Effects and risk assessment of the polycyclic musk compounds Galaxolide ® and Tonalide ® on early life stages of marine organisms

The current research investigated the environmental risk of the polycyclic musk compounds, Galaxolide ® (HHCB) and Tonalide ® (AHTN), in the marine environments. These substances are lipophilic, bioaccumulated and potentially biomagnified in aquatic organisms. To understand the toxicity of HHCB and AHTN, we performed acute toxicity tests by exposing marine microalgae ( Phaeodactylum tricornutum , Tretraselmis chuii and Isochrysis galbana ), crustaceans ( Artemia franciscana ), echinoderms ( Paracentrotus lividus ), bivalves ( Mytilus galloprovincialis ), fish ( Sparus aurata ) and a candidate freshwater microalga ( Raphidocelis subcapitata ) to environmentally relevant concentrations (0.005 - 5 µg/L) following standardized protocols. The effects of both substances on microalgae growth were incipient and only I. galbana was sensitive to HHCB and AHTN, with IC 10 values of 5.22 µg/L and 0.328 µg/L, respectively. Significant (p < 0.01) concentration dependent responses were measured in P. lividus and M. galloprovincialis developments as well as S. aurata mortality tested with HHCB. The effect of HHCB on P. lividus larvae development was the most sensitive endpoint recorded, producing an EC 50 value of 4.07 µg/L. Our results show that HHCB represents a high risk to P. lividus larvae development for early life stages in marine environments.


Introduction
Polycyclic musk compounds (PMCs) have been included in the priority lists of the European Commission existing substances regulation (OSPAR, 2004). PMCs are synthetic chemicals, produced in large quantities as a replacement for nitro musks, which have been banned because of their environmental persistence and adverse effects on humans and the environment (Gao et al., 2016a). They have wide applicability in household and personal care products such as detergents, shampoos, lotions, perfumes, as well as additives in cigarettes and fish baits (Gao et al., 2016b;Heberer, 2002). Due to their high water solubility, inherent lipophilicity and biological stability, coupled with external application (they do not undergo biotransformation), it is not surprising to find them as contaminants in aquatic ecosystems at concentrations ranging from ng/L to µg/L (Dietrich and Hitzfeld, 2004).
In particular, the PMCs marketed as Galaxolide ® (HHCB) and Tonalide ® (AHTN) represent about 95% of total fragrance materials in the perfume industry (Balk and Ford, 1999a) and are the most commonly detected PMCs in environmental compartments and biological tissues (Bester et al., 1998;Díaz-Garduño et al., 2017;Fromme et al., 2001;Kannan et al., 2005;Nakata et al., 2007;Sumner et al., 2010;Villa et al., 2012). Fromme et al. (2001) detected HHCB and AHTN in surface water in Berlin (Germany) at concentration values ranging from 70 -1590 ng/L and 20 -530 ng/L, respectively. Similarly, Heberer (2002) also reported high levels of HHCB and AHTN in surface water in Berlin at concentrations ranging from 20 -12500 ng/L and 30 -1100 ng/L, respectively. Moreover, these substances have been detected in surface waters in almost every country in Europe (Dsikowitzky et al., 2002;Fromme et al., 2001;Villa et al., 2012). Although most studies were for the freshwater environment, there are data confirming their presence in the marine environment (Bester et al., 1998;Díaz-Garduño et al., 2017;Pintado-Herrera et al., 2013;Sumner et al., 2010), and reported values over 5 µg/L depending on the proximity to a sewage treatment plants (STP). Sumner et al. (2010) studied the transport of PMCs from an STP effluent into coastal waters and reported that the concentrations of HHCB and AHTN in the open sea depends on the distance from STP and the dilution power towards the open sea. For instance, the authors measured the concentration of HHCB in STP effluent ranging from 987 -2098 ng/L, being diluted towards the sea over 2 km, to 6 -28 ng/L. However, this level could be higher near urban coasts where STP effluents are directly discharged into coastal waters, which could represent a potential threat to the survival of nearshore organisms, an important part of the marine ecosystems.
