Grass carp (100–250 g) juveniles were obtained from a fish farm located in Guangdong Province, China, and maintained at 28–29°C in a recirculating freshwater system for at least 2 weeks of acclimation to the laboratory condition before carrying out the experiments.

The hemolysis model was established as described previously, with minor modification (28). Fish were randomly divided into two groups: one group of fish given intraperitoneal injection of PHZ at 40 mg/kg body weight and a control group of fish injected with an equal volume of phosphate-buffered saline (PBS). In time course experiments, fish were sacrificed at 12, 24, 48, and 96 h after PHZ injection.

The Hb concentrations in plasma were evaluated as described previously, with slight modifications (10, 26). Briefly, the blood was collected from the caudal vein of each group of fish using heparinized syringes and mixed with 0.7% buffered saline. After centrifugation at 400 × g for 10 min at 4°C, the supernatant was obtained and quantified spectrophotometrically at a wavelength of 404 nm using a microplate reader (Molecular Devices, San Jose, CA, USA).

The total iron concentrations in the head kidney and middle kidney were measured according to previous reports (11). Briefly, the head kidney and middle kidney (100 mg) were homogenized in double deionized H₂O at 1:10 (w/v). The samples were mixed with 500 µl of an acid mixture containing 1 mM HCl and 10% trichloroacetic acid (TCA) and then incubated at 50°C for 1 h. After centrifugation at 15,000 × g for 15 min at room temperature (RT), 750 µl of the supernatant was mixed with 250 µl of 20 mg/ml ascorbic acid, followed by 200 µl of ferrozine (0.85% w/v in hydroxylamine hydrochloride). The samples were developed over 30 min. The absorbance was measured at 560 nm using a microplate reader. A standard curve was generated using an iron solution standard (500 µg/dl).

To reveal the effect of hemolysis on tissues, histopathology analysis was performed. The collected samples were fixed in 4% paraformaldehyde (Servicebio, Wuhan, China) for at least 12 h. The tissues were subsequently stained with hematoxylin and eosin (H&E). The iron deposition in tissues was determined with the Perls’ iron stain assay, and the Sirius red staining assay was performed to verify the effect of hemolysis on tissues. Finally, the tissue sections were observed under a microscope (Axiostar Plus, Carl Zeiss, Oberkochen, Germany), and microphotographs were taken using a Canon PowerShot G6, 7.1 megapixels (Canon 219 Inc., Tokyo, Japan).

The apoptosis assay was performed as described previously, with minor modifications (29, 30). TdT-mediated dUTP nick-end labeling (TUNEL) staining was carried out on 5-μm-thick paraffin-embedded sections using a one-step TUNEL apoptosis assay kit (DAPI; Guge Biology, Wuhan, China). The cell nucleus was counterstained with 4′,6-diamidino-2-phenylindole (DAPI) (Servicebio, Wuhan, China) at 25°C for 5 min. Finally, images were captured under a fluorescence microscope (Leica DMI8, Wetzlar, Germany). The caspase-3 activity in the collected tissues was determined by following the manufacturer’s instructions (Solarbio Biology, Beijing, China).

The Western blotting (WB) process followed a previous procedure, with slight modifications (26, 31). Briefly, following the membrane transfer of proteins by SDS-PAGE, the membranes were blocked in Tris-buffered saline with Tween 20 (TBST) containing 5% skim milk at 37°C for 2 h, followed by incubation with 100 μl of anti-grass carp Hb and heme oxygenase-1 (HO-1) rabbit serum (diluted 1:1,000) at 4°C overnight. After washing three times with TBST, the membranes were incubated with 100 μl of the secondary antibody horseradish peroxidase (HRP)-linked goat anti-rabbit immunoglobulin G (IgG) antibody (diluted 1:10,000) at RT for 1 h. The signals from HRP conjugates were detected using Clarity Western ECL Substrate (Solarbio Biology, Beijing, China). Finally, the membranes were imaged and analyzed using the ChemiDoc™ MP System (Bio-Rad, Hercules, CA, USA).

