Sexually mature male zebrafish (outbred) (4–5 months-old) were maintained in the aquarium facility of the Department of Structural and Functional Biology, Institute of Biosciences, Botucatu, São Paulo State University (UNESP) in 6-L tanks in the recirculating system under constant temperature conditions (28°C), and proper photothermal conditions (14-h light/10-h dark). The following water parameters were monitored in all tanks every other day: pH, salinity, dissolved oxygen, and ammonia concentration. The animals were fed twice a day with commercial food (Sera Vipan Flakes^®). Handling and experimentation were performed according to the Brazilian legislation regulated by National Council for the Control of Animal Experimental (CONCEA) and Ethical Principles in Animal Research (Protocol n. 1031-CEUA) and University of Calgary animal care committee and in agreement with the procedures of the Canadian Council of Animal Care (Protocol #AC19-0161).

In this study, we used methimazole (1-methyl-3H-imidazole-2-thione) (CAS 60-56-0; MW, 114.17 g/mol; purity, ≥99%; Sigma-Aldrich, St. Louis, MO, United States) to chemically generate hypothyroidism in adult zebrafish males. Exposure concentration of 1 mM was prepared following the methodology described in Rodrigues et al. (2021). The working concentration of 1 mM methimazole and 100 μg/L T4 (L-Thyroxine) (CAS 51-48-9; MW, 776.87 g/mol; purity ≥98%; Sigma-Aldrich, St. Louis, MO, United States) were chosen based on previous studies (Sharma and Patiño, 2013; Sharma et al., 2016; Rodrigues et al., 2021). In this study, adult male zebrafish (n = 144) were divided into four replicate tanks per experimental group: control [only filtered water (n = 36)]; group I [filtered water with T4 (100 μg/L) (n = 36)]; group II [filtered water with methimazole (1 mM) (n = 36)]; group III [1 mM of methimazole followed by addition of T4 (100 μg/L) (n = 36) as reversal treatment group (methimazole + T4)]. In the T4 group, males were exposed to T4 (100 μg/L) from the second week until the end of treatment. In the methimazole group, males were exposed to 1 mM methimazole for 21 days. In addition, zebrafish males were exposed to 1 mM methimazole for 21 days, and T4 (100 μg/L) was added in the water from the second week until the end of treatment. The reversal treatment was performed to assess whether the apparent effects were due to lowering thyroid hormone levels. After euthanasia, plasma T3 levels were measured on the treatments and the heads were sampled for histology (control, methimazole and methimazole + T4), while the testes were dissected and immediately used for in vivo experiments (histomorphometric analysis and gene expression); androgen plasma levels and ex vivo organ culture experiment [short-term (18 h) incubation for androgen release by zebrafish testicular explants] were available on the treatments (control, T4, methimazole and methimazole + T4 groups). The brain and pituitary from the control and methimazole groups were sampled for gene expression.

Blood from adult male zebrafish were collected in different conditions (control, methimazole and methimazole + T4) (n = 5 per condition) to confirm the hypothyroidism status. Animals were euthanized, and the caudal peduncle was cut for blood sampling using heparinized syringes and tubes. Plasma fractions were isolated after blood centrifugation at 4°C for 10 min at 800 × g (Eppendorf Centrifuge 5424 R) for thyroid hormone analysis. Plasma Triiodothyronine (T3) levels were measured by Enzyme-Linked Immunosorbent Assay (Competitive ELISA kit) (Invitrogen, TX061-1 EA, Carlsbad, CA, United States). This assay is designed to detect and quantify the levels of T3 in different sample types such as serum, urine, and plasma. Here, we prepare plasma sample according to the manufacturer’s procedure. Briefly, high binding 96-well strip-well plate was coated with donkey anti-sheep IgG. Plasma samples were extracted with ethyl acetate (5:1) (v/v) solvent: sample ratio. Samples were frozen and solvent solution was collected (this step was repeated for maximum extraction); and dry pooled solvent extracts down in a speedvac for 2–3 h. Then, samples were reconstituted at room temperature in the 1X Assay Buffer, and 100 μl of either standards or samples were added to the wells in duplicate. 25 μl of T3 conjugated and T3 antibody were added to each well. Plate was incubated for 2 h, shaking, at room temperature. The plate was washed with 1X Wash Buffer, followed by the addition of the detection reagent (TMB substrate solution). After, 30 min, the reaction was stopped with 1 M HCL, CAUSTIC and read at 450 nm using a microplate reader (Epoch, Agilent, Santa Clara, CA, United States). Data were evaluated as nanograms of Triiodothyronine (T3) per milliliter of plasma.

