of transporters like OCT1, OCT2, OATP2B1, OAT1 and OAT3. This suggests that taking into consideration various descriptors together, including not only basic descriptors but also 2D-and 3D-molecular descriptors, would help to precisely and more extensively characterize the molecular basis of the specific interactions of some pyrethroids with drug transporters. In particular, this may lead to a better knowledge of the molecular features involved in cis-stimulation of OATP2B1 by pyrethroids like allethrin and tetramethrin. Cis-stimulation of OATP transporters has already been reported for other chemicals [60]. Similarly, other transporters like MRP2, OAT1, OAT3 and MATE1 are also subjected to cis-stimulation by some chemicals [61,62,66]. The molecular mechanisms that underlie such cis-stimulatory effects remain however yet poorly characterized.
Among transporters targeted by allethrin and tetramethrin, OCT1 is likely a major one. Indeed, allethrin and tetramethrin inhibited OCT1-mediated transport of DAPI with IC 50 in the 2-5 μM range. Moreover, OCT1-mediated transport of the dye was similarly markedly inhibited by 100 μM imiprothrin and prallethrin, whereas other pyrethoids also impacted OCT1 activity, although in a weaker manner. Allethrin and tetramethrin also blocked at 100% and indicated by dotted lines on graphs; they are the means ± SEM of at least three independent assays, each being usually performed in triplicate. *, p<0.05 when compared to untreated control cells.
doi:10.1371/journal.pone.0169480.g008 Table 3. Summary of allethrin and tetramethrin effects towards drug transporters. OCT1-mediated transport of the reference OCT1 substrate TEA, as well as that of the endogenous substrate dopamine. Inhibition of OCT1-related transport of dopamine by the two pyrethroids was however rather less marked than that of DAPI; indeed, allethrin used at 10 μM reduced OCT1-mediated DAPI and dopamine accumulation by 73.7% and 51.6%, respectively, whereas 10 μM tetramethrin decreased DAPI uptake by 66.4%, but failed to significantly alter that of dopamine ( S7B Fig and Fig 5). This likely illustrates the fact that inhibitory profiles for a transporter can vary according to the nature of the substrate and of drug binding sites, as already established for various transporters, including organic cation transporters [67][68][69].

Drug transporter Allethrin Tetramethrin
The mechanism by which pyrethroids like allethrin and tetramethrin down-regulate OCT1 activity remain to be determined. Because the two pyrethroids failed to trans-stimulate OCT1 activity, but rather trans-inhibited it, they are likely not substrate for OCT1. This may suggest that they do not act through a competitive mechanism toward the transport drug binding site on OCT1. High lipophilicity has been previously shown to be positively correlated to OCT1 inhibition, whereas a high molecular dipole moment and many hydrogen bonds were negatively correlated and molecular weight was not correlated [63]. In agreement with these data, we have found that molecular weight was also not correlated with OCT1 inhibition by pyrethroids. By contrast, LogP was negatively correlated for pyrethroids, whereas the number of hydrogen bonds as well as those of hydrogen bond acceptors and donors were not correlated. The reason for these discrepancies between our study and that of Ahlin et al. [63] is unclear, but could be linked to the small number and the relative structural homology of pyrethroids analyzed in the present study (n = 14) when compared to the larger number of structurallydiverse chemicals (n = 191) previously investigated [63]. Moreover, all the analyzed pyrethroids are rather lipophilic and non-charged compounds. Nevertheless, a great number of molecular descriptors, notably 2D-matrix based descriptors, has been found to be correlated to OCT1 activity inhibition by pyrethroids; some of them, notably some walk and path counts, were moreover demonstrated to exhibit a highly significant linear relationship with the percentage of OCT1 activity inhibition. Beyond pyrethroids, further studies are likely required to evaluate the relevance of these molecular descriptors for predicting OCT1 inhibition by diverse structurally-unrelated chemicals. The putative relevance of the transporter inhibitions described in vitro in the present study to human exposure to environmental pyrethroids is probably a key-point that has to be considered. In this context, it is first noteworthy that only some of the pyrethroids tested, i.e., allethrin, tetramethrin, imiprothrin and prallethrin, exhibit significant inhibitory effects towards transporters; this indicates that such effects on drug transporters are likely idiosyncratic offtarget effects of a limited number of compounds rather than representative of effects relevant to the toxicology of pyrethroids as a class. Moreover, blood concentrations of pyrethroids are very low or not detected in human subjects environmentally exposed to these pesticides (<5μg/L, i.e., <12.8 nM for cyfluthrin, cypermethrin, deltamethrin or permethrin [70]). This makes very unlikely that pyrethroid concentrations required to inhibit drug transporters, which range from 2.6 