Duration of chilling phase, but not thermal condition, influence the gonad maturation of male and female domesticated pikeperch (Sander lucioperca)

1 INTRODUCTION

The pikeperch (Sander lucioperca) is among the species with the highest potential for the development and diversification of European inland aquaculture (Mylonas & Robles, 2014). The development of pikeperch aquaculture is highly limited due to, among other factors, the seasonality of reproduction. This can be overcome by the environmental control of broodstock reproduction which can be reproduced at any time of the year in recirculated aquaculture systems (RAS). But until now, a very high variability of gonad maturation and reproductive performances is observed in pikeperch broodstock reared in RAS (Khendek et al., 2018; Zarski et al., 2015). In particular, lack of spawning is often observed when females are not exposed to hormonal treatments (Zarski et al., 2019). To this end, a well-established photothermal programme allowing for the induction of complete gonadal development is necessary.

The reproductive cycle of temperate teleosts, including percids, is strictly controlled by the fluctuation of photothermal conditions (Taranger et al., 2010; Wang et al., 2010). In Eurasian perch Perca fluviatilis, exposure to continuous photophase or constant photoperiod (16L/8D) from the start of the reproductive cycle inhibits or slows down the initiation of gametogenesis in both sexes (Migaud et al., 2004). In pikeperch, gonad maturation is inhibited by a continuous lighting while optimal maturation is only achieved under a naturally stimulated photoperiod (Ben Ammar et al., 2015). Investigations showed that temperature is also a major factor to modulate the gametogenesis in percids. In Eurasian perch, exposure to high temperature from the initiation of gonad development inhibits the progress of oogenesis (Abdulfatah et al., 2013; Fontaine et al., 2015). Temperature is also important to induce the spawn in walleye Stizostedion vitreum (Malison et al., 1998) and to induce puberty or to regulate gonadal development in domesticated pikeperch (Hermelink et al., 2011, 2013). Thus, both temperature and photoperiod seem to be crucial factors for initiation and development of gonad maturation in percids. But the data on the proper parameters of temperature and photoperiod to be applied in pikeperch broodstock are limited, and the actual photothermal programmes used are based mostly on the protocols developed for Eurasian perch (Perca fluviatilis). These programmes involve a long phase (around 4.5 months) of chilling period at a low temperature (6°C) with a short (8 h) photoperiod (Abdulfatah et al., 2011, 2013; Fontaine et al., 2015; Khendek et al., 2018). Another study reported that a chilling period lasting at least 3 months at 12°C should be fully capable of inducing gonadal development (Hermelink et al., 2011). This suggests that the photothermal programmes suitable for pikeperch gonadal development may include, by comparison to Eurasian perch, a much shorter chilling period (only 3 months) with a much higher thermal regime (12°C) during the chilling period. In this context, we hypothesize that contrary to Eurasian perch, the success of pikeperch gametogenesis advancement is not dependent neither on the rearing temperature nor on the duration of the chilling period.

To address this question, we compared the gonadal maturation and the sex-steroid production in domesticated pikeperch males and females submitted to different combinations of chilling durations (long-L or short-S) and temperatures (6 or 12°C). In this way, the morpho-anatomical indexes, gonadosomatic index (GSI) and hepatosomatic index (HSI), the profile of gametogenetic stages, 17 beta-estradiol-E2 and testosterone-T plasma concentrations were measured as these physiological endpoints belong to the most widely used parameters to evaluate the gonad maturation in fish including pikeperch.

2 MATERIALS AND METHODS

All experimental manipulations were carried out in agreement with the European and French national legislations on animal welfare after evaluation and approval of the experimental project (protocol number: APAFIS#5862-2016062720134545) by the local ethic committee in France (Name: CELMEA).

