E ﬀ ects of Glyphosate-Based and Derived Products on Sea Urchin Larval Development

: Despite the widespread use of herbicide glyphosate in cultivation, its extensive runo ﬀ into rivers and to coastal areas, and the persistence of this chemical and its main degradation product (aminomethylphosphonic acid, AMPA) in the environment, there is still little information on the potential negative e ﬀ ects of glyphosate, its commercial formulation Roundup ® and AMPA on marine species. This study was conducted with the aim of providing a comparative evaluation of the e ﬀ ects of glyphosate-based and its derived chemicals on the larval development of the sea urchin Paracentrotus lividus , thus providing new data to describe the potential ecotoxicity of these contaminants. In particular, the e ﬀ ects on larval development, growth and metabolism were assessed during 48 h of exposure from the time of egg fertilization. The results conﬁrm that AMPA and its parent compound, glyphosate have similar toxicity, as observed in other marine invertebrates. However, interestingly, the Roundup ® formulation seemed to be less toxic than the glyphosate alone.


Introduction
Since the 1990s, the presence of emerging contaminants in the aquatic environments has been studied in increasing depth [1]. These molecules are ubiquitously distributed in aquatic ecosystems, where'they are continuously released and are found at medium-low (µg/L) and low levels (ng/L) [2,3]. The concentrations of pollutants are likely to increase due to the increasing number of new molecules with commercial relevance, and the inadequacy of wastewater treatment plants and waste management [4]. Therefore, a serious threat is posed to organisms that live in riverine, estuarine and coastal environments.
Numerous ecotoxicological studies have demonstrated that the first developmental stages (gametes, embryos and larvae) in the life cycle of marine invertebrates represent critical phases that are the most susceptible to significant and persistent alterations of environmental parameters and the presence of xenobiotics [2,5]. In fact, compared to adults, the early life stages have a greater surface/volume ratio and are less developed with regard to the ability to metabolise absorbed chemical substances. These characteristics facilitate the uptake of chemical products on the one hand, and result in the poor efficiency of the detoxification processes on the other [6]. As the larval stages represent a potential bottleneck in the maintenance of natural populations, they are commonly used to evaluate the toxicity of a large variety of contaminants and in the biomonitoring of marine pollution [5,7].
Glyphosate (N-[phosphonomethyl]-glycine) is a non-selective herbicide that is able to prevent the growth of plants through the inhibition of the 5-enolpyruvyl-shikimate-3-phosphate synthase (EPSPS),

Collection of Adult Specimens
P. lividus adult specimens (mean test diameter of 5.02 cm ± 0.31 s.d.) were collected by scuba divers at a 2-5 m depth in a coastal area of the North Adriatic Sea (45 • 13 42.9" N, 12 • 19 22.0" E). After collection, the organisms were immediately transported to the Hydrobiological Station, "Umberto D'Ancona" in Chioggia, kept for 5 days in tanks with a continuous flow of seawater and natural variation in the environmental parameters (mean temperature: 26.6 • C ± 1.8 s.d.; mean salinity: 33.8 ± 1.8 s.d.), and were fed ad libitum with Ulva sp.

Experimental Conditions Tested
Thirteen experimental conditions were tested: the control was represented by artificial sea water (ASW) (35 salinity, pH 8.2; Ocean Fish, Prodac International S.r.l, Padova, Italy) filtered with a 0.45 µm Sartorius filter, and four environmentally relevant concentrations (1, 10, 50, 100 µg/L) of glyphosate, AMPA and Roundup in ASW. For Roundup, the values refer to the concentration of the active substance, glyphosate. Glyphosate and AMPA were purchased from Sigma-Aldrich (Milano, Italy). Roundup (Roundup Power 2.0, composition: 360 g pure glyphosate/L, Monsanto Europe N.V., Antwerp, Belgium) was purchased from a local commercial vendor. ASW was chosen to prevent the interference of other contaminants that might be present in natural sea water.
The experiment was replicated three times, each time using a pool of eggs and sperm from three female and three male sea urchins.