Thyroid follicles from the different experimental groups were analyzed histologically. As thyroid follicles in fish appear distributed among the afferent branchial arterioles (Patinõ et a., 2003; Van der Ven et al., 2006), head was separated from the trunk and fixed in 2% glutaraldehyde and 4% paraformaldehyde in Sorensen buffer [0.1 M, pH 7.2] for at least 24 h at room temperature. The material was dehydrated, embedded in glycol methacrylate (GMA) resin (Technovit 7100 - Heraeus Kulzer, Wehrheim, Germany), and serial sections (3 μm thickness) were stained with 0.1% toluidine blue in 1% sodium borate and examined and documented using a Leica DMI6000 microscope (Leica, Heidelberg, Germany).

After exposure to methimazole or methimazole + T4, zebrafish testes (n = 8 per treatment) were dissected and immediately fixed in 2% glutaraldehyde and 4% paraformaldehyde in Sorensen buffer [0.1 M, pH 7.2] at 4°C overnight. Subsequently, testes were dehydrated, embedded in GMA resin (Technovit 7100—Heraeus Kulzer, Wehrheim, Germany), sectioned at 3 μm thickness, and stained with 0.1% toluidine blue. The slides were evaluated, and the proportion of section surface area of spermatogenic cysts containing different germ cell types were determined: type A undifferentiated spermatogonia (A_und* and A_und), type A differentiated spermatogonia (A_diff), type B spermatogonia (SpgB), spermatocytes (Spc), and spermatids (Spt). Intersection points were counted on the histologic fields, for which five fields per slide (n = 8 slides per treatment) were quantified using a grid of 540 (54 × 10) intersections under 100x objective lens. The proportion of section area occupied by spermatogenic cysts containing different germ cell types were represented as fold-change of control value.

For the quantification of the relative number of spermatozoa, twenty different histological fields were captured at 100x objective lens and analyzed by IMAGEJ Software (available at http://imagej.nih.gov/ij/index.html) according to Fallah and collaborators (2019, 2020), Tovo-Neto and collaborators (2020) and Rodrigues and collaborators (2021).

Total RNA from testes (control, methimazole, and methimazole + T4 groups) was extracted using TRIzol™ (Invitrogen, Carlsbad, CA, United States), according to the manufacturer’s instructions, and quantity and purity were checked with a NanoDrop™ One Spectrophotometer (Thermo Scientific, Madison, WI, United States). cDNA synthesis was performed as described previously (Nóbrega et al., 2010). qPCR reactions were conducted using 5 μL of 2x SYBR-Green Universal Master Mix, 1 μL of forward primer (9 mM), 1 μL of reverse primer (9 mM), 0.5 μL of DEPC water, and 2.5 μL of cDNA. The relative mRNA levels of thrα and thrβ (thyroid hormone receptors), fshr (follicle-stimulating hormone receptor), cyp17a1 (17α-hydroxylase/17,20 lyase/17,20 desmolase), insl3 (insulin-like peptide 3), cx43 (testicular connexin), igf3 (insulin-like growth factor 3), amh (anti-Müllerian hormone), gsdf (gonadal somatic cell derived factor), nanos2 (marker of undifferentiated spermatogonia), dazl (deleted in azoospermia-like), sycp3l (synaptonemal complex protein 3) and odf3a (outer dense fiber protein 3) were measured in the different treatments.

Brain (n = 8) and pituitary (n = 4 pools of 4 pituitaries for each pool) were collected from control and methimazole groups. Brain of each fish was kept separate. Total RNA was extracted from the brain using TRIzol™ (Invitrogen, Carlsbad, CA, United States) method. At least four pituitary glands were pooled per group (n = 4 pools per treatment), and total RNA was extracted using a commercial kit (PureLink^TM RNA Mini Kit, Ambion, Life Technologies, Carlsbad, CA, United States). After RNA extraction, the usual downstream methods were followed according to procedures described above. The relative mRNA levels of gnrh2 and gnrh3 (gonadotropin-releasing hormones), gnih (gonadotropin-inhibitory hormone), and crf (corticotropin-releasing hormone) were analyzed in the brain, and the lhb (luteinizing hormone), fshb (follicle-stimulating hormone), and tsh (thyroid-stimulating hormone) mRNA levels were determined in the pituitary gland. mRNA levels of the targets (Cts) were normalized by the transcript levels of β-actin and expressed as relative values of the control group. Primers were designed according to zebrafish sequences (Supplementary Table S1).

Blood from adult male zebrafish in different conditions (control, T4, methimazole, and methimazole + T4 groups) were collected (n = 8 per condition). Fish were euthanized, and the caudal peduncle was cut for blood sampling using heparinized syringes and tubes. Subsequently, blood was centrifuged at 4°C for 10 min at 800 ×g (Eppendorf Centrifuge 5424 R), and 11-Ketotestosterone (11-KT) plasma levels were quantified by ELISA (582751, Cayman Chemical, Ann Arbor, MI, United States), following the manufacturer’s instructions. The results were evaluated as nanograms of 11-KT per milliliter of plasma.

An ex vivo testis culture system described previously (Leal et al., 2009) was used to culture zebrafish testes. For short-term incubations (18 h for 11-KT release analysis), testes were submerged in a culture medium, whereas for long-term exposure (7 days for histomorphometrical analysis and gene expression), testes were placed on a nitrocellulose membrane on top of a cylinder of agarose (1.5% w/v, Ringer’s solution, pH 7.4) and exposed to 1 ml of medium culture in 24-well flat-bottom plates, as described by Leal and collaborators (2009).

Zebrafish testes were collected from eight animals per condition (control, T4, methimazole, methimazole + T4) post-dissection. One testis was cultivated in the Lebovitz medium (L-15), whereas its contra-lateral one in L-15 containing recombinant zebrafish Fsh (rzfFsh; 100 ng/ml). The rzfFsh protein was obtained from ImmunoPrecise Antibodies (Europe) B.V. Science Park Utrecht, Netherlands. Following incubation, testes were individually weighed, and the medium was collected and stored at −20°C for androgen release (11-KT) assay (see Sub-Section 2.9).

This technique was used to examine whether treatment conditions (T4, methimazole, methimazole + T4) modulated rzfFsh (100 ng/ml)-induced androgen release by zebrafish testis. The androgen (11-KT) release capacity of zebrafish testes into culture medium was measured after 18 h incubation as described previously (García-López et al., 2010). The levels of 11-KT released by zebrafish were quantified by ELISA using a commercial kit (Cayman Chemical) following manufacturer’s instructions.

To examine the effects of 100 nM T3 (3,3’,5- Triiodo-L-Thyronine) (CAS 6893-02-3; MW, 650.97 g/mol; purity ≥95%; Sigma-Aldrich, St. Louis, MO, United States) (n = 8) on zebrafish spermatogenesis, long-term incubations were performed according to Leal and collaborators (2009). For that, one testis was incubated in the presence of T3 alone and its contralateral one in a basal culture medium. The proportion of section area occupied by different germ cell types were represented as fold-change of basal value. The relative number of spermatozoa per area was quantified as described above (see Sub-Section 2.4). Also, this technique was used to analyze if different concentrations of T3 and T4 (10, 100 and 1,000 nM) modulate expression of selected genes in zebrafish testis. For that, total RNA from testis explants (n = 8) was extracted and the relative mRNA levels of nanos2, sycp3l, 3β-HSD (3-beta (β)-hydroxysteroid dehydrogenase), and cyp17a1 were evaluated as described above (see Sub-Section 2.5) (Supplementary Table S1).

