Introduction

Marine waters with a depth greater than 1000 m represent 88% of oceanic areas (Jannasch and Taylor 1984). Until few decades, they were isolated from the direct footprint of anthropic activities. However, due to both technological improvements and an increase in resources demand (oil, gas, seafloor massive sulfide deposits), the exploitation of deep seas is more and more economically viable. Nowadays, offshore deep waters represent an important part of the new oil exploration/production areas (Maddahi and Mortazavi 2011). Discharge of chemicals, whether chronic or accidental, could upset the balance of these sensitive ecosystems. Consequently, the need for relevant data giving a better understanding of the potential environmental risks of anthropic activities in deep-sea areas is becoming increasingly crucial.

Ecological studies are difficult to carry out in these environments. However, they provide fundamental information after deep-sea contaminations, as in 2010 in the case of the Deepwater Horizon accident in the Gulf of Mexico: about 780,000 m3 of oil were released into the environment at 1500 m depth, and dispersant (7000 m3) was injected directly into the leaking Macondo well, reducing the rise of oil to the sea surface. This technical response led to the sedimentation of an estimated 3 to 14% of the 780,000 m3 released at sea (Chanton et al. 2014; Fingas 2017). Biological impacts of this sedimented oil were reported in many ecological studies (Beyer et al. 2016). Additionally, experimental approaches can provide relevant data to assess and understand the new field of deep-sea ecotoxicology. To date, few data are available, but several approaches could be used (Lemaire et al. 2012; Lemaire 2017; McConville et al. 2018). One of these is the study of the impact of chemicals on piezotolerant species at atmospheric pressure. This approach has been carried out on cold-water corals (DeLeo et al. 2016), on the benthic giant amphipod Eurythenes gryllus (Olsen et al. 2016), or on deep-sea fish species Coryphaenoides rupestris (Lemaire et al. 2012) and Anoplopoma fimbria (McConville et al. 2018). Nevertheless, in order to obtain a better understanding of the potential biological impact of a chemical in the deep-sea, it is necessary to develop experimental protocols to recreate contamination and assess its effects under high hydrostatic pressure (HP) (Lemaire et al. 2012). Ecobarotoxicology, i.e., the use of a hyperbaric chamber to assess the potential impact of chemicals at simulated depth, offers a relevant way to understand this new domain of ecotoxicology.

The ideal experiment would be the contamination of a deep species at depth or simulated depth (Lemaire et al. 2012). This type of experiment is difficult to achieve, but alternatives exist. It is possible to analyze the effects of HP on animals prior exposed to chemicals at atmospheric pressure. This kind of study was performed by exposing juvenile fishes contaminated with dispersed oil (Dussauze et al. 2017) or shallow shrimps exposed to different cadmium concentrations (Domingues 2015) to HP. Another way is to perform a comparison of toxicity at the surface and at simulated depth. In this case, the use of a reference species for scientific and legislative purposes is a good compromise. In our study, the turbot Scophthalmus maximus was selected as a model species. The originality of our experiment was to conduct contamination concomitantly at a simulated depth of 1000 m and at atmospheric pressure in order to compare the lethal concentrations (LC) of dispersed oil. The LCs were determined using the total hydrocarbon concentration (THC) during the exposure time. The LC results are presented for 10 and 50% of mortality (respectively, LC10 ad LC50) both at atmospheric pressure and at high hydrostatic pressure.

Materials and methods

Animals

Juvenile turbot (Scophthalmus maximus) (n = 792; weight: 2.89 ± 0.02 g mean ± SEM, France turbot -Noirmoutier, France) were acclimatized in a 300 L seawater tank for 2 weeks. The photoperiod was according to the season (June to August). Oxygen saturation (> 95%), pH (8.1 ± 0.2), and temperature (maintained at 15.2 ± 0.3 °C) were measured daily. The turbot were fed daily with dried commercial pellets provided by the hatchery.

Chemicals

The petroleum oil used in the study was a crude oil from the Lula oil field. This offshore oil field is located in ultra-deep water at approximately 2000 m depth. This oil contained 54% saturated hydrocarbons, 10% polar compounds, and 36% aromatic hydrocarbons. The dispersant was the Finasol OSR52, a third-generation formulation from the TOTAL Company.

Experimental systems

In this study, the oil concentration range (0 to 300 mg L− 1 of THC) and the oil exposure duration (24 h) were defined according to a pilot experimentation using an ecotoxicological bench of twelve replicates (see below). The complete procedure of LC50 determination, used for the pilot experimentation, was previously described in Dussauze et al. (2015).