Gamete Preparation and Larval Exposure
To induce gamete emission, each sea urchin was injected with 1 mL of 0.5 M KCl solution using a sterile syringe through the peristomal membrane. Sperm from three males was collected with a micropipette and maintained on ice until use. After KCl injection, the eggs of three females were collected in 250 mL beakers filled with ASW by placing each female upside down over the beaker mouth. Dams were allowed 30 min to release eggs, which were then filtered with a 200 µm nylon mesh to remove impurities (e.g., fragments of algae, prickles and pedicels), concentrated using a 20 µm nylon mesh and re-suspended in ASW. Egg density was assessed in subsamples under a Leica DM750 optical microscope. Eggs from the three females were pooled in a single beaker in equal proportions. Prior to preparing the sperm pool, the density of the ejaculate of each male was estimated in diluted subsamples using a Z2 Coulter Counter (Beckman Coulter Inc., Miami, FL, USA), in order to define an equal contribution of sperm from all males. Sperm were activated in ASW and activation was checked after one minute under optical microscope.
The two pools of gametes thus constituted were brought together by adding an adequate volume of activated sperm to the egg suspension to obtain a sperm:egg ratio of 1250:1 [39]. After 15 min, the raising of the fertilization membrane in at least 90% of eggs was checked under the microscope. Fertilized eggs were collected with a 20 µm nylon mesh in order to eliminate sperm and then resuspended in ASW. Five 24 mL glass vials were prepared for each experimental condition. Equal volumes of the fertilized egg suspension were distributed into the vials at a final concentration of 60 eggs/mL. Exposures were carried out for 48 h at a temperature of 22 • C.

Larval Development and Growth Assessment
In each experiment, two replicates per experimental condition were fixed with 10% neutralized formalin at 24 hpf, and the other three were fixed at 48 hpf. Larval development was assessed at both 24 and 48 hpf on 100 larvae per replicate. The following developmental stages were assigned: blastula (the pre-hatching stages), gastrula, prism, early pluteus and pluteus. At 48 hpf, larval abnormalities ( Figure 1) were also assessed in 100 echinoplutei per replicate, in accordance with the method proposed by His and colleagues [40]. The remaining larvae in the 48 hpf replicates were then filtered on a Sartorius acetate cellulose filter (25 mm in diameter and 8 µm porosity) and gently washed with deionized water to remove formalin and salts. The filter was placed on a histology slide in an oven at 60 • C for 1 h and then it was clarified with cedar oil and covered with a coverslip for preservation. Italy). Roundup (Roundup Power 2.0, composition: 360 g pure glyphosate/L, Monsanto Europe N.V., Antwerp, Belgium) was purchased from a local commercial vendor. ASW was chosen to prevent the interference of other contaminants that might be present in natural sea water.
The experiment was replicated three times, each time using a pool of eggs and sperm from three female and three male sea urchins.

Gamete Preparation and Larval Exposure
To induce gamete emission, each sea urchin was injected with 1 mL of 0.5 M KCl solution using a sterile syringe through the peristomal membrane. Sperm from three males was collected with a micropipette and maintained on ice until use. After KCl injection, the eggs of three females were collected in 250 mL beakers filled with ASW by placing each female upside down over the beaker mouth. Dams were allowed 30 min to release eggs, which were then filtered with a 200 μm nylon mesh to remove impurities (e.g., fragments of algae, prickles and pedicels), concentrated using a 20 μm nylon mesh and re-suspended in ASW. Egg density was assessed in subsamples under a Leica DM750 optical microscope. Eggs from the three females were pooled in a single beaker in equal proportions. Prior to preparing the sperm pool, the density of the ejaculate of each male was estimated in diluted subsamples using a Z2 Coulter Counter (Beckman Coulter Inc., Miami, FL, USA), in order to define an equal contribution of sperm from all males. Sperm were activated in ASW and activation was checked after one minute under optical microscope.
The two pools of gametes thus constituted were brought together by adding an adequate volume of activated sperm to the egg suspension to obtain a sperm:egg ratio of 1250:1 [39]. After 15 min, the raising of the fertilization membrane in at least 90% of eggs was checked under the microscope. Fertilized eggs were collected with a 20 μm nylon mesh in order to eliminate sperm and then resuspended in ASW. Five 24 mL glass vials were prepared for each experimental condition. Equal volumes of the fertilized egg suspension were distributed into the vials at a final concentration of 60 eggs/mL. Exposures were carried out for 48 h at a temperature of 22 °C.

