Comparative study of the sensitivity of two freshwater gastropods, Lymnaea stagnalis and Planorbarius corneus, to silver nanoparticles: bioaccumulation and toxicity

Metal-based nanoparticles (NPs) are considered detrimental to aquatic organisms due to their potential accumulation. However, little is known about the mechanisms underlying these effects and their species-specificity. Here we used a stable silver (Ag) NPs (20 nm) with a low dissolution rate (≤ 2.4%) to study the bioaccumulation and biological impacts in two freshwater gastropods: Lymnaea stagnalis and Planorbarius corneus . Ag bioaccumulation showed dose-related increase with enhanced concentration in both species after 7d exposure. L. stagnalis displayed a higher accumulation for AgNPs than P. corneus which could due to the more active L. stagnalis have greater contact with suspended AgNPs. Furthermore, hepatopancreas and stomach were more targeted organs than kidney, mantle and foot. The hemolymph rather than hepatopancreas appeared more susceptible to oxidative stress elicited by AgNPs. Comparison with impacts elicited by dissolved Ag revealed that the effects observed on AgNPs exposure were mainly attributable to NPs. These results highlighted the relationship between the physiological traits, bioaccumulation, and toxicity responses of these two species to AgNPs and demonstrated the necessity of species-specificity considerations when assessing the toxicity of NPs.


Introduction
Invertebrates comprise more than 99% of global species richness with huge diversity (Ruppert et al., 2004). In aquatic ecosystems, invertebrates are crucial components of food webs and fulfil many ecosystem services, such as being food sources and aiding nutrient cycling (Ruppert et al., 2004). Their species distribution, abundance and diversity convey information about environmental health and ecosystem sustainability (Prather et al., 2013).
Gastropods, in particular, are increasingly used as the excellent biomonitors for environmental pollution due to their 1) ubiquitous presence in the aquatic system; 2) capability of accumulating rather high levels of contaminants; and 3) sedentary lifestyle (Baroudi et al., 2020;Caixeta et al., 2020).
Among this taxonomic group, the freshwater gastropods Lymnaea stagnalis and Planorbarius corneus are ubiquitous across Europe and Asia, with the former being widespread also in North America. Those two gastropods are similar in size (up to 5 cm), and as pulmonate gastropods, they both rely on cutaneous (via skin) and aerial (via pneumostome, which transfers oxygen into the lung) respiration (Meshcheryakov, 1990). However, they have different shell morphology and P. corneus presents a deeper and longer diving behaviour compared with L.
stagnalis (Bekius, 1971). Moreover, these species are considered good environmental pollution indicators given their wide geographic distribution, ease of availability, and sensitivity toward chemicals (Amorim et al., 2019;Châtel et al., 2020;Herbert et al., 2021). The L. stagnalis has been used as an excellent biological model for ecotoxicological studies and applied in OECD test guidelines (Amorim et al., 2019;Ducrot et al., 2014). Notably, interspecific variation should be considered when comprehensively evaluating environmental risks. For example, freshwater gastropods Biomphalaria glabrata showed much higher nerve injuries than P. corneus upon pesticide exposure (Garate et al., 2020), and the freshwater amphipod Gammarus pulex accumulated more diclofenac than Hyalella azteca (Fu et al., 2020), another freshwater amphipod. Although those researches enriched information for risk assessments, how the specific species variations exactly contribute to the disparities to pollutants are little known.
Manufactured nanoparticles (NPs) are extensively used in various commercial products.
Silver (Ag) NPs, in particular, display great antimicrobial and antiseptic properties, constituting over 55% of the world's NPs production (Walters et al., 2014). The massive utilization of AgNPs raises the probability of their release into the aquatic environment, which could threaten ecological resources (Kaegi et al., 2010). The physicochemical behaviour of NPs in the aquatic environment follows a complex process. For metallic NPs, dissolution and aggregation are critical transformation processes of AgNPs in the aquatic environment (Zhang et al., 2016).
One recent study demonstrated that AgNPs exhibited lower toxicity to Daphnia magna in surface water than in the M4 culture medium (Hu et al., 2018). Thus, investigating AgNPs' behaviour in conditions relevant to natural water is critical to understanding their bioavailability and biological effects. For AgNPs that can release silver ions (Ag + ), another crucial question is whether the toxicity is mediated by the particle reactivity and characteristics or via the released ions (Shen et al., 2015;Walters et al., 2014). On the one hand, some studies demonstrated that Ag + are exclusively responsible for the toxicity of AgNPs to the nematode Caenorhabditis elegans (Yang et al., 2012), D. magna (Shen et al., 2015) and the copepod Amphiascus tenuiremis (Sikder et al., 2018). On the other hand, the AgNPs-induced toxicity appeared to originate from the particles, rather than Ag + for the freshwater gastropod Bellamya aeruginosa (Bao et al., 2018) and the oyster Saccostrea glomerata (Carrazco-Quevedo et al., 2019;Yang et al., 2012). Thus, particle-specificity should be considered besides the dissolved ion's effect when thoroughly evaluating NPs impacts on organisms. Furthermore, the interspecific variations may also be one of the reasons for the differing outcomes, which have hardly been addressed thus far.
In this study, we used AgNPs as model NPs to compare the sensitivity of L. stagnalis and P. corneus to environmental chemical compounds. The Volvic ® mineral water was used to simulate natural pond water due to the similar chemistry (Auffan et al., 2018;Auffan et al., 2020). Bioaccumulation is critically important to link the environmental chemistry of NPs to biological responses (Croteau et al., 2014;Lekamge et al., 2018). First, we compared the bioaccumulation of AgNPs in the two species i.e. L. stagnalis and P. corneus at different concentrations of AgNPs environmental exposure (from 0 to 500 μg/L). Second, we elucidated potential biological responses (immune responses and oxidative stress) caused by different Ag forms (i.e., NPs and Ag + ) in the two species. Given that the two selected gastropods exhibit different species traits, we hypothesized that they will show different AgNPs bioaccumulation patterns and thus dissimilar accumulation in tissues and translocation to the circulatory fluid, followed by consequences of alterations of the metabolic state of hepatopancreas and hemocyte functions.

Nanoparticle Chemical and Colloidal Stability
Citrate-coated spherical AgNPs (Biopoure) with a primary size of 18.4 ± 2.4 nm at 1.01 mg Ag per mL in 2 mM citrate was purchased from Nanocomposix (San Diego, USA). To better understand the physical-chemical transformation of AgNPs and determine the colloidal stability of AgNPs and their kinetics of dissolution over time (2h, 1d, 2d, and 7d) without organisms, a preliminary exposure to Cit-AgNPs in 100 mL Volvic ® water in 400 mL glass beakers was performed. AgNPs were suspended in Volvic ® water at predicted environmental and supra-environmental concentrations (100 and 500 μg/L, respectively) (Zhao et al., 2021) and beakers were covered with aluminium foils to prevent evaporation and exclude dust.
At each sampling time, the hydrodynamic size and ζ-potential of the AgNPs suspensions were measured using a NanoZetaSizer (Malvern Instruments Inc., UK). Moreover, silver released from the particles was measured at different time points using inductively coupled plasma-mass spectrometry (ICP-MS, Agilent 7700x, Agilent Technologies, USA). To this end, 4 mL of the particle suspensions were sampled and placed in an Ultra Centrifugal Filter Unit (Amicon ® ) and centrifuged at 7500 g for 30 min at 4°C in a microcentrifuge (Sigma 3-16KL; Sigma). The 2 ml collected supernatant was diluted two folds in ultrapure water with 1% v/v nitric acid before ICP-MS analysis.

Gastropod collection and acclimation
L. stagnalis and P. corneus were collected from a protected aquatic garden (47° 22' 18.413" N,5° 48' 24.439" E,Autoreille,France). Gastropod individuals were allowed to acclimatize in the laboratory a climate-controlled room at 20 °C under a 16:8 h light/dark artificial photoperiod regime for two weeks. A quarantine period was applied for several days before acclimatization. Fresh lettuce was fed to gastropods once daily during the acclimation period. Before exposure, L. stagnalis and P. corneus of comparable sizes (length: 2.95 ± 0.20 cm and 2.88 ± 0.29 cm, respectively) were collected from among free-floating gastropods and transferred to aquariums filled with Volvic ® water for one day; food was withheld during this period to depurate their guts.

AgNPs exposure and sample collection
To assess the gastropods' responses with increasing duration of exposure at environmentally relevant AgNPs concentrations, gastropods were exposed to 0, 10, 50, 100, 250, 500 μg/L or Ag + (11.96 μg/L, Ag + released from AgNPs of 500 μg /L at 7d, determined as described in section 2.1). The exposure medium was prepared in 1 L glass beakers with Volvic ® water. Six gastropods were added per exposure condition (42 gastropods for each species in total). Two types of endpoints were assessed in the gastropods following exposure to AgNPs: 1) bioaccumulation and 2) biological responses. The exposure time (7d) was chosen because this time had previously been shown to allow for Ag accumulation (Croteau et al., 2011b;Silva et al., 2022) and could trigger biological responses of freshwater gastropods to metal NPs (Ma et al., 2017). Gastropods were not fed during the whole exposure to minimize faecal scavenging. No mortality in both species of gastropods was observed for all treatments during the exposure.
After exposure, the six gastropods of the same species from each exposure condition were rinsed with deionized water and wiped with facial tissue. Firstly, hemolymph (approximately 500 µL per gastropod) was collected with the micropipette tip by tickling the gastropod's foot sole (Boisseaux et al., 2016). The hemolymph was pooled and collected in conical tubes and kept on ice to minimize cellular adhesion. The pooled hemolymph preparations were then divided into two portions for bioaccumulation and biochemical assay analysis. For the biochemical assay, once sampled, hemolymph was immediately processed for measurement of total hemocyte counts, hemocyte mortality, and reactive oxygen species (ROS) generation, while the remaining hemolymph was stored at -80 ℃ for the enzymatic activity measurement.
After hemolymph sampling, three gastropods were randomly selected for tissue Ag determination. To prevent potential inadvertent metal contamination, labware and vials were soaked for at least 24 h in acid (15% HNO3 and 5% HCl), rinsed several times with ultrapure water and dried under a laminar-flow prior to use (Croteau et al., 2011a). The gastropods were plunged into hot water (70°C) for 90s to prevent contraction before dissection (Schols et al., 2021). Hepatopancreas, stomach, kidney, mantle and foot were dissected and frozen at -20 °C before Ag bioaccumulation analysis. The remaining three gastropods per treatment were placed individually and snap frozen at -80℃ for the biochemical assays.

Ag bioaccumulation
Ag concentrations in the different tissues and hemolymph were measured using ICP-MS (Liu et al., 2017). Gastropod tissues were freeze-dried in a vacuum freezing machine at -80 °C for 72h (Christ® Gamma 1-16/2-16 LSC, Germany). Dried samples were ground to a homogenous powder, then weighed and digested with 1 mL 65% HNO3 and 1 mL 30% H2O2 in Teflon vessels at 120 °C overnight. For hemolymph, 100 μL were digested with 1 mL HNO3 and 1 mL H2O2 in Teflon vessels at 120 °C overnight. After cooling down to room temperature, the remaining solution was collected and diluted with MillQ water to 3 mL and stored at 4 °C before ICP-MS determination.

Biomarkers of oxidative stress
Hemocytes were diluted in 500 μL Phosphate Buffer Saline solution (pH 7.4, Gibco, UK), and used for cytometry analysis. Intracellular ROS generation was detected using the CellROX ® Green Reagent (Life Technologies Europe B.V., Zug, Switzerland) via Microplate Reader (Synergy H1, Bio Tek) with fluorescence excitation at 485 nm and emission at 520 nm.
The total protein content was measured using Protein Quantification Kit-Rapid (51254, Sigma-Aldrich, USA).

Immune response
Total hemocytes were counted on the hemocytometer under the microscope (Olympus BX-UCB). Hemocyte mortality was assessed using the propidium iodide (PI, P3566, Invitrogen, USA) exclusion method (Boisseaux et al., 2016). Briefly, cells were first adjusted to 50 000 cells/well with a final volume of 200 μL using Snail Saline Buffer (5 mM HEPES, 3.7 mM NaOH, 36 mM NaCl, 2 mM KCl, 2 mM MgCl2, 4 mM CaCl2; 145 mOsm; pH = 7.9) prewarmed at room temperature, in a 96-well U bottom sterile polystyrene plate (Greiner,Germany) (Boisseaux et al., 2016). A volume of 2 µL of propidium iodide (1 mg/ mL) were then added for the discrimination of dead cells. The plate was shielded from light and placed on a shaker under slow agitation for 10 min. Hemocyte mortality was calculated as the mean fluorescence intensity (MFI) using the Microplate Reader (Synergy H1, Bio Tek) with fluorescence excitation at 535 nm and emission at 617 nm.

Statistical analysis
Statistical analysis was performed using GraphPad Prism 9.3.1. For comparison with the control, a one-way analysis of variance (ANOVA) was performed followed by a Dunnett's multiple comparisons test. Regression analysis in bioaccumulation (concentration-response calculation) was performed using a four-parameter non-linear regression equation. For comparison between AgNPs and Ag + effects, an unpaired t-test was applied. Results are presented as the mean and standard error of the replicates of each experiment. Multivariate analysis was conducted by Principal Component Analysis (PCA) to identify factors explaining observed variance, and correlation analysis to identify associations between multiple variables.
The R project 4.1.2 was used to visualize PCA and correlation matrices, using the packages Factoshiny and Corrplot, respectively.

Physicochemical properties and ion release of AgNPs
The characteristics of the AgNPs in the stock suspension as provided by the manufacturer can be found in Fig.S1. The Z-averaged hydrodynamic diameters of the AgNPs after 2h, 1d, 2d and 7d of suspension in Volvic ® water are shown in Fig.1a. Generally, the hydrodynamic sizes ranged from ~25 nm to ~50 nm, where smaller sizes occurred at 500 μg/L. The ζ-potential of AgNPs in all the suspensions was around -10 mV (Fig.1a). AgNPs are negatively charged (-39 mV in the stock solution) but rich divalent cations (e.g. Ca 2+ and Mg 2+ ), present in Volvic ® water ( Fig.S2), can neutralize the surface charge of the NPs.
In the present study, the increased dissolution (from 0.69% to 2.4%, 11.97 μg/L at 7d) was observed in 500 AgNPs μg/L suspensions over time (Fig.1b). However, the dissolution of 100 μg/L remained largely unchanged (from 0.49 % to 0.65 %). At each time point, the % of released Ag + in 500 AgNPs μg/L suspensions was higher than that in 100 μg/L (Fig.1b). In contrast, previous studies reported that the dissolution rate was much higher for 10 μg/L AgNPs (5.5%) than for 100 μg/L (2.47%) in deionized water after 96 h exposure, respectively (Ali, 2014;Ali et al., 2014). Notably, studies have suggested that Ag + could complex with Clin aqueous solutions, which decreases the measured Ag + (Chen et al., 2013). Hence, the lower dissolution at 100 μg/L in the present study was possibly due to the low ratio of Ag + to Cl -(from Volvic water medium) compared with that at 500 μg/L. Furthermore, in our case, the smaller aggregate size at 500 μg/L likely contributed to the maintenance of a higher specific surface area of AgNPs and greater dissolution.

Bioaccumulation in tissues and hemolymph
Generally, NPs bioaccumulation studies so far focused on the whole soft tissue of gastropods, i.e., CuONPs in L. stagnalis and Potamopyrgus antipodarum, and AgNPs in Peringia ulvae (Croteau et al., 2014;Khan et al., 2015;Ramskov et al., 2014). In our present study, the Ag content was measured in various tissues (hepatopancreas, stomach, kidney, mantle and foot) and biofluid (i.e., hemolymph) after 7d exposure, allowing to examine Ag's distribution profile in different tissues (range from 0 to 500 µg/L). The tissues were chosen on the basis that they have previously been shown to be target organs for NPs accumulation (Caixeta et al., 2020;Kuehr et al., 2021) and because of the important role of hemolymph served in the immune system (Iwanaga and Lee, 2005).
In the present study, the bioaccumulation differences between two species was observed, all tissues of L. stagnalis showed a higher Ag content than that in P. corneus (Fig.2). For example, at 50 µg/L AgNPs exposure concentration, L. stagnalis accumulated around 14-fold and 9-fold more Ag compared to P. corneus in hepatopancreas and stomach, respectively ( Fig.2a and b). This difference in Ag bioaccumulation could be attributed to differences of their species' traits. Indeed, we observed that the P. corneus prefers to stay at the bottom of the baker during the experiment, while the L. stagnalis spends most of the time climbing higher or floating to the surface (Video S1 in the online SI). When L. stagnalis moves to the water surface, they open the pneumostome fully for aerial respiration and contract mantle muscles to expel gas from the lung (Nargeot and Puygrenier, 2019). Notably, AgNPs could remain suspended in natural freshwaters, stabilizing particles against agglomeration (Chinnapongse et al., 2011), which is also the case in our experimental condition (Fig.1a). Therefore, the more active L. stagnalis would be expected to have greater contact with NPs, which might explain the differential accumulative result. Indeed, different physiological traits might contribute to the bioconcentration outcomes variance between and even for the same species. For example, different residence times on the surface resulted in the considerable variance of the pesticide chlorthion to accumulate within L. stagnalis (Legierse et al., 1998), confirming the substantial role of physiological traits in bioaccumulation processes.
In addition to the difference of bioaccumulation between the two species, the varied pattern among organs was also observed. The dose-response curve of bioaccumulation in different tissues as a function of AgNPs exposure concentration shown in Fig.2, demonstrates that all the tissues showed increasing bioaccumulation with increasing AgNPs concentration.
Notably, for the hepatopancreas and stomach, significant increases were observed from 50 µg/L on ( Fig. 2a and b). However, for the remaining organs (kidney, mantle and foot), higher concentrations (starting from 100 µg/L) of the AgNPs were required to obtain significant bioaccumulation compared to that in control (Fig. 2c, d and e). The disparity of bioaccumulation among organs might be related to their AgNPs uptake order. Ingestion has been commonly recognized as a pathway for NPs uptake into aquatic organisms (Croteau et al., 2014;Kuehr et al., 2021). For gastropods, after ingestion, salivaries and enzymes from intestine aid in the extracellular digestion of ingested feed particles in stomach (Carriker, 1946). Small particles (< 4 µm) are capable of penetrating the pyloric filter and entering the hepatopancreas for subsequent intracellular digestion (Carriker, 1946;Dillon, 2000). As the major metabolic tissue, the hepatopancreas is responsible for xenobiotics biotransformation, redistribution to other tissues and elimination (Bao et al., 2018;Livingstone, 1998). Thus, AgNPs might migrate from the hepatopancreas epithelia to adjacent connective tissue (e.g., mantle or kidney) (Robinson and Ryan, 1988). Hence, compared with other tissues, the hepatopancreas and stomach might be the first organs in contact with the AgNPs. The bioaccumulation and tissue distribution of metal might also be related to the amount of metal-binding proteins and chelates in different tissues (Kuehr et al., 2021). Previous studies reported that hepatopancreas was the major target tissue for NPs accumulation in freshwater gastropods, exemplified by AgNPs and CuONPs exposure to Bellamya aeruginosa, presumably due to a high content of metal-binding proteins. (Bao et al., 2018;Ma et al., 2017).
Studies on bioaccumulation are particularly important to link the environmental chemistry of NPs to biological responses (Garner et al., 2018). Although numerous studies showed that NPs generate toxicity to hemolymph of gastropods (Caixeta et al., 2020), few of them addressed the NP accumulation in this biofluid, with the exception of a recent study where TiO2NPs was visualized in the hemolymph of the land snail Cornu aspersum . It is generally accepted that NPs and their aggregates or agglomerates are able to cross the gut barrier and translocate to hemolymph; endocytic pathways were identified as the primary mechanism of NPs internalization at the hemocyte level (Moore, 2006;Sendra et al., 2020;Shao et al., 2020). In the present study, along with the increase of AgNPs concentration, the Ag content in hemolymph increased, this increase being significant at 500 μg/L in both species (Fig.2f). The Ag content in L. stagnalis was higher than in P. corneus, e.g. 453.34 ± 104.99 μg/L at the highest AgNP exposure, which is nearly 15-fold greater than in P. corneus (30.52 ± 1.47 μg/L). The Ag accumulation difference tendency in the hemolymph of the two gastropods is consistent with that of the tissue profile, which confirms the migration of AgNPs from the tissues to hemolymph.
Although ICP-MS analysis does not allow to distinguish how much of this accumulation originated from AgNPs or Ag + , the low dissolution of AgNPs in the exposure medium ( 2.4%) suggests that most of the accumulated Ag originated from the AgNPs. A previous study showed that polyvinylpyrrolidone-coated platinum nanoparticles (PtNPs) and CuONPs displayed greater bioavailability than their ionic form in L. stagnalis (Croteau et al., 2014;Sikder et al., 2021), which indicates that dissolution of NPs had a negligible influence on metal uptake by freshwater gastropods.

Oxidative stress in hepatopancreas and hemolymph
Chemical bioaccumulation in organisms is generally considered to be a prerequisite for adverse effects (Luoma and Rainbow, 2008). Our findings demonstrate oxidative stress induction by AgNPs in the hepatopancreas and hemolymph of both gastropods (Fig. 3). The hepatopancreas and hemolymph have been used for assessing NPs' effects in gastropods as a detoxifying organ and its important role in the immune system of the invertebrates, respectively (Baroudi et al., 2020;Bhagat et al., 2016;Caixeta et al., 2020;Lekamge et al., 2018). It is known that NPs can induce oxidative stress in aquatic organisms by producing ROS such as superoxide radicals (O2¯) (Caixeta et al., 2020;Lekamge et al., 2018). The antioxidant defense systems has been shown to be important in eliminating ROS (Bhagat et al., 2016). SOD catalyzes the dismutation of O2¯ to oxygen (O2) and hydrogen peroxide (H2O2), and then CAT converts H2O2 to O2 and water. GST not only participates in the transition process from H2O2 to water, but also in the defense against lipid peroxides. MDA (a marker of lipid peroxidation) has been suggested as a sensor against NPs in gastropods . AChE activity is widely used as a biomarker of neurotoxicity for marine invertebrates after metal exposure (Deidda et al., 2021). Notably, no published data on the effect of AgNPs exposure on AChE activity in freshwater gastropods exists.
In hepatopancreas, our results showed that AgNPs exposure induced no lipid damage to both gastropods after 7d exposure, demonstrated by insignificant changes in MDA contents compared to control ( Fig. 3d and i). Similar results were also observed in the hepatopancreas of B. aeruginosa after CuONPs exposure (Ma et al., 2017). SOD, in particular, played an essential role for L. stagnalis and P. corneus to react to AgNPs stress. Results showed that SOD activity in all treatments was lower than those in control, with a significant decrease at 250 and 500 μg/L for both snails (Fig.3a and f). For P. corneus, GST activity was inhibited at all the exposure concentrations (Fig. 5h), indicating this enzyme was also critical to help to respond to oxidative stress. A significant increase in AChE activity was only observed in L.
stagnalis upon exposure to 250 and 500 μg/L AgNPs (Fig.5e). Similar AChE activation after AgNPs exposure was also observed in D. magna (Ulm et al., 2015). Considering the high accumulation of Ag in the tissues, the binding of Ag to AChE might stimulate the catalytic function of AChE (Najimi et al., 1997).
In hemolymph, significant lipid damage (increased MDA contents) was observed in all AgNPs-exposed (except 10 μg/L) L. stagnalis and P. corneus ( Fig. 3o and t). In consonance with our study, previous studies reported that MDA levels of the hemolymph of freshwater gastropods Biomphalaria alexandrina (Fahmy et al., 2014) and L. luteola (Ali, Daoud, 2014) increased after ZnONPs and AgNPs exposure, respectively. AgNPs exposure caused a significant increase in ROS levels at 500 μg/L ( Fig. 3f and p), which is in agreement with previous studies where ROS levels were estimated to increase in gastropod cells after iron oxide NPs treatment (Sidiropoulou et al., 2018). SOD activity significantly increased in the hemolymph of L. stagnalis and P. corneus exposed from 0 to 250 μg/L and 0 to 50 μg/L, followed by a decrease for the remaining concentrations to near the control level ( Fig. 3i and   q). The subsequent decrease of SOD in the present study might be due to the involvement of other antioxidant defence system constituents (e.g., GST increased significantly at the highest concentration, which will be discussed below), which helps to cope with oxidative stress. GST activity was generally increased in the hemolymph of both species at 500 μg/L ( Fig. 3n and s).
Compared with L. stagnalis, more antioxidative enzymes of P. corneus participated in the AgNPs stress responses. CAT decreased at the higher concentrations (250 and 500 μg/L) in P. corneus (Fig. 3r), while no significant change occurred in L. stagnalis. Previous studies also found decreases in CAT activities in freshwater gastropods L. luteola exposed to CuONPs (21 μg/L) for 5d (Ali et al., 2015).
The above results indicated that SOD and GST are more involved in the oxidative stress response of L. stagnalis and P. corneus to AgNPs exposure than other enzymes. Although no lipid damage was observed in the hepatopancreas of both gastropods, AgNPs exposure generated neurotoxicity to L. stagnalis. Furthermore, the hemolymph are more susceptible to oxidative stress originated by AgNPs other than hepatopancreas in L. stagnalis and P. corneus.

Immune responses in hemolymph
The hemocyte count is considered an indicator of innate immunological status of invertebrates (Kacsoh and Schlenke, 2012). Experimental exposure to AgNPs for 7d yielded a significant decrease in total hemocyte count (Fig. 4a), which indicates immune suppression in both species of gastropods. Our results are in line with those reported in the freshwater gastropod, Lymnea luteola, which showed a reduction in hemocyte counts after exposure to AgNPs and TiO2NPs (Ali et al., 2015;Ali et al., 2014). The fluctuation of hemocyte count can be attributed to the alteration of immunocompetence due to exposure of NPs that adversely affected the immune surveillance capability in the organisms. The levels of ROS generated by AgNPs were of biological significance at high concentrations, as shown by a significant increase of hemocyte mortality in 500 μg/L in both species (Fig. 4b). The increased hemocyte mortality as observed in the present study is in accordance with the finding in mussels of L. luteola after AgNPs exposure (Ali, Daoud, 2014).
Proteins are responsible for hemolymph osmotic balance which regulates water distribution in intravascular compartments and directly influences the dynamics of hemolymph flow (Rawi et al., 1995). In the present study, significant decreases in protein content were observed at higher concentrations (250 and 500 μg/L) (Fig. 4c), which is in accordance with the finding in the freshwater gastropod B. alexandrina, after three weeks of ZnONPs exposure (Fahmy et al., 2014). The reduction in protein content may reflect damage in the hepatic parenchyma, which is the origin of protein generation (Rawi et al., 1995). Fig 5, our finding shows that exposure to AgNPs led to the accumulation of total Ag in different tissues and hemolymph and subsequent changes in the biochemical parameters of hepatopancreas (the detoxification organ) and hemolymph. The AgNPs exposure impairs the immune biomarkers and generates oxidative stress in the hemolymph. The highest concentration exposure induced AChE enhancement indicating that neurotoxicity occurred.

Toxic mechanism of AgNPs and Ag +
There is controversy over the toxic mechanism of AgNPs. As some studies indicated (Gao et al., 2021;Shen et al., 2015;Yan et al., 2021), the Ag + was a major contributor to the AgNPs toxicity; however, other studies suggested that the toxicity of AgNPs was attributed to the particles (Bao et al., 2018;Carrazco-Quevedo et al., 2019). Some metal and metal oxide NPs, such as AuNPs and TiO2NPs, even if not capable of releasing ions, could also trigger toxic effects by inducing oxidative stress (Arini et al., 2020;Girardello et al., 2016), suggesting a particle-specific effect. In our study, there were significant differences in biological responses between 500 μg/L and Ag + treatments. Excepting the hemocyte mortality for P. corneus, all the immune and oxidative responses in hemolymph are different ( Fig. 3 and 4). These findings suggested different toxic mechanisms between AgNPs and Ag + . Ag + may be complexed various organic/inorganic ligands abundant in natural water, substantially reducing Ag + bioaccumulation (Azimzada et al., 2017). However, it is undeniable that Ag + in natural water could exert toxicity on organisms (Gao et al., 2021;Shen et al., 2015;Yan et al., 2021). AgNPs' toxicity may originate from the Ag + effect and the particle-specific effect (Xu et al., 2019).
Considering the higher toxicity (e.g., enhanced ROS production and MDA content) induced by AgNPs compared with Ag + , we supposed that the particle-specific effects might play a major role for AgNPs in exerting toxicity on freshwater gastropods.

Principal component and correlation data analysis
PCA was performed on all accumulation and biological variables and on all the treatments for both gastropod species (Fig. 6, Table S1). The biplot of the PCA (Fig. 6a) demonstrated that L. stagnalis and P. corneus form distinct clusters along with the first principal component (PC1), which explained 52.07% of the total variance. The variables contributing the most to this separation are shown in the PCA plot of variables (Fig. 6b) where the direction and length of the vectors display the magnitude and correlation among vectors. The high values in the color scale indicate a high contribution to the PCA. More variations were explained for L. stagnalis than P. corneus, illustrating from the coverd ellipse, which verified L. stagnalis are more sensitive to exposure than the other. It was concluded that Ag accumulation, ROS and AChE are grouped indicating their positive correlation with each other. ROS and MDA change of hemolymph and CAT and AChE of hepatopancreas were the most contributing variables to the separation observed between the two species (Fig. 6b, Table S1).
The correlation analysis between various measured attributes of L. stagnalis and P. corneus is depicted in Fig.7 and Table S2. In both species, a significant positive relationship between Ag content in hemolymph was observed with ROS and MDA, whereas a negative correlation was noted with total hemocyte counts and protein content. In hepatopancreas, the negative correlation of the Ag content with SOD and CAT was recorded. Those findings indicate a great linkage between NPs accumulation and oxidative stress. Positive correlations were found between Ag accumulation in hemolymph and tissues (i.e., hepatopancreas, stomach, kidney, mantle and foot), proving that Ag could be translocated through the body after initial exposure.

Conclusions
The present study provided deeper insight into the differential AgNPs bioaccumulation and immunological and molecular responses of two freshwater gastropods L. stagnalis and P.
corneus using AgNPs as model NPs. The fitness of these two freshwater gastropods was investigated concerning bioaccumulation, antioxidant defence and immunity. As a general trend in both species, Ag bioaccumulation is augmented with an increase of exposure concentration. However, L. stagnalis accumulated higher Ag than P. corneus, which could be reasoned with the differential physiological traits. Hepatopancreas and stomach showed significant increase from 50 µg/L on, which was not the case in the other studied tissues (kidney, mantle and foot). As for the biological response, the hemolymph of L. stagnalis and P. corneus are more susceptible to AgNPs exposure-induced oxidative stress compared to the hepatopancreas. Although no lipid damage was found in the hepatopancreas, neurotoxicity was observed in this tissue. Furthermore, although the effects produced by Ag + should not be underestimated, the Ag particulate form better explained the biological responses, indicating that the NPs themselves play a significant role in exerting toxicity. The results provide valuable information for the toxicity of NPs toxicity on freshwater gastropods, demonstrating that bioaccumulation and the underlying mechanisms of NPs vary depending on the species' sensitivity related to physiological traits. This work also highlights that further studies are needed to better understand the species-specific stress responses.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have influenced the work reported in this paper.  (mean ± SE, n = 3). Different numbers of asterisks represent statistical differences compared with the control: *P < 0.05, ** P < 0.01, ***P < 0.001, ****P < 0.0001. Concentration-response curves were calculated by four-parameter non-linear regression analyses. Fig. 3 The levels of biomarkers in the hepatopancreas (a-j) and hemolymph (k-t) of L. stagnalis and P.

Table S1
Original scaled covariance matrix generated by PCA analysis using R.

Table S2
Full correlation matrix of (a) Lymnaea stagnalis and (b) Planorbarius corneus generated by correlation analysis using R.