2.1 Experimental setup

In this experiment, 120 domesticated pikeperch (F1, males and females) reared in RAS were studied. The hatching occurred at Asialor Company (Pierrevillers, France), and larvae were transferred at the same day to the Aquaculture Experimental Platform (registration number for animal experimentation C54-547-18) belonging to the URAFPA facilities of the University of Lorraine (Nancy, France). Since the juvenile stage, fish were kept in two 2 m3 RAS tanks (renewal rate of 150% per hour) supplied originally with the city water. Fish were fed daily ad libitum with commercial dry pellets (Le Gouessant pellets, proteins 47%, lipids 13%, ash 6.90%, fibre 2.2%) under lighting (16L/8D) at 22°C. At a body weight ranging between 580 and 924 g, fish (age = 25 months) were distributed into eight tanks of 2 m3 for 3 weeks at 22°C with each tank containing 15 fish. During this acclimation period, the same photoperiod of 16L/8D was applied and all environmental factors remained constant until the start of the experiment (day 0). From day 0, the four groups of fish (two tanks and 30 fish per group, statistical experimental unit = fish) were then subjected to the start of different thermal regimes (Figure 1) where long (L, 4 tanks in total) or short (S, 4 tanks in total) chilling periods were applied either at 12°C (four tanks in total) or 6°C (four tanks in total). The photoperiod kinetics followed the programme described in Fontaine et al. (2015). As no apparent sexual dimorphism exists in pikeperch, both sexes were combined in the tanks without individual knowledge of the sex at this experiment start.

Thermal regimes applied in the groups subjected to long (L) and short (S) chilling period. Yellow targets correspond to samplings with organ and blood extractions. Red targets correspond to samplings with blood extractions only. D0 corresponds to the start of the photoperiod decrease. D30 corresponds to the start of the temperature decrease in all thermal regimes. D90 and D140 correspond to the start of the chilling period in the 12°C (red) and 6°C (blue) groups, respectively. D0, D30, D155, D185, D215 and D275 correspond to the sampling times T0, T1, T2, T3, T4 and T5, respectively

In each tank, water quality was assessed three times per week. pH and dissolved oxygen were measured with a multi-parametric probe ph/Oxi 340 i (WTW, Weinheim, Germany). pH was maintained above 7.4 by addition of NaCO3. The dissolved oxygen was maintained over 6 mg/L. The concentrations of total ammonia and nitrite-nitrogen were measured using indophenol blue colorimetric methods in which nitroprusside was used as the catalyst and the water sample containing the ammonium ions from each tank was treated with a solution of chlorine and phenol to give indophenol blue which can be detected by the spectrophotometer (CARY I). The concentrations of total ammonia and nitrite nitrogen always remained below 1 mg/L. The temperature and the photoperiod were controlled automatically and continuously by a centralized technical management. Potential temperature and photoperiod changes are operated once a day and, if change, the temperature variation is fixed at 1° whatever the stage of the photothermal programme. The light intensity was fixed at 10 lx at the water surface. The water flow rate remained constant at 3 m3 per hour.

2.2 Fish sampling

The first sampling of fish was performed at day 0 (T0) of the experiment, when the different photothermal programmes initiated marked by the decrease of photoperiod (Figure 1). The second sampling (only blood) occurred at day 30 (T1) before the temperature drop. The third sampling occurred at day 155 (T2) when the lowest temperature was applied in each group. The fourth sampling (only blood) was made at day 185 (T3). The fifth sampling was performed at day 215 (T4) before the resumption of the temperature increase in the two S groups. The last sampling of the experiment was performed at day 275 (T5) before the resumption of the temperature increase in the two L groups. T5 is thus the only sampling time which characterizes the long chilling period (L). At all times, fish were anaesthetised in MS-222 (150 mg/L, Sigma) and then the blood was sampled from all groups within 5 min after netting by caudal vein puncture using a 2-mL heparinized syringe. Then, blood was centrifuged at 3000 g for 10 min to collect plasma. At times T0, T2, T4 and T5, an average of 10 males (minimum: six, maximum: 18) and six females (minimum: three, maximum: 10) were weighed and measured to calculate the condition factor (fish weight*100/length3). Next, fish were euthanized by an overdose of MS-222 (300 mg/L; Sigma). After death, gonads and liver were extracted and weighted in order to calculate the gonadosomatic (GSI) and hepatosomatic (HSI) indices (GSI = 100 × gonad weight/total fish weight, %; HSI = 100 × liver weight / total fish weight, %). At all times, sampled fish were healthy and no lesion was observed.

2.3 Gonad histology

For females and males, samples of gonads were fixed in a Bouin–Holland solution for 1 week, washed once with water, twice with 70% ethanol and stored in absolute ethanol (Abdulfatah et al., 2011). Fragments of gonads were cut into 3-mm-thick slices, dehydrated/embedded in paraffin with successive baths (two baths of 1 h in alcohol 95°, two baths of 1 h in absolute ethanol, one bath of 1 h in acetone, two baths of 1 h in toluene and three baths of 1 h in a paraffin substitute dubbed Diawax). For females, sections of 6 μm were cut from the Diawax block and slides were stained with a Masson’s trichrome–haematoxylin Gill III (Merck, Darmstadt, Germany), 0.5% phloxine B (VWR, California, USA) and 0.5% light green (Sigma, Saint Quentin Fallavier, France). For males, sections of 4 μm were cut from the Diawax block and slides were stained with a trichrome: Regaud iron haematoxylin (Merck), 1% Ponceau-Fushine of Masson (Sigma) and 0.1% light green (Sigma). Gonadal development was evaluated according to the following procedure.

