M. galloprovincialis and R. philippinarum were collected along the southern coast of the Lagoon of Venice (near Chioggia, Italy) and immediately transferred to the laboratory. Organisms acclimated for 1 week before exposure to Cu(I) ions were reared in aquariums filled with filtered seawater (FSW, 35 psu, pH 8.1), which was changed every other day. They were kept in thermostatic rooms at 19°C, and the water temperatures in the experimental tanks were maintained at constant values using electronic thermostats. The bivalves were fed with Microbe-Lift^®/Phyto-Plus B (Ecological Laboratories, Inc., Cape Coral, FL, United States of America) and microalgae (Isochrysis galbana). The collected bivalves were carefully checked for shell damage (damaged animals were not used in the experiments), and epibionts (i.e., barnacles and algae) were removed from the mussel shells. The experiments were performed outside the period of sexual maturity for mussels (winter to late spring, Da Ros et al., 1985) and clams (late spring to summer, Meneghetti et al., 2004) to reduce the potential for additional stress related to spawning during the experiments and to favor the energy balance towards the growth rate rather than the reproductive rate.

Copper(I) chloride (purified, >99%, code 22433–2) was purchased from Aldrich Chem. Co., Milwaukee, WI, United States of America, as a water-soluble form of copper(I). Various nominal concentrations were obtained from a freshly prepared 10 mM stock solution in FSW, i.e., 0.05, 0.1, 0.5, and 1 μM, corresponding to 4.9, 9.8, 49.5, and 98.9 μg L^-1 (0.9 × 10⁻⁵, 1.8 × 10⁻⁵, 0.9 × 10⁻⁴, and 1.8 × 10⁻⁴% by weight of Cu). Working solutions were always freshly prepared by dissolving drop-by-drop stock solutions in FSW to avoid the formation of precipitates and appeared clear for the entire experimental time. Bathocuproine (2,9-dimethyl-4,7-diphenyl-1,10-phenanthroline sulfonate; 96% pure, code 140910, Aldrich) was used at a concentration of 1 µM as a specific chelator of Cu(I) ions (White et al., 2001) to distinguish between the effects of Cu(I) and Cu(II) ions in the medium.

Hemolymph was collected from the anterior adductor muscle with a 1-mL plastic syringe in the presence of 0.38% Na-citrate in FSW, pH 7.5, as an anticlotting agent. Once collected, the hemolymph was placed into a 1.5-mL Eppendorf tube and immediately centrifuged at 780 × g for 10 min. The pellet was resuspended in FSW and temporarily stored on ice. Sixty microlitres of hemocyte suspension was placed in the centre of a culture chamber or on a Superfrost Plus slide (Thermo Scientific, Gerhard Menzel B.V. and Co. KG, Braunschweig, Germany) depending on the assay. Culture chambers were set up with a Teflon ring (15 mm internal diameter and 1 mm thick), glued to the centre of a siliconized glass slide, and then smeared with Vaseline to be covered with a coverslip (Cima, 2010). Hemocytes were left to adhere to the glass substratum (in the case of the culture chambers, the glass substratum was represented by the coverslip after keeping the chambers upside down) for 30 min at room temperature. Culture chambers were used only for the LC₅₀, lysosome stability index, phagocytosis index and glutathione index assays. For each experiment, three pools of hemocytes from five to six individuals were used.

This assay (Strober, 2015) was employed to determine the median lethal concentration (LC₅₀), which is the compound concentration that is lethal for 50% of the cultured cells exposed for the experimental time (60 min). After adhesion, hemocyte monolayers were incubated with CuCl at 0.1, 1, 10, 20, and 50 μM, corresponding to 9.89, 98.99, 989.97, 1979, and 4929 μg L^-1, respectively. At the end of exposure, the contaminant was discharged by two washes with FSW, and the hemocytes were incubated for 5 min with 0.25% Trypan Blue in FSW. This vital dye is excluded from functional/vital hemocytes but is retained in the cytoplasm of any senescent/dead hemocyte with altered plasmalemma permeability (Gorman et al., 1997). Observations, counting and micrographs of the hemocytes were performed directly on the culture chambers with an Olympus CX31 light microscope (LM) equipped with a Lumenera Infinity 2 digital camera and Infinity Capture Application v. 5.0.0 software (Lumenera Co. 2002–2009, Ottawa, ON). The percentage of dead hemocytes was determined by counting the number of cells with blue-colored cytoplasm, i.e., those unable to exclude the dye, divided by the total number of cells × 100. For the LC₅₀ evaluation, the probit method (SPSS 11.0, SPSS Corp., Chicago, IL, United States of America) was performed. Based on the results obtained, the concentrations used in the subsequent acute toxicity experiments were considered sublethal.

Each experiment was replicated three times (n = 3), and the results are expressed as the average ±SD. The percentages of positive cells were evaluated by counting the hemocytes (at least 200 cells for each monolayer) in 10 optic fields at a magnification of 1000× (0.21 mm view field diameter). Values were normalized to control in FSW (100%). Data from the treated and control hemocytes were compared using one-way ANOVA followed by Dunnett’s test for multiple comparisons with DSAASTAT v. 1.1 2011 (Onofri, 2007). Differences were considered statistically significant when p < 0.05.

Pollutants can act on cytoskeleton proteins that ensure cell motility based on shape changes, cytosol fluidity and pseudopodium formation. Actin microfilaments consist of F-actin that assembles from globular monomeric actin or G-actin according to a model of dynamic instability that requires energy such as ATP (Romet-Lemonne and Jégou, 2021). Even in bivalve hemocytes, microfilaments are responsible for the formation of the cellular projections involved in cell motility and phagocytosis processes, such as pseudopodia and lamellipodia (Olabarrieta et al., 2001).

Phalloidin is a toxin produced by Amanita phalloides that selectively binds to microfilaments, blocking their depolymerization and therefore stabilizing them (Dancker et al., 1975). Therefore, phalloidin can be used in laboratory assays when conjugated with fluorochromes to mark F-actin (Figures 1A,B). These assays highlight the disappearance of cytoskeletal microfilaments due to their disassembly provoked by toxic substances. After exposure to CuCl, a highly significant decrease (p < 0.001) in fluorescence beginning from a concentration of 0.5 μM (49.5 μg L^-1) was found in both M. galloprovincialis and R. philippinarum (Figure 2A). Data on the effects of pollutants on the actin cytoskeletons of bivalves are generally scarce. The effects of exposure to Cu(II) ions have been well documented on only M. galloprovincialis hemocytes. Fagotti et al. (1996) first reported a cytoskeleton F-actin damage effect at concentrations of 5 μg L^-1. Lethal and sublethal concentrations of Cu(II) ions from CuCl₂ dissociation in water (3.17 × 10⁵ to 12.72 × 10⁵ μg mL^-1) (Gómez-Mendikute and Cajaraville, 2003) and 0.02, 0.2, 1, 1.5 and 2 mg L^-1 (Katsumiti et al., 2018), caused deleterious effects on F-actin. According to these authors, the harmful effects consist of the irregular organization of the cytoskeletal filaments of F-actin, which are poorly arranged and lead to decreased phagocytosis activity.

In the cell-mediated innate immune responses of bivalves, phagocytosis is the main defense mechanism against pathogens and foreign materials (Cheng, 1981). This multi-step process involves recognition and adhesion of the target particle to the phagocyte plasma membrane, the ingestion of it in a heterophagic vacuole (phagosome), the fusion of lysosomes with phagosome containing the phagocytic material and the release of the lysosomal hydrolases contained within for intracellular digestion of the engulfed material (Cima et al., 2000).

The phagocytosis index describes the ability of hemocytes to phagocytize target particles such as yeast cells. These target cells are easily phagocytized by control (unexposed) hemocytes, which retain their original size but withdraw pseudopods to assume a spherical shape (Figures 1C,D).

In hemocyte cultures exposed to CuCl, the percentages of M. galloprovincialis and R. philippinarum cells that engulfed the target yeast cells significantly decreased beginning from a concentration of 0.5 μM (49.5 µg L⁻¹) (p < 0.001) (Figure 2B). Similar results have been reported after bivalve hemocyte exposure to Cu(II) ions beginning from concentrations of 50 μg L^-1 for M. galloprovincialis (Torres-Duarte et al., 2019) and 60 μg L^-1 for R. philippinarum (Matozzo et al., 2001). These authors interpreted these observations by analyzing the relationship between the changes in phagocytic activity and alterations in cytoskeleton organization.

