Introduction

Domestic, industrial, and hospital wastewaters constitute complex mixtures that will undergo treatments directly through wastewater treatment plants (WWTPs) before they are released in surface waters. All this anthropogenic activity generates a wide variety of chemical micropollutants coming from multiple usages. The few examples encompass food products (e.g., caffeine), pharmaceuticals (Ferrari et al. 2003, Ferrari et al. 2004) (Besse and Garric 2008, Leclercq et al. 2009, Besse et al. 2012, Aminot et al. 2016), pesticides (Gerecke et al. 2002, Köck-Schulmeyer et al. 2013), personal care products (Miege et al. 2009a), surfactants (Schröder 1993) (Ahrens et al. 2009, Yu et al. 2009), composite materials (e.g., PCBs (polychlorinated biphenyl)) (Urbaniak and Kiedrzyńska 2015), polycyclic aromatic hydrocarbons (PAHs) (Mailler et al. 2015, Ozaki et al. 2015), and metal contamination (Gagnon et al. 2014). Treated effluents are then discharged in rivers, and by extension, aquatic wildlife is under chronic chemical pressure stemming from anthropogenic activities. Among the xenobiotics likely encountered, a wide range of known ecotoxic effects is now being investigated by the scientific community. Effluents from WWTPs are known to contain endocrine disruption contaminants (Fent et al. 2008, Miege et al. 2009b, Dagnino et al. 2010, Cédat et al. 2016). Adverse effects on life history traits of ecotoxicological relevant organisms have been evidenced, in particular reprotoxic (Brion et al. 2004, Gust et al. 2010, Gust et al. 2014) and genotoxic effects (Lacaze et al. 2013, Wigh et al. 2016).

A growing body of evidence highlights prospective and retrospective hazards of contaminants on the immune system of non-target organisms (Fournier et al. 2001, Galloway and Depledge 2001, Auffret et al. 2002, Auffret et al. 2004, Gagnaire et al. 2007, Russo et al. 2009, Müller et al. 2009, Segner et al. 2012, Gust et al. 2013a, Boisseaux et al. 2014, Renault 2015, Le Guernic et al. 2016), notably from WWTP effluents (Gagné et al. 2002, Akaishi et al. 2007, Salo et al. 2007, Gagné et al. 2008, Farcy et al. 2011, Gust et al. 2013b). In natural environments, animals encounter multiple stress; among which are chemical pressure and microbiological threats. In this context, the immune system is of particular importance. The immunocompetence of organisms plays a determinant role in their vulnerability to contracting infectious diseases and thus in their health condition. Additionally, immune cells (i.e., hemocytes) are involved in a panel of physiological features of non-immune nature (e.g., tissue and shell repair, clearance of cells, xenobiotic metabolization). Additionally, immune cells (i.e., hemocytes) are involved in a panel of physiological features of non-immune nature (e.g., tissue and shell repair, clearance of cells, xenobiotic metabolization). Furthermore, analysis of the immune parameters of surrogate species may be used as biomarkers of environmental pollution exposure.

The Lymnaea stagnalis snail is a freshwater gastropod living in natural environments surrounding Lyon, France. It is a promising model organism in environmental immunotoxicology, notably because the hemolymph can be collected mechanically with relative confidence as to the non-detrimental effect of the method (Boisseaux et al. 2016a). It has been previously observed that river samples downstream of a WWTP in the Montreal area (i.e., < 0.1, 1, 5, and 10 km) had notable effects on the immune parameters of L. stagnalis following 7 days of exposure (Gust et al. 2013b). The mixing zone ratio of WWTP effluents to the river flow is often higher than 1:500. Hence, the present study was designed to investigate the potential immunotoxicity of WWTP effluents (i.e., before mixing zones in rivers) with the intent of maximizing the anthropogenic multicontamination pressure that wildlife encounters chronically. To that aim, we exposed L. stagnalis to the treated effluents (100%, i.e., not diluted) directly in a flow-through system for 29 days. An intermediate exposure consisting in a 1:8 dilution of treated effluent with groundwater control was tested to determine whether dose responses occurred. Individual time-course assessments of immunocompetence markers were conducted throughout the experiment. All analyses were done on hemocytes, which are the circulating immunocompetent cells of this species. The objective of the study was to both (i) investigate the potential immunotoxicity of WWTP effluents on the L. stagnalis snail and (ii) provide greater insight into the ecotoxicological relevance of the immunomarkers assessed in a realistic multicontamination context.