Bioaccumulation of PMCs, particularly HHCB and AHTN, have also been reported in marine organisms including crustaceans, bivalves, fish, marine birds and mammals (Kannan et al., 2005;Moon et al., 2012;Nakata et al., 2007;Rüedel et al., 2006) at levels of environmental concern. The concern over the potential effects are not only related to the environment, but also the impacts on food safety and, consequently, public health. Therefore, Vandermeersch et al. (2015) reviewed emerging contaminants in seafood, acknowledging that HHCB and AHTN were the most commonly detected PMCs, with concentrations reaching 160 and 45 µg/g lipid weight in mollusks and fish, respectively (Gatermann et al., 2002). Similarly, in the framework of the FP7 ECsafefood project, where 62 commercial seafood samples (mackerel, tuna, salmon, seabream, cod, monkfish, crab, shrimp, octopus, perch and plaice) in the European Union were analysed for residues from personal care products, HHCB was detected at concentrations ranging from 2.5 -414.4 µg/kg dry weight, with the highest concentration measured in fish (sole) and AHTN found at concentrations ranging from 2.5 -12.2 µg/kg dry weight (Cunha et al., 2018). Furthermore, these compounds have the potential to elicit adverse effects in marine organisms due, respectively, to the bioconcentration factors (based on lipid content) of 3504 and 5017 (Fromme et al., 2001) and high octanol water partition coefficient (LogKow) of 5.9 and 5.7 (Balk and Ford, 1999a).
Data on the acute, sub-chronic and chronic aquatic toxicity of HHCB and AHTN are available for algae, crustaceans, mollusks, bivalves and various fish Carlsson et al., 2000;Carlsson and Norrgren, 2004;Gooding et al., 2006;Liu et al., 2011;Luckenbach and Epel, 2005;Parolini et al., 2015;Pedersen et al., 2009;Wollenberger et al., 2003;Yamauchi et al., 2008). Although the majority of the studies were focused on freshwater ecosystems, Breitholtz et al. (2003) and Wollenberger et al. (2003) investigated the acute, sub-lethal and lethal effects of these substances on the marine copepods, Nitocra spinipes and Acartia tonsa, respectively. While Wollenberger et al. (2003) concluded that they inhibited larval development in A. tonsa at low concentration and should be considered very toxic, Breitholtz et al. (2003) reported low adverse effects in N. spinipes.
Several attempts have been made to assess the environmental risk of HHCB and AHTN in the aquatic environment. Balk and Ford (1999b) provided an insight by using acute and chronic toxicity data from freshwater and terrestrial organisms to estimate the environmental risk, but with no reference to the marine environment. Other ERAs performed so far (European Commission, 2008a, 2008bHERA, 2004;OSPAR, 2004) revealed a paucity of ecotoxicity data for the marine environment needed to effectively assess the risk of these substances.
The aim of this research was to examine the toxicity of HHCB and AHTN in the marine environment by exposing organisms from different trophic levels at early life stages to environmental concentrations of HHCB and AHTN and to estimate the level of risk they might exert on the marine ecosystems. Therefore, we performed acute toxicity using marine organisms -microalgae (Phaeodactylum tricornutum, Tretraselmis chuii and Isochrysis galbana), crustaceans (Artemia franciscana), echinoderms (Paracentrotus lividus), bivalves (Mytilus galloprovincialis) and fish (Sparus aurata); and a candidate freshwater alga -Raphidocelis subcapitata. Furthermore, understanding the toxicity of HHCB and AHTN in lower trophic organisms will help to predict possible bottom-up and top-down effects, which could lead to a functional and structural disruption of the ecosystems (De Lange et al., 2010;Fleeger et al., 2003;Johnston and Roberts, 2009;Van Den Brink, 2008). The species used in the current study have been used to examine the effects of effluents from sewage treatment plants (Diaz-Garduno et al., 2016;Meric et al., 2005), metals (Moreno-Garrido et al., 2005;Yu et al., 2007), organic solvents (Okumura et al., 2001) and inorganic chemicals (Latała et al., 2009;Mariani et al., 2006) in marine environments. Also, these organisms have been endorsed by international organisations for ERA of contaminants due to their sensitivity (OECD, 2013;Environment Canada, 2011;European Comission Joint Research Centre, 2003;US EPA, 2002, 1992.

