Ethics
The study was conducted in accordance with the European Directive 2010/63/EU of 22nd September on the protection of animals used for scientific purposes; the Spanish Royal Decree 53/2013 of February 1st on the protection of animals used for experimentation or other scientific purposes; the Catalan Law 5/1995 of June 21th, for protection of animals used for experimentation or other scientific purposes and the Catalan Decree 214/1997 of July 30th for the regulation of the use of animals for the experimentation or other scientific purposes. The procedures used were evaluated by IRTA’s (Institute of Agrifood Research and Technology) Committee of Ethics and Experimental Animal (CEEA) and the Catalan Government and procedures authorized with ID V7MH4802M. The authors complied with the ARRIVE guidelines.
Broodstock maintenance
A flathead grey mullet broodstock was formed with individuals originally obtained from the Ebro River (Spain) or from a semi-extensive pond fish farm (Finca Veta La Palma, Isla Mayor, Spain) which had been held for 1.5 to 3.5 years in IRTA facilities (Sant Carles de la Ràpita, Spain). Thirty females with weights ranging from 0.9 to 2.2 kg and standard length (SL) from 37 to 53 cm, and 15 males ranging from 0.7 to 1.3 kg and 34 to 43.5 cm SL were used. All fish were larger than the reported SL for first maturation in this species (27–35 cm for females, 25–30 cm for males)12. To identify individuals, each fish was tagged intramuscularly with a Passive Integrated Transponder (PIT) tag (Trovan®, ZEUS Euroinversiones S.L. Madrid, Spain). The sex of the individuals was determined by the presence or absence of oocytes obtained through slight suction with a 1.67-mm plastic catheter inserted through the genital opening. Fish were maintained in 10 m3 covered tanks in a recirculating system (IRTAmar®) supplied with 36‰ salinity water under natural conditions of light and controlled winter temperatures (≥ 14 °C) during the last year. Before the study, conducted from early August to early November, the selected broodstock was transferred to another 10 m3 tank, water temperature was controlled at 23.1 ± 0.2 °C and photoperiod was ambient (14L:10D–11L:13D). Fish were fed 7 days a week at the rate of 1.5% of their body weight with a mix of two commercial marine fish diets; 90% mix of Le-2 and Le-5 Europa RG (Skretting, Spain) and 10% Brood Feed Lean (Sparos, Portugal). During all experimental procedures, for hormone administration and sampling, fish were anaesthetised with 73 mg L−1 of MS-222. When required for the study, males were euthanised with an overdose of MS-222 (250 mg L−1) and death was confirmed by a cut in the gills to exsanguinate the fish.
Hormonal induction
Induction of vitellogenesis
Females were assigned randomly to rGth and control groups taking care to have similar distribution of females in different initial maturation stages in both groups. The control group (total n = 9) was formed with 6 females in previtellogenesis (5 in primary growth and 1 in cortical alveoli stage) and 3 in early vitellogenesis. A total of 21 females were assigned to receive the hormonal treatment; 12 females initially were in previtellogenesis (8 in primary growth and 4 in cortical alveoli stage) and 9 in early vitellogenesis. Those females in early vitellogenesis had longer time in intensive captive conditions (> 2.25 years) although not all females that were held for this time had started vitellogenesis.
Single chain recombinant Mugil cephalus rFsh and rLh produced in Chinese Hamster Ovary (CHO) cells were purchased from Rara Avis Biotec S.L. (Valencia, Spain). Mugil cephalus rFsh was supplied with a concentration of 12 μg mL−1 and rLh with concentration of 8 μg mL−1. A methodology based on the protocol described by Ramos-Júdez et al.9 was followed. The pattern of application of rFsh and rLh aimed to mimic the physiological variations of Fsh and Lh associated with natural reproductive development; initially only rFsh was administered during the early stages of gametogenesis and, subsequently, a decrease in rFsh with an increase of rLh to regulate late gametogenesis16. The protocol was applied according to ovarian development (Fig. 1). Increasing weekly doses of 6, 9 and 12 µg kg−1 rFsh were administered intramuscularly to induce previtellogenesis (~ 100 µm oocyte diameter) to vitellogenesis (> 200 µm). Weekly doses were maintained at 12 µg kg−1 during vitellogenesis. As vitellogenesis progressed and when mean diameter of the most developed oocytes was ≥ 300 µm, females received in addition a weekly administration of rLh at rising doses. Initially an rLh dose of 2.5 µg kg−1 was given and maintained until females entered into late-vitellogenesis (≥ 400 µm)17 and was then increased to 4 and 6 μg kg−1. At an oocyte diameter of ~ 500 µm, weekly rFsh doses were reduced to 4 μg kg−1 while rLh was increased to 9 μg kg−1. A combination of 4 μg kg−1 rFsh and 12 μg kg−1 rLh per week was administered until vitellogenic growth was completed. The completion of oocyte growth was determined when oocytes were deemed approaching maturation; microscopic examination showed that the most developed oocytes were nearing 600 µm in diameter. The nine females in the control group were also manipulated each week and were injected with saline solution (1 mL) a total of twelve times.