Material and methods

Experimental design

The experiment lasted for 29 days. The system was run for 1 week before starting the experiment (i.e., before adding snails to their respective conditions).

Three exposure conditions were tested (one per aquarium): (i) water control consisting in clean groundwater (see supplementary material, SM), (ii) 12.5% effluent (mixture with clean groundwater), and (iii) 100% effluent. Effluent refers to the (treated) effluent coming directly from the output of the WWTP on the site before it is discharged into the river. To that aim, the WWTP effluent output pipe disposal was shunted into a 60-L storage tank from which the effluent was pumped (flow-through system). With respect to the groundwater control, a 1-m3 capacity storage tank was brought on the site and regularly filled by transporting water from the laboratory. This water was the same as used for breeding the snails (see SM).

Exposures took place in an experimental hall located next to the WWTP. Three 15-L glass aquaria were disposed in a water bath set at 20 ± 1 °C (monitored with Tynitag recorders) and they were adjusted at 10 L with an outflow disposal. The snails were exposed in a flow-through system. A peristaltic pump (Ismatec BVP) was used to deliver a water flow of 23 ± 1 mL min−1 into each aquarium in order to achieve approximately 3.3 renewals of each aquarium’s water body per day. Neon-tube lighting provided a photoperiod to the snails (16/8 h light/dark daily cycles).

Information on the WWTP and the effluents

With approximately 2.24 million people, Lyon is the second largest urban area in France in terms of population (Insee). The WWTP, from which the municipal effluents came from, is implanted in this urban area. It was built at a treatment capacity of 300,000 population equivalents (i.e., 91,000 m3 j−1). Briefly, this WWTP consists in pretreatment of the municipal wastewater influents (screening, grit removal, fat skimming, and primary lamella settling) and then the biological treatment reduces carbonated pollution by oxygenation and the nitrogen pollution by nitrification/denitrification steps. Sludges are treated separately. To date and to our knowledge, no thorough chemical micropollution analysis of effluents has been done in this WWTP, but it is classified as being in accordance with French regulations for no alert classification (see details in SM).

Snails

A total of 72 snails were selected from the breeding stock of our laboratory according to shell length (37 ± 2 mm) and wet weight (3 ± 0.6 g) calibration criteria. Animals were split into three cohorts of 24 snails of similar calibration criteria, each one assigned to an aquarium (i.e., one cohort per condition, no replicates). The snails’ wet weights were recorded (blotted dry) at day 0 and day 29 (see details in SM). The fold increase in wet weight of each animal was then calculated as the weight at day 29 divided by the weight at day 0. Note that during the experiment, all organisms were fed lettuce ad libitum.

Markers of immunocompetence

The hemolymph (n = 72 snails) was collected at different days by stimulation of the shell retraction reflex provoking extrusion of hemolymph through the hemal pore. Note that for each measurement, if not sufficient quantity of hemolymph was collected, cell density was not determined. Thus, taking into account snail mortality and sample constraints, each measurement was n = 65 ± 5. Hemolymph was thus collected with a micropipette into 500-μL polypropylene tubes (Dutscher, Dumath, France). The lapse of time between collection and further biomarker assessments was minimized and the order of snails analyzed was randomized between conditions and days in order to troubleshoot possible artifacts coming from hemocyte adhesion to the tube, as discussed in Boisseaux et al. (2016a).

Several markers of immunocompetence were evaluated. Immunocapacity was defined as the combination of hemocyte density and viability (i.e., non-functional markers of immunocompetence) and immunoefficiency was defined as the combination of hemocyte-related oxidative activities, phenol oxidase activities, and phagocytosis capacity (functional markers of immunocompetence).

Hemocyte density and viability were assessed at days 0, 3, 10, 20, and 29. The methods employed are described in Boisseaux et al. (2016a). Briefly, they were separately assessed in flow cytometry (Guava easyCyte™ 6-HT apparatus, Millipore). Hemocyte density was determined directly in the tubes without addition of any extra medium in order to be close to the physiological status. For hemocyte viability, cells were adjusted to 50,000 per well replicate (Greiner, polystyrene, U-bottom) and propidium iodide (PI) was used to discriminate dead hemocytes.

