Sperm quality of artificially matured shortfinned eel is not affected by human chorionic gonadotropin dose and route of administration

Animals

Forty male silver eels were captured from Lake Ellesmere, New Zealand, in autumn by commercial fishermen using fyke nets. The eels were packed into chilled polystyrene boxes and transported to Dunedin, where they were transferred directly into a 1,000 L recirculating tank of 35 ppt salt water, in keeping with the euryhaline abilities of these fish (Damsteegt, Wylie & Setiawan, 2018). Water temperature was gradually increased from 10 to 20 °C during a one-week period. The fish were held under a 12:12 (L:D) photoperiod with the tank covered in shade cloth to decrease light penetration and were monitored for activity and wellbeing at least once daily throughout the acclimation and experimental periods. PVC pipes were provided for shelter and refuge. Average body weight of the males at Day 0 was 138.6 g (± SE 2.6 g), with a range of 120–184 g. Fish were not fed during the experiment. The fish and manipulations used in this study were approved by the University of Otago Animal Ethics Committee (AEC 36-11) in accordance with national welfare regulations.

Experimental design and induction of spermatogenesis

The 40 fish, each separate experimental units, were assigned randomly by the same researcher (SLD) to one of four hCG dose groups. After an acclimation period of two weeks, the fish were anaesthetised in 0.15 g/L benzocaine (Sigma-Aldrich, St Louis, MO) before receiving their initial hCG injection (Day 0) of 0 (control), 250 (low dose), 500 (medium dose), or 1,000 IU/fish (high dose). Within each dose group, the fish were randomly allocated to an injection administration route of either IM (dorsal muscle) or IP (ventral midline, halfway between pectoral fins and anus), so that each dose × administration route contained five fish (c.f., Lokman et al., 2016) of comparable body weights (group weights 142–168 g; P > 0.13).

To allow for tracking of individuals, a PIT-tag was implanted into the muscle just anterior of the dorsal fin at the time of the first injection. At that time, control fish received an initial injection of 250 µL of ice-cold saline, whereas low-dose fish were injected with 250 µL of ice-cold purified hCG (Pregnyl, Batch # 128300; Merck, Sharp and Dohme Corp., Auckland, New Zealand) at 1 IU/µL, medium-dose fish at 2 IU/µL, and high-dose fish at 4 IU/µL. The initial injections were administered according to the randomly designated route, i.e, either IM or IP. Following hCG administration, all fish were placed into a second 1,000 L tank, that was part of the same recirculation system; accordingly, on treatment days, animals were randomly netted from the tank, injected, and then moved to the ‘other’ tank on the same system, thus switching back and forth between both 1,000 L tanks on injection days.

Induction of sperm maturation

Every seven days from Day 31 onwards (six treatment days in total, designated Weeks 5–10), all eels received a booster hCG injection of 150 IU/fish in 250 µL of ice-cold saline, administered according to the specific route allocated to the individual. Injections were made after inducing anaesthesia in benzocaine. The purpose of this booster was to maximise milt production and quality as described previously. Weekly booster injections continued until Week 10, when, as expected, controls started spermiating, signalling the end of the experiment and rendering comparisons of sperm parameters across initial hCG doses no longer appropriate—indeed, spermiation in controls was assumed to not be due to the key focus of the experiment (injection of hCG at Time 0), but rather, on subsequent spermiation-induction from Week 5 onwards. Once spermiation started in the control group, eels were therefore placed in overdose benzocaine (0.3 g/L) and euthanised by spinal transection.

Milt collection

Milt was collected from males one day after administration of the hCG booster while under anaesthesia in 0.15% benzocaine. The posterior portion of the abdomen and the vent were wiped gently with a paper towel before rinsing the same area with distilled water. The same area was again dried lightly with paper towel before gentle abdominal pressure was applied to express milt. A modified 15 mL plastic tube and three mL Pasteur pipette coupled to a venturi was used to collect expressed milt via suction. The collected milt was then decanted into pre-weighed microcentrifuge tubes for assessment of total milt weight (c.f., Koumpiadis et al., 2021), as weight of a small sample is much more easily and accurately determined than the corresponding volume.