During phagocytosis, lysosomes have an important role because lysosomal hydrolases degrade foreign material engulfed inside heterophagic vacuoles. The hemocytes of bivalve mollusks may accumulate high levels of metals, mainly in the lysosomes (Pauley and Nakatini, 1968; Viarengo et al., 1981; Moore, 1990; Bordin et al., 1996). Alterations in the integrity of lysosomal membranes can lead to the release of such enzymes into the cytosol, resulting in damage to the cell itself (Lowe et al., 1995). For these reasons, the stability of lysosomal membranes can be evaluated as a useful biomarker of possible cellular damage through their capacity to retain the NR basic vital dye (Matozzo et al., 2001). NR is a cationic dye that can enter viable cells by pinocytosis or passive diffusion across the plasma membrane and accumulate in acidic intracellular compartments such as lysosomes (Coles et al., 1994). Healthy cells retain NR in their lysosomes that appear with a red-stained content (Figure 1E). Cells that exhibit lysosomal membrane alterations display an entirely red cytosol, since the dye diffuses outside the lysosomes (Figure 1F). Severe alterations in NR uptake have been reported in bivalve hemocytes of the genera Mytilus, Dreissena and Ruditapes collected from animals exposed to cadmium, lead and tributyltin (Coles et al., 1994; Gorman et al., 1997; Matozzo et al., 2001). These alterations to lysosomal membranes might be attributable to the high levels of metals that can accumulate in lysosomes, leading to the inactivation of Mg²⁺-ATPase, a proton pump that maintains the acid gradient inside the organelles. Dysfunction of this pump allows NR to pass from the inside of the lysosome to the cytosol. Alternatively, the accumulated metals can cause lipid peroxidation, leading to the destabilization of lysosomal membranes (Matozzo et al., 2001).

In the present study, a significant (p < 0.001) increase in the release of NR towards the cytoplasm from lysosomes was observed in the presence of 1 μM (98.9 μg L^-1) and 0.5 μM (49.5 μg L^-1) CuCl (Figure 2C) in M. galloprovincialis and in R. philippinarum, respectively. The results of previous studies utilizing this assay demonstrated a significant increase in toxicity by Cu(II) ions to M. galloprovincialis hemocytes, starting from a concentration of 1 mg L^-1, which corresponds to a concentration of 7.43 µM (Katsumiti et al., 2018).

β-Glucuronidase is a lysosomal enzyme that hydrolyses carbohydrate polymers. It is widespread in bivalve mollusks, and its activity is used as a biomarker of pollutant toxicity (Cima et al., 1999). The cytoenzymatic assay (Hayashi et al., 1964) allows the in situ localisation of the activity of this enzyme (Figure 1G). β-Glucuronidase acts on the synthetic substrate naphthol AS-BI β-glucuronic acid by hydrolysing it. This reaction generates a product that binds with the diazonium salt pararosaniline, forming an insoluble red precipitate that makes the occurrence of the reaction identifiable within the cell (Cima, 2017). After exposure to CuCl, a significant (p < 0.001) decrease (Figure 2D) in the in situ enzyme activity was observed in the hemocytes of both species beginning from a concentration of 0.5 μM (49.5 μg L^-1).

The actions of the lysosomal transport chain and lytic enzymes on the degradation of pollutants produces reactive chemical species (Galloway and Depledge, 2001), which are unstable and may have strong oxidizing actions on other cellular structures (Livingstone, 2001; Valavanidis et al., 2006). Reactive oxygen species (ROS) are molecules such as hydrogen peroxide (H₂O₂), the hydroxyl radical (^·OH), the superoxide anion radical (O·^- ₂) and the peroxyl radical (ROO^·) that are generally produced during basal cell metabolism. However, a small percentage (2%–3%) of ROS can leave the compartments where these reactions occur and cause oxidative damage to various molecules essential for cell viability, such as phospholipids, proteins and nucleic acids. Therefore, cells develop a series of molecules and enzymes with antioxidant capacity that are able to react with ROS and prevent oxidative damage. Some pollutants may cause increased ROS production, exceeding the antioxidant capacities of the cell. The imbalance between the production and neutralization of ROS in the cell is called oxidative stress, and has become an important evaluation parameter in environmental toxicology (Kaloyianni et al., 2009). Consequently, changes in the activities of antioxidant molecules and enzymes are useful indexes in the evaluation of the toxicity of an environmental contaminant. In the hemolymph of M. galloprovincialis, extensive oxidative stress was indicated by various immunohistochemical parameters after exposure to metals like cadmium (Koutsogiannaki et al., 2014) and copper (Torres-Duarte et al., 2019) and a significant increase in the production of ROS, followed by oxidative damage identified by DNA fragmentation, protein carbonylation and lipid peroxidation, has been reported.