Hemocyte oxidative activities were assessed at days 3 and 29. The protocol was adapted from Boisseaux et al. (2016b). Briefly, 50 μL of hemolymph per snail was deposited in duplicate in a 96-well plate (Greiner, polystyrene, flat bottom). After centrifugation, the lymph was replaced with buffer solution. This corresponded to the basal oxidative activity. Additional analysis (other duplicates) was performed by challenging the hemocytes ex vivo with zymosan particles (125 μg well−1). This addition was intended to stimulate the NADPH oxidase activity, which corresponds to the immunocompetent oxidative activity of L. stagnalis hemocytes involved in microbe destruction.

Phenol oxidase activity was assessed at day 10. The method was adapted from Seppälä and Leicht (2013). Briefly, 50 μL was deposited in duplicate in a 96-well plate (Greiner, polystyrene, flat bottom). The activity was monitored by absorbance measurement (480 nm, TECAN microplate reader, 60 min kinetic with 2-min interval measurements) after addition of 50 μL of l-DOPA at 5 mg mL−1 dissolved in SSB prewarmed to room temperature.

Phagocytosis capacity was assessed at day 29. First, hemolymph samples were diluted at 500 cells μL−1 with a final volume of 200 μL per well to reach 100,000 hemocytes well−1. Dilutions were done using SSB prewarmed to room temperature, in a 96-well plate (Greiner Bio-One, for cell cultures, sterile, F-bottom, polystyrene). Samples with a hemocyte density lower than 500 cells μL−1 were not diluted and so 200 μL of the raw sample was added per well. Then, 7.2.105 fluorescent microbeads (polyscience YG 2.0-μm latex microsphere) were gently mixed with each sample, which corresponds to a hemocyte-to-bead ratio of 1:7 for 100,000 hemocytes. The plate was shielded from light and slowly agitated for 24 h in a heat-controlled chamber (Aqualytic, TC Serie 140G). Afterwards, hemocytes were fixed (SSB containing 0.5% formaldehyde (v/v) and 0.2% (w/v) sodium azide) and samples were read in the green photomultiplicator of the flow cytometer (10,000 events were acquired per sample). Phagocytosis capacity was finally calculated as the percentage of hemocytes that engulfed at least one bead.

Statistical analysis

The R software was used (Team 2016) for statistical analysis. Survival curves were plotted using the survival package (Therneau 2015). Data were right-censored because all organisms did not die during the experiment. The logrank test was run to enquire about statistically significant survival differences between conditions. To analyze the weight increase of the snails between exposure conditions, the Kruskal-Wallis test was run. Linear mixed effects models (LMEMs) were fit to the markers of immunocompetence data sets using the lmer function from the lme4 package (Bates et al. 2014). For hemocyte density and viability, models were first built with one fixed effect, the day, and an interaction term of the day with the condition (coded as a quantitative variable). The snail was included as a random effect. Note that beforehand, the data of hemocyte viability—expressed in proportions—were transformed into logit values for normalization of their distribution. For the phenol oxidase activity, models were built with the condition as the fixed effect and the plate series in which kinetics were monitored by spectrofluorometry as the random effect. For the phagocytosis capacity, models were built with the condition as the fixed effect and as the random effect the plate series in which microbeads were mixed with hemocytes. Note that beforehand, the phagocytosis capacity data set was logit-transformed for normalization. For oxidative activities, two independent analyses were done for day 3 and day 29. Models were built with the factor “zymosan” (two modalities, “presence” or “absence”), the condition, and an interaction term between zymosan and the condition. The plate series was included as a random effect. In each model, an effect was considered statistically significant when the 95% confidence interval (CI) of its regression coefficient did not contain the value 0. The unbiased pattern of models, their homoscedasticity, and normality assumptions were verified by looking at the residuals and normal quantile-quantile plots of the respective effects (SM).