Larval Development and Growth Assessment
In each experiment, two replicates per experimental condition were fixed with 10% neutralized formalin at 24 hpf, and the other three were fixed at 48 hpf. Larval development was assessed at both 24 and 48 hpf on 100 larvae per replicate. The following developmental stages were assigned: blastula (the pre-hatching stages), gastrula, prism, early pluteus and pluteus. At 48 hpf, larval abnormalities ( Figure 1) were also assessed in 100 echinoplutei per replicate, in accordance with the method proposed by His and colleagues [40]. The remaining larvae in the 48 hpf replicates were then filtered on a Sartorius acetate cellulose filter (25 mm in diameter and 8 μm porosity) and gently washed with deionized water to remove formalin and salts. The filter was placed on a histology slide in an oven at 60 °C for 1 h and then it was clarified with cedar oil and covered with a coverslip for preservation.  Larval growth was evaluated by measuring the somatic and the post-oral skeletal rod of 100 plutei per replicate. All samples were observed using an optical microscope (Leica DM 750) provided with a Leica DFC 295 digital camera; and the computer acquisition of images was performed using Leica Application Suite 3.8 software. Larval skeletal rods were measured using the ImageJ image analysis program. Due to technical issues, in one out of three experiments the results from larvae exposed to 1 µg/L of AMPA were not available.

Respiration Rate
Oxygen consumption was measured in only one of the three experiments. Additional vials were prepared and equipped with an O 2 sensor spot (OXSP5, Pyro Science GmbH, Aachen, Germany): 3 replicates for each contaminant in a concentration equal to 100 µg/L, 3 replicates for ASW (control) and 2 replicates for the blank containing only ASW, i.e., in the absence of larvae. Fertilized eggs were added at the same final concentration reported above for the other bioassays. Before starting the measurements, all vials were completely filled and then sealed with a screw cap. To detect the concentration of O 2 , the Fiber-Optic Oxygen Meter Piccolo2 (Pyro Science GmbH, Aachen, Germany) was used. This measurement was performed every 8 h, starting from 0 hpf up to 48 hpf. Oxygen saturation never fell below 80%. Values are calculated as the difference in O 2 concentration between 0 hpf and 48 hpf and expressed as the oxygen consumption rate in pmol/larva/h, after correcting the values with the blank.

Statistical Analysis
The differences in larval development among the exposure conditions were assessed using a non-parametric permutational multivariate analysis of variance, PERMANOVA [41], applying the Euclidean distance matrix on the raw data. First, a PERMANOVA was performed on the overall data set, then various PERMANOVA were used to perform pair-wise comparisons between exposure conditions ("adonis" function in "vegan" package [42]) with "experiment" as the random factor. To assess the differences in larval rod lengths, measurements were analysed through a linear mixed model fit by maximum likelihood ("lmer" function in the "lmerTest" package [43]) with "experiment" as the random factor, followed by a Tukey post-hoc test for the pairwise comparisons. For the oxygen consumption analysis, a one-way ANOVA followed by a Tukey multiple comparison post-hoc test was carried out, and the normality and homoscedasticity were checked in the residuals. Significant pair-wise comparisons were shown with respect to (i) control (ASW), (ii) different concentrations of the same contaminant, and (iii) different contaminants at the same concentration. All data were analyzed using R and RStudio software (version 3.5.3, [44]). The significance threshold was fixed at p < 0.05.

Larval Development, Abnormalities and Growth
The PERMANOVA and linear mixed model results highlighted significant differences among the experimental conditions tested at both 24 hpf and 48 hpf for all endpoints considered (Table 1). In detail, at 24 hpf, the developmental stage was significantly affected in AMPA 50 µg/L and Roundup 100 µg/L with respect to the control (p < 0.05) (Figure 2). Although not significant, a dose-dependent pattern of delay in development can also be observed in glyphosate exposure. At 48 hpf, differences in development between treatments and control were always significant, except for AMPA 1 and 10 µg/L (Figure 3). With regard to exposures to the same contaminant, significant differences between concentrations were found for glyphosate and Roundup (Figure 2) at 24 hpf, for'AMPA at 48 hpf (Figure 3).  Overall, the percentage of the pluteus stage decreased with the increase in contaminant concentration for all contaminants tested (Figure 3). When comparing contaminants at the same concentration, the only significant difference was between glyphosate and AMPA 10 µg/L, at 24 hpf ( Figure 2), with AMPA showing a greater percentage of blastulae and gastrulae.
As shown in Figure 4, the exposure to increasing concentration of glyphosate and AMPA was positively related to the percentage of abnormal echinoplutei. This value reached its maximum (~26%) in larvae exposed to 100 µg/L of glyphosate, with significant differences compared to the control and all the other experimental conditions tested (p < 0.01). Conversely, the exposure to an increasing concentration of Roundup did not cause significant changes in the percentage of abnormalities compared to the control, except for a significant increase at 10 µg/L (p < 0.05).
In particular, the somatic rod was significantly different from the control in the presence of 10, and 100 µg/L of glyphosate (p < 0.05 and p < 0.001, respectively), and 50 and 100 µg/L of AMPA (p < 0.01 and p < 0.05, respectively) ( Figure 5). In addition, the somatic rod in larvae exposed to 100 µg/L of glyphosate was significantly shorter than in larvae exposed to 1 and 50 µg/L of glyphosate (p < 0.01 and p < 0.05, respectively), and to 100 µg/L of Roundup (p < 0.05). A significant decrease in somatic rod length was also found in larvae exposed to 50 µg/L of AMPA compared to those exposed at 10 µg/L of the same chemical (p < 0.05).
With respect to the control, the post-oral rod length was significantly affected by the presence of 10 and 100 µg/L of glyphosate (p < 0.05), 50 µg/L of AMPA (p < 0.01), and 50 and 100 µg/L of Roundup (p < 0.05) ( Figure 6). Moreover, significant decreases in post-oral rod length were observed after exposure to both glyphosate 10 µg/L and AMPA 50 µg/L compared to AMPA 10 µg/L treatment (p < 0.05 and p < 0.01, respectively), and after exposure to Roundup 50 and 100 µg/L compared to Roundup 1 µg/L treatment (p < 0.05).