In females, five stages were considered:

  • Protoplasmic stage: protoplasmic oocytes with vacuole free cytoplasm
  • Cortical alveoli stage : appearance of yolk vesicles, occupying two or three rings in the cytoplasm periphery and then oocytes are full of yolk vesicles (green colour). Follicular and cellular layers are differentiated.
  • Early vitellogenesis: appearance of yolk globules in the oocytes (purple colour).
  • Late vitellogenesis: oocytes accumulate yolk globules and oil droplets in all the oocytes.
  • Atresia: the oocyte becomes atretic.

At each sampling time, three to six females per temperature and two slides per female were analysed. For each slide, 30 oocytes were randomly selected and assigned to the above oogenetic stage. The number of oocytes belonging to each stage was counted and was expressed relatively to the total number of oocytes.

In males, five stages were considered:

  • Differentiated spermatogonia A (SPGA): the largest cells of the germ line with voluminous, irregular nucleus surrounded by voluminous cytoplasm.
  • Spermatogonia B (SPGB): similar to spermatogonia A but smaller.
  • Spermatocyte (SPC): smaller than spermatogonia and identified by densely staining chromatin.
  • Spermatid (SPD): larger than spermatozoa with regularly spherical and dense nucleus.
  • Spermatozoa (SPZ): recognized by their kidney-shaped cell with light-coloured hair-like tails.

At each sampling time, six males per temperature and two slides per male were analysed. For each slide, two pictures were taken and the areas corresponding to each spermatogenetic stage were summed and expressed relatively to the whole area of the pictures.

For both sexes, the histological analysis used an optical microscope (Leica CME) equipped with an ocular camera (Jeulin, 1.3MP) and the ImageJ software. Such analysis was also used to differentiate the fish sex for T0, T1 and T3 samples while the sex was determined visually for T2, T4 and T5 during gonad extraction.

2.4 Hormone assays

The concentration of 17β-estradiol (E2, ng/mL) was assayed on 50 μL of plasma, diluted or not, using the DIAsource E2-ELISA kit (DIAsource, KAP0621). Sensitivity was 5 pg/mL. The coefficient of variation (CV) intra-assays varied between 3.1% and 2.6% and CV inter-assays varied between 4.7% and 2.4% for low and high levels, respectively. Testosterone (T, ng/mL) was assayed on 25 μL of plasma using the DIAsource Testosterone ELISA Kit (DIAsource, KAPD1559). Sensitivity was 83 pg/mL. CV intra-assays varied between 2.6% and 8.4%, and CV inter-assays varied between 9.4% and 12.1% for low and high levels, respectively.

2.5 Statistical analysis

After checking the normality and homogeneity of variances (Shapiro–Wilk and Levene tests, respectively), the effects of sampling time and temperature were analysed, for each sex independently, by a two-way analysis of variance (ANOVA) followed, when significant, by Tukey (HSD) test with different N. The statistical experimental unit was the fish. The two factors were (1) the temperature (two modalities: 6 and 12°C; four tanks per modality; from 21 to 41 statistical repetitions depending on the sex and the variable) and (2) the time (six modalities: from T0 to T5; eight tanks per modality from T0 to T4 and four tanks at T5; from 8 to 44 statistical repetitions depending on the sex and the variable). Regarding the temperature*time interaction (12 modalities), there were four tanks per modality from T0 to T4 and two tanks per temperature at T5 (from 3 to 18 statistical repetitions depending on the sex and the variable). If the assumptions of the ANOVA failed, non-parametric Kruskal–Wallis followed by Mann–Whitney or bilateral multiple comparisons tests were used. The level of significance used in all tests was p < 0.05. All statistical analyses were performed using the STATISTICA software (StatSoft, Tulsa, OK, USA).