Schematic representation of the rFsh and rLh treatment applied to flathead grey mullet females (n = 21). The protocol was applied according to the development of female gonads determined by ovarian biopsies. Weekly doses of 6, 9 and 12 µg kg−1 rFsh were applied to induce previtellogenesis (~ 100 µm oocyte diameter) to vitellogenesis (> 200 µm) and vitellogenic growth was maintained with a dose of 12 µg kg−1 per week. When the mean diameter of the largest oocytes was ≥ 300 µm, in addition to rFsh, rLh was administered in increasing doses. Initially an rLh dose of 2.5 µg kg−1 was given and maintained until females presented ≥ 400 µm oocytes and was raised to 4 and 6 μg kg−1. At ~ 500 µm diameter, weekly rFsh doses were reduced to 4 μg kg−1 whereas rLh was increased to 9 μg kg−1. A combination of 4 μg kg−1 rFsh and 12 μg kg−1 rLh per week were administered until vitellogenic growth was completed (~ 600 µm). Each point corresponds to a weekly administration. This scheme represents the longest pattern of administration of those females that required a total of thirteen weeks to complete vitellogenic growth.
Assessment of gonadal development was undertaken on alternate weeks by ovarian biopsies obtained by slight suction through a plastic cannula. Fresh ovarian samples were examined under a microscope (× 40 magnification), to measure the mean diameter of the most advanced oocytes (n = 20 per female) and a sample fixed for histology. Blood samples were obtained before the initial treatment (week 0) and when vitellogenic growth was completed in the treated group and at the end of the experiment in the control group.
Hormonal administration in males
In parallel, males were assigned randomly to rGth and control groups taking care to have similar distribution of males in different initial maturation stages in both groups. The control group (total n = 6) was formed with 5 males with no presence of sperm and 1 with a presence of sperm (sperm index 1, presence of sperm, but not fluid, see below). A total of 9 males were assigned to receive the hormonal treatment; 7 males with no presence of sperm and 2 with presence of sperm (sperm index 1). Males treated with rGth were split into two groups that received the same treatment, but at different times in the experimental period. The reason was to assure the availability of spermiating males for spawning induction when the females completed vitellogenic growth. Group 1 of males (n = 4 treated, n = 3 control) initiated the treatment on week 1 and the group 2 (n = 5 treated, n = 3 control) on week 3 (Fig. 2). The aim of the treatment was to apply rGths according to their described role in spermatogenesis. Increasing doses of rFsh (6, 9 and 12 µg kg−1) were administered at early spermatogenesis and high levels (12 µg kg−1) during testicular growth (Fig. 3). No rLh was administered during early spermatogenesis, both rFsh and rLh (12 µg kg−1 of both rFsh and rLh) during the middle stages of testicular growth and only rLh (12 µg kg−1) was administered at late stages to induce sperm maturation18. Group 1 received the treatment for a total of 12 weeks whereas group 2 for 10 weeks. At week 9, group 1 received a dose of 12 µg kg−1 rFsh instead of rLh as a reduction in the spermiation stage was observed (see “Induction of spermatogenesis and spermiation” section). Doses were in the range of previous studies with males using rGths produced in CHO cells6,7,8,9. Control males were injected with 1 mL of saline solution per week.
Overall scheme of the experiment. Females receiving the rGths treatment were injected weekly until completion of vitellogenic growth which took place in a maximum period of 13 weeks, following 13 administrations. Arrow heads indicate the moment in which individual females were induced to spawn (number of females indicated by the number on the top). Control females received a total of 12 saline injections. Group 1 of males started the treatments (rGth treatment and saline) on week 1, whereas the group 2 on week 3.
