Healthy specimens of gilthead seabream were from the Oceanographic Center of Murcia (Mazarrón, Spain), where they were kept in running seawater aquaria (dissolved oxygen 6 ppm, flow rate 20% aquarium vol/h) with a natural temperature and photoperiod. They were fed three times per day with a commercial pellet diet (44% protein and 22% lipids; Skretting) at a feeding rate of 1.5% of fish biomass. The environmental parameters, mortality, and food intake, as well as behavior, were recorded daily.

Two-month-old gilthead seabream specimens (n = 100/treatment), with a body weight of 26.6–63.2 (from the beginning to the end of the experiment), were exposed to EE2 (Figure 1A) in 170 l aquaria. Briefly, EE₂ (5 μg/g food, 98% purity; Sigma) was incorporated in the commercial food using the ethanol evaporation method (0.3 l ethanol/kg of food), as described elsewhere (40). The specimens were fed three times a day ad libitum with the pellet diet supplemented with EE₂ (treated fish) or non-supplemented (untreated fish) for 76 days (days of treatment, dt). Following this time, all the specimens were fed with the commercial food for a further 23 days (days posttreatment, dpt) (Figure 1A). In order to evaluate the effect of EE₂ on the immune response, specimens were intraperitoneally (i.p.) injected with keyhole limpet hemocyanin (KLH) (45 μg/fish; Sigma-Aldrich) and Imject Alum adjuvant (4 mg/fish; Thermo Scientific) (vaccinated/immunized fish) or phosphate-buffered saline (PBS) (control/unvaccinated fish) at the end of the treatment period, 76 dt. Samples were taken 40 and 76 dt and 1, 9, and 23 days postimmunization (dpi) or dpt. Food intake was similar in all groups. Specimens (n = 6 fish/treatment/time of sampling) were fasted for 24 h before sampling. They were tranquilized by 8 μl/l of clove oil, immediately anesthetized using 40 μl/l of clove oil, weighed and decapitated before the head kidneys and spleens were removed, and processed for gene expression and/or flow cytometry analysis, as described later. Serum samples from trunk blood were obtained by centrifugation and immediately frozen and stored at −80°C until use. Cell suspensions from head kidney and spleen were obtained as described elsewhere (41, 42).

Another set of experiments was performed for serum Ab titer determination by ELISA: (1) 2-month-old gilthead seabream specimens were dietary treated with EE₂ (5 μg/g food) or the GPER1 agonist G1 (5 μg/g food; Sigma-Aldrich) for 110 days (30), (2) adult specimens (650 g mean weight) were treated with 0, 5, and 50 ng/l EE₂ for 2 months by bath immersion (31), and (3) adult specimens (225 g mean weight) were exposed to dietary G1 (0, 2, and 20 μg/fish/day treatment) for 50 days (27).

The serum Vtg levels were determined by the enzyme-linked immunosorbent assay using a commercial kit (Cayman Chemical), following the manufacturer’s instructions. In brief, an aliquot of 1:500 diluted serum from untreated and EE₂-treated fish (n = 6 fish/treatment/time of sampling) was added to flat-bottomed 96-well plates, followed by a commercial polyclonal Ab against gilthead seabream Vtg (1:100) and an antirabbit IgG (whole molecule)-peroxidase Ab (1:1,000) (Sigma-Aldrich). Finally, the chromogen tetramethylbenzidine (TMB) was added and the absorbance was read at 450 nm using an FLUOstart luminometer (BGM; LabTechnologies).

Total RNA was extracted from head kidney from untreated and EE₂-treated (vaccinated or not) fish (n = 6 fish/treatment/time of sampling) at 1, 9, and 23 dpi/dpt with TRIzol Reagent (Thermo Fisher Scientific), following the manufacturer’s instructions, and quantified with a spectrophotometer (NanoDrop, ND-1000). The RNA was then treated with amplification grade DNase I (1 U/μg RNA; Thermo Fisher Scientific) to remove genomic DNA traces that might interfere with the PCRs, and the SuperScript III RNase H Reverse Transcriptase (Thermo Fisher Scientific) was used to synthesize first-strand cDNA with oligo-dT18 primer from 1 mg of total RNA for 50 min at 50°C. The β-actin (actb) gene was analyzed by PCR performed with an Eppendorf Mastercycle Gradient Instrument (Eppendorf) to check cDNA quality. Reaction mixtures were incubated for 2 min at 94°C, followed by 30 cycles of 45 s at 94°C, 45 s at the specific annealing temperature (55°C), 45 s at 72°C, and finally, 10 min at 72°C.

In the same samples, the expression levels of the gene coding for the proinflammatory cytokine interleukin-1β (il1b) were analyzed by real-time PCR performed with an ABI PRISM 7500 instrument (Applied Biosystems) using the SYBR Green PCR Core Reagents (Applied Biosystems). Reaction mixtures were incubated for 10 min at 95°C, followed by 40 cycles of 15 s at 95°C, 1 min at 60°C, and finally, 15 s at 95°C, 1 min at 60°C, and 15 s at 95°C. The gene expression was corrected by the ribosomal protein S18 gene (rps18) content in each sample using the comparative cycle threshold method, Ct method (2^−ΔΔCt). The gilthead seabream-specific primers used are listed in Table 1. In all cases, samples were analyzed in triplicate.