The tissue immunofluorescence assay was performed according to a previously described method (32). After dewaxing in Safeclear II, the tissue sections (5 μm) were rehydrated in graded ethanol and then treated in a microwave for 15 min in 10 mM sodium citrate buffer. The tissues were blocked for 2 h with 5% non-fat milk at RT, the grass carp primary antibodies anti-Hb, anti-HO-1, and malondialdehyde (MDA) (1:1,000 in 2% non-fat milk) for 2 h at RT, and then washed three times using PBS with Tween 20 (PBST) buffer. Subsequently, the fluorescein isothiocyanate (FITC)- or Cy3-conjugated anti-rabbit IgG antibody was added (1:10,000 in 2% non-fat milk) and incubated in darkness for 1 h at 37°C. After washing three times, the tissues were stained with DAPI for 5 min at RT. Finally, they were observed under a fluorescence microscope (Olympus BX51, Tokyo, Japan) after rewashing three times with PBST.

After the treatment of tissue sections as above, the sections were blocked in 5% non-fat milk at RT for 2 h and then incubated overnight at 4°C with antibodies against 4-hydroxynonenal (4-HNE) diluted 1:1,000 in PBST containing 2.5% milk. Signal was developed using polymeric peroxidase-conjugated secondary antibody and DAB. All images were acquired under an Olympus BX51 inverted microscope.

To further explore the effect of hemolysis on tissues, we examined the contents of β-GAL, LPO, MDA, and three kinds of antioxidases (namely, glutathione peroxidase, superoxide dismutase, and catalase) according to the manufacturer’s instructions (Solarbio Biology, Beijing, China).

The method for total RNA extraction was mainly according to our previous reports (27, 32). Briefly, total RNA of the collected tissues was extracted using the TRIzol reagent (Takara, Dalian, China). Approximately 1 μg of total RNA was used to synthesize the first-strand cDNA using HIScript^® Q Select RT SuperMix for qPCR (Vazyme, Nanjing, China) following the manufacturer’s instruction and then stored at −20°C.

The expressions of the detected genes in the collected tissues were studied using quantitative real-time PCR (qRT-PCR) in a real-time thermal cycler (Roche, Basel, Switzerland) following the protocol of the SYBR Real-Time PCR Premixture (Vazyme, Nanjing, China). The procedures were as follows: pre-incubation at 95°C for 30 s, followed by 40 cycles of 95°C for 5 s, 55°C for 20 s, and 72°C for 20 s, and finally at 4°C for 5 min. The qRT-PCR for each gene was performed in triplicate and normalized to the values of β-actin. Statistical analysis used the 2^−ΔΔCT method. The qRT-PCR for each gene was performed in triplicate and normalized to the values of b-actin, the specific primers are listed in ( Table 1 ).

Statistical analysis was performed using SPSS software (version 17.0). All data were represented as the mean ± standard deviation (SD) of at least three independent experiments. The statistical significance of the data was assessed with one-way analysis of variance (ANOVA), and the figures were drawn in GraphPad Prism 7. Differences were considered significant at *p < 0.05 or **p < 0.01.

Here, hemolysis in grass carp was established after intraperitoneal injection of PHZ. The extent of hemolysis at the different treatment time points was evaluated using plasma. The PHZ-treated groups showed obvious Hb accumulation relative to the non-treated groups at 12 and 24 h, and the accumulation of Hb was markedly attenuated after treatment for 48 h ( Figures 1A, B ). To assess whether hemolysis in grass carp affects the antioxidant system, we determined the activity of three antioxidases: glutathione peroxidase (GSH), superoxide dismutase (SOD), and catalase (CAT). After treatment with PHZ, the content of GSH in serum was significantly increased from 12 to 24 h. A significant attenuating effect of GSH was obtained after treatment for 48 and 96 h ( Figure 1C ). Compared to GSH, we observed a significantly increased effect on the SOD contents at all tested time points ( Figure 1D ). For the content of CAT, we found that hemolysis caused the highest effect at 24 h. After treatment for 96 h, the content of CAT showed no significant difference compared to the control groups ( Figure 1E ).