The experimental system (Fig. 1) is composed of one source tank, two exposure tanks and one hyperbaric chamber. The 100 L source tank was equipped with a water homogenizing system similar to the system described in detail in Milinkovitch et al. (2011). The two exposure tanks (15 L cylindrical plexiglass tanks) had continuous renewal (1.2 L min− 1) of their water from the source tank. One of these tanks was placed in the hyperbaric chamber and immediately next to it. The 130 L hyperbaric chamber (fully described in Theron et al. 2000) has hydraulic hull passageways allowing continuous water flow under pressure. This water circulation allows temperature and oxygen saturation to be controlled during the pressure exposure. The exposure tanks had a gas impermeable deformable membrane at one end, allowing the transmission of pressure to the water without gas dissolution. The configuration of the experimental device made it possible to directly compare contaminations carried out at atmospheric pressure and at a simulated depth of 1000 m.

Fig. 1
figure 1

Hyperbaric chamber and high-pressure water circulation system from Theron et al. (2000) coupled with one source tank and two exposure tanks

Full size image

Experimental conditions

Thirty-three fish were randomly assigned to each of the two exposure tanks. Then, the tank used for pressure was placed in the hyperbaric chamber. The animals were acclimated for 24 h to the experimental system. Then, the pressure was gradually increased in the hyperbaric chamber at a speed of 0.2 MPa/h to reach 10.1 MPa. After 24 h of acclimatization, the oil-dispersant mixture was poured into the source tank every 6 h for 24 h. The exposure tanks were continuously supplied with contaminated water from the source tank. The total hydrocarbons concentrations (THC) were measured in water samples collected every 2 h. At the end of the 24 h exposure time, decompression of the hyperbaric chamber was performed at 0.1 MPa/min, and then survival of fish was assessed.

For this experimentation, ten oil exposures were performed concomitantly at 0.1 and 10.1 MPa, and four oil exposures were performed at 0.1 MPa. During the experimentation, 462 fish were assigned to the 0.1 MPa condition and 330 fish to the 10.1 MPa condition.

Measurement of seawater Total hydrocarbon concentrations (THC)

For the two exposure tanks, the THC was measured every 2 h during the contamination period in triplicate from 30 ml samples. Following the method described by Fusey and Oudot (1976), a separation of the aqueous and the organic phases was performed. Then, the absorbance of the organic phase was measured and quantified at a wavelength of 390 nm (UV-Vis spectrophometer, Unicam, France). The results are expressed in mg L− 1.

Statistics

Using the Excel™ macro: REGTOX© version 7.0.6, the LC10 and LC50 were calculated on the basis of the mean measured THC concentrations over the contamination period. LCs were expressed in mg.L−1 and given with their 95% confidence interval. The results were considered as statistically different when their confidence interval at 95% did not overlap.

Results and discussions

This study aims to compare the acute toxicity of dispersed oil on a model fish, at atmospheric pressure (0.1 MPa) and under high HP (10.1 MPa). As reported by Mestre et al. (2013), there is an important need to obtain information from ecotoxicological trials for environmental risk assessments in the deep-sea environment and particularly under high hydrostatic pressure. However, HP is a difficult parameter to control, and therefore, only few laboratories are able to produce data in this particular ecotoxicology context (Vevers et al. 2010; Dussauze et al. 2016, 2017; Lemaire et al. 2012; Lemaire 2017). Due to the difficulty in bringing and maintaining deep-sea species, the experimental protocol was carried out on a model fish, the turbot Scophthalmus maximus. This species was chosen for two main reasons: (1) this fish is one of the model species for ecotoxicological tests for the OSPAR Commission (OSPAR Protocols on Methods for the Testing of Chemicals Used in the Offshore Oil Industry),(2) this species is present from the surface to at least 70 m of depth (Muus and Dahlstrøm 1989) in European marine waters, from the Mediterranean Sea to the Arctic Circle (Imsland et al. 2001). This broad distribution area illustrates the variation of habitat that this species is able to cope with. These strong adaptive capacities of shallow species have previously been used in other ecobarotoxicological studies: on seabass Dicentrarchus labrax (Dussauze et al. 2017; Lemaire et al. 2016) or on shrimp Palaemon varians (Domingues 2015). In ecotoxicology, determination of the LC is the first step in the understanding of the toxicity of a chemical. It makes it possible to determine the sensitivity of the studied species to a given chemical, and it can be used to compare this sensitivity with other species or other chemicals. This approach was used in our study to compare the toxicity of a chemically dispersed oil both at simulated depth (1000 m) and at atmospheric pressure, and therefore to have information about the pressure dependence of toxicity responses.