Chemical selections
Analytical grade HHCB and AHTN were purchased from Sigma Aldrich, Spain. The physiochemical properties, preparations in organic solvent and determination of exposure concentrations followed details found in Ehiguese et al. (2019). In brief, HHCB and AHTN were dissolved in dimethyl sulfoxide (DMSO) (0.001% v/v) in glass vials to form a stock solution. Concentrations (0.005, 0.05, 0.5 and 5.0 μg/L) of each substance were prepared by diluting the stock solutions in Nanopure water.

Microalgae growth inhibition test
Microalgae growth inhibition tests were performed using three marine and one freshwater species, following the procedure reported by Garrido-Perez et al., (2008). Inocula of Phaeodactylum tricornutum, Tetraselmis chuii and Isochrysis galbana (marine species) were provided by the Laboratory of Marine Culture of the University of Cádiz, Spain, and Raphidocelis subcapitata (strain 61.81) was provided by the Culture Collection of Algae at Goettingen University, Germany. Seawater microalgae culture media was prepared by adding nutrients (macro-and micronutrients and vitamins) according to f/2 medium (Guillard and Ryther 1962) to synthetic seawater according to the formula from USEPA (2002a). For the freshwater species, the culture media was prepared according to the concentrations of COMBO Media (Kilham et al. 1998) in Nanopure water. Before exposure to HHCB and AHTN, an inoculum from each species of microalgae was cultured in fresh media and in the same chamber of assay (19±1 ºC; 11000 lux; photoperiod 14/10 light/day). Inocula were maintained for three days to reach the exponential growth phase, in order to provide acclimatised and healthy cells for the tests.
An aliquot of 0.25 mL of each inoculum was added to 3.75 mL of the culture media spiked with different concentrations of contaminants. The exposure was performed in triplicates including two controls: (1) culture media without contaminants (normal growing of the microalga in a fresh medium), and (2) culture media without contaminants but including DMSO (to test the toxic effects of the organic solvent). The addition time of the microalgae was as short as possible, to ensure that they had similar biomass at the initial time. After 30 min inoculation, the initial biomass (B0) was measured, representing time zero (T0). The biomass was measured indirectly through absorbance at 680 nm (maximum chlorophyll peak in a fresh culture of microalga), with a TECAN 2000 micro-well plate reader. This measurement was repeated at 24, 48 and 72 h. The endpoint observed in this test was the inhibition of biomass growth at 72 h compared to the control. The minimal growth acceptable for the control was 16 times higher than the initial density.

Artemia toxicity test
The cysts of A. franciscana are certified biological material (AF450, INVE) and were acquired from Acuazul, S.C. (Spain). Before starting the trial, dehydrated artemia cysts were hatched. In a one-litre Erlenmeyer flask containing 300 mL of natural seawater (35 g/L salinity), approximately 100 mg of cysts were resuspended and left with aeration and continuous illumination throughout the hatching time, which occurs within 24-36 h. Once the artemia were hatched, the aeration was removed. The hatched nauplii were placed in clean seawater to avoid moving empty or unhatched cysts. Therefore, all nauplii used in the test belong to the same cohort. The test was carried out in triplicates using glass Petri dishes of 60 mm diameter. On each plate 10 nauplii were placed using a glass Pasteur pipette to avoid stress to the nauplii. Subsequently, seawater spiked with each contaminant was added to a total volume of 10 mL. The test was conducted at 20±0.4 o C and in the dark to minimize the swimming of the nauplii and the consumption of energy. The test lasted for 72 h and records were taken at 24, 48 and 72 h. The Artemia were counted using a stereomicroscope at a magnification of 2x and 4x, identifying dead specimens (mortality was recorded when they showed no sign of any movement of their limbs for 30 seconds) and those that presented motility problems such as movements in circles, asynchronous, among others. The test was considered acceptable when survival of over 80% was recorded in the control group after 72 h.