The superoxide anion radical is a main type of ROS produced in oxidative stress induced by pollutants and is the precursor of most ROS. Under normal conditions, the superoxide anion radical plays a positive role because it is a mediator in oxidative chain reactions (Bus and Gibson, 1982), as it can reduce cytochrome c in the intermembrane space of mitochondria (Wegerich et al., 2013). The mitochondrial matrix contains a specific form of superoxide dismutase (SOD), which then eliminates the formed superoxide anion radical (Turrens, 2003), but an excess of this radical due to pollutants can overpower its scavenging activity. Increased mitochondrial reactive oxygen species production, superoxide anions included, is responsible for the disruption and dysfunction of (COX), which is the terminal oxidase in the mitochondrial electron transport chain and is pivotal for ATP synthesis (Srinivasan and Avadhani, 2012). In the present paper, intracellular superoxide anion production revealed as formazan precipitates (Figure 1H) significantly (p < 0.001) increased after exposure to 1 μM (98.9 μg L^-1) and 0.5 μM (49.5 μg L^-1) CuCl (Figure 3A) in M. galloprovincialis and in R. philippinarum, respectively. The effects on ROS production by pollutants often show contradictory results depending on the exposure conditions used.

As an example, in previous experiments exposing M. galloprovincialis hemocytes to CuCl₂ for 24 h, a significant decrease in intracellular superoxide anion production was reported (Gómez-Mendikute and Cajaraville, 2003). Most ROS are very reactive, have a very short lifetime and decompose very quickly even in the absence of external agents. In this scenario, the presence of low levels of superoxide anion radicals found might be considered in connection with a long exposure time to pollutants. Nevertheless, the progressive increase in oxidative damage is related to self-sustaining and amplifying actions triggered by ROS. A free radical can give up its unpaired electron to a nonradical, which, in turn, becomes a free radical capable of extending and propagating oxidative damage through a chain process. Therefore, the formation of a free radical often represents the initial, triggering event of the production of various molecules capable of damaging the normal redox state of the cell (Zorov et al., 2006).

The enzymatic index of COX was evaluated by determining the percentage of positive cells containing mitochondria that turned brown in the presence of the active enzyme (Figure 1I). After exposure to CuCl, this enzyme was significantly (p < 0.001) inhibited (Figure 3B) beginning from a concentration of 0.5 µM (49.5 μg L^-1) in both species, indicating oxidative stress induced by the overproduction of ROS. This result is in agreement with previous observations on R. philippinarum hemocytes exposed to CuCl₂, which revealed a decrease in enzymatic activity at the concentration of 110 μg L^-1 (Matozzo et al., 2001). The mechanism of action of copper in mitochondrial metabolism and function is complex and only recently has been explained (Ruiz et al., 2021). At low concentrations, copper is essential for the assembly of COX and SOD inside mitochondria, which regulate a physiological ROS production promoting mitochondrial turnover and biogenesis. However, high levels of ROS like that provoked by Cu(II) complexes used in antitumor treatment in mammals are responsible of mitochondrial toxicity and DNA interactions, inhibiting cell proliferation and producing apoptosis.

Among molecules with antioxidant capacity, reduced glutathione (GSH) is a cytoplasmic tripeptide present in all animal cells. Its protective function against oxidative stress is due to the presence of a thiol group, which is reactive with ROS, decreasing their toxicity to the cell. By reducing other molecules, GSH is converted into its oxidized form, GSSG, which is able to react with electrophilic oxidants through a redox reaction. Thus, cell survival depends on the GSH/GSSG ratio. Glutathione can be considered an important biomarker that signals oxidative stress in the cell because the amount of GSSG depends on stressful conditions accompanied by the high production of ROS. When the cells are exposed to pollutants, the enzymes used to reduce GSSG to GSH, such as glutathione reductase and glutathione transferase are unable to convert the large amount of the oxidized form into the reduced form, and the cell undergoes oxidative stress (Liu et al., 2022). Chlorobimane allows the evaluation of GSH content in hemocytes, as the cytoplasm emits blue fluorescence when excited by UV light (Figure 1J). This assay showed a significant (p < 0.001) decrease (Figure 3C) in the number of fluorescent cells beginning from a concentration of 0.5 μM (49.5 μg L^-1) CuCl in both M. galloprovincialis and R. philippinarum, suggesting that this contaminant can alter the oxidation state of cytoplasmic GSH with an increase in ROS production.