Milt volume was extrapolated from milt weight (assumed 1:1 volume/weight) and expressed as a proportion of body weight to account for variation in fish size where: ${\displaystyle \text{Milt volume \% BW}=\left(\text{milt weight (g)/body weight (g)}\right)\ast 100.}$

Milt was stored on ice until assessment for sperm quality by CASA which was done as soon as practicable on the same day, within 2 h of collection. Limited by equipment availability, not all time points were analysed for motility parameters (Weeks 6 and 9 were not done).

CASA

Previous research (Gallego et al., 2013) has established that sperm quality does not change between ejaculate portions within a collection (e.g., the first portion of an ejaculate is of no better quality than the last portion of milt expressed by hand-stripping). Therefore, CASA analysis for this experiment was conducted on a sample from the whole milt collected.

Sperm activity parameters were quantified using milt diluted in fresh sea water (sea water not previously used for fish husbandry). Pilot trials established an optimal dilution and suspension ratio of 1 µL milt in 199 µL of sea water, similar to that reported by Gallego et al. (2013). The milt was ice cold and the seawater was at culture temperature of 20 °C. The suspension was quickly mixed and 4 µL loaded onto a 20 µm Leja slide (Leja Products BV, Nieuw-Vennep, Netherlands). A stopwatch was started at the time of milt addition to the water (sperm activation), the slide being loaded as soon as possible thereafter, and, comparable to other studies on anguillid eels (Butts et al., 2020; Koumpiadis et al., 2021), the first measurement of sperm activity being taken at 30 s post-activation. Pilot trials (data not shown) established that there was no difference in sperm activity measured at 15, 20 and 30 s post-activation, 30 s being chosen as a practical time point for repeatability. Video footage was recorded from 20 s post-activation onwards, using a camera (XC-ST50, Sony, Tokyo, Japan) mounted upon an external phase-contrast microscope at 10× magnification (CX41 Olympus, Melville, NY, USA). Each sample was analysed in duplicate (dual chamber slides) using a computerised sperm analyser (CEROS, Hamilton Thorne, USA). On average, 203 spermatozoa were tracked for each sample (range: 106–255) for 0.5 s at 60 frames/s. Alongside sperm concentration (CONC), sperm activity measures included percent motility (MOT), percent progressive motility (PROG), straight line velocity (VSL), curvilinear velocity (VCL), average path velocity (VAP), while path character was determined by path straightness (STR) and linearity. The lower threshold to determine static cells was predetermined at 20 µm/s for VAP and 17 µm/s for VSL. Progressive cells were pre-determined to have a minimum threshold VAP of 50 µm/s and 50% STR. Pilot trials established VCL as the best potential indicator of eel sperm activity, due to the spermatozoa swimming a slightly curved trajectory.

Statistical analysis

Only data from responding eels were analysed, i.e., non-responders (no sperm collected) were omitted a priori from analysis. Accordingly, animals in the control groups were not used in any of the analyses from Weeks 5–9, as sperm could not be obtained from these fish. Sperm that was collected from controls in Week 10 was deemed to have been produced not in response to treatment with saline, but in response to the weekly hCG booster injections starting in Week 5. These controls were, therefore, also excluded from statistical analysis, but data from controls were graphed for contrast against responders in hCG-injected groups.

Prior to analysis, all data were tested for normal distribution and homogeneity of variance (Levene’s test), and subjected to log-transformation as required. For all motility data, angular transformations (arcsin $\sqrt{x}$; Claringbold, Biggers & Emmens, 1953) were carried out, as percent data are not normally distributed and two-way non-parametric tests are often problematic (e.g., Mundry & Fisher, 1998). Using IBM SPSS Statistics v.24, data were subsequently analysed using a series of full factorial two-way univariate ANOVAs, with route of administration (two levels: IM, IP) and hCG dose (three levels: 250, 500, 1000 IU hCG/fish) as fixed factors; Tukey’s multiple comparisons were used to identify which contrasts differed when significant main effects were detected. The differences between groups for the dependant variables were analysed within week only, not between weeks. Data from Week five were excluded from statistical analysis, as only few eels responded in each treatment group; data from these fish have been included for graphical representation only. To determine whether there were differences in the number of fish that responded to each treatment, a Fisher’s Exact Test was employed. Differences between means for all ANOVAs were considered to be significant for P < 0.05. In order to correct for Type I errors associated with running multiple ANOVAs (for each of three time points) for each sperm quality trait, the Benjamini–Hochberg’s false discovery rate (FDR) was set at 5%.