Results

Survival and weight increase

Snail mortality was low since all exposure conditions ended with a snail survival higher than 80% (survival curves in Fig. S1, SM). The logrank test did not suggest significant differences among the exposure conditions. Snails exposed to effluents had a higher weight increase than water controls (Fig. 1). The median fold increase in wet weights was approximately 1.15, 1.30, and 1.40 for water control, 12.5% effluent, and 100% effluent exposure conditions, respectively. However, the Kruskal-Wallis test did not suggest any statistically significant difference between conditions (p > 0.05).

Fig. 1
figure 1

Weight increases. This figure corresponds to the fold increase in wet weights of individual snails of respective exposure conditions between day 0 and day 29. At starting the experiment, snails in each condition were homogeneously calibrated by weights (Fig. S2, supplementary material)

Full size image

Immunocapacity

Figure 2 shows the time-course of the snails’ immunocapacity. Over the whole experiment, the mean hemocyte density was 536 cells μL−1 (95% CI, 500–571). These values are similar to those of other immune studies carried out on L. stagnalis (Boisseaux et al., 2016a). No strong effect of effluent exposure was observed over time for hemocyte density (Fig. 2a). However, the 95% CI of the regression coefficient of the interaction term “100% effluent” with “day” (LMEM) was − 0.07 and 7.67 and the corresponding t value was 1.91 (see Table S1 in SM). This is at the extreme limit of the statistical significance chosen (the 95% CI of the regression coefficient contains the value 0). Consequently, it is highly suspected that “effluent 100%” had an effect by increasing the snails’ hemocyte density over time compared to the water control. Complementary analysis was done by coding the day within the model as a qualitative variable. The results suggest that the day effect was significant only at day 3 and day 10. Furthermore, one interaction was detected between day 29 and the 100% effluent condition with a mean (95% CI) of 109.0 cells μL−1 (− 1.8; 220.2). Hence, this strengthens the likely increased effect of the 100% effluent compared to the water control, particularly at day 29 (t-value = 1.89).

Fig. 2
figure 2

Immunocapacity of snails over the course of the experiment. a The hemocyte density and b hemocyte viability which were measured at an individual scale over the course of the experimentation into the three exposure conditions. The vertical gray and dashed lines indicate the days of hemolymph sampling and immunocapacity quantification

Full size image

Regarding hemocyte viability, no strong effect of effluent exposure was observed (Fig. 2b). The mean logit-transformed hemocyte viability was 1.06 (95% CI, 0.90; 1.20), which corresponds to 74.3% of cell viability (95% CI, 71.1; 76.8). These values are consistent with those of other immune studies carried out with L. stagnalis (Boisseaux et al. 2016a). LMEM analysis confirmed the absence of a significant effect of effluents compared to controls. However, a significant negative day effect is suggested over the course of the experiment, but this was independent of the condition (Table S2, SM).

Immunoefficiency

The phenol oxidase-like activity—evaluated at day 10—was higher for the “12.5% effluent” and 100% effluent exposure conditions (Fig. 3a) than for the water control. However, LMEM analysis showed that these differences were not statistically significant.

Fig. 3
figure 3

Immunoefficiency explorations. a The phenol oxidase-like activity of hemolymph samples at day 10. b The capacity of phagocytosis of hemolymph samples at day 29

Full size image

The phagocytosis capacity at day 29 (Fig. 3b) was lower for the effluent 100% condition than for 12.5% effluent and the water control. LMEM analysis showed that there were no significant differences between exposure conditions. Nevertheless, the 95% CI of the 100% effluent coefficient was estimated at − 0.61 and 0.05 (logit-transformed values) (t value, − 1.7). Hence, a negative impact of the 100% effluent condition is also highly suspected.