The percentage of Zap70-, IgM-, and Pax5-positive cells was determined by flow cytometry (Figure S1A in Supplementary Material). In brief, cytoplasmic Zap70 and Pax5, and surface and total IgM, were detected in aliquots of 0.5 × 10⁶ head kidney leukocytes at 1, 9, and 23 dpi/dpt and of spleen leukocytes at 23 dpi/dpt (n = 6 fish/treatment/time of sampling) from untreated and EE₂-treated fish (vaccinated or not). The leukocytes were washed with PBS containing 2% fetal calf serum (FCS; Thermo Fisher Scientific) and 0.05% sodium azide (FACS buffer). Cells were fixed with 4% paraformaldehyde for 15 min at room temperature. After three rinses, cells were incubated in ice-cold PBS containing 1% BSA and saponin (Thermo Fisher Scientific) at 4°C to permeabilize the plasma membrane. Cells were then stained with 0 and 2 μg/ml (0, 1:100) of the 99F2 rabbit monoclonal Ab to Zap70 (Cell Signaling) (43) or mouse monoclonal Ab specific to seabream IgM (Aquatic Diagnostics) (44), in PBS containing 2% FCS for 30 min at 4°C. After washing, cells were incubated with a 1:1,000 dilution of a phycoerythrin-conjugated antirabbit or -mouse Ig Ab, respectively, for 30 min at 4°C, and washed again twice. In other experiments, cells were also stained with 2 μg/ml (1:100) of the D19F8 rabbit monoclonal Ab to Pax5 conjugated with Alexa 488 (Cell Signaling) (43) in PBS containing 2% FCS for 1 h at room temperature. After washing, cells were analyzed in a flow cytometer (BD Biosciences).

To assess the proliferative activity of head kidney and spleen Zap70⁺ and IgM⁺ cells at 1, 9, and 23 dpi/dpt (n = 6 fish/treatment/time of sampling), 5-ethynyl-2′-deoxyuridine (EdU) (Thermo Fisher Scientific) was i.p. injected 2 h before sampling. Then, aliquots of 0.5 × 10⁶ head kidney and spleen cells were used to determine the percentage of double positive cells, i.e., Zap70⁺/EdU⁺ and IgM⁺/EdU⁺. EdU⁺ cells were detected by fluorescent-azide coupling reaction with EdU according to the manufacturer’s protocol (Click-iT; Thermo Fisher Scientific). Zap70⁺ and IgM⁺ were labeled as described earlier and analyzed by flow cytometry (Figure S1 in Supplementary Material). The percentage of positive cells is given on head kidney R2 region (FSC^low/SSC^low), which includes macrophages, lymphocytes, and hematopoietic precursor cells and excludes acidophilic granulocytes (R1: FSC^high/SSC^high) (44).

Total IgM titers and those specific to KLH, lysozyme (unrelated antigen), Shewanella putrefaciens strain Pdp11 (a skin microbiota bacterium from gilthead seabream) (45), or Vibrio anguillarum strain R82 (a fish pathogen) were determined in serum by ELISA (Aquatic Diagnostics), following the manufacturer’s instructions. In brief, serial dilutions of pooled serum samples (n = 6 fish/treatment/time of sampling) were added to flat-bottomed 96-well plate precoated with KLH (1 μg/well), chicken egg white lysozyme (1 μg/well; Sigma-Aldrich), and S. putrefaciens or V. anguillarum (10⁶ bacteria/well), followed by a monoclonal Ab specific to gilthead seabream IgM (1:100) and then an antimouse IgG peroxidase Ab (1:1,000) (Sigma-Aldrich). Finally, the chromogen TMB was added and the absorbance was read at 450 nm using an FLUOstart luminometer (BGM; LabTechnologies). As negative controls, the serum or the primary Ab was omitted.

Vibrio anguillarum strain R82 was grown in tryptic soy agar plates at 25°C. Fresh single colonies were diluted in 5 ml of tryptic soy broth, cultured for 16 h at 25°C on an orbital incubator at 300 rpm, and adjusted to 10⁷ bacteria/ml. Aliquots of 50 μl containing 10⁶ bacteria in HBSS with Ca²⁺/Mg²⁺ were placed in flat-bottomed 96-well plates and incubated for 2 h with 50 μl of pool gilthead seabream serum samples diluted 1/10 in PBS. Moreover, several controls were introduced: blank, without bacteria; positive control, with 10 μg/ml gentamicin (0% growth or 100% bactericidal activity); negative control, without serum (100% growth or 0% bactericidal activity); and decomplemented serum control (pretreated at 50°C, 20 min) to asses complement-mediated killing. The bactericidal activity of serum samples was determined using a 2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide inner salt (XTT) assay where the yellow tetrazolium salt XTT is reduced to a colored formazan dye by dehydrogenase enzymes in metabolically active cells (46). The bactericidal activity was expressed as the percentage of bacterial growth inhibition.

