Sexually mature zebrafish (Danio rerio) (6–12 months old) were maintained in 60 L glass aquaria, under 28°C ± 1°C and 14 h light: 10 h dark with the light signal at 06:00 (Maitra et al., 2014). Fish were fed ad libitum thrice daily with live blood worms and acclimatized for at least 10 days to laboratory conditions prior to their use in experiments following standardized laboratory protocols (Westerfield, 2000). All animal experiments were executed following the guidelines of the Institutional Animal Ethics Committee of Visva-Bharati University (1819/GO/Re/S/15/CPCSEA dated 31.08.2018) and approved (IAEC/III-08/2020) by the committee according to Indian law.

Recombinant human IGF1, human chorionic gonadotropin (hCG; LH analog), PI3-Kinase inhibitor wortmannin, proteinase-K, alkaline phosphatase-conjugated anti-rabbit IgG, and anti-rabbit IgG-FITC were obtained from Sigma-Aldrich, India. Dihydroethidium (DHE) and JC-1 reagent were procured from Cayman, United States. Rabbit polyclonal anti-PGC1 beta: A17258; anti-NRF1: A5547; anti-TFAM: A13552; anti-SDHA: A2594; anti-COX IV: A6564; anti-UQCRC2: A4181; anti-ATP5A1: A5884; anti-STAR: A16432; anti-LHCGR: A6266 were from ABclonal (Woburn, MA, United States). Rabbit polyclonal anti-p-PI3 Kinase p85 (Tyr458): #4228 and rabbit monoclonal anti-p-IGF-1 Receptor β (Tyr1135/1136): #3024; anti-p-Akt (Ser473): #4060; anti-p-Akt (Thr308): #9275; anti-p-AMPKα (Thr172): #2535; anti-p-GSK-3β (Ser9): #9336; anti-GSK-3β: #39315 were from Cell Signalling Technology (Danvers, MA, United States). Rabbit polyclonal anti-p-IRS1 (Tyr612): ZRB09432; anti-actin: A2066 were from Sigma-Aldrich, India. Anti-SIRT1 antibody: ARP32386 was from Aviva Systems Biology, India. TRIzol reagent was from Invitrogen, Life Technologies. Details of the antibodies used in this study and their evaluation of putative binding to the host target protein have been enlisted in Supplementary Table S1. Anti-IGFIRβ was raised in rabbit against a synthetic peptide corresponding to the C-terminus of IGF1Rβ of human origin and is identical to the corresponding region of zebrafish IGF1 receptor (Maures et al., 2002). Unless otherwise specified, hormones, inhibitors and other reagents were from Sigma-Aldrich, India.

Sexually mature males and females were maintained in separate aquaria at 28°C ± 1°C and 14 L:10 D cycle for at least 7 days prior to their use in the experiment (Westerfield, 2000). On the day before sampling, fish were released together in the breeding tanks at dusk (18:00) at the ratio of 1 female: 2 males (5 females with 10 males/experimental time-point) in 40 L aquaria. The maintenance and breeding of zebrafish were as described elsewhere (Nagel, 2002; Braunbeck et al., 2005; Das et al., 2016). Briefly, the bottom of the breeding tanks was replaced by a stainless-steel grid of mesh size 2 mm and placed on top of egg trays of similar dimensions for the collection of eggs after spawning. Plant imitations of green plastic wire materials were used as spawning substrates. As reported earlier (Braunbeck et al., 2005) as well as in our setup, females start releasing eggs within 30 min of morning lights on at 06:00. To cover the pre-ovulatory time frames, the females (n = 5/clock hour in triplicate) were autopsied at 04:00 (−2 h before light signal on), 05:00 (−1 h before light signal on), 06:00 (0 h with the light signal on), and 06:30 (+0.5 h after light signal on and just prior to ovulation). These time points provided us with ovarian follicles of immature fully grown stage with intact GV (04:00), onset of oocyte maturation (05:00), matured with complete GVBD (06:00), and peri-ovulatory follicles in the abdominal cavity (06:30) (Das et al., 2016; Liu et al., 2017). For the determination of mitochondrial parameters, ovaries were extirpated at indicated clock hours and divided into two lobes. One ovarian lobe (mid-portion) from each sampled fish was processed immediately for the quantification of ATP content and RNA isolation, whereas the other lobe was utilized for obtaining full-grown (FG) follicles for the determination of mitochondrial membrane potential.