The phenomenon of apoptosis, or programmed cell death, is characterized by cell shrinkage and nuclear pyknosis and is accompanied by alterations in the integrity of the plasma membrane in the absence of inflammatory responses (Elmore, 2007). This irreversible phenomenon is considered to be related to both alterations in the homeostasis of cytosolic Ca²⁺ (Marinovich et al., 1996) and the induction of oxidative stress (Kannan and Jain, 2000).

The degree of chromatin fragmentation, i.e., the last event before cell blebbing, was evaluated by means of the TUNEL reaction (Figure 1K). Significant increases in the number of M. galloprovincialis and R. philippinarum cells with fragmented DNA were observed after exposure to 1 μM (98.9 μg L^-1) CuCl (Figure 4A). At this concentration, the hemocytes became small, shrank, and underwent blebbing.

Evidence from several species of mollusks has supported the hypothesis that the induction of apoptosis in hemocytes is a cellular response to exposure to environmental pollutants such as herbicides, insecticides and drugs (Mitali et al., 2013; Ewere et al., 2021; Martyniuk et al., 2022). As an example, the gastropod Lymnaea stagnalis responds to exposure to various toxic substances by activating an apoptosis program involving lectins and increased ROS production (Russo and Madec, 2007). In particular, regarding the effects of Cu(II), a dose-dependent increase in apoptosis induced by oxidative stress has also been reported in the hemocytes of the bivalve Perna canaliculus after in vitro exposure (Nguyen et al., 2018).

At the same CuCl concentration able to induce apoptosis, a sustained increase in cytosolic calcium was observed in both M. galloprovincialis and R. philippinarum (Figure 1L; Figure 4B). This event consequently gives rise to a series of Ca²⁺-dependent toxic mechanisms in the cell, such as cytoskeleton disassembly, calmodulin inhibition, and endonuclease activation (Orrenius et al., 2003). These alterations could be caused directly by the binding of metal particles to proteins, followed by protein denaturation, or indirectly by a change in calcium metabolism due to oxidation of the thiol groups of cytoskeletal proteins by the increased presence of ROS. Viarengo et al. (1994) assumed that impairment of Ca²⁺ homeostasis observed in the hemocytes of Mytilus edulis is the consequence of a direct effect of the metal on Ca²⁺ channels. However, heavy metals, including copper, have affinities for the SH groups of proteins such as Ca²⁺-ATPase. Inactivation of this protein could lead to imbalances in calcium homeostasis and increased cytosolic Ca²⁺, causing severe alterations in the cytoskeleton and affecting important functions such as adhesion, motility and phagocytic capacity (Viarengo et al., 1993; Matozzo et al., 2001).

In the literature, data on the immunotoxic effects of copper to mussel and clam hemocytes have only considered Cu(II). A comparison of the toxic effects observed with Cu(I) in this study with previously reported results on the same specific biomarkers after bivalve hemolymph exposure to Cu(II) is summarized in Table 1.

The comparison of the responses after exposure to various Cu(I) concentrations of the two bivalves reveals that the immune system of M. galloprovincialis appears more resistant than that of R. philippinarum. In particular, the toxic effects of ROS production and lysosomal membrane permeability alterations were higher in R. philippinarum (0.5 µM compared to 1 µM of M. galloprovincialis). Concerning the differences in the mechanism of immunotoxicity at the subcellular level, Cu(I) generally shows significant detrimental effects on the cytoskeleton, stability of the lysosomal membranes, and increases in cytosolic Ca²⁺ at much lower concentrations than Cu(II). These different behaviors suggest the potential enhanced toxicity on molecular events by copper ions in the lowest oxidation state (Beswick et al., 1976; McBrien, 1980). The absence of toxic effects in the presence of bathocuproine, a Cu(I) chelator, confirms that toxicity is due to the presence of Cu(I) ions, and not Cu(II), in the medium, possibly due to its oxidation.