3 RESULTS

No mortality occurred during the acclimation period and during the experimental time course. For both sexes, an elevation of growth and condition factor was observed between T0 and T2 (females: p < 10–3 and p < 10–2, respectively; males: p < 10–3 and p < 10–3, respectively) without significant differences between the two sexes (p > 0.05). Then, these parameters remained steady along the experimental time course (Tables 1 and 2). In females, the temperature significantly influenced the condition factor with a significant reduction among those kept at 12°C compared to those at 6°C (p < 0.05).

TABLE 1.
The effect of thermal condition (6 and 12°C) and chilling period on weight (mean ± SD, n = 3–10 in males, n = 6–18 in females; expressed in grams) of pikeperch. Sampling times T0, T2, T4 (S = short chilling period) and T5 (L = long chilling period) correspond, respectively, to days 0, 155, 215 and 275 after the beginning of the photothermal programme. For each sex, different letters indicate significant difference between the sampling times (p < 0.05)
Sex T0 Temperature T2 T4 (S) T5 (L)
Males 767 ± 137 a 6°C 1538 ± 308 b 1677 ± 317 b 1243 ± 144 b
12°C 1481 ± 291 b 1319 ± 196 b 1466 ± 256 b
Females 785 ± 87 a 6°C 1467 ± 262 b 1453 ± 219 b 1484 ± 188 b
12°C 1531 ± 292 b 1388 ± 276 b 1757 ± 136 b

TABLE 2.
The effect of thermal condition (6 and 12°C) and chilling period on the condition factor (mean ± SD, n = 3–10 in males, n = 6–18 in females) of pikeperch. Sampling times T0, T2, T4 (S = short chilling period) and T5 (L = long chilling period) correspond, respectively, to days 0, 155, 215 and 275 after the beginning of the photothermal programme. For each sex, different letters indicate significant differences between the sampling times (p < 0.05). In females, the asterisks indicate a significant temperature effect over the whole sampling period except T0 (p < 0.05)
Sex T0 Temperature T2 (D155) T4 (D215) T5 (D275)
Males 0.87 ± 0.13 a 6°C 1.09 ± 0.11 b 1.12 ± 0.05 b 1.06 ± 0.09 b
12°C 1.06 ± 0.08 b 1.02 ± 0.08 b 1.02 ± 0.16 b
Females 0.89 ± 0.06 a 6°C 1.11 ± 0.1 b,* 1.15 ± 0.13 b,* 1.23 ± 0.13 b,*
12°C 1.05 ± 0.12 b,* 1.02 ± 0.15 b,* 1.06 ± 0.12 b,*

In females, a progressive increase of GSI was observed along the experimental time course (p < 10–4) associated with a significant difference between T4 and T5 (p < 10–2, Figure 2a). However, the temperature did not significantly influence the GSI (p > 0.05). The temperature (p < 10–3), time (p < 10–4) and the interaction between temperature and time (p < 0.05) significantly influenced the HSI (Figure 2b) marked by a steady and elevated level from T1 in pikeperch at 6°C contrary to those at 12°C which returned to the initial level.

image
The effect of thermal condition (6 and 12°C) and chilling period on gonadosomatic index GSI (Figure 2a, mean + SD, n = 3–10) and hepatosomatic index HSI (Figure 2b, mean + SD, n = 3–10) values of female pikeperch. Sampling times T0, T2, T4 (S = short chilling period) and T5 (L = long chilling period) correspond, respectively, to days 0, 155, 215, and 275 after the beginning of the photothermal programme. Different letters indicate significant differences between the sampling times or between the interactions time*temperature (p 

Like in females, only the length of the chilling period affected the male GSI (p < 10–4) but no difference was noticeable between T4 and T5 (Figure 3a). The temperature (p < 10–3) and time (p < 10–4) significantly influenced the HSI, but the interaction between both factors was not significant (Figure 3b). The HSI was lower among males kept at 12°C compared to those kept at 6°C.

image
The effect of thermal condition (6 and 12°C) and chilling period on gonadosomatic index GSI (Figure 3a, mean + SD, n = 6–18) and hepatosomatic index HSI (Figure 3b, mean + SD, n = 6–18) values of male pikeperch. Sampling times T0, T2, T4 (S = short chilling period) and T5 (L = long chilling period) correspond, respectively, to days 0, 155, 215 and 275 after the beginning of the photothermal programme. Different letters indicate significant differences between the sampling times (p p 

The histological analysis did not reveal any effect of temperature for either sex. However, a marked difference was observed between T4 and T5 showing a much more advanced gametogenesis when the chilling period was lengthened (Figure 4). At T4, females and males reached the early vitellogenesis stage and the spermatid stage, respectively, while at T5 they reached the late vitellogenesis stage and the spermatozoa stage, respectively (Figure 5).

