Deficiency in Neuroserpin Exacerbates CoCl2 Induced Hypoxic Injury in the Zebrafish Model by Increased Oxidative Stress

Protective strategy against hypoxic-ischemic (H/I) induced injury has been intensively discussed. Neuroserpin, an inhibitor for tissue plasminogen activator (tPA), has been proved a vital neuroprotective agent in cerebral ischemia mouse model and oxygen-glucose deprivation and reoxygenation (OGD/R) cell model. Neuroserpin is a promising therapeutic hint for neonatal hypoxic-ischemia injury. Here, we established a neuