Trace elements such as cadmium (Cd) and copper (Cu) occur in the environment either naturally (Bubach et al. 2015) or due to emissions from anthropogenic activities (Vareda et al. 2019). Since Cu has a key role in catalysis and electron transference as a cofactor of enzymes (Messerschmidt 2010), it is readily transported into cells, where it can cause cellular damage via oxidative stress or displacement of cofactors in proteins (Solioz 2018). Conversely, Cd is not an essential element, but it interferes with biochemical pathways by mimicking other essential elements (Bridges and Zalups 2005). These elements can accumulate in living organisms, which in turn act as vectors of contamination through bioaccumulation processes (Primost et al. 2017).

Gastropods have been widely used in biomonitoring studies (Primost et al. 2017) due to their ecological diversity and role in connecting terrestrial and aquatic ecosystems through food chains (Dodd et al. 2016). Physa acuta (Physidae, Draparnaud 1805) is a pulmonate freshwater gastropod found in lentic and lotic ecosystems (Saadi et al. 2020), commonly feeding on detritus and periphyton (Lowe and Hunter 1988). Native to North America, P. acuta has been introduced to other continents, and has expanded its distribution as the most cosmopolitan freshwater gastropod in the world (Dillon Jr. et al. 2002). P. acuta is an outcrossing protandrous simultaneous hermaphrodite. Eggs are laid in gelatinous coiled/curved masses, and individuals hatch as fully developed juveniles (Nuez 2010).

Despite having been used in many ecotoxicological experiments, most publications with chronic sensitivity assessments derived from mortality or growth of the shell of P. acuta were reported in the form of NOEC and LOEC endpoints (Arthur and Leonard 1970; Cheung and Lam 1998; Gao et al. 2017). Furthermore, when LC/EC values were estimated, these usually represented the sensitivity of adult forms (Cheung and Lam 1998; Spehar et al. 1978; Woodard 2005).

Earlier life stages of organisms are generally more sensitive to contamination (Hoang and Klaine 2007), and NOEC/LOEC statistics are progressively considered undesirable statistical endpoints (Warne and van Dam 2008). Considering this, our objective was to perform a comparative assessment between the toxicity of Cd and Cu to neonates of P. acuta based on regression-derived endpoints. We conducted 28-day chronic bioassays to evaluate the effect of Cd and Cu exposure on the mortality and shell length of P. acuta neonates. Furthermore, we conducted a relative sensitivity analysis which allowed us to observe how the sensitivity of P. acuta to Cd and Cu compares with other freshwater snail species sensitivity data.

Materials and Methods

Individuals of P. acuta used in this study came from a well-established population in laboratory aquaria at the Faculty of Exact Science, National University of La Plata. Individuals were originally collected close to La Balandra (34\(^{\circ }\) 56\('\)6.27\(''\)S, 57\(^{\circ }\) 44\('\)31.35\(''\)W), Buenos Aires province in Argentina in 2012. Egg masses were collected from the aquaria and separated under controlled conditions: dechlorinated and oxygenated tap water (conductivity 1.0 mS/cm; hardness 180 mg/L CaCO\(_{3}\); pH range 7.8 ± 0.2; temperature 20 ± 2°C), and photoperiod 16:8 light:darkness. Only 72-h-old neonates (shell length: 0.09 ±0.01 mm) were used to conduct 28-day chronic experiments. Seven concentrations (C1–C7) and a negative control (Ctrl) were tested and run in triplicate with 10 snails per 300 mL polyethylene container (i.e. 30 snails per concentration). Dilutions were renewed every 48 h. Negative controls contained the same water that was used to make treatment dilutions. Snails were fed with 1.5 mL of blended lettuce leaves on each renewal day.

Stock solutions were prepared on each renewal day with double distilled water, using 3CdSO\(_{4}\)·8H\(_{2}\)O (Merck) and CuCl\(_{2}\)·2H\(_{2}\)O (Cicarelli), both ACS analytical grading. Nominal concentrations for Cd were: 0.005, 0.009, 0.022, 0.035, 0.048, 0.061, 0.075 mg/L, and for Cu were: 0.005, 0.007, 0.011, 0.019, 0.026, 0.030, 0.034 mg/L. Sets of samples were randomly collected at different days of renewal for later analysis. Since not every assay replicate and concentration was sampled at each medium renewal event, data of measured and nominal concentrations from each set of samples was used to calculate estimated concentrations by linear regression methods for each assay. Metal concentrations were measured using a Varian Spectr AA 330 atomic absorption spectrophotometer with air-acetylene flame (APHA 1998 Method 3111 B). Quality assurance and control comprised freshly prepared dilutions at the moment of renewal, and the calibration of equipment using certified reference solutions from Accustandard Inc. (Cd:AA08N-1, Cu: AA15N-5). The equipment detection limit for both metals was 0.005 mg/L. All labware was previously cleaned in a 10% HNO\(_{3}\) bath, and stock solutions as well as assay samples were all refrigerated and acidified with HNO\(_{3}\) (Analar) analytical grade for storage. Lethal effects were assessed by recording the number of dead snails every 48 hours, and shell length was measured every 7 days to assess chronic sublethal effects. Snails were considered dead if they did not move and no heart beat was observed after stimulation. Shell length was measured from the tip of the protoconch to the anteriormost end of the aperture with a 1/100 mm micrometer ocular. Endpoints were assessed with a stereoscopic microscope.

Statistical analysis was performed in R Core Team (2013). Lethal concentrations (LC) and effect concentrations (EC) for shell length were derived by fitting log-logistic models with the package drc. Control group mortality of < 10% was considered acceptable. Relative toxicity between Cd and Cu was estimated utilising the function EDcomp(), which models the ratio EC\(_{Cd}\)/EC\(_{Cu}\) as a function of the response level. EC and LC model parameters were compared with the function compParm() (Ritz et al. 2019). We retrieved acute LC values from Bulus Rossini (2013).

Additionally, we estimated LT\(_{50}\) values by fitting a two-parameter multinomial log-logistic model (LL.2, type event in drc), assuming all individuals would die at some infinite time interval. The trend of LT\(_{50}\) as a function of metal concentration in chronic exposure was analysed implementing the metaphor package via weighted linear regression of the pooled mortalities and their respective squared standard errors with the function rma() (Ritz et al. 2019). Model selection was performed based on Lack-of-fit test and Akaike Information Criterion (AIC) against Weibull models. Normality of residuals was assessed via Shapiro test or by visual assessment with QQ-plots or residual plots and homoscedasticity with Levene’s test. When needed, power transformations, or the use of robust standard errors estimates were applied to address model misspecifications.

Relative sensitivity (RS) of P. acuta compared to other freshwater gastropods was calculated as RS = Log(LC\(_{50{Pa}}\)/LC\(_{50{i}}\)), where LC\(_{50{Pa}}\) is the geometric mean of all available data for this species and its synonyms (including results from this study), and LC\(_{50{i}}\) represents the lethal concentration estimates for each i species. Data was collated from the USEPA ECOTOX database, and only compatible endpoints were used for comparisons. LC\(_{50}\) values were normalised to water hardness (100 mg CaCO\(_{3}\)/L) using the pooled regression slopes for Cd (USEPA 2016) and Cu (USEPA 2007). We are aware that Cu can be predicted more accurately via the Biotic Ligand Model (BLM), but given the lack of information on other required parameters, we considered the use of pooled slopes a convenient approach. All graphics were built with the ggplot2 and cowplot packages in R.

Results and Discussion

On average, estimated concentrations (C1–C7) departed 26% and 19% from nominal values for Cd and Cu, respectively, and mortality in both experiments was dose dependant (Table 1).

Table 1 Estimated concentrations of heavy metals in mg/L and their respective mortality (%) per group. Estimated values were derived by linear regression from measured concentrations and nominal values
Full size table

Chronic LC values were derived from a four-parameter model with a free lower limit due to the incidence of one death in one control replicate of each metal assay (i.e. lower limit \(\ne\) 0, mortality < 10%). EC values were derived by fitting a four-parameter log-logistic model with shell length data at day 28. The lack of fit test was not significant (\(p\) > 0.05) in all fitted models, and AIC was the lowest against Weibull models. Shapiro tests for residuals of EC models were not significant for Cd (p = 0.7379) or Cu (p = 0.1839). LC and EC values at 28 days and their respective 95% confidence limits are shown in Table 2. Bulus Rossini (2013) 4-day LC\(_{10}\) for Cd was 0.1720 (CI 0.1314/0.2125 mg/L), LC\(_{50}\) was 0.3849 (CI 0.3325/0.4373 mg/L), and LC\(_{90}\) was 0.8614 (CI 0.5875/1.1354 mg/L). Cu 4-day LC\(_{10}\) was 0.0922 (CI 0.0756/0.1088 mg/L), LC\(_{50}\) was 0.1624 (CI 0.1444/0.1804 mg/L), and LC\(_{90}\) was 0.2860 (CI 0.2358/0.3362 mg/L).

Data compilation from the ECOTOX database for Cu included 109 entries of 36 freshwater gastropod species, and 21 entries for species whic

Table 2 LC and EC values for different response levels at 28-day chronic exposure to Cd and Cu with their respective confidence limits
Full size table

h are now considered synonyms with P. acuta (e.g. P. heterostropha) (Lydeard et al. 2016). Cd data comprised 52 entries for 16 species of freshwater gastropods, of which 5 entries were for P. acuta and synonyms (e.g. P. integra) (Lydeard et al. 2016). Most data comprised values for 2-, 3-, and 4-day acute exposure. Chronic values for Cu included 7 entries from four different species of gastropods, whereas Cd only had 3 chronic LC values for which only one entry belonged to a species other than P. acuta (Das and Khangarot 2010).

Relative toxicity analysis between Cd and Cu for acute and chronic endpoints show that metals are not equally toxic to P. acuta at most response levels (i.e. Cd/Cu ratio \(\ne\) 1) (Table 3), and the toxicity ratio is not constant, increasing with the response level in different ways for acute and chronic effects (Fig. 1a, b).

Table 3 Comparisons between Cd and Cu LC and EC values for different response levels with p values, and their respective confidence limits
Full size table
Fig. 1
figure 1

Relative toxicity between Cd and Cu as a function of response level for a acute and chronic LC, and b chronic EC. Black dots and guide lines show where metals were equally toxic, and arrows show divergence from that point. Subfigure c shows the relationship of LT\(_{50}\) as a function of metal concentration

Full size image

Acute exposure data revealed that Cu had a more pronounced toxic effect than Cd. The Cd/Cu ratio was > 1 for all response levels with an increasing trend (Fig. 1a). The relative sensitivity analysis for acute exposure shows that most available data correspond to 2-day and 4-day exposure times. P. acuta sensitivity to Cu falls in the higher end of sensitivity compared to other freshwater gastropods, with the majority of the LC\(_{50}\) values falling behind that of P. acuta (Fig. 2b). Cd data showed a difference in spread between 2-day and 4-day LC\(_{50}\) values. Most freshwater gastropod species showed a similar level of sensitivity for a 2-day Cd exposure, with P. acuta falling in the middle range of sensitivity compared to other species. Conversely, 4-day exposure values displayed a wider spread, and P. acuta sensitivity sits in the middle-high end of the range compared to other freshwater snail species (Fig. 2a).

With respect to chronic lethal effects, we can observe an increasing mortality with an incipient higher toxicity from Cd, which eventually equates to that of Cu at the 25% response level. For higher response levels, Cu exerted a more toxic effect than Cd (Fig. 1b). This larger toxic effect was reflected in the significant difference detected for the slope parameter between curves (p < 0.001) for which Cu showed a steeper one (Slope\(_{Cd}\)/Slope\(_{Cu}\) = 0.29). Furthermore, LT\(_{50}\) estimates were significantly affected by Cd and Cu exposure (\(p\) < 0.001), with a decrease in lethal time of 5.90% and 8.28% per increment of 0.001 mg/L of Cd and Cu, respectively. Untransformed fitted models can be observed in Fig. 1c, and LT\(_{50}\) estimates are shown in Table 4. Woodard (2005) reported LT\(_{50}\) values of 13.07 days for P. acuta exposed to their lowest tested concentration (0.4 mg/L Cd), however, those assays included larger snail size classes (shell length > 5.6 mm) and softer water conditions. Conversely, Williams et al. (1985) estimated a LT50 of 19.58 days for 0.4 mg/L of Cd in moderately hard water. With regard to LT\(_{50}\) values for Cu, Shuhaimi-Othman et al. (2012) found a similar pattern for Cu toxicity (LT\(_{50}\) 6.81 days for 0.081 mg/L of Cu), which decreased the lethal time to a greater degree than Cd (LT\(_{50}\):11.81 days for 0.61 mg/L Cd) in experiments with Melanoides tuberculata. Lastly, relative chronic sensitivity analysis showed that P. acuta sensitivity to Cd is in the higher range compared to other freshwater gastropods (Fig. 2a), whereas its sensitivity to Cu sits in the middle range of the available data for freshwater snails. Nonetheless, most values fell behind P. acuta LC\(_{50}\) (Fig. 2b). The lowest 28-day LC50 value for P. acuta adults exposed to Cd was reported by Spehar et al. (1978), which after normalisation resulted in a lower dose (0.0104 mg/L Cd) compared to our results. With respect to Cu, Besser et al. (2016) reported 28-day LC\(_{50}\) values for P. gyrina juveniles (0.79-0.85 mm), with values of 0.0420 and 0.0340 mg/L Cu under similar water hardness conditions to our tests. However, Das and Khangarot (2011) obtained a 4-week LC\(_{50}\) value of 0.0159 mg/L Cu for adults of Lymnaea luteola in very hard water. The adjusted LC\(_{50}\) to 100 mg/L CaCO\(_{3}\) was 0.0078 mg/L Cu; a much higher sensitivity than the one estimated from our experiments and from Besser et al. (2016).