Diagram of the hormonal treatment of flathead grey mullet males. Asterisk indicates the moment that group 1 received 12 µg kg−1 of rFsh instead of rLh. Green indicates rFsh administration and blue indicates rLh administration.
In order to evaluate the progression of maturation, at the beginning of the treatment and on a weekly basis, a gentle squeeze on the ventral abdomen toward the urogenital opening was applied to release the sperm. Spermiation stage was determined on a scale from 0 to 3 (0 = not fluent, 1 = fluent but no sample can be obtained, 2 = fluent, 3 = very fluent). By this method, only mature cells are released together with the seminal plasma. Therefore, to determine the stage of development in males, some males were sacrificed at the beginning (n = 2) and the end of the treatment (n = 2 per group) and the testes removed for histological procedures and the measurement of gonadosomatic index (GSI: testes weight/fish weight × 100). Towards the end of the experiment (weeks 10 and 13), sperm samples from running males were collected, total volume recorded and stored at 4 °C for sperm quality analysis. Blood samples were also taken from treated and control males each 2 weeks of the treatment.
Oocyte final maturation and spawning induction
Three different treatments were followed for maturation and spawning induction. Females that completed vitellogenic growth (as determined by gonadal biopsy) and males with running milt (spermiating stage 2 or 3) were selected for spawning induction. The individuals were separated from the main group and stocked in a separate 10-m3 tank per treatment. The three separate tanks had the same conditions as the holding tank. One male in each spawning tank received a dose of 24 µg kg−1 of rLh while the others followed the hormonal treatment previously described, receiving 12 µg kg−1 of rLh. The selected females were injected intramuscularly with either (i) priming rLh and resolving Progesterone (P4) (Prolutex, IBSA Group, Italy) as in Ramos-Júdez et al.9 (ii) rLh for priming and resolving injections or (iii) P4 for priming and resolving injections (doses in Table 1). Priming and resolving injections were administered 24:05 ± 0:40 h apart. Ovarian samples were taken with a cannula and examined before each administration. Maturation and ovulation induction of females from each treatment group was staggered on different consecutive days in order to separate the different spawning events. The sex ratio was 1:2 or 1:3 (female:male) per spawning event depending on availability of males. After all the spawning events, males were returned back with the main group to the initial tank. In the cases where females ovulated and showed a swollen belly, but did not release the eggs, manual stripping was applied.
Parallelly, oocytes obtained from cannulation from females (n = 6) that had received the priming injection of rLh (30 µg kg−1) were incubated in vitro with different hormones. A total of 58.7 ± 29.9 oocytes were incubated in a well of a 96-well plate with 200 µL of Leibovitz’s L-15 medium, with either no hormone (control), rLh (concentrations of 400, 200, 100, 50 and 10 ng mL−1), rFsh (400, 200, 100, 50 and 10 ng mL−1) or P4 (4000, 1000, 500 and 50 ng mL−1). Each treatment was applied in triplicate to oocytes from each female. After 48 h of incubation at 21 °C, the follicles were examined under a binocular microscope and oocytes without the follicular layer (ovulated) and intact follicles (un-ovulated) were counted for each well.
Egg collection and incubation
Surface out-flow egg collectors (mesh size of 500 μm) were placed to receive eggs from the tanks and were frequently inspected for eggs. When spawning was observed, eggs were transferred into a 10 L bucket. The number of eggs per spawn (fecundity) was estimated by counting the eggs in triplicate subsamples. A sample of eggs (n = 50–100) was observed under a microscope and the percentage (%) fertilization for each batch of spawned eggs was determined by calculating the % of eggs that reached the 2- to 16-cell stage. Eggs for incubation were collected after careful agitation of the eggs in the 10 L bucket. The eggs were incubated at a density of 13,230 ± 7273 eggs L−1 in 30 L incubators with the same conditions as the broodstock holding tanks. Each incubator was supplied with an air stone placed down in the centre to maintain the eggs in suspension and prevent accumulation of the eggs at the surface or bottom of the incubator. The number of eggs in each incubator was estimated by mixing the incubator homogenously and taking three 100 mL samples and counting the eggs in each sample. The eggs were left one day to develop and survival rate (percentage of eggs with embryos) was estimated as for percentage fertilization with a sample of eggs (n = 50–100) taken from the incubator. The following day, the number of hatched larvae in each incubator was estimated volumetrically as for the eggs. In parallel to the 30 L incubators, fertilised eggs were transferred into individual wells filled with sterile seawater in a 96-well cell culture plate (EIA plate) and placed in a refrigerated incubator at 21 °C in duplicate for each spawn and revised daily until the last larva died, to estimate hatching rate and larval survival under starvation as in Giménez et al.19. Percentage of survival was calculated as the number of larvae alive/total hatched larvae.