Sea urchin toxicity test
Sea urchin (P. lividus) fertilization and larval development tests were performed following procedures described by Fernandez and Beiras (2001), and Environment Canada (2011). Individuals were collected from the uncontaminated rocky subtidal environment off the Bay of Cádiz (Spain) at 1.5 -2 m depth. They were immediately transported to the laboratory in a cooler box. Matured individuals were dissected, and eggs and sperm were collected using a micropipette. For the fertilization test, 10 µL of sperm were added to the aliquots containing 10 mL of the test concentrations arranged in sequence for 10 min; then, 1 mL of eggs was added to each test tube, swirled gently and allowed to proceed for 10 min. The endpoint for fertilization success was the presence of a fertilization membrane. A larval development test was performed by adding 1 mL of fertilized eggs to beakers containing test solutions in triplicate, including the controls (seawater and DMSO), in dark conditions for 48 h. The test was considered valid when development in the control was ≥ 80% and the result expressed as percentage of normal pluteus stage, normalized to the corresponding seawater control. Both fertilization and larval development tests were conducted in a controlled chamber at a temperature of 20 º C.

Mussels larvae development test
The mussels M. galloprovincialis were purchased from an aquaculture farm in north-western Spain. The depurated mussels were transported to the laboratory at 8 o C. Animals with matured gonads were held at 9 o C for two weeks to acclimate to the laboratory condition. The toxicity test was performed following the ASTM protocol for acute toxicity of saltwater bivalves (ASTM, 2012). Mussels were placed each in a beaker containing filtered seawater and induced to spawn by thermal stimulation at 19 o C. Eggs and sperm were filtered to remove debris using 75 µm and 37 µm screens, respectively. Before fertilization, the egg and sperm quality and density were evaluated under a microscope. Aliquots of 10 µL of sperm were added to the eggs for fertilization (10 6 /egg) and fertilization success was assessed under the microscope. Fertilized eggs were added to test solutions at 50 eggs/ml and incubated for 48 h at 16±0.2 o C. The test was performed in triplicate including seawater and solvent controls (DMSO). Samples were fixed with 40% formalin at the end of the test and 100 larvae were counted under the microscope (X 40 magnification) per replicate distinguishing between normal developed larvae (D-shaped) and malformed larvae. Test results were accepted when normal developed larvae in control was ≥80%.

Fish larva mortality test
The effects of environmental concentrations of HHCB and AHTN on the survival of seabream (Sparus aurata) larvae was tested following OECD guidelines for fish early-life stage toxicity test (OECD, 2013). Yolk-sac larvae (3 -5 h post hatch) of S. aurata were obtained from the laboratory of Marine Culture, Faculty of Marine and Environmental Sciences, University of Cádiz, Spain. Individuals (n = 70) of the yolk-sac larvae were added to each beaker (Pyrex © ) containing 600 ml of seawater spiked with the concentrations of contaminants, and each treatment was run in triplicate. Seawater and solvent (DMSO) controls were also tested; the latter was to ensure that there were no solvent driven effects. The exposure proceeded for 96 h, physiochemical parameters were monitored using the CRISON CM35+ and 40MM+ multiparameter probes (Crison-Hachs Lange S.L.U., Spain) and values recorded during the exposure were: Temperature (16.01±0.2 o C), salinity (34.2±0.3), pH (7.7±0.2) and oxygen (>5 mg/L). The test was considered valid if mortality did not exceed 10% in the control group. Mortality in each treatment group was recorded and data were expressed as the percentage of survived larvae at each experimental condition.

Statistical analysis
Statistical analysis was performed using IBM SPSS Statistics, version 24.0. Significant differences between organisms exposed to the different concentrations of PMCs or to the organisms from control were checked using a one-way ANOVA followed by multiple comparisons of Turkey's or Dunnett's test. Statistically significant differences were set at p < 0.05. Spearman's rank order of correlation was calculated between the measured effects and the exposure concentrations. In addition, the ICPIN statistical programme, provided by USEPA on its website, was used for the analysis of toxicity data and to calculate the ICx/ECx (concentrations that cause growth inhibition or any effect, respectively, to x% of the population).
The Risk Quotient (RQ) was derived from the ratio of the measured environmental concentrations (MECs) listed in Table 1 to the PNEC (predicted no effect concentration). PNEC is the effect concentrations (EC) or inhibition concentration (IC) obtained from the tests divided by an assessment factor (European Comission Joint Research Centre, 2003).