Gravid females were anaesthetized by brief cold shock, ovaries extirpated aseptically and placed immediately in ice-cold, oxygenated zebrafish Ringer (116 mM NaCl, 2.9 mM KCl, 1.8 mM CaCl₂, 5 mM HEPES, pH 7.2), supplemented with streptomycin (100 μg/mL) and penicillin (100 IU/mL) as described elsewhere (Rezanujjaman et al., 2020). This preparation also served as the culture medium in subsequent in vitro experiments. Individual ovarian follicles were separated manually using a pair of fine-tip forceps or gentle pipetting under a dissecting microscope as described earlier (Das et al., 2013). The protocol for selecting fully grown intact follicles was according to our standard laboratory protocols (Maitra et al., 2014). In brief, freshly isolated healthy FG (mean diameter ∼650 µm) follicles from independent donor females were pooled and cultured in 1 mL of Ringer’s solution in a 24-well tissue culture plate (∼50–60 follicles/well) for indicated time intervals at 25°C ± 1°C under gentle agitation (40–50 rpm). Recombinant human IGF1 and hCG were dissolved in sterile water. PI3K inhibitor wortmannin (Wort, 10 µM) was dissolved in DMSO and added to the incubation medium 1 h prior to IGF1 addition. The effectiveness of the dose selected for Wort (10 µM) was confirmed through attenuated Akt phosphorylation. All chemicals were prepared at 1000-fold stock, and oocytes in control wells received an equivalent volume of solvent only. Triton X-100 mediated delivery of anti-IGF1Rβ or anti-NRF-1 into the oocytes was performed following the method described previously (Maitra et al., 2014; Das et al., 2016). Briefly, Triton X-100 diluted (0.1%, v/v) with 10 mM PBS; pH 7.2 was sonicated for 1 min and complexed with anti-IGF1Rβ or anti-NRF-1 antibody (20 μg/mL) for 30 min each at 25°C ± 1°C and at 4°C. The resultant complex (10 µL) was layered on FG follicles in culture. The medium was replaced with a fresh Ringer after 1 h followed by hCG (10 IU) or IGF1 (10 nM) stimulation for indicated time. NaN₃ (6.25 mM) and LiCl (15 mM) were dissolved in sterile water. Doses of hormones and inhibitors used in this study were as described earlier (Stackley et al., 2011; Das et al., 2016; Martin et al., 2018; Fallah and Habibi, 2020; Biswas and Maitra, 2021). Viability was assessed by the 0.1% trypan blue dye exclusion method. All subsequent in vitro experiments were carried out in triplets using pooled ovaries from separate donor females (n = 5).

The mitochondrial membrane potential was determined using the fluorescent probe JC-1 (5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethyl-imidacarbocyanine iodide) staining following the manufacturer’s protocol (Cayman Chemical, United States). JC-1 depicts the mitochondrial metabolic status; while the red fluorescence due to the formation of J-aggregates indicates high membrane potential and energized mitochondria, the green fluorescence of JC-1 monomers is indicative of low membrane potential (Sivandzade et al., 2019). Briefly, follicle-enclosed oocytes harvested from independent donor females (n ≥ 5) were cultured in triplicate for indicated treatments and time points, followed by incubation with JC-1 dye for 30 min in the dark. Subsequently, the stained follicles were washed thrice with freshly prepared cell-based assay buffer (Cayman Chemical, United States), assembled onto groove slides, and visualized under confocal microscopy (Leica, Switzerland) through a series of optical sections. While J-aggregates were detected with filter sets designed to detect rhodamine (λ_ex = 516), detection of JC-1 monomers was achieved with filter sets designed to detect FITC (λ_ex = 488). Objectives, pinhole, filters, gain and offset, were kept constant throughout all in vitro experiments. The fluorescent intensity was analyzed for each oocyte by ImageJ software taking around 30 oocytes in each group from three independent experiments.