Respiration Rate
The oxygen consumption in P. lividus larvae exposed to four experimental conditions (control and 100 µg/L of glyphosate, AMPA and Roundup), measured for 48 h from fertilization, is shown in Figure 7. Exposure to the tested contaminants caused a statistically significant alteration in the oxygen consumption rate (F (3,8) = 7.27, p = 0.011). The respiration rate of the larvae was significantly higher in glyphosate and AMPA treatments than in the control. No significant variations in oxygen consumption were found in larvae exposed to Roundup with respect to the control.

Discussion
Although some previous studies have assessed the effects of glyphosate and Roundup on the very early life stages of sea urchins [40,45], this is the first that compares three glyphosate-related substances in 48 hpf exposures of P. lividus embryos. In addition, a standard "one male/female pair" toxicological test was not chosen [39,46] in order to enhance the ecotoxicological relevance of the study. In particular, the use of pools of three females and three males provides an opportunity to investigate biological responses at the population level [47,48], thus removing the possible bias of inter-individual incompatibility [49]. As suggested in a recent review [5], the multiple parent approach should be preferred when the goal of the research is to make future predictions using an external spawner as a model species.
The concentrations tested in this study were selected based on the information available on the levels of glyphosate and AMPA in coastal waters [14,15,50]. Although the concentration of most emerging contaminants is relatively low [51,52], the potential impacts of these substances on exposed biota represent a global concern, owing to their continuous release into marine and estuarine areas worldwide [53].
Since the 1980s, echinoderm early life stages have been used to test the toxicity of pollutants [54,55]. In particular, several sea urchin species have been used as useful bioindicators to evaluate the harmful effects of pollutants and other environmental stressors [36,46,[56][57][58].
In the present study, alterations in larval development were revealed by focusing on various endpoints: larval development, larval growth, morphological anomalies and metabolic variations measured as oxygen consumption. Based on the frequencies of the various developmental stages, a delay in larval development was mainly evident at 48 hpf, with a dose-dependent trend and statistically significant differences with respect to the control in most of the cases, even at the lowest concentrations tested for glyphosate and Roundup. Similarly, the measurement of the skeletal rods proved to be a sensitive tool for assessing the various exposure conditions. Skeletal rod growth was significantly reduced by all contaminants at the highest concentration tested. However, the somatic rod length was most affected in glyphosate and AMPA treatments, which also caused a significant decrease at lower concentrations (10 and 50 µg/L, respectively). Conversely, the post-oral rod length showed a more pronounced effect in Roundup treatment, with a significant reduction starting from 50 µg/L. Since the post-oral rod originates later from the somatic rod [59], some differences in the mode of action of Roundup can be hypothesized. Interestingly, similar results were obtained in P. lividus larvae exposed to the surfactant linear alkylbenzene sulfonate (LAS) [57]. This suggests a key role for the surfactant component in the Roundup formulation.
Significant alterations in larval morphology can have direct effects on larval performance. In several sea urchin species (P. lividus, Heliocidaris tuberculata, Centrostephanus rodgersii, Dendraster excentricus), it has been demonstrated that survival of larvae depends on their ability to swim and feed, which is determined by the length of the arms and therefore of the skeletal spicules that support them [60][61][62]. Reductions in the post-oral rod can result in less efficient food assimilation [63] or even lead to a reduction of the pluteus stomach and swimming ability, thus exposing larvae to higher predation risks [64].
His and colleagues [40] used experimental conditions similar to those used in our work, but no significant increase in larval abnormalities was found up to 200 µg/L of glyphosate. Unlike those results, in our study 100 µg/L of the same contaminant caused a significant increase in the percentage of abnormalities, mostly at the skeletal level, with incomplete, deformed or apically crossed skeletal rods. The higher frequency of abnormal larvae could be related to a reduction in acetylcholinesterase (AChE) activity [65] since it has a central role in primary mesenchyme cell movement beginning with gastrulation, and thus, in the spiculae formation [66]. In fact, glyphosate and AMPA significantly inhibited AChE activity in the gills of Mytilus galloprovincialis exposed in vivo to the contaminants for 7 and 21 days [67]. Similar results were obtained in other bivalve species, crustaceans and fish for glyphosate [68][69][70]. Surprisingly, in our study, an increased presence of larval abnormalities was not found in the Roundup 100 µg/L exposure, even if the concentration of the active substance was the same.
Other substances in the Roundup formulation probably acted antagonistically to the active substance.
In fact, the oxygen consumption results suggested that Roundup was less toxic than the glyphosate alone. Both glyphosate-and AMPA-exposed larvae, but not those exposed to Roundup, showed a significantly increased oxygen consumption rate, which highlights increased energy expenditure to support larval development. Alterations in respiration rates have been widely reported in various life stages and taxa after exposure to anthropogenic stressors, such as pollutants and seawater acidification [71][72][73][74][75]. To our knowledge, no data regarding respiratory alterations due to pollutants in the larval stages of sea urchins are available and a limited number of studies have been carried out in larvae of marine invertebrates [76]. Nonetheless, in a recent study analyzing the effect of ocean acidification on larvae of the sea urchin Heliocidaris crassispina, a decrease in pH caused a gradual increase in oxygen consumption, doubling the value at pH 7.3 compared to the control at pH 8.0 [77]. Similar increases in respiratory rates at reduced pH have also been observed in other studies on larvae of different sea urchin species, Strongylocentrotus droebachiensis and Strongylocentrotus purpuratus [78,79]. In sea urchin larvae exposed to reduced pH, increased metabolic rates are likely responsible for the slower development and the decreased body length. Similarly, in larvae exposed to glyphosate and AMPA increased respiration rate can be explained by increased energy demand necessary to maintain homeostasis in stressful conditions, to the detriment of growth. Although oxygen consumption rate was not affected in Roundup treated larvae, a significant reduction in larval size compared to the control was observed, as found in both glyphosate and AMPA treatments.
The secrecy behind the ingredients that compose Roundup makes difficult to understand if these compounds act antagonistically or synergistically. A recent study showed that Roundup causes renal damage in rats, while the glyphosate active substance alone does not [80]. In some studies, the synergistic activity of the other ingredients in the Roundup formulation and the main component, glyphosate, has been revealed. Additive substances, such as POEA, were shown to boost glyphosate toxicity or to cause per se lethal [81] or sublethal effects [82,83], namely, disruption of the very first cell divisions in sea urchin embryos and a delay in the cell cycle.
Very little is known about the effects of AMPA on invertebrates, and even less on marine invertebrates. It has been recently demonstrated that both glyphosate and AMPA can disrupt the bacterial community in the digestive apparatus of the mussel M. galloprovincialis [84]. These contaminants caused changes in the expression of genes related to the immune defense system, which enhances the propagation and growth of pathogens. Moreover, exposure to AMPA affected the hemocyte parameters of M. galloprovincialis [67,85], and the response was even greater when animals were exposed to a mixture of glyphosate and AMPA. These findings match our results and suggest that AMPA has a similar toxicity to glyphosate.
To summarise, glyphosate-related chemicals are harmful to sea urchin populations as smaller larvae are produced, which have a higher frequency of abnormalities, and reach the pluteus stages in lesser amounts and with higher energy requirements.
The present study fills a gap in the knowledge about the effects of glyphosate-based and derived chemicals as single stressors on sea urchin early life stages, even though the still unknown composition of Roundup impedes a definite interpretation of its adverse effects. The results highlight that the selected pollutants, in environmentally relevant concentrations, induce significant effects on larval development and metabolic consumption in the sea urchin P. lividus, raising further concerns about the presence of glyphosate-related chemicals in marine ecosystems. Future studies could focus on the effects of mixtures of these emerging contaminants or combinations of contaminants and other ecologically relevant human-driven stressors.