If RQ < 1 = no risk expected, and no further evaluation required If RQ > 1 = potential risk and further evaluation is required 3. Results

Microalgae growth inhibition tests
Data of growth inhibition for microalgae are presented in Figure 1. There was no critical difference between microalgae exposed to seawater control and DMSO. In brief, it is possible to observe that toxicity of both compounds was incipient for the four microalgae, although AHTN seemed to be slightly more toxic. For P. tricornutum, growth inhibition for both compounds were observed in some concentrations. However, significant differences (p < 0.05) in growth in relation to the control only occurred with exposure to AHTN at 0.005 μg/L by 16.2% ( Figure 1A). On the one hand, T. chuii growth was inhibited by AHTN and the decrease in biomass was significantly (p < 0.01) concentration dependent (Table S1 -Supplementary Materials). On the other hand, the exposure to HHCB produced a stimulation (hormesis) in the growth ( Figure 1B) and similar biphasic responses was seen in I. galbana biomass after exposure to both musk compounds ( Figure 1C). The freshwater microalgae, R. subcapitata growth decreased after exposure to HHCB and AHTN. In the case of AHTN, significant (p < 0.05) concentration dependent growth inhibition of R. subcapitata was observed (Table S1 -Supplementary Materials), and the highest inhibition occurred at 5 μg/L by 23.5%. In contrast, R. subcapitata growth inhibition by HHCB was more severe at 0.05 μg/L by 14.5% in relation to the control ( Figure 1D).

Invertebrate and fish early life stage toxicity tests
For all the organisms tested for fertilization, larvae development, motility and mortality, the threshold values for the control experiments (seawater control and solvent control -DMSO) were not exceeded.   Figure 1A-1D: Microalgae exposed to galaxolide and tonalide for growth inhibition tests. % Enrichment of (A) P. tricornotum, (B) T. chuii, (C) I. galbana and (D) R. subcapitata (negative values corresponding to % growth inhibition). Asterisks (*) indicate significant differences (p<0.05) in relation to control.
Moreover, there was no significant difference (p < 0.05) in responses detected in organisms exposed to seawater control and DMSO.
The effects of HHCB and AHTN on survival and motility of A. franciscana are shown in Figure 2A and 2B, respectively. Both responses showed no significant difference (p<0.05) compared to the control. For the A. franscana mortality test, percentage mortalities in the controls were 3.33% and 6.67% for seawater and DMSO, respectively. The effect of HHCB on the survival of A. franciscana was very minimal with the highest mortality of 10% recorded at a 0.5 µg/L HHCB. The effect of AHTN on A. franciscana mortality was also very low, with only 3.33% mortality at 0.05 and 0.5 µg/L. Other concentrations of AHTN tested had no effect on A. franciscana survival. However, Artemia motility test was not sensitive to either compound ( Figure 2B).
The effects of HHCB and AHTN on P. lividus fertilization tests are presented in Figure 2C and 2D. The percentage of sea urchin that were able to fertilize in the controls were 89.50% and 90.00% for seawater and DMSO, respectively. Percentage fertilization of sea urchin tested with 0.005 µg/L, 0.05 µg/L, 0.5 µg/L and 5 µg/L HHCB were 86.33%, 82.33%, 81.00% and 82.33%, respectively. Under exposure to AHTN, sea urchin fertilization success recorded at 0.005 µg/L, 0.05 µg/L, 0.5 µg/L and 5 µg/L were 89.00%, 83.33%, 86.33% and 86.33%, respectively. Although the effect is incipient, P. lividus fertilization failure induced by both substances was more pronounced for HHCB with significant effects at 0.05 -5 µg/L, while a significant difference (p < 0.05) was only observed at 0.05 µg/L AHTN compared to control ( Figure 2C). The results for the sea urchin larvae development tested with HHCB and AHTN are presented in Figure  2D. The percentages of larvae development in the seawater and solvent controls were 83.50% and 81.50%, respectively. The number of oocytes that were able to develop to pluteus stage was significantly (p < 0.01) concentration dependent (Table S2 -Supplementary Materials) and a significant (p < 0.05) decrease in larvae development was recorded at 5 µg/L HHCB compared to the control ( Figure 2D). The percentage of larvae that were able to develop from morula stage to normal pluteus stage for 0.005 µg/L, 0.05 µg/L, 0.5 µg/L and 5 µg/L HHCB were 85.00%, 85.00%, 68.00% and 44.33%, respectively, while similar concentrations of AHTN reduced larvae development by 63.00%, 75.33%, 54.67% and 62.33%, respectively.