image

The effect of thermal condition (6 and 12°C) and chilling period on oogenesis and spermatogenesis stages (n = 3–6 per histogram) of pikeperch. Sampling times T0, T2, T4 (S = short chilling period) and T5 (L = long chilling period) correspond, respectively, to days 0, 155, 215 and 275 after the beginning of the photothermal programme. Vtg , vitellogenesis. SPG, spermatogonia; SPC, spermatocytes; SPD , spermatids; SPZ, spermatozoa

image

The effect of chilling period on the most advanced gametogenesis stage of pikeperch reached at the end of this period. Top left: early vitellogenesis stage reached at T4 (S = short chilling period). Top right: late vitellogenesis stage reached at T5 (L = long chilling period). Bottom left: spermatid stage reached at T4 (S = short chilling period). Bottom right: spermatozoa stage reached at T5 (L = long chilling period). Vtg, vitellogenesis; SPD, spermatids; SPZ, spermatozoa

The hormonal analysis did not show any temperature effect on plasma E2 and T levels (p > 0.05; Figure 6) for either sex. But they revealed strong fluctuations along the experimental time course for both females and males (p < 10–4 for each group). In females, E2 and T plasma levels peaked at T3 and T5, respectively, with a significant marked elevation of T between T4 and T5 (P < 0.01; Figure 6a and b). In males, the highest E2 and T levels were measured at T3 and T4 and T5, respectively (Figure 6c and d).

image
The effect of thermal condition (6 and 12°C) and chilling period on 17 beta-estradiol (E2) and testosterone (T) values of female (Figures 6a and 6b respectively, mean + SD, n = 3–44) and male (Figures 6c and 6d respectively, mean + SD, n = 6–23) pikeperch. Sampling times T0, T1, T2, T3, T4 (S = short chilling period) and T5 (L = long chilling period) correspond, respectively, to days 0, 30, 155, 185, 215 and 275 after the beginning of the photothermal programme. Different letters indicate significant differences between the sampling times (p 

4 DISCUSSION

Regardless of the sex of the fish, changes in temperature during the chilling period neither caused a change in the GSI level or caused any marked alteration of the gametogenesis stages. In contrast, in the same species, the females reared at least 3 months at 12°C during the chilling period exhibited a higher GSI than those kept at 6°C (Hermelink et al., 2011). This apparent discrepancy may stem from five major differences between both studies. First, the initial average weights highly differ between both experiments (340 g vs. 775 g). Also, the absence of weight change during the chilling period (Hermelink et al., 2011) contrasts strikingly with the twofold weight increase observed in our experiment, possibly due to differences in the diet composition. Second, in the study of Hermelink et al. (2011), the rate of GSI increase in females from the initial time was in the range of three to five fold (initial GSI: 0.9%; final GSI: 4%). This rate is much lower than the rate of GSI observed herein found to have increased in the range of 15–30 fold (initial GSI: 0.45%; final GSI from 7% to 12%). These different zootechnical performances suggest a contrasted response to temperature between the pikeperch populations, which can be partly explained by different features in the experimental setup. Third, the duration of the programme was much longer in our study (9 months vs. 6 months). This comes from the fact that, to reach the temperature plateau during the chilling period, a slow temperature drop was applied in our experiment (from 60 to 110 days). This kinetics is in accordance with data on programme optimization obtained for the Eurasian perch Perca fluviatils (Abdulfatah et al., 2011) and contrary to the quick drop applied in the study of Hermelink et al. (2011). Fourth, the level and the variations of photoperiod applied during our photothermal programme (Fontaine et al., 2015) contrast with the constant photoperiod maintained at 12 h/12 h during the studies of Hermelink et al. (2011, 2016). Given the effect of photoperiod on oogenesis in pikeperch female maturation (Ben Ammar et al., 2015), we hypothesize that natural changes of photoperiod are necessary to optimize the GSI. Last, the contrasting results may come from the features of the domesticated fish populations, notably the fish life history. Indeed, the maturation success in response to such a photothermal programme depends on the captive pikeperch population used (Khendek et al., 2018).