Data were analyzed by the analysis of variance. An unpaired Student’s t-test was applied to determine differences between two groups. The critical value for statistical significance was taken as p < 0.05. The asterisks *, **, and *** refer to p < 0.05, p < 0.01, and p < 0.001, respectively. All statistical analyses were carried out using the GraphPad Prism 5 program.

The survival of gilthead seabream specimens was 100% during the trial (data not shown). As a control for estrogenic endocrine disruption, serum Vtg levels were analyzed. The results showed that the dietary intake of EE₂ significantly increased serum Vtg levels, as we have previously described for hepatic Vtg transcript levels in adult and juvenile gilthead seabream fish treated with EE₂ (29–31, 47, 48). Vtg levels peaked at the end of the treatment time and then gradually decreased, although they were still altered at 23 dpt (Figure 1B). Furthermore, when the expression of the gene coding for il1b was analyzed in head kidney at 1, 9, and 23 dpi to assess the activation of innate immunity in response to the immunization, increased il1b transcript levels were only evident 1 and 9 dpi as expected (Figure 1C). However, dietary EE₂ failed to significantly affect il1b mRNA levels (Figure 1C).

Both the dietary intake of EE₂ and immunization led to an increase in the percentage of T lymphocytes (i.e., Zap70⁺ cells) in head kidney at 1 dpi (Figure 2A), whereas no statistically significant differences were observed at 9 and 23 dpi with any of the treatments (Figures 2B,D). Although the percentage of proliferating T lymphocytes (i.e., Zap70⁺/Edu⁺ cells) in head kidney was unaffected by the treatments at 9 dpi (Figure 2C), EE₂ slightly increased the percentage of proliferating T lymphocytes in the head kidney at 23 dpi (Figure 2E).

Although the percentage of IgM⁺ B lymphocytes in the head kidney was slightly higher in EE₂/immunized fish than in the control at 9 dpi/dpt, it was lower in all treated group than in the control at 23 dpi/dpt (Figures 3A,C,F). Interestingly, although the percentage of IgM⁺/Pax5⁻ B lymphocytes, i.e., plasma cells, was unaltered at 1 and 23 dpi/dpt, it was higher at 9 dpi/dpt in EE₂/non-immunized fish but not in immunized fish treated with EE₂ (Figures 3B,D,G). In addition, the dietary intake of EE₂ and immunization resulted in a decreased proliferation of IgM⁺ B lymphocytes at 23 dpi/dpt (Figures 3E,H). Neither immunization nor EE₂ exposure was seen to modulate IgM⁺ B-lymphocyte abundance, proliferation, or differentiation in the spleen at 23 dpi/dpt (data not shown).

First, the KLH-specific IgM titer was analyzed by ELISA in serum from untreated and EE₂-treated fish, both control and immunized, at 1, 9, and 23 dpi/dpt. Unexpectedly, the dietary intake of EE₂ significantly increased the KLH-specific IgM titer at 1 dpi/dpt (Figure 4A). However, immunization failed to elicit a specific IgM response at any of the times analyzed (Figures 4A–C), presumably reflecting the poor primary Ab response to KLH.

These results prompted us to analyze the IgM titer to an unrelated antigen, lysozyme, in the serum of non-immunized fish exposed to EE₂ by different routes (in this case dietary exposure and bath immersion). Strikingly, EE₂ significantly increased lysozyme-specific IgM titers during treatment (40 dt) (Figure 4D) and posttreatment (1 and 9 dpt) (Figures 4E,F), but the effect did not last until 23 dpt (Figure 4G). Similar results were obtained in fish bath-exposed to EE₂ for approximately 2 months (Figure 4H). The induction of natural antibodies by the administration of EE₂ was further confirmed by determining total serum IgM levels (Figures 4I–M). Furthermore, EE₂-induced natural IgM antibodies were not only polyreactive but also neutralizing, since they reacted against the commensal S. putrefaciens (Figures 5A–C) and the pathogen V. anguillarum (Figures 5D–F), inducing the death of the latter (Figures 5G,H). In all cases, the natural antibodies show low affinity, since high dilution of serum had to be used to detect them (Figures 4–6).

As previous studies demonstrated the key role of estrogen signaling through GPER1 in the regulation of innate and adaptive immunities in gilthead seabream, we next evaluated the effect of GPER1 activation on natural Ab production using the specific agonist G1. The results showed that G1-treated fish had higher lysozyme-specific (Figure 6A) and total IgM (Figure 6B) serum titers and higher bactericidal activity against V. anguillarum at 30 dt than the untreated controls (Figure 6E). Moreover, the effects of EE₂ and G1 on natural Ab production in fish treated for 110 days were similar (Figures 6C,D).