In this study, we used T4 for the long-term experiments as the hormone is converted into T3 over time (see above), while to assess the rapid effects of thyroid hormones (short-term experiments), we used the active hormone (T3). Therefore, adult zebrafish were intracelomically injected with 0 and 250 ng of T3 per fish (n = 16). The dose was selected according to its ability to stimulate deiodinase type 3 mRNA levels as described previously (Nelson and Habibi, 2008; Nelson et al., 2010). A stock solution of T3 was dissolved in sodium hydroxide (0.02 M) and further diluted in physiologic saline. The control group was injected only with the physiologic saline solution. After 12 h, the animals were euthanized, and brain and pituitary glands were sampled. Brain of each fish (n = 8) was kept separate. The pituitaries were pooled (n = 4 pools of 4 pituitaries per condition) for RNA extraction. mRNA levels of gnrh2 and gnrh3, gnih, and crf were quantified in the brain, and lhb, fshb, and tsh were quantified from pituitary glands (Supplementary Table S1). RNA extraction and downstream procedures were followed as described in Sub-Section 2.5 (see the Figure below).

All data were subjected to normality Shapiro-Wilk test followed by the Bartlett homogeneity variance test. Significant differences between two groups were identified using unpaired or paired t-tests, while for three or more groups, one-way ANOVA followed by the Student–Newman–Keuls or Dunnett’s tests was used. Significance level (p) was considered ≤0.05 in both cases. Data are represented as mean ± SEM (Standard Error of Mean). All statistical analysis was performed using Graph Pad Prism software 7.04 (Graph Pad Software, Inc., San Diego, CA, United States, http://www.graphpad.com).

To confirm that basal thyroid hormone levels were affected after methimazole or methimazole + T4 treatment, plasma samples were collected and T3 levels were measured in the different experimental groups (Figure 1). The analysis showed that plasma T3 levels were significantly decreased following methimazole treatment (approximately 59 ng/ml) when compared to control group (approximately 182 ng/ml) (Figure 1B). In contrast, the observed effect on T3 levels after methimazole treatment was recovered by co-treatment with T4, and plasma T3 levels (approximately 137 ng/ml) were significantly similar to the levels found in control animals (Figure 1B).

Thyroid gland follicles were examined histologically in the control group and following treatments with methimazole and methimazole + T4 (Figures 1A,C,D,E). The control group thyroid follicles consisted of squamous or cuboidal epithelial cells filled with colloid (Figure 1C). The results demonstrate a significant modification in the histological condition of thyroid follicles in fish treated with methimazole. Three-week exposure to 1 mM methimazole resulted in thyroid gland follicles with colloid depletion, columnar epithelium and follicle cell hypertrophy (Figure 1D). These morphological changes were consistent with previous studies in which adult male zebrafish were exposed to methimazole (Rodrigues et al., 2021), and other goitrogens, such as perchlorate (Patiño et al., 2003) and 6-n-propyl-2-thiouracil (PTU) (Van der Ven et al., 2006). The observed effect of methimazole was reversed by co-treatment with T4, in which the thyroid follicles were found to be morphologically similar to the control group (Figure 1E). The results demonstrate that methimazole-induced hypothyroidism in zebrafish had altered thyroid function following treatment with the goitrogen. Also, the results demonstrate that co-treatment with T4 restored the zebrafish thyroid follicular structure.