μM to 77.6 μM for allethrin/tetramethrin according to IC 50 values, and are therefore much higher than those targeting ion channels [56], may be reached in vivo. Moreover, following systemic absorption, pyrethroids are rapidly and extensively metabolized through oxidation, ester hydrolysis and conjugation, which precludes their accumulation in any specific tissues or organs [71]. Transporter inhibition in response to environmental exposure to pyrethroids such as allethrin and tetramethrin is therefore unlikely to occur at first view, which discards any drug-drug interactions due to drug transport impairment or any alteration of endogenous substrate transport by environmental pyrethroids. Putative consideration of the inhibitory effects of pyrethroids like allethrin and tetramethrin towards drug transporters under a regulatory point of view for pyrethroid risk assessment can therefore be excluded. It should be however kept in mind that humans are likely exposed not only to a single pyrethroid, but to other pesticides and xenobiotics, that may also interact with drug transporters [55], as recently demonstrated for notably organochlorine pesticides [43], polychlorinated biphenyls [72], diesel exhaust particle components [41] and perfluorinated surfactants [73]. Plasma levels of pyrethroids, especially allethrin and tetramethrin, associated with those of other pollutants, may therefore be sufficient to contribute to synergic or additive inhibitory effects towards drug transporters, as already described for pesticide combinations [74]. In addition, concentrations of pyrethroids may be much higher in the gastro-intestinal tract (before first-pass metabolism) than in plasma, and activities of drug transporters notably expressed at the intestinal level, especially P-gp, MRP2, BCRP, OATP2B1 and OCT1 [22,75], may therefore be hypothesized to be impacted by ingested pyrethroids. Moreover, characterizing concentration-dependence of transporter activity inhibition by allethrin and tetramethrin in the present study was based on IC 50 values; because IC 50 values rely on substrate concentrations [76], the relevance of potential transporter inhibition has likely to be challenged according to the nature and in vivo concentrations of putative substrates. Finally, elevated concentrations of pyrethroids are likely to occur during acute poisoning [77] and, in this case, inhibition of transport of endogenous substrates like dopamine by allethrin or tetramethrin may contribute to the toxicity of these pesticides.
Another points that should be considered are the putative handling of pyrethroids by drug transporters, notably by intestinal transporters which in case of food and drinking water exposure would first be encountered [29], and the consequences for toxicokinetics of these insecticides. In this context, it is noteworthy that whether pyrethroids or their metabolites can be substrates for drug transporters remains to be determined. As already discussed above, transport of allethrin and tetramethrin by OCT1 is however unlikely to occur, as demonstrated by the failure of these two chemicals to trans-stimulate OCT1 activity. In the same way, deltamethrin has been recently shown to be not effluxed by P-gp [78], which agrees with the fact that this pyrethroid did not inhibit this ABC efflux pump in the present study and also failed to alter P-gp ATPase activity [28]. Conflicting results have however been reported with respect to interactions of pyrethroids with P-gp. Cypermethrin, esfenvalerate, fluvalinate and permethrin have thus been shown to inhibit P-gp activity or ATPase activity [26, 27], but they did not alter P-gp activity in the present study. Similarly, permethrin and resmethrin inhibited BCRP ATPase activity [28], without inhibiting BCRP activity in our study. The reasons of such discrepancies are unclear, but they could be linked to the high concentrations of pyrethroids used in some previous studies [26] or to the different methodological approaches retained for investigating ABC efflux pump, i.e., ATPase activity versus effective transport of reference substrates.
Finally, the hypothesis that pyrethroids may alter expression of drug transporters, notably in the liver, has likely to be additionally considered. Indeed, pyrethroids can activate drug sensing receptors like PXR and CAR [17,18], which are well-known to be implicated in drug transporter expression regulation [79,80]. Activation of PXR or CAR by pyrethroids may consequently result in up-regulation of transporters targeted by these receptors such as P-gp [81], MRP2 [82] or MRP3 [83]. Whether such putative regulations may occur in response to environmental concentrations of pyrethroids and may have functional consequences would deserve further studies.
In summary, the pyrethroids allethrin and tetramethrin were found to inhibit activity of various ABC and SLC drug transporters. Such inhibitions of drug transporters however occurred for concentrations of the two pyrethroids much higher than those commonly expected in response to environmental exposure, making unlikely any relevant contribution of transporter inhibition to pyrethroid toxicity in environmentally exposed humans.