FG follicles (∼60 follicles/well) from indicated treatment groups (at least three replicates per treatment group) were fixed in chilled acetone overnight at 4°C, washed (x 3) with cold PBS (10 mM, pH 7.4), permeabilized with 0.5% Triton-X-100 containing PBS, and blocked in 2% BSA in PBST (PBS + 0.1% Tween 20) for overnight at 4°C. Subsequently, the follicles were washed (x 3) and incubated with NRF-1 antibody (1:50; in blocking solution) in a micro-centrifuge tube overnight at 4°C. Follicles were washed (x 3) and incubated with anti-rabbit IgG-FITC (1:250; in blocking solution) for 2 h at RT, washed and counter-stained with 1 μg/mL of 4′, 6-diamidino-2-phenylindole (DAPI; Sigma) for 1 min followed by visualization under confocal microscopy. Follicles incubated without primary antibody served as the negative control. A series of optical sections were captured for around 30 follicle-enclosed oocytes to examine NRF-1 antigen localization in individual follicular cells.

The determination of follicular ROS was carried out using DHE (10 μM) reactive oxygen probe followed by fluorescence microscopy as described earlier (Biswas and Maitra, 2021). Briefly, FG follicles cultured in triplicate for indicated treatments and time points were incubated with oxidation-sensitive fluorescent probe (DHE) for 30 min in the dark and observed under a fluorescence microscope (Victory FL, Dewinter Optical Inc.) using an excitation filter G (Green): 510–550 nm. The fluorescent intensity was analyzed for each oocyte by ImageJ software taking around ten fields from each replicate. The experiment was repeated thrice using FG follicles harvested from different donor females.

Total RNA was extracted from FG follicles (∼60 follicles/treatment well in triplicate) using TRIzol solution following manufacturer’s protocol. Reconstitution of RNA pellet was followed by quantification, and integrity check was performed using Multiskan Go microplate spectrophotometer (Thermo Scientific, India). 1 μg of total RNA from each treatment group pre-treated with 1 U of DNase I was reverse transcribed at 42°C for 1 h using Verso cDNA synthesis kit (#AB1453, Thermo Scientific), and qPCRs were set up on a Quant-Studio-5 Real-Time PCR System (Applied Biosystems) (Biswas et al., 2020) using gene-specific primers from earlier reports (Supplementary Table S2). The reaction mix for both target and reference genes contained cDNA template (diluted 1:10), primers (250 nM each for forward and reverse) and SYBR Green PCR Master Mix (Applied Biosystems). An initial enzyme activation step at 95°C for 10 min was followed by 45 cycles of thermal cycling with denaturation at 95°C for 30 s, annealing at 60°C for 30 s, and elongation at 72°C for 45 s. PCR conditions were optimized to produce primer amplification efficiency of 90%–110%. Briefly, each primer pair was tested using a serial dilution series of cDNA of the control group (undiluted, 1:10, 1:100, 1:1000, 1:10000). A standard curve was generated and the primer efficiency was calculated using the equation: Efficiency % E = 10 ^ − 1 / s l o p e * 100 . Inclusion of no template control (NTC) and a melt-curve analysis followed by agarose gel electrophoresis was executed to confirm the absence of contamination and specificity of amplification, respectively. To quantify the relative expression of each gene, the comparative threshold cycle (Ct) values for all samples were normalized for the reference gene ef1a as it showed stable expression over time as well as between different treatment groups (Tang et al., 2007; McCurley and Callard, 2008; Nelson and Van Der Kraak, 2010a). Relative fold-change of mRNA levels over the calibrator (quantity = 1) was calculated using the 2 − ∆ ∆ C t method. The error bars display the calculated maximum (RQmax) and minimum (RQmin) expression levels representing SEM of RQ value (n = 3 technical replicates; n = 3 biological replicates) (Livak and Schmittgen, 2001).

DNA was extracted from cultured follicles from indicated treatment groups at 2 h of incubation in vitro. The follicle-enclosed oocytes were transferred to a sterile micro-centrifuge tube, homogenized with DNA extraction buffer (pH 8) containing proteinase K (10 mg/mL), and incubated at 55°C for 2 h. After checking the integrity of DNA samples by UV absorbance (A260/280 between 1.8 and 2), mtDNA copy number was determined by qPCR using ndi, coxI, coxII as the mitochondrial gene target versus a nuclear-encoded gene ef1α (Supplementary Table S3). After an initial denaturation step for 10 min at 95°C, PCR amplification was executed using 100 ng DNA in a Quant Studio 5 Real-Time PCR platform and a cycling profile of 30 s at 94°C, 1 min at 60°C, 30 s at 72°C for 40 cycles followed by a melt curve analysis to check the specificity of amplification. The mtDNA content was calculated by measuring the threshold cycle ratio (ΔCt) of a mitochondrial-encoded gene (ndi, coxI, coxII) versus a nuclear encoded gene ef1α.