Methimazole-induced hypothyroidism promoted histomorphometrical alterations in the proportion of germ cell cysts compared to the control (Figures 2A–D). There was a significant increase in the proportion of the area occupied by type A undifferentiated spermatogonia (A_und*), type A differentiated (A_diff) and spermatogenic cysts containing type B spermatogonia (SpgB) in the methimazole group as compared to control (Figure 2D). The number of meiotic cells (Spc) and post-meiotic haploid cell population (Spt) did not change between control and methimazole-induced hypothyroidism group (Figure 2D). However, the relative number of spermatozoa reduced drastically when compared to control as clearly evidenced in the photomicrographs and analysis of spermatozoa number by field (Figures 2A,B,E). Co-treatment with T4 rescued the proportion of A_und*, A_diff and SpgB types returned to its basal values, while the proportion area occupied by Spc and Spt significantly increased (Figure 2D). Remarkably, the production of spermatozoa was recovered in the co-treatment with T4 (as viewed in the fields of Figures 2A,C,E).

Transcript levels of selected genes involved in reproduction were measured by qPCR in the testis from methimazole-induced hypothyroidism and rescued groups (methimazole + T4) (Figure 3). In this study, we measured transcript levels of two thyroid hormone receptor subtypes (thrα, thrβ). The thrα transcript level was higher in the methimazole-treated group than control, but the difference was not statistically significant (Figure 3A). The thrα transcript level was further increased significantly in the methimazole + T4 treated group (Figure 3A). Similarly, the thrβ was increased in the methimazole and methimazole + T4 treated groups, compared to the control (Figure 3B).

We also measured mRNA levels of fshr which was not altered in the methimazole treated group, but considerably increased in the methimazole + T4 treated fish, compared to the control (Figure 3C).

In the present research, we quantified transcript levels for cyp17a1 and insulin-like peptide 3 (insl3) genes. Treatment with methimazole significantly reduced the cyp17a1 and insl3 transcript levels compared to control (Figure 3D,E). Co-treatment with T4 did not influence the methimazole induced response on the expression of these transcripts (Figure 3D,E). We also observed significant reduction in the testicular connexin (cx43) mRNA levels in the methimazole treated group, compared to control (Figure 3F). Co-treatment with T4 in this case increased the cx43 mRNA to a level not significantly different from the control (Figure 3F).

Among others, igf3, amh, and gsdf genes are known to be expressed in the Sertoli cells. Treatment with methimazole or methimazole + T4 were without significant effects on igf3 and amh transcript levels (Figure 3G,H). The gsdf mRNA level, however, was higher in the methimazole and methimazole + T4 treated groups (Figure 3I).

With regard to germ cell marker genes, such as marker for type A undifferentiated spermatogonia, nanos2 mRNA level was considerably higher in the methimazole-treated group compared to control (Figure 3J). Co-treatment with T4, significantly reduced the methimazole-induced response to a level higher than the basal control (Figure 3J). The dazl (deleted-in azoospermia-like) transcript level expressed by A_diff and SpgB, increased in the methimazole treated group compared to the control (Figure 3K). Co-treatment with T4 did not influence the methimazole-induced response (Figure 3K). In our study, synaptonemal complex protein 3 (sycp3l) which is a marker for spermatocytes was not significantly affected by methimazole or methimazole + T4 treatments (Figure 3L). However, the transcript level of the outer dense fiber protein 3 (odf3a) which is a marker for spermatids was increased following treatment with methimazole (Figure 3M). Co-treatment with T4 reduced the methimazole-induced response to a level not significantly different from the basal control (Figure 3M).

In the present research, we measured the 11-KT concentration to partially assess the effect of methimazole-induced hypothyroidism on steroidogenesis, in vivo and in vitro (Figure 4A). Four treatment groups were studied in this experiment, including control, control + T4, methimazole and co-treatment (methimazole + T4) groups (Figure 4A). Animal exposure with T4 (control + T4) significantly reduced the plasma 11-KT level as compared to control levels (Figure 4B). Also, treatment with methimazole significantly reduced the plasma 11-KT concentration to almost undetectable level compared to the control level of over 40 ng/ml (Figure 4B). However, co-treatment with T4 significantly increased and nullified the methimazole-induced response to a level not significantly different from the control (Figure 4B).