Number of responders to treatment with hCG

There was no difference in the number of fish that responded to the various treatments (Fisher’s Exact Test, P range = 0.42–0.96). Aside from control fish, every eel spermiated at least once during the trial. Only two fish did not spermiate every week once they had started; one fish in the IM medium-dose group spermiated in Week 5, but then only once more after that, in Week 9. A second fish, belonging to the IM high-dose group, first spermiated in Week 7, but then skipped a week before continuing to spermiate at the Week 9 and 10 time points. The first control fish to spermiate was encountered in Week 10, one day after the 6th booster injection of 150 IU (Fig. 1A).

Milt volume

No significant main or interaction effects were seen when comparing relative milt volumes (Fig. 1B). Only during Week 10 was a strong trend (P_route = 0.059) detected, relative milt volume (roughly 1.5% of body weight) in IM-injected eels being almost double that of IP-injected eels; injection dose, on the other hand, did not have any effect (P_dose = 0.256). During Weeks 9 and 10, particularly large variation was observed in the amount of milt expressed between individual fish within each time point (Fig. 1B), which more than likely accounted for the lack of a significant treatment effect. The largest volume of milt recovered was 4.45% of body weight from one fish during Week 10 (high dose, IM), and from another fish (4.06%; medium dose, IM) in Week 9, which was in both cases several-fold greater than the average milt volume for IM-treated eels. This latter fish also produced a milt sample of 3.11% during Week 10, more than 2.5 times the average for IM-treated eels.

Sperm concentration

Throughout the experiment, sperm concentration remained relatively stable at 8–12 ×10⁹ ml⁻¹ in all treatment groups, although a near-significant main effect of injection route was evident at Week 10 (F_1,23 = 4.136, P = 0.054; Fig. 1C); thus, independent of dose, fish injected IP (Week 10: 11.48 ± 0.50 ×10⁹ ml⁻¹) produced sperm with a spermatozoa concentration around 15% higher than that observed in fish injected in the muscle (Week 10: 10.19 ± 0.43 ×10⁹ ml⁻¹). HCG dose and route of injection did not interact at any of the time points.

Percentage of motile sperm

Weak treatment effects were seen when looking at percentage of motile sperm in a sample, an initial dose of 250 IU hCG appearing to be the most effective in terms of sperm motility throughout the experiment, particularly when administered IP (Fig. 2A)—this is especially evident from the near-significant effect of dose at Week 7 (F_2,15 = 3.223, P = 0.068), when low-dose IP fish showed a relative motility twice that of fish in the remaining treatment groups (44.1 ± 3.5% as opposed to 17.0–26.7% in other groups). There was no evidence for any significant interaction between hCG dose and route of administration (all P > 0.05).

Progressive motility

The progressive motility data trends closely mirrored those seen for total motility percentage, low-dose IP fish displaying a generally higher level of progressively motile sperm throughout the experiment (Fig. 2B). Only during Week 7, a significant dose effect (F_2,15 = 3.717, P = 0.049) was seen, the fish injected with the smallest amount of hCG showing more progressively motile sperm (21.5 ± 4.8%) compared to eels in the medium-dose (9.6 ± 2.4%) and high-dose (12.2 ± 4.6%) groups. However, this effect was no longer significant when taking into account that multiple tests were conducted. Significant interaction effects, similarly, were not observed.

VCL

Finally, when focusing specifically on sperm swimming speed represented as curvilinear velocity (Fig. 2C), the only significant effect detected (F_2,15 = 4.569, P = 0.028) was seen during Week 7, when fish injected with the lowest dose of hCG showed a higher average VCL (108.9 ± 6.1 µm s⁻¹) than in response to the higher doses (medium dose 91.1 µm s⁻¹, high dose average 89.7 µm s⁻¹). When applying the Benjamini–Hochberg correction for multiple statistical testing, this effect was no longer significant.