Intact follicles were harvested at appropriate time intervals, washed (x 3), and lysed in ice-cold oocyte extraction buffer as described earlier (Biswas and Maitra, 2021), supernatant separated by spinning at 17,500×g for 20 min at 4°C, and stored at 80°C until further use. Following the quantitation of protein concentration, oocyte lysates (50 µg/lane) from each treatment group were resolved through SDS-PAGE and immunoblotted following our standard laboratory protocols (Das et al., 2013; Maitra et al., 2014). The electroblotted Hybond-P PVDF membrane was blocked in 5% BSA in TBST (pH 7.6, Tris 50 mM, NaCl 150 mM, Tween-20 0.1%), incubated with primary antibody (1:1000) overnight at 4°C followed by secondary (1:2000) antibody incubation for 2 h at RT (for details see Supplementary Table S1). BCIP-NBT was used as the substrate to develop the bands, visualized, and recorded in Gel Doc apparatus (Bio-Rad). The signal intensity of target bands in each lane was measured through ImageJ software and neutralized by endogenous factor (anti-actin or respective total protein) for that lane. The normalized target signal for each sample is then divided by the normalized target signal observed in the experimental control sample to generate the fold change in target protein (n = 3 technical replicates; n = 3 biological replicates) (Pillai-Kastoori et al., 2020).

Estimation of ATP levels from indicated treatment groups (∼80 follicles/well in triplicate) was performed using ATP Colorimetric Assay Kit (Elabscience, Texas, United States) following the manufacturer’s protocol. The assay relies on the determination of phosphocreatine formed by the catalysis of ATP and creatine phosphate by creatine kinase. Absorbance was recorded at 636 nm using a microplate reader, and results (ATP content) are expressed as nmol/mg tissue.

After harvesting and separating ovarian follicles as described in previous sections, FG follicles with mean diameter ∼650 µm and without any morphological deformity were selected under a stereo zoom microscope (Dewinter, Italy) and cultured in 24 well plates (50 oocytes/well) in the presence of hCG (10 IU) or IGF1 (10 nM) either alone or in combination with αIGF1R and NaN₃ at 25°C ± 1°C under gentle agitation to study oocyte maturation in vitro. The germinal vesicle breakdown (GVBD) was monitored at 6 h through immersion in clearing solution (3% ethanol, 6% formaldehyde and 1% acetic acid) and visualizing ooplasmic clearance under an inverted microscope (Victory FL, Dewinter Optical, Inc., Italy) fitted with phase-contrast optics. These conditions were optimal in our model system and as reported earlier (Nelson and Van Der Kraak, 2010b; Zhou et al., 2016; Biswas and Maitra, 2021).

All the experimental data presented are mean ± S.E.M. of at least three independent experiments. Data were analyzed by Student’s t-test (for single-factorial designs) and one-way analysis of variance (ANOVA) followed by post hoc multiple comparison tests (Tukey HSD and Bonferroni) (for multi-factorial designs) using GraphPad Prism8 software. p < 0.05 was considered statistically significant.

The dynamics of mitochondrial bioenergetics machinery during the natural course of meiotic resumption and ovulation triggered by early morning light signals was examined in vivo. Ovaries were harvested from gravid females (n = 5) at four consecutive time intervals just prior to the release of eggs (spawning): 04:00 (−2 h before light signal on), 05:00 (−1 h before light signal on), 06:00 (0 h with the light signal on), and 06:30 (+0.5 h after light signal on) (Figure 1A). These time points provided us with ovarian follicles mostly at stages of G2-arrest (04:00), initiation of maturation (05:00), matured with complete GVBD (06:00), and peri-ovulatory follicles (06:30), without using any exogenous hormonal stimulation.