The results for the larval development toxicity test for M. galloprovincialis exposed to HHCB and AHTN are presented in Figure 2E. The number of fertilized eggs that developed after 48 h to D-veliger stage in the control was 98%. However, the embryotoxicity of HHCB to mussels was significantly (p < 0.01) concentration dependent (Table S2 -Supplementary Materials). From 0.05 -5 µg/L HHCB, the percentage of abnormal larvae development was significantly (p<0.05) different from the control ( Figure 2E) and the percentage effect increased to 19.88% in the highest concentration tested (5 µg/L). Similarly, significant toxicity of AHTN to embryo of M. galloprovincialis was observed in 0.5 µg/L and 5.0 µg/L with percentages of abnormal larvae of 8.36% and 11.63%, respectively ( Figure 2E).
The results of the effect of HHCB and AHTN on the survival of yolk-sac larvae of S. aurata larvae after 96 h exposure was significantly (p < 0.01) concentration dependent (Table S2 -Supplementary Materials) and the measured effect is presented in Figure 2F. After 96 h, the percentage mortality of yolk-sac larvae of S. aurata in seawater and DMSO controls was 5.33% each. Percentages of mortality in fish exposed to 0.005 µg/L, 0.05 µg/L, 0.5 µg/L and 5 µg/L of HHCB were 8.67%, 12.00%, 10.67% and 13.33%, respectively. On the other hand, the percentage mortality of fish tested with the same range of concentrations of AHTN were 10.00%, 7.33%, 12.67% and 13.33%, respectively. galloprovincialis larval development and S. aurata larval mortality tests exposed to galaxolide and tonalide. Asterisks (*) indicate significant differences (p<0.05) in relation to control.

Risk Quotient (RQ)
Quantitative risk estimation of chemicals in the environment is achieved using monitoring data of measured environmental concentration (MEC) and predicted no effect concentration (PNEC), giving a risk quotient necessary for risk characterisation. A literature survey of MEC for both polycyclic musk compounds were undertaken, and the values are presented in Table 1. Due to incipient effects detected in this study for all the species of microalgae and animals, it was not possible to calculate the EC50 values and their respective confident intervals, except for P. lividus larvae development tested with HHCB, producing an EC50 value of 4.07 µg/L. For aquatic environments (freshwater and marine), PNEC is estimated by dividing the EC50 value by an assessment factor of 1000 (EC TGD, 2003). Where possible, IC10/EC10 was evaluated and only I. galbana produced an IC10 values for HHCB and AHTN (5.22 and 0.33 µg/L, respectively) that fall within the range of the concentration tested. The IC10 values were also considered for risk estimation and the estimated risk for each substance tested are presented in Table 2. Our results indicated that HHCB posed a high risk to P. lividus larvae development with RQ between 1.18 -56.53 and both contaminants were potential threat to I. galbana growth with RQ of 0.92 -44.06 and 5.64 -609.76 for HHCB and AHTN, respectively.

Discussion
The aim of this study was to evaluate the impacts of environmentally relevant concentrations of HHCB and AHTN on marine organisms after short-term exposure and, where possible, characterise the risk following the EU technical guidance document for ERA (European Comission Joint Research Centre, 2003). Several in vitro and in vivo toxicity tests have been performed with these compounds using freshwater species (Balk and Ford, 1999b;Carlsson and Norrgren, 2004;Gooding et al., 2006;Parolini et al., 2015;Schreurs et al., 2002;Składanowski et al., 2005) and the only tests with marine organisms used high concentrations Wollenberger et al., 2003), making it difficult to compare such results with actual environmental impacts. Furthermore, for evaluation of toxicity of substances, it is imperative to use a varied battery of tests because organisms are not equally susceptible to the same toxic substance.