A preliminary trial was made to examine the larval growth and development. The trial did not focus on survival as the facilities and staff were not available to provide optimal conditions for the larvae. Larval rearing was carried out using mesocosm conditions20, with low larval density in a large tank (6 larvae L−1, 1500 L tank) under more natural or, at least, less strict conditions than those used in intensive rearing, and using an endogenous bloom of wild marine zooplankton, mostly harpacticoid copepods, together with periodic addition of rotifers and Artemia nauplii. The trial was carried out from November 11th to December 18th 2020 using larvae 4 days post hatch (dph) that hatched on November 7th, from a spawn obtained with the rLh + rLh spawning treatment. The larvae were stocked in a 1500 L tank at 20 °C, 12hL:12hD photoperiod and fed on rotifers for 26 days (4 to 30 dph) followed by newly hatched A. nauplii (24–39 dph). Phytoplankton (a mixture of Tetraselmis suecica and Isochrysis galbana) was added every day in order to maintain a green medium, and every two days the rotifer concentration was assessed to maintain a density of 5 rot mL−1. Artemia nauplii were added when the larvae reached 4.3 mm TL (20 dph) being the main food for larvae after 5 mm TL as suggested by Hagiwara et al.21.
Ten larvae arbitrarily chosen were sampled at 4, 6, 9, 11, 17, 23, 27, 32, 37 and 39 dph and anaesthetised with MS-222. Standard length was measured using a digital camera connected to an image analyser (AnalySIS, SIS Gmbh, Germany). Photographs were also used to estimate the presence or absence of food in the gut and to examine swim bladder inflation as well as other indicators of larval development.
Histological analysis
Ovarian samples obtained by cannulation and testis portions were preserved in Bouin’s fluid for 24 h and stored in 70% ethanol until processed. The dehydrated tissues were embedded in paraffin and 3 μm sections cut. The testes portions (from the anterior, middle and posterior part) were oriented to obtain horizontal sections. Cut sections were stained with hematoxylin and eosin (Casa Álvarez, Spain) for morphological evaluation. The slides were examined under a light microscope (Leica DMLB, Houston, USA).
Oocytes were classified as previously described by Ramos-Júdez et al.9. Oocytes were classified as previtellogenic: with primary growth (PG) oocytes, which presented multiple nucleoli situated in the germinal vesicle, or with cortical alveoli (SGca) oocytes that had small oil droplets and granular vesicles in the peripheral ooplasm. The incorporation of yolk globules indicated vitellogenic stages: early secondary growth (SGe), late-secondary growth (SGl) when oocytes reached ≥ 400 μm17 and full-grown secondary growth oocytes (SGfg) when the vitellogenic growth was completed with the fusion of yolk granules and thickening of vitelline membrane. Oocytes were classified as oocyte maturation stage (OM) when oil droplets were coalescing and the nucleus was positioned to one side of the oocyte, indicating the initiation of the germinal vesicle migration (GVM). Oocytes with disintegrating structure and hypertrophy were in atresia22. Maturation stage of females was determined according to the most developed stage of oocytes present. Additionally, the percentage of oocytes in different stages in the ovaries among weeks was calculated through the identification of ≥ 50 random oocytes per female each week.
To evaluate testes samples, the number of spermatogonia type A and B (SPG), spermatocytes (SPC), spermatids (SPD) and spermatozoa (SPZ) were scored in 12 seminiferous tubules randomly selected from different areas (anterior, middle and posterior) per sample and the percentage abundance of each germ cell type was determined.