We also measured the androgen (11-KT) release capacity of zebrafish testicular tissue using the ex vivo culture system (Figure 4A). In this experiment, we compared testis taken from control and those exposed to T4 (100 μg/L), methimazole (1 mM) and methimazole co-treated with T4. We tested the 11-KT release response into culture media following in vitro treatment for 18 h to recombinant zebrafish Fsh (100 ng/ml) (Figure 4A). In the basal medium condition, the treatments (T4, methimazole and methimazole + T4) did not change the basal androgen (11-KT) release capacity of zebrafish testicular tissue (Figure 4C). However, as expected, treatment with Fsh significantly increased the 11-KT level in the control group (Figure 4C). Also, treatment with T4 was responsive to Fsh but to a level not significantly different from the control (Figure 4C). In comparison, the isolated testis from methimazole-treated zebrafish was completely unresponsive to Fsh (Figure 4C). However, co-treatment with T4 restored the Fsh-induced response by increasing the 11-KT concentration to the level observed following treatment of the control group with Fsh alone (Figure 4C).

In this experiment, we measured brain transcript levels of a number of neurohypothalamic peptides, including gnrh2 and gnrh3, gnih, and crf, as well as pituitary gonadotropin hormone subunits, fshb, lhb, and tshb in the methimazole-induced hypothyroidism in fish (Figure 5A,C). As shown in Figure 5C, we observed a significant increase in the gnih transcript level compared to the control shown as the dotted line. Transcript levels of the other neuropeptides measured including gnrh2, gnrh3 and crf remained unchanged. In the same group of methimazole-treated fish, the results the pituitary glycoprotein hormone subunits, demonstrate significant increase in fshb mRNA and a massive increase in tshb transcript levels in the methimazole-induced hypothyroid fish (Figure 5C). No change was observed in the lhb transcript level (Figure 5C).

We also measured the same transcript levels following 12 h acute treatment with 250 ng of T3/fish in vivo (Figures 5B,D). In this experiment T3 injection did not significantly alter gnih, gnrh2 and crf, but significantly increased the gnrh3 transcript level (Figure 5D). In the same group of fish, acute treatment with 250 ng of T3/fish significantly reduced the pituitary lhb but was without effect on the fshb and tshb transcript levels (Figure 5D).

In this research, we also examined the direct action of T3 at 100 nM on ex vivo culture of zebrafish testis for 7 days (Figure 6A) and results provide information on the proportion of germ cells in the cultured testis. Histomorphometrical evaluation of zebrafish testis revealed that treatment with T3 significantly stimulated the type A undifferentiated spermatogonia (A_und) abundance with no effect on type A differentiated spermatogonia cells as compared to basal medium incubation (Figure 6B). In the same tissue, we observed a reduction in type B spermatogonia (SpgB) abundance (Figure 6B). As for meiotic and post-meiotic cells, treatment with T3 reduced the proportion of area occupied by Spc and Spt when compared to basal condition (Figure 6B). Interestingly, the number of spermatozoa was stimulated in zebrafish testis treated with T3 as clearly shown by the analysis of spermatozoa number by field (Figure 6C). In addition to these results, the effect of thyroid hormones on testicular gene expression was analyzed (Figures 6D–G) (Supplementary Figure S1). Several concentrations of T3 and T4 (10, 100 and 1,000 nM) were tested in zebrafish testis. T4 did not change the expression of any transcript (nanos2, sycp3l, 3β-HSD and cyp17a1) (Supplementary Figure S1). On the other hand, expression analysis revealed that T3 (100 and 1,000 nM) increased nanos2 (Figure 6D), and up-regulated sycp3l transcript levels at 100 nM (Figure 6E). Levels of 3β-HSD and cyp17a1 were also quantified and did not change within any levels of T3 (Figures 6F,G).