Hemocyte oxidative activities at day 3 were not strongly affected by exposure conditions (Fig. 4). The stimulation of the activity by zymosan particles was significant at day 3, but it was even more successful at day 29. At this day, the basal oxidative activity of hemocytes was increased for 100% effluent exposure compared to the water control. Both the 12.5% effluent and 100% effluent exposure conditions had a higher zymosan-stimulated oxidative activity than the water control. Nevertheless, statistical inferences suggest that the effects of exposure conditions with effluents are not significantly different from the water control, whether we consider the basal or the zymosan-stimulated oxidative activity of hemocytes. LMEMs suggest very weak interaction effects between exposure conditions and the success of stimulation with zymosan. However, the 100% effluent condition at day 3 had a t value of 1.54, which is notable and in coherence with the boxplots, although not statistically significant. Furthermore, at day 29, LMEM confirms the even more pronounced increase of 100% effluent in oxidative activity compared to control. The estimated 95% CI of the corresponding coefficient was estimated at − 0.01 and 0.24 (log10-transformed values) (t value, 1.80). Again, inferences are at the extreme limit of statistical significance for this effect and so the 100% effluent condition is likely to increase the oxidative activity of hemocytes significantly at day 29.

Fig. 4
figure 4

Oxidative activities from hemocytes. This graph represents the oxidative activity of hemocytes from snails assessed at days 3 and 29 into the respective exposure conditions. The activity was assessed in the basal status as well as in the stimulated status consisting in an ex vivo challenge of hemocytes with zymosan particles in order to stimulate the immunocompetent oxidative activity (i.e., likely NADPH oxidase)

Full size image

Discussion

Snails successfully survived effluent exposures (survival rate higher than 80% for all conditions at the end of the 29 days). The water parameters matched the OECD recommendations for ecotoxicological tests with L. stagnalis (Ducrot et al. 2014).

We made the choice to focus at an individual level because the quantity of hemolymph extruded per snail was big enough, and individual information about immunocompetence can be linked with individual traits. Besides, individual data allows more robust statistical analysis by taking into account the individual variability and thus avoiding false conclusions. However, due to the high inter-individual variability observed for immune parameters, we can still propose to pool hemolymph samples in further studies with L. stagnalis in order to provide an “average trend” with less variability. Nevertheless, this would require more snails for statistical analysis because one pool of several snails will be considered as n = 1.

Snails exposed to effluent had a higher weight increase over 29 days than water control snails. Since all the organisms were fed lettuce ad libitum, it can be suggested than organic matter and microbial loads in effluents are partly responsible for the results observed. Markers of immunocompetence were not strongly affected by the effluents. Nonetheless, a trend toward an increase in hemocyte density and oxidative and phenol oxidase activity and a reduced phagocytosis capacity resulting from exposure to 100% effluents are highly likely. Many inferential analyses were very close to statistical significance.

This subtle inflammatory effect can have devastating consequences for wild specimens since it is well known that the immune system is a process with high energy demands. Maintaining the immune system under abnormally high activity for long periods will likely induce life history trade-offs. To better corroborate this, further research is needed, particularly for longer exposure periods because this is the only way to picture the likely in natura effects of a subtle interaction of WWTP effluents with the L. stagnalis immune system.

Even if the micropollutants were not chemically quantified, one can reasonably consider the effluents to be contaminated by multiple xenobiotics. Furthermore, by hypothesizing the dilution factor of effluents in the mixing zone of the river higher than 500, the present study intended to elicit a stronger concentration exposure of micropollutants (from this source) than in natura reality downstream of the WWTP. A previous study had investigated the effect of municipal effluents from a Montreal WWTP to the immune responses of L. stagnalis (Gust et al. 2013b). Contaminated mediums were collected directly downstream of the mixing zone of a WWTP discharge (from < 0.1 up to 10 km). Immune responses at the cellular level consisted in statistically significant increases in hemocyte density, ROS production, thiol groups, and phagocytosis capacity and efficiency. In the present study, by collecting effluents directly from the WWTP (i.e., before discharge into the river), the induced immune responses were expected to be stronger. The results contradicted this expectation, but direct comparison is tricky since (i) snails were not from the same population, (ii) snail calibration was different, (iii) the WWTP effluent was not the same, (iv) chemical micropollution is likely different as is (v) microbial effluent loads, and (vi) the measurement methodologies were different. Nonetheless, similar trends were observed with regard to immunostimulation or immune inflammation of effluent-exposed snails, except for phagocytosis capacity.

By exposing the snails to chemical multicontamination at environmentally relevant concentrations, a high probability of exposure was expected to at least one xenobiotic interacting specifically with the L. stagnalis immune system or that was eventually indirectly immunotoxic after altering other physiological functions at an earlier stage.