Interestingly, the remarkable increment in ATP levels (Figure 1B) and formation of J-aggregates (Figure 1C) at 06:30 (i.e., just prior to the release of eggs) prompted us to examine the status of the endocrine and growth factor axis during these sampling hours. As shown in Figure 1D, lhcgr expression showed a sharp increment from 05:00 (−1 h) onwards when the oocytes underwent GV migration and GVBD. While igf1 mRNA expression gradually increased from 05:00 (−1 h) onwards (∼3.8 folds), reaching the peak at 06:30 (+0.5 h) (∼9.4 folds), the transcript abundance of igf3 displayed a marginal but significant increase (∼1.2 folds) over 04:00 (−2 h) followed by its sharp decline during the ovulatory phase (06:30) (Figure 1E). These data prompted us to hypothesize the possible involvement of gonadotropin signaling in modulating oocyte energy metabolism.

LH regulation of igf1 expression and mitochondrial bioenergetics was checked in FG follicles in vitro. Congruent with elevated ATP levels, hCG (10 IU) stimulation (2 h) could trigger a significant increase in NRF-1 (nuclear transcription factor) and TFAM (mitochondrial transcription factor) expression in vitro (Figure 2A). Interestingly, a sharp increase in igf1 transcript abundance corroborated well with IGF1R phosphorylation in hCG-stimulated follicles at the indicated dose and duration (Figures 2B, C). Though TFAM expression remained largely unaltered, pre-incubation with anti-IGF1Rβ resulted in a sharp decline in NRF-1 mRNA abundance (Figure 2D). While hCG administration promotes, priming with IGF1Rβ or NRF-1 antibodies could attenuate ATP levels and meiotic maturation even after hCG induction in vitro (Figures 2D, E), indicating the importance of gonadotropin regulation of oocyte bioenergetics, potentially through IGF axis, on the maturational process.

Priming with graded levels of recombinant IGF1 (10, 100 nM) could significantly increase the expression of PGC-1β and NRF-1 at both transcript and protein levels. However, TFAM expression did not undergo increment in IGF1-stimulated follicles at the indicated dose levels and time point (2 h) in vitro (Figures 3A, B). Nevertheless, a robust increase in PGC-1β and NRF-1 transcripts at the lower dose (10 nM) prompted us to select this as the lowest effective dose for subsequent experiments.

As shown in Figure 3C, IGF1 administration induced a significant increase in ATP content in FG follicles at both the time points tested over the unstimulated control. In a line of agreement, a sharp increase in red fluorescence concomitant with reduced green fluorescence was observed in IGF1-stimulated cells over the control at 2 h of incubation in vitro (Figures 3D, E). While the red fluorescence is characteristic of higher polarized mitochondria due to the formation of J-aggregates, less active mitochondria fluoresce green owing to the retention of JC-1 in its monomeric form. Notably, the corresponding differential interference contrast (DIC) images revealed that compared to untreated control follicles, which appeared nearly opaque (prophase I arrest), IGF1 stimulation prompted the resumption of meiotic maturation (evident from the appearance of germinal vesicle as white spot and startling clarity in the ooplasm) (Figure 3E), indicating a correlation between mitochondrial polarization (higher J aggregates), ATP synthesis and onset of meiotic maturation in IGF1-stimulated follicles in vitro.

GSK3β is a negative regulator of the PGC-1 protein stability in somatic tissues, whether this axis is conserved in zebrafish oocytes was investigated next. As shown in Figure 4A, the selective inhibition of GSK3β by LiCl in zebrafish FG follicles could significantly upregulate PGC-1β immunoreactive protein. Correspondingly, the kinetics of GSK3β phosphorylation (Ser 9) (inhibition) revealed a sharp increase between 15 and 45 min of IGF1 stimulation, whilst the PGC-1β protein level was at its peak between 60 and 120 min of IGF1 induction (Figure 4B) indicating a close correlation between GSK3β phosphorylation (inhibition) and PGC-1β stabilization. Since GSK3β is a direct target of Akt, which phosphorylates it on Ser 9 (inactivation), we further examined the role of Akt signaling on PGC-1β upregulation in a manner sensitive to GSK3β inhibition. Immunoblot analyses revealed a rapid increase in p-IGF1Rβ (Tyr 1135/1136), p-IRS1 (Tyr 612), p-PI3K (Tyr 458), and p-Akt (Ser473, Thr 308) as early as 15 min of IGF1 (10 nM) stimulation (Figure 4C). Intriguingly, compared to IGF1-treated positive controls, pre-incubation with either anti-IGF1Rβ or wortmannin (a PI3K inhibitor) could abrogate GSK3β phosphorylation (Ser 9) as well as PGC-1β upregulation (Figure 4D) indicating that Akt signaling is essential for PGC-1β stabilization through GSK3β inactivation in zebrafish follicle-enclosed oocytes (Figure 4E).