Sperm quality evaluation
Sperm samples from running males were collected, for quality evaluation, on week 10 and week 13 at the end of the experimental period. Samples were collected in a 1 mL syringe avoiding the contamination by urine, faeces and water. Sperm was divided into two aliquots, one was maintained as undiluted sperm and one was diluted 1:10 (1 part sperm + 9 parts diluent) in Marine Freeze® (IMV Technologies, L’Aigle, France) extender23 and both samples were maintained in Eppendorf tubes at 4 °C until evaluation. Sperm was activated by pipetting 5 μL of the sperm sample (undiluted or diluted) into an Eppendorf with, depending on the concentration of the sperm, 195, 295, 495 or 995 μL of sea water. Immediately, the Eppendorf was agitated to thoroughly mix, then 2 μL containing activated sperm was pipetted into an ISAS counting chamber (Integrated Sperm Analysis System, Spain), and videos of tracks of the activated spermatozoa were recorded 15 s after activation with the CASA system SCA-VET-01 (Microptic, Barcelona, Spain). Videos were recorded using a digital camera (Basler Ace ACA1300-200UC, Basler AG, Ahrensburg, Germany) connected to an optical phase-contrast microscope (Nikon Eclipse Ci, Tokyo, Japan) with × 10 negative phase contrast objective. The following sperm parameters were determined: (1) sperm concentration (spz mL−1), (2) sperm motility (%), (3) rapid progressive sperm (%), and (4) sperm velocity (μm s−1): the curvilinear velocity (VCL), straight-line velocity (VSL) and average path velocity (VAP). The CASA program was set to classify motile sperm to have a VCL of > 25 μm s−1 and fast progressive sperm to have a straightness (SRT = VSL/VAP × 100) of > 80% and a VCL of > 80 μm s−1. All samples were analysed on the day of the collection and 48 h after collection. Samples collected at the end of the experiment (week 13) were also analysed on days 1, 4, 6, 8, 11, 13 and 15 after collection.
Steroid hormone determination
Blood samples were collected and centrifuged at 3000 rpm at 4 °C for 15 min and the plasma stored at − 80 °C until analysis. Plasma levels of 17β-estradiol (E2) and 11-ketotestosterone (11-KT) were analysed using commercially available enzyme immunoassays (EIA) (Cayman Chemical Company, USA). Steroids were extracted with methanol that was evaporated and extracts were re-suspended 1:10 (E2) or 1:100 (11-KT) in the EIA buffer.
Statistical analysis
Data is expressed as the mean ± standard deviation (SD). A Chi-square test was used to examine the distribution between groups of fish that had different maturational stages at the start of the experiment, and whether fish that ovulated or spawned had a determinate maturity status at the beginning of the experiment. Shapiro–Wilk and Levene tests were used to check the normality of data distribution and homogeneity of variance, respectively. Mann–Whitney U test or Kruskal–Wallis test, followed by Dunn’s pairwise comparison, were used in non-normally distributed data to compare oocyte diameter between treated and control group within a week and between weeks during the experiment, respectively. One-way ANOVA followed by Holm–Sidak post hoc test was used to separately examine differences in the independent variables of diameter of full-grown oocytes, percentage of OM, total eggs per females, and fecundity between the three spawning treatments. Variances across the groups were not equal for the GSI data, which was log-transformed and groups compared using the Brown–Forsythe test and Games–Howell post-hoc multiple comparisons test. Spawning data from rLh + rLh and rLh + P4 spawning treatments (i.e., latency period, fertilization and hatching percentages) was compared using a student t-test. Differences in percentage of OM and oocyte diameter before and after the priming or resolving injections, and percentage ovulation of oocytes incubated in vitro were examined by one-way repeated measures (RM) ANOVA with individual females as the subject. Two-way RM ANOVA with pairwise comparison by the Holm–Sidak test was used for comparing E2 and 11-KT between weeks and treatment groups. Sperm quality parameters were compared with a one-way and two-way RM ANOVA. In the one-way RM ANOVA, the male was the subject, day of storage the independent variable and sperm quality parameters the dependent variables. In the two-way RM ANOVA, the male was the subject, time of storage (0 or 48 h) and sample dilution (undiluted or Marine Freeze®) were the independent variables and sperm quality parameters the dependent variables. Significant differences were detected at a significance level of P < 0.05. Statistical analyses were performed with SigmaPlot version 12.0 (Systat Software Inc., Richmond, CA, USA), with the exception of the Brown-Forsythe test that was conducted with SPSS software version 20.0 (Armonk, NY: IBM Corp).