To evaluate the effect of hemolysis on tissues, multiple staining assays were performed. The results of H&E staining of the head kidney showed an obvious Hb deposition in the PHZ-treated groups, especially in the 24- and 48-h treated groups. Large amounts of Hb were contained in leukocyte-like cells, and the interstitial spaces were enlarged significantly ( Figure 2A ). A similar observation was found in treated middle kidney, where abundant deposits of Hb were mostly engulfed by leukocyte-like cells. Several leukocyte-like cells aggregated together to phagocytose the excess Hb. The hemolysis caused serious damage to the middle kidney, which especially destroyed the integrity of the kidney tubules ( Figure 2A ). To further assess whether the excess Hb caused by hemolysis led to iron accumulation in tissues, we performed the Prussian blue staining assay in the head and middle kidney. The staining results confirmed that the hemolysis resulted in a large amount of ion accumulation both in the head and middle kidney ( Figure 2A ). The accumulation of Hb in the head and middle kidney was also detected by the microplate reader, with the results showing that the PHZ-induced hemolysis in grass carp significantly increased the accumulation of Hb in all tested time points ( Figures 2B, C ). Determination of the deposition of iron using the kit showed that the content of iron in the head kidney was significantly higher at 24 and 48 h compared to the control ( Figure 2D ), and the iron content in the middle kidney was significantly increased in all tested time points ( Figure 2E ). Subsequently, indirect immunofluorescence of Hb and HO-1 was performed to examine whether the iron was derived from the degradation of Hb. Compared to the control groups, the contents of Hb in both the head and middle kidney were significantly higher; however, after treatment for 48 h, the deposition of Hb showed a pronounced attenuating trend ( Figure 2F ). Results of the determination of HO-1 clearly displayed that the hemolysis in grass carp caused a sharp increase in the HO-1 content of both the head and middle kidney, which corresponded to the attenuating trend of Hb ( Figure 2F ). The WB results further verified that Hb increased significantly from 24 to 96 h in the head kidney after treatment with PHZ, which was similar to that determined for the content of HO-1 ( Figure 2G ). In the middle kidney, the WB results showed that the accumulation of Hb rapidly increased at 24 and 48 h after injection of PHZ; however, no significant effect was observed in the hemolysis-induced release of HO-1 in the middle kidney ( Figure 2G ). In addition, we also detected the mRNA expression of HO-1 in the head and middle kidney. As shown in Figures 2H, I , the expression level of HO-1 coincided with the results of the WB.

Accordingly, the PHZ-induced hemolysis caused Hb deposition and iron accumulation in both the head and middle kidney. qRT-PCR was used to further assess whether the PHZ-induced hemolysis modulated the iron metabolism-related genes. The results revealed that the hemolysis significantly increased the expression of divalent metal ion transporter 1 (DMT1) at all tested time points in the head kidney and at 48 and 96 in middle kidney ( Figures 3A, H ). Transferrin (Tf) was significantly upregulated at 12 and 48 h in the head kidney and at 24 and 48 h in the middle kidney ( Figures 3B, I ). A similar result to DMT1 was found for the expression of transferrin receptor 1 (TfR1), which was significantly upregulated from 12 to 96 h in the head kidney and from 24 to 96 h in the middle kidney ( Figures 3C, J ). The expression level of ferritin was significantly increased from 24 to 96 h both in the head and middle kidney ( Figures 3D, K ). Interestingly, the mRNA expression of ferroportin 1 (FPN1) was significantly decreased at all tested time points in both head and middle kidney ( Figures 3E, L ). However, the expression of haptoglobin (Hap) was significantly increased at all tested time points in both head and middle kidney ( Figures 3G, N ). Lastly, the expression data for hepcidin showed a big difference in the head kidney and middle kidney. The expression of hepcidin was downregulated at 48 and 96 h in the head kidney, but upregulated in the middle kidney ( Figures 3F, M ).

To assess whether hemolysis affects the level of ROS in tissues, we used the dihydroethidium (DHE) assay to detect ROS production after PHZ injection for 48 h. As shown in the data in Figure 5A , the hemolysis in grass carp significantly increased the production of ROS in the head kidney compared to the control group. In the middle kidney, the data also revealed that the PHZ-induced hemolysis improved the production of ROS, especially in kidney tubules ( Figure 4A ). To further reveal the effect of hemolysis on the production of ROS, the mean fluorescence intensity of ROS was analyzed. The analysis showed that the PHZ-induced hemolysis significantly increased the mean fluorescence intensity of ROS both in the head and middle kidney ( Figure 4B ).