The effect of blocking IGF1 action on the expression of NRF-1, the binding partner of PGC-1β, was investigated next. Confocal fluorescence microscopy revealed a sharp increase in NRF-1 immunolocalization in IGF1-stimulated FG follicles, more specifically at the follicular layer, compared to the unstimulated group (Figures 5A, B). Serial optical sections of IGF1-stimulated follicle-enclosed oocytes showed a well-organized peripheral aggregation of NRF-1 immunofluorescence in individual follicular cells (right panel, Figure 5A). Priming with either anti-IGF1Rβ or pharmacological inhibition of PI3K/Akt signaling cascade could diminish NRF-1 expression in IGF1-treated follicles (Figures 5A, B). Corresponding immunoblot analysis revealed that compared to IGF1-treatment alone, FG follicles pre-incubated with anti-IGF1Rβ or wortmannin showed a significant downregulation in NRF-1 expression at both mRNA and protein levels (Figure 5C).

Whether IGF1 could modulate electron transport chain subunits was examined next. While IGF1 administration (10 nM) could elevate the transcript levels of nuclear-encoded (sdha, atp5α1) (Figure 5D) and mitochondrial-encoded (ndi, coxI, coxII) (Figure 5H) respiratory chain subunits, immunoblot data revealed a marked increase in SDHA, COXIV, and ATP5α1, but not UQCRC2, in the treated group over the untreated control (Figure 5E). Notably, a sharp decline in nuclear- and mitochondrial-encoded respiratory chain subunits due to the blockade of the IGF1 axis (Figures 5F–H) might indicate reduced OXPHOS activity and corroborated well with reduced NRF-1 immunolocalization under similar conditions (Figures 5A–C).

Next, the specificity of IGF1 action and the importance of NRF-1 on mitochondrial metabolic status (membrane polarization) were examined. As shown in Figures 6A, B, priming with either anti-IGF1Rβ or anti-NRF-1 prior to IGF1-stimulation resulted in the loss of red fluorescence of J-aggregates and elevated green fluorescence (JC-1 monomers), indicating low polarized organelles. Congruently, similar to the effect of NaN₃ (an inhibitor of the respiratory chain), pre-incubation with IGF1R or NRF-1 antibodies could lower ATP levels compared to IGF1 alone (Figure 6C), indicating the importance of IGF1-mediated signaling cascade in oocyte energy metabolism.

The relevance of IGF1 action or the attenuation of it on follicular redox homeostasis, steroidogenic potential and the meiotic maturational response was investigated next. As shown in Figure 7A, the pharmacological inhibition of PI3K/Akt signaling heightened follicular levels of ROS synthesis as evident through DHE staining followed by fluorescence microscopy. Congruently, the transcript abundance of SOD2 (a mitochondria-specific antioxidant enzyme) and HSP70 (a ubiquitous molecular chaperone) underwent a sharp decline in wortmannin-treated follicles (Figure 7B). Besides, while IGF1 alone could elicit Lhcgr and StAR expression at mRNA and protein levels (Figure 7C), priming with anti-IGF1Rβ or Wort prevented IGF1-induced StAR, P450scc, 3β-hsd, and 20β-hsd expression (Figure 7D). Intriguingly, similar to the attenuated meiotic maturational response upon NaN₃ treatment, IGF1R immunodepleted follicles failed to undergo meiotic resumption upon IGF1 stimulation in vitro (Figure 7E). Thus, IGF1 regulation of NRF-1/PGC1β expression and mitochondrial bioenergetics strongly indicate the physiological relevance of functional and active mitochondria in maintaining redox homeostasis, steroid biosynthesis, and oocyte maturation in zebrafish FG follicles.