In the present experiment, the maintenance at 12°C during the chilling phase did not disrupt the gonad maturation in pikeperch. Wang et al. (2010) suggest that in percid fishes a given temperature threshold exists, above which gametogenesis would be impaired. In Eurasian perch females, a temperature in the range of 6–14° degrees does not strongly impact the progress of gametogenesis and gonadogenesis (Abdulfatah et al., 2013). In yellow perch, the maintenance at constant 23°C, instead of applying a chilling phase à 10°C, inhibits the GSI increase and the gametogenesis advancement (Kolkovski & Dabrowski, 1998; Shewmon et al., 2007). Our data thus support that the achievement of percid gametogenesis in RAS conditions is not markedly influenced by the rearing temperature below 12–14°C. At higher temperatures, the slowdown or even blockage of gametogenesis is often accompanied with a decrease of sex-steroid production (e.g. Abdulfatah et al., 2013). In our study, the absence of temperature effect of sex-steroid levels might explain why gametogenesis was not disrupted by the tested thermal conditions.

Despite no apparent effect on pikeperch gametogenesis, the 12°C temperature induced the drop of HSI in both sexes associated with a decrease in the condition factor, at least in females. This suggests that the whole energetic metabolism was affected by such temperature. These data are consistent with what observed in other fish species for which an elevated temperature during vitellogenesis is linked to a decrease of HSI (e.g. Im et al. 2016; Akhoundian et al. 2020). Still, pikeperch is thought to be one of the most thermophilic percid fish species studies so far (Geay & Kestemont, 2015). Thus, this thermophilic feature does not seem to apply at such low temperature for maturing broodstock. From the perspective of the fish production, this drop of HSI at 12°C raises questions regarding the long-term consequences on the broodstock growth and reproductive performance. However, the maintenance of fish at a higher temperature is obviously beneficial for the fish farmers in terms of energetic costs.

Additionally, in both sexes, a further degree of gonadogenesis and/or gametogenesis was observed when the chilling period was extended by 2 months (60 days). This phenomenon was not clearly demonstrated in the experiment of Hermelink et al. (2011), and it may also be explained by the longer duration of the programme in our study. This shows that in our photothermal conditions, the chilling period should exceed the length of time (135 days at 6°C or 185 days at 12°C; see Figure 1) that is necessary for the completion of gametogenesis and gonadogenesis. This further advancement may be due to the maintenance or even increase of the plasma sex-steroids E2 and/or T during this extension of chilling period. Indeed, E2 and T are both involved in the regulation of fish hepatic vitellogenesis and spermatogenesis (Nagahama & Yamashita, 2008; Lopes et al., 2020). In females, the vitellogenin is incorporated into the oocytes during vitellogenesis which contribute to the oocyte growth and the advanced gonadogenesis. Besides E2 (or T), vitellogenin and GSI levels are often highly positively correlated with each other during the oogenesis (Koya et al., 2003; Martyniuk et al., 2013; Khendek et al., 2017). In males, the increase of T during spermatogenesis is also often linked to the increase of GSI (Melamed et al., 2000; Khendek et al., 2018). In that study, we assume that the persistence or increase of the plasma concentrations of E2 and T from T4 to T5 was necessary to still progress in the gonadogenesis and gametogenesis processes by achieving vitellogenesis in females and the transition from the spermatid stage to the spermatozoa stage in males.

In conclusion, to optimize the advancement of pikeperch gonadal development, we recommend to apply a slow temperature decline of 2 months (60 days) after the programme start followed by temperature stabilization at 12°C for a minimum of 6 months (185 days). Still, further studies are necessary to test the consequences on gamete quality, reproductive success and broodstock growth.

ACKNOWLEDGEMENTS

The authors acknowledge the Eurostars Programme (E!9390 TRANSANDER) that supported financially that experiment. This study was partly supported by the Eurostars project (E!9390 TRANSANDER).

CONFLICTS OF INTEREST

The authors declare that they have no conflict of interest.

ETHICS APPROVAL STATEMENT

All experimental manipulations were carried out in agreement with the European and French national legislations on animal welfare after evaluation and approval of the experimental project (protocol number: APAFIS#5862-2016062720134545)

AUTHOR CONTRIBUTIONS

Sylvain Milla: conceptualization; funding acquisition; project administration; writing – review and editing. Amine Khendek: conceptualization; investigation; writing – review and editing. Daniel Zarski: conceptualization; funding acquisition; project administration; writing – review and editing. Imen Ben Ammar: investigation; writing – review and editing. Pascal Fontaine: funding acquisition; project administration; writing – review and editing.

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

REFERENCES