The effects of HHCB and AHTN on P. tricornutum, T. chuii, I. galbana and R. subcapitata showed that both compounds have limited effects on microalgae growth (Figure 1). Microalgae have been used in water quality assessments as in-situ bio-monitors because they are primary producers at the base of the ecological trophic arrangement and the basic supplier of oxygen in aquatic ecosystems (Sibila et al., 2008). Microalgae toxicity tests are useful in ERA and have gained international recognition leading to development of test guidelines for reliable and relevant toxicity data (OECD, 2002). The potential of HHCB and AHTN to inhibit the growth of microalgae in the aquatic ecosystem have been incipiently reported. Previous studies revealed that significant acute toxicity of microalgae occurred at concentration greater than 100 µg/L or 500 µg/L depending on the species (Balk and Ford, 1999b;Ding et al., 2020) but our data suggest that even at low concentrations microalgae growth might be inhibited.
The microalgae growth inhibition tests were mildly sensitive to HHCB and AHTN; IC10 values for most of the microalgae were out of range of the concentrations tested, except for I. galbana with IC10 values of 5.220 and 0.328 µg/L for HHCB and AHTN, respectively (Table 2). In contrast, previous studies have reported higher EC50 values of 0.050 mg/L for Navicula spp. and 0.336 mg/L for Scendesmus quadricauda exposed to HHCB (Ding et al., 2020). This is because microalgae are not equally sensitive to contaminants. For example, in all the microalgae tested, we found that differential sensitivity was exhibited, the order of sensitivity to HHCB was P. tricornutum > I. galbana, and to AHTN was T. chuii > P. tricornutum > I. galbana. Similarly, differential sensitivity of two microalgae to HHCB was recently reported, whilst the microalgae, Navicula spp. are more sensitive to HHCB than S. quadricauda (Ding et al., 2020).
Artemia was not sensitive to either substance since the endpoints measured were not critically different from the control (Figure 2: A & B). This is not surprising because Artemia is known to be fairly resistant to toxic substances in comparison to other invertebrates and they respond differently to different environmental contaminants (Ferreira et al., 2007;Hadjispyrou et al., 2001).
HHCB significantly reduced the fertilization success of P. lividus as the concentrations increased. Meanwhile, AHTN had no significant effects on sea urchin fertilization success ( Figure 2C). Sea urchin fertilization success depends on sperm fitness, motility potentiated by the axonemal engine in the flagellum, morphology and chemotaxic navigation (Hussain et al., 2016). For organoleptic substances, such as HHCB and AHTN, the latter might be the most compelling factor inducing reduction in sea urchin fertilization success since the exposure procedure involved prior treatment of sperm with the contaminants before the introduction of eggs. Between the two contaminants tested, HHCB had more critical effects on P. lividus fertilization by 19% reduction at 0.5 µg/L, and we observed that only 0.05 µg/L AHTN significantly reduced sea urchin fertilization by 16.67% ( Figure 2C). Importantly, the effect exerted by both contaminants on sea urchin fertilization success was below 20%, presenting less toxic effects compared to other contaminants of emerging concern. For example, 500 ng/L propanol, 500 ng/L 17α-ethinylestradiol and 5000 ng/L gemfibrozil reduced sea urchin fertilization success by 24.1%, 36.9% and 26.9%, respectively (Capolupo et al., 2018). Similarly, other contaminants of emerging concern belonging to pharmaceutical and personal care products have been reported to significantly affect sea urchin fertilization success at concentrations detected in the environment (Aguirre-Martìnez et al., 2015;Paredes et al., 2014).
Ecotoxicity studies with early life stages of aquatic organisms have been recommended as a faster and more cost-effective means of examining chemicals and environmental samples, because newly hatched larvae are sensitive to exogenous substances as the embryos lose their protective membranes and are fully exposed to potential xenobiotics (Beiras et al., 2003). P. lividus and M. galloprovincialis are well recognized in toxicity bioassays and are applied globally for the evaluation of toxicity of marine contaminants by the exposure of gametes to aqueous phases, such as surface waters and pore waters (Aguirre-Martìnez et al., 2015;Bellas et al., 2005) and elutriates (Losso et al., 2007). Critical effects were detected in P. lividus and M. galloprovincialis larvae exposed to HHCB and AHTN compared to the control ( Figure 2D and E). The effects of HHCB on sea urchin larvae development was significantly (p < 0.01) concentration dependent (Table S2 -Supplementary Materials) and significantly different (p < 0.05) compared to the control, with only 44.33% of larvae able to develop to pluteus stage after 48 h exposure to 5 µg/L HHCB. The sensitivity of sea urchins to environmental chemicals is widely reported and significant evidence showed that they represent an important biomonitoring tool for ecosystems health. Similarly, HHCB and AHTN significantly (p < 0.05) affected the development of M. galloprovincialis larvae when compared to the control ( Figure  2E). However, when considered in relation to the number of oocytes exposed to each contaminant, the effects were incipient with the highest percentage of deformed and undeveloped oocytes being 19.88% and 17.60% for HHCB and AHTN, respectively. Although P. lividus and M. galloprovincialis larvae development tests were similar, the effects recorded were more pronounced in sea urchin larvae than mussels.