Fig. 2
figure 2

Relative sensitivity (RS) calculations for LC\(_{50}\) values from ECOTOX database for a Cd and b Cu. P. acuta chronic LC values were included for Cd comparisons (28*). Positive values represent a greater sensitivity, and negative values correspond to a lesser sensitivity than P. acuta (dashed line)

Full size image
Table 4 LT\(_{50}\) estimations, and their respective standard errors in days for Cd and Cu tested concentration groups
Full size table

In regard to chronic sublethal effects, Cu also showed a significantly steeper slope parameter than Cd (\(p\) < 0.05, Slope\(_{Cd}\)/Slope\(_{Cu}\) = 0.11), implying a more pronounced toxic effect. Results from our 28-day exposure tests showed the highest concentrations of Cd and Cu affected shell length to the same extent. Mean shell length at the highest concentrations at day 28 were 0.9230 ±0.0619 mm and 1.0238 ± 0.0478 mm for Cd and Cu, respectively. Therefore, no significant differences were found (p = 0.2119) between lower limits for Cu and Cd model parameters. Upper limits (i.e. control treatments mean response) did not show significant differences (p = 0.2185), with values of 2.8017 ± 0.0647 mm and 2.9048 ± 0.0415 mm for Cd and Cu experiments, respectively (See Fig. 3a, b). However, we found significant differences between EC values for Cd and Cu (Table 3), suggesting a more toxic effect exerted by Cd, particularly at lower response levels (Fig. 1b). Literature reported 28-day EC\(_{50}\) for growth (shell length) of Lymnaea palustris and L. stagnalis were 0.1420 mg/L and 0.0580 mg/L of Cd and Cu, respectively (Coeurdassier et al. 2003). Furthermore, Dorgelo et al. (1995) estimated an EC\(_{50}\) for the growth (shell length) of Potamopyrgus jenkinsi under Cd (0.016 mg/L) and Cu (0.013 mg/L) exposure. Unfortunately, there were no details about water hardness values for the experiments, but their EC\(_{50}\) estimations seem to suggest a greater sensitivity than those of P. acuta.

Fig. 3
figure 3

P. acuta mean shell length for 28-day exposure of each concentration group for a Cd and b Cu tests

Full size image

Inhibitory effects of Cd and Cu on the growth of freshwater snails have been previously reported (Arthur and Leonard 1970; Cheung and Lam 1998; Das and Khangarot 2011; Gao et al. 2017). Inhibition of the shell mineralisation is a known effect of metal ion exposure in molluscs, since they displace enzyme cofactors that participate in the process of biomineralization, and induce acidosis (Lionetto et al. 2016). Although it is beyond the scope of this study, it would useful to explore the ecotoxicology of biomineralisation in order to understand whether different mechanisms of action could explain differences in toxic response in P. acuta (e.g. EC Cd/Cu ratios as seen in Fig. 1b). For example, Brix et al. (2011) and Grosell and Brix (2009) found that the sensitivity of L. stagnalis to Cu and lead was the result of calcium being outcompeted at the apical membrane level where channels regulate intake. Although Cd toxicity is hindered by calcium in solution, it has been found to interfere with carbon anhydrase (CA) function and the homeostasis of calcium in mussels (Machado and Lopes-Lima 2011). Furthermore, Cu studies show different degrees of inhibition of CAs and specificity for other binding sites too (Lionetto et al. 2016), and histopathological effects such as desquamation and necrosis of the mantle epithelium (Otludil and Ayaz 2020). In saying this, growth inhibition could also be explained by a decrease in feeding rates (Das and Khangarot 2010). Further study is required to understand the influence of water chemistry and speciation on toxicodynamic processes (USEPA 2007, 2016).

The extensive distribution of P. acuta, its suitability to laboratory culture, and its sensitivity range, demonstrate the usefulness of this freshwater gastropod for chronic biomonitoring projects. The results from this study show that P. acuta presents a good range of sensitivity for the assessment of lethal and sublethal effects under the exposure to metal ions such as Cd and Cu. Neonates of P. acuta showed a higher acute sensitivity for Cu than Cd, and compared to other freshwater snails. Conversely, chronic sensitivity to Cd was higher than Cu for sublethal effects, whereas for lethal effects, it sits within the middle range compared to other freshwater gastropods. On a final note, this study also contributes to the generation of regression based sensitivity data, particularly for chronic exposure assessments which are mostly reported as NOEC/LOEC values in freshwater gastropods studies.