In the pilot experiment, the LC50 at 24 h of exposure was 113.1 (104.1–126.2) mg L− 1. 0%, and 100% of mortality were reached respectively at 87 ± 22 and 209 ± 37 mg L− 1 of dispersed oil. A 48 h observation period showed that no mortality occurred after the first 24 h of contamination. This pilot experiment provided an exposure time (24 h) and a concentration range (0 to 300 mg L− 1 of THC) of interest to carry out the main experiment (i.e., comparison of the toxicity at the surface and at depth). In our experimental system, no mortality was observed in the control groups (without oil exposure) either at 0.1 and 10.1 MPa. Then, for an equivalent THC, a higher mortality was observed at 10.1 MPa when compared to fish kept at 0.1 MPa (Fig. 2). The LC10 were not statistically different at 0.1 and 10.1 MPa (Table 1). However, the LC50 for the surface experiment (90.8 mg L− 1) was statistically higher than the LC50 measured at 10.1 MPa (50.9 mg L− 1).

Fig. 2
figure 2

Mortality curves calculated according to the Total Hydrocarbon Concentrations after 24 h of exposure for experimentations performed at the surface (0.1 MPa) and under pressure (10.1 MPa) and during the pilot experimentation. Letter differences indicate statistical differences between groups

Full size image
Table 1 Lethal concentrations 10 (LC10) and 50 (LC50) for the surface and high pressure experiments
Full size table

The differences in mortalities according to the experimental condition could be explained by the proteiform strong biological impacts of HP on:

  • (1) on molecular system with biological membranes and specially their fluidity (Cossins and Macdonald 1989; MacDonald 1997) or in the aryl hydrocarbon receptor pathway and enzymatic antioxidant systems involved in the defense against organic chemicals in marine fish (Lemaire 2017).

  • (2) the nervous system with high pressure nervous syndrome (HPNS) (Bennett and Towse 1971; Somero 1992). In our study, fish exposed to pressure was observed using the porthole of the hyperbaric chamber. From our observations, no erratic fish behaviors (swimming acceleration, convulsion or loss of balance), which are characteristic of HPNS (Dussauze et al. 2017), were observed. The only difference in behavior between fish exposed to atmospheric pressure and hydrostatic pressure was a trend to swim to the surface during the acclimatization of fish to pressure.

  • (3) energetic metabolism (Sébert et al. 1998; Sébert and Theron 2001). This last point is of particular interest since oil or hydrocarbons exposure is known to induce a reduction in the maximal O2 consumption of the respiratory chain (Stabenau et al. 2008; Dussauze et al. 2014). This effect is possibly mediated via reactive oxygen species (ROS) which have an impact on several complexes (I, III, and IV) of the respiratory chain (Musatov and Robinson 2012). Furthermore, oil is known to induce deleterious effects on the cardiovascular function of fish (Brette et al. 2014; Tissier et al. 2015). In our experiment, it is possible that oil exposure (through its impact on mitochondrial activity and cardiovascular function) induces a reduction of the fish metabolic scope and consequently limits energy availability at a time when HP induces an increase in the energetic demand. This assumption is in good coherence with a recent study showing that oil contamination has a significant deleterious impact on the adaptive capacities of juvenile Dicentrarchus labrax to an increase in HP (Dussauze et al. 2017).

Conclusion and outlook

Deep-sea ecotoxicology is an emerging phenomenon caused by the direct or indirect impact of human activities (oil and mining industries, red mud, microplastics, fisheries…) on deep-sea ecosystems. The scope is broad to understand these anthropic impacts, which were highlighted by the Deepwater Horizon oil spill. Future studies need to focus on true deep species. Several species could be good candidates: fish species such as sablefish (Anoplopoma fimbria) or Atlantic halibut (Hipoglossus hipoglossus) due to their presence at great depth and their controlled life cycle; Deep-sea corals such as Lophelia pertusa are also a possible and interesting species to study. Finally, species performing diel vertical migration, such as some calanoid zooplankton species can offer a wide range of possibilities to study the impact of oil on deep species.