Significant (p < 0.01) concentration dependent responses (Table S2 -Supplementary Materials) were observed in S. aurata mortality tests with HHCB and AHTN (Table S2 -Supplementary Materials). Notwithstanding, the percentage mortality of S. aurata exposed to both fragrances were below the toxic threshold, the highest effect being 13% for HHCB and AHTN at the highest concentration of 5 µg/L. Although this species and endpoints proved to be sensitive to other contaminants at low concentrations (Capolupo et al., 2018;Díaz-Garduño et al., 2016), the impacts of environmental concentrations of the tested fragrances in this study were low.
Analysing the RQs, HHCB represented potential high risk for the marine environment based on the EC50 estimated for P. lividus larvae development. P. lividus seems to be very sensitive to chemical exposure because previous studies of environmental contaminants in coastal waters have reported significant toxicity of industrial and domestic effluent (Díaz-Garduño et al., 2016), pharmaceutically active products (Aguirremartínez et al., 2015;Capolupo et al., 2018), UV-filters (Paredes et al., 2014) and organic pollutants (Bellas et al., 2005;Mariani et al., 2006) to sea urchin larvae development. The quantitative risk estimate of HHCB obtained in this study for P. lividus (1.18 -56.53) was comparable to that of propanol (0.02 -17.29), previously reported by Capolupo et al. (2018). We also found that both compounds posed high risk to I. galbana based on IC10 values recorded (Table 2). Microalgae have broadly been used in evaluation of ERA for other emerging contaminants. For example, I. galbana have been reported to be significantly affected by UV filters and pharmaceutical active ingredients (Aguirre-Martìnez et al., 2015;Paredes et al., 2014). Fragrances have been demonstrated to pose high risk to microalgae in the marine environment, of which HHCB and AHTN were more toxic than musk xylene and musk ketone (Díaz-Garduño et al., 2017). The adversity of measured environmental concentrations of HHCB and AHTN to I. galbana deserve attention because microalgae occupy the lowest trophic level of the marine food chain and, therefore, serve as food for higher trophic organisms, supporting a balance in the ecosystem. Again, HHCB and AHTN inhibition of microalgae growth represents a bottom-up effect that might result in structural and functional disruption of the ecosystems (Fleeger et al., 2003).
Due to the instability of these substances under laboratory exposure, it is difficult to estimate the exact concentration causing toxic effect. Although we could not measure the concentrations of the exposure water and bioaccumulation because of practical constraints, studies have shown that after 3 h of exposure, over 30% was lost and this reduced to ca. 80% after 96 h Ehiguese et al., 2019;Tumová et al., 2019;Wollenberger et al., 2003). Again, there are currently no techniques to measure the concentrations of these substances in-situ; therefore, there is a probability that reported environmental concentrations are a fraction of the actual concentration in the marine environments eliciting biological effects, given that some amount could be lost before or during sample analysis in the laboratory. As a result, the present study should be considered relevant for marine protection goals due to the evidence of potential adversity of low concentrations of HHCB and AHTN, especially to I. galbana biomass and P. lividus larvae development.

Conclusion
This study evaluated the toxicity of comparable measured environmental concentrations of HHCB and AHTN to microalgae and early life stages of marine organisms, adding to the dearth of information regarding the adverse effects of fragrances in the marine environment. For all the species of microalgae, invertebrates and fish tested, differential sensitivity was observed. Although the effects of both fragrances were generally incipient, P. lividus larvae development was the most sensitive species/endpoint. From our data, the environmental risk of HHCB and AHTN was characterized as high for P. lividus larvae development and I. galbana growth, suggesting the necessity for further research into possible sublethal effects of both substances in the marine environment.