To directly demonstrate the damage caused by PHZ-induced hemolysis to the head and middle kidney, we firstly performed the Sirius red stain assay. The results showed that hemolysis in grass carp caused an obvious deposition of collagen fiber both in the head and middle kidney ( Figure 5A ). The oxidative markers 4-HNE and MDA were also determined in both head and middle kidney, with the results showing that the PHZ-induced hemolysis significantly increased the content of 4-HNE at 24 and 48 h in the head kidney; similar results were found in the middle kidney ( Figure 5A ). For MDA, the detection results showed that the accumulation of MDA in the head kidney was significantly increased at 24 and 48 h, with similar results revealed for the middle kidney ( Figure 5A ). Furthermore, oxidative-related enzymes were also detected. It was shown that hemolysis significantly increased the content of β-GAL at all tested time points both in the head and middle kidney ( Figures 5B, E ). For the determination of LPO, the results revealed that hemolysis increased its content in both head and middle kidney ( Figures 5C, F ). Similar results were also found in the determination of MDA ( Figures 5D, G ).

To further investigate the effect of hemolysis on tissues, we evaluated apoptosis in the head and middle kidney after injection of PHZ for 24 and 48 h. The detection data for the head kidney clearly showed that the hemolysis markedly increased the TUNEL signal after PHZ treatment for 24 and 48 h, which confirmed that it induced apoptosis in the head kidney ( Figure 6A ). A similar result was also found for the middle kidney, in which the PHZ-induced hemolysis significantly increased the apoptosis both at 24 and 48 h compared to the control group ( Figure 6A ). In addition, we also detected the mRNA expression and enzyme activity of caspase 3 in the head and middle kidney after treatment at different time points. The data in Figure 6B revealed that the hemolysis markedly increased the expression of caspase 3 from 12 to 48 h in the head kidney. Besides, compared to the control head kidney, the enzyme activity of caspase 3 in the treated groups was significantly increased from 12 to 48 h ( Figure 6C ). In the middle kidney, the PHZ-induced hemolysis significantly increased the mRNA transcript and enzyme activity level of caspase 3 at 12, 24, and 48 h ( Figures 6D, E ).

Under hemorrhagic venereal infection or pathophysiological conditions, hemolysis occurs, which releases large amounts of Hb into the tissues and induces inflammation (6). To evaluate the effect of hemolysis on cytokines, we measured the expressions of several cytokine-regulated genes in the head kidney. Figure 7A shows that the hemolysis significantly increased the expression of TNF-α at 12 and 24 h in the head kidney. Compared to the control group, the expression levels of IL-1β and IL-8 were both significantly increased after PHZ treatment from 12 to 48 h in the head kidney ( Figures 7B, C ). For IFN-1γ, it was found that the mRNA transcript level at 12 h was significantly higher compared to the control group, and no significant effect was observed at other time points ( Figure 7D ). The mRNA transcript levels of IL-4 at 12 and 24 h were significantly increased ( Figure 7E ); however, the expression of IL-12 was more noticeably attenuated from 24 to 96 h ( Figure 7F ). In addition, we also detected the expressions of three chemokine genes, with the detection data revealing that PHZ-induced hemolysis activated their expressions to varying degrees. The detection data for CCL4 revealed that its mRNA transcript level reached the highest point after 24-h treatment, followed by that at 48 and 24 h ( Figure 7G ). The expression of CCL11 was significantly increased at 24 h compared to the control group ( Figure 7H ). For CCL20, its mRNA transcript level was significantly upregulated at 12 and 24 h; however, its expression at 48 h was downregulated ( Figure 7I ). Finally, we also detected the expressions of TLR4 and Nrf2. Their data revealed that the hemolysis significantly upregulated the mRNA transcript level of TLR4 at 12 and 24 h ( Figure 7J ). The hemolysis significantly increased the expression of Nrf2 at all tested time points ( Figure 7K ).