The present study reports that the raw WWTP effluents studied did not substantially jeopardize—from a conservative statistical significance viewpoint—the markers of immunocompetence assessed in L. stagnalis.

Our results do not support the great relevance of assessing the analyzed immunomarkers (phagocytosis, NADPH-like activity, phenol oxidase-like activity, density and viability of hemocytes) in the present studied site context. Since many studies report the modulation of these immune parameters in response to pollutants (Coles et al. 1994, Galloway and Depledge 2001, Gagné et al. 2002, Russo and Lagadic 2004, Auffret et al. 2006, Vijayavel et al. 2009, Gust et al. 2013b), their relevance cannot be ruled out. However, our study encourages strongly to analyze other endpoints in L. stagnalis and for example to focus at lower biological levels (e.g., genomics, proteomics). This would allow (i) increasing the detection sensitivity of (immuno)biological responses, (ii) targeting more physiological pathways in multiplexed biomarkers analysis, and (iii) obtaining patterns of molecular signature in such multipollution context.

Controls were not exposed to significant loads of microbes. On the contrary, it is well known that WWTP effluents contain multiple microbes. Even if the microbial load in raw effluents was not quantified, we can confidently assume that snails from these conditions were continuously exposed to bacteria throughout the experiment. Consequently, we pictured their a posteriori immunocompetence that we could define as the immunocompetence of animals following microbe exposure. For example, in the study reported by Gust et al. (2013b), heterotrophic bacteria (HB) were quantified at 92.105 and 820.105 HB CFU g−1 (colony-forming units) in snail tissues exposed to upstream WWTP water and downstream (< 0.1 km) WWTP water, respectively. It is well known that some microbes can stimulate some immune parameters of aquatic molluscan species (Knaap et al. 1987, Oubella et al. 1993, Mitta et al. 1999, Allam et al. 2000, Bianchi et al. 2015). As often in environmental immunotoxicology, immune responses are highly correlated to both micropollutants and microbes (Gagné et al. 2002, Akaishi et al. 2007, Farcy et al. 2011, Gust et al. 2013b).

Apart from micropollution, particles in suspension (e.g., organic matter, nanomaterials, microplastics) other than microorganisms may interact with the immune system of aquatic species. For instance, human carbon manufactured nanomaterials (i.e., graphene nanoplatelets in this example) are known to be internalized in model cell lines (Lammel et al. 2013) and to play a Trojan horse role for xenobiotics (Lammel et al. 2015). In aquatic molluscan species, immune systems are thought to be an important target of nanomaterials, and immunotoxicity has already been reported in bivalves (Canesi et al. 2008, Canesi et al. 2012). These studies contribute ideas on further hemocyte-related endpoints worth assessing in a WWTP effluent exposure context (e.g., microscopy analysis for materials, lysosomal and mitochondria activities, defense enzyme activities).

Last, in another study (unpublished data), effluents of the WWTP studied herein did not induce effects on the locomotion capacity of a freshwater amphipod, whereas effluents from another WWTP did significantly impact their locomotion indicating a low ecotoxic effect of the effluents. To delve further into the characterization of the immunotoxic hazard of WWTP effluents with the immunocompetence markers selected, it would be of great interest to compare immune responses of L. stagnalis exposed to effluents from a WWTP for which the quality is clearly identified as alarming and investigate it over a longer period (e.g., a few months).

Conclusion

The immune system of snails—characterized by a battery of hemocyte endpoints—successfully dealt with exposures to urban WWTP effluents lasting 29 days. Urban WWTP effluents are a well-known source of xenobiotics and multiple microbial contaminations and therefore it is highly relevant to study the immunity status of organisms in this context. Herein, the treated effluent studied, even without dilution, did not significantly affect the selected immunomarkers, indicating a low hazard on immunocapacity and immunoefficiency of snails exposed to such a discharge. Nonetheless, it is highly suspected that the effluent did induce subtle inflammatory responses, and if so, consequences over the long term can be important for these animals (life history trade-offs). Further work should focus on longer periods of exposure and with effluents known to be highly polluted. Last, other hemocyte-related biomarker approaches should be considered, particularly for mechanistic insights or exploration of sensitive immunotoxic signature patterns (e.g., transcriptomics, proteomics).