Discussion
Data reported in this study showed that the impairment of transduction pathways caused by low concentrations of pollutants have the potential to affect the ability of animals to elaborate strategies of defense or adaptation. Indeed, exposure of mussels to low environmental range concentrations of the pharmaceutical fluoxetine affected cell signaling mediators such as cAMP steady-state levels, PKA activity, and mRNA levels for a 5-HT1 receptor in haemocytes, and also altered mRNA levels of an ABCB gene encoding the membrane transporter P-glycoprotein, which displays a peculiar function in cellular defense machinery [5].
The mode of action of FX as a selective serotonin reuptake inhibitor (SSRI) is to increase serotoninergic neurotransmission at mammalian synapses by blocking removal of 5-HT by reuptake transporters ([28] and reference therein). As such, FX mimics the action of a continuous exposure to increased extracellular levels of 5-HT ([30] and reference therein). As discussed in detail below, the physiological functions of serotonin are mediated by multiple membrane receptors coupled to the stimulation of phospholipase C (PLC) or to the induction/inhibition of adenylyl cyclase (AC) in mammals, and several lines of evidence suggest a similar scenario also in bivalves, although the pharmacological classification of 5-HT receptors in these organisms is incomplete and yet limited to what reported by Tierney [14]. The reduction of cAMP levels and PKA activities below control values observed in this study are, therefore, consistent with the inhibition of AC activity by a putative increase of 5-HT extracellular levels induced by FX. As a further support to this hypothesis, FX effects were completely abolished when mussels were exposed to FX in the presence of PROP. Indeed, although PROP is a prototypical β-AR antagonist used in human therapies to counteract cardiovascular pathologies [16], it is also an effective antagonist for type 1 5-HT receptors (5-HT1), which are coupled to AC inhibition [14], [26], [27]. Therefore, both compounds can affect the same molecular target indirectly (FX) by increasing the availability of the agonist molecules in the extracellular medium, directly (PROP) by binding to the receptor and avoiding its occupation by the agonist.
Data from in vivo exposures also showed that ABCB mRNA expression profile is consistent with variations of cAMP-related parameters. Considering the specific effect displayed by FX and PROP on cAMP signaling, these data provided further hints about a cAMP/PKA involvement in the transcriptional regulation of the mussel ABCB gene encoding the P-glycoprotein. Besides present results, other in vivo investigations provided some indirect evidence towards this connection Indeed, mussels exposed to environmental concentrations of carbamazepine, whose therapeutic mode of action in humans is at least in part linked to the cAMP/PKA signaling pathway, showed significant variations of intracellular cAMP levels and PKA activities, changes which were correlated with the regulation of ABCB mRNA expression in different tissues [7], [8]. Further environmental pollutants, including metals or algal products, are reported to alter both cAMP levels and MXR-related mRNA expression in mussels [9], [31].
To provide clear demonstration about the involvement of cAMP and PKA in the modulation of the mussel MXR system, and more specifically in the transcriptional regulation of the ABCB gene, in vitro investigations using haemocytes as a cell model were performed. To our knowledge, the data represent the first direct evidence that the regulatory framework, well documented in mammals, is evolutionary-conserved in bivalve mollusks ( Fig. 8 ).
10.1371/journal.pone.0061634.g008 Figure 8  Schematic representation of the cAMP-dependent pathway leading to ABCB transcriptional regulation in mussel haemocytes.
5-HT1: type 1 serotonin receptor; 5-HT: serotonin; PROP: propranolol; Gi: inhibitory G-protein; AC: adenylyl cyclase; FSK: forskolin; cAMP: cyclic-AMP; PKA(i): inactivated cAMP-dependent protein kinase (PKA); PKA(a): activated PKA (catalytic subunit); dbcAMP: dybutyril cAMP; H89: N-[2-((p-Bromocinnamyl)amino)ethyl]-5-isoquinolinesulfonamide dihydrochloride; CRF: cAMP-responsive factors.   Haemocytes were treated with noradrenaline (NOR) or 5-HT as the physiological agonists known to modulate cAMP levels through adrenergic and serotoninergic receptors, respectively, and both are reported as the main neuromodulators in mussels [15].
Data on the expression and distribution of adrenergic receptors in bivalves is pending, although indirect evidence suggest the occurrence of α-AR and β-AR in all bivalve tissues investigated to date (reviewed in [15]). NOR is the main catecholamine found in bivalves [32]–[35], and NOR levels readily increase in response to different stressors commonly faced by bivalves in their natural habitat [34]. NOR concentrations used in this study (1 µM) are 10-time higher than that reported in circulating fluids under stress conditions [34]. Nevertheless, NOR was ineffective on haemocyte cAMP signaling and ABCB mRNA expression, although Lacoste et al. [34], [35] showed that NOR concentrations ranging from 0.1 to 10 µM decreased phagocytosis and increase hsp70 promoter activity in oyster haemocytes probably through a β- and a α-AR signaling pathway, respectively. It must be stressed that the effect of NOR on the signaling pathway was not assessed in the above studies. According to our results, AR receptors coupled to cAMP regulation are either absent or not responsive to NOR in mussel haemocytes. Conversely, 5-HT significantly decreased cAMP levels in haemocytes.
Mammalian 5-HT receptors are classified into seven distinct classes (5-HT1, 5-HT2, etc.), subdivided in subclasses and coupled to an enormous array of behavioral and biochemical effects [36]. In particular, mammalian 5-HT1 and 5-HT7 are coupled to the inhibition and activation of AC, respectively, while 5-HT2 receptors are coupled to the stimulation of PLC, leading to Ca2+ increases and PKC activation. Phylogenetic studies reported that the main subfamilies of 5-HT receptors diverged early in evolution to form three major subclasses: 5-HT1 (which includes 5-HT5 and 5-HT7), 5-HT2, and 5-HT6 receptor subclasses [37]. This divergence occurred before the branching of vertebrates from invertebrates, and thus finding members of these major subclasses in invertebrate species would be expected [38]. Further division within the 5-HT1 and 5-HT2 subfamilies seemed to have occurred after the branching of vertebrates and invertebrates, and these subtypes have evolved independently within vertebrates and invertebrates [14], [39]. Therefore, it is not surprising to find that pharmacological profiles of molluscan and mammalian 5-HT receptors are difficult to compare, considering the large phylogenetic distances involved [39]. Studies examining both functional and molecular characterization of 5-HT receptors in mollusks reported that 5-HT1 subtypes are reminiscent of an ancestor 5-HT gene that existed before the divergence of the 5-HT receptor subtypes in vertebrates, and kept characteristics of more than one receptor subtype [39], [40]. This particular receptor subtype seems negatively coupled to the cAMP pathway in invertebrates as well as in mammals [14], which is in line with the results from the present experiments. A partial sequence encoding a 5-HT receptor structurally homologous to the mammalian 5-HT1 sub-group was recently obtained from the gonads of another bivalve mollusk, Mytilus edulis [41], and found to be expressed also in digestive gland of M. galloprovincialis [42]. Gene expression analyses reported in this study and that of Ciacci et al. [43] observed the expression of this putative 5-HT1 receptor in mussel haemocytes. Moreover, mRNA levels for this receptor were increased in response to 5-HT stimulation, in agreement with a previous study in the brain of a shrimp by Tiu et al. [44]. These authors hypothesized that in the continuous presence of the agonist, a feedback regulatory mechanism for 5-HT1 mRNA expression in invertebrate cells may exist as new receptors may be needed to replace old receptors, or alternatively, new receptors are needed when all other receptors are occupied [44]. In agreement with the reported findings, both the in vivo effects of FX and the in vitro effects of 5-HT examined in this study lead to an increase of 5-HT1 mRNA levels.
Besides lowering cAMP levels, 5-HT also reduced PKA activities in mussel haemocytes, again in agreement with the occupation of 5-HT1 receptors. Concomitantly, the agonist reduced ABCB mRNA expression. Interestingly, the antagonist PROP prevented the up-regulation of 5-HT1 mRNA expression in vitro as well as in vivo, counteracting all the downstream effects induced by the agonist.
Pharmacological ligands known to increase cAMP levels and/or PKA activities increased also ABCB mRNA expression. These included FSK, which directly activates the enzyme adenylyl cyclase, and dbcAMP, a lipophilic analog of cAMP, which specifically activates PKA activities. The latter effect was blocked by haemocyte pre-incubation with H89, the selective inhibitor of PKA. These findings clearly indicate that the cAMP/PKA pathway exerts a significant control over ABCB transcription. Indeed, in mammalian cells transcriptional regulation of the ABCB1 gene requires the PKA-mediated activation of several transcription factors (TFs) that are phosphorylated by PKA after cAMP activation of the holoenzyme, and bind to consensus sequences in the promoter region of the mammalian ABCB1 gene [28], [45], [46]. PKA has been extensively studied in mussels, showing a high degree of structural and functional homology with its mammalian counterpart [47], [48]. In particular a PKA isoform (namely PKAmyt1) homologous to the human type-I PKA was reported in mussel haemocytes [48]; this protein isoform corresponds to the one involved in ABCB1 transcriptional regulation in mammals [28].
Furthermore, sequence analysis of the proximal 5′-untranslated region of ABCB gene from two mussel species revealed several putative binding sites for PKA-regulated TFs, which also occur within the promoter of the human ABCB1 gene [28]. Although specific studies on TF binding activities and ABCB promoter function in mussels are not available, these findings do support the hypothesis that activity of the mussel ABCB promoter is linked to the cAMP/PKA pathway. Furthermore, the identification of heat shock elements (HSE) in the promoter region of mussel ABCB gene provides structural evidence that the ABCB transcript is co-expressed with HSP70 gene products in response to elevated temperatures or further stressors triggering HSF transactivation [49]–[52].
In conclusion, the current study provides the first clear evidence for the cAMP/PKA-mediated regulation of the ABCB mRNA expression in mussels ( Fig. 8 ). This is a significant finding given the crucial role displayed by the Pgp transporter encoded by the ABCB gene in the response of mussels to environmental stressors and their adaptation to these stressors.
Various studies reported that cAMP levels are increased by metals or other environmental contaminants [9], [53], [54], therefore the induction of cAMP-dependent ABCB transcription may serve as a defense strategy. Nevertheless, this study also demonstrated that ABCB transcription may be negatively modulated by at least an endogenous regulator, i.e. 5-HT, through a highly evolutionally conserved pathway. Therefore, pollutants able to interact with the 5-HT pathway have the potential to negatively affect the cAMP-signaling pathway and impair the ability of mussels to cope with environmental stressors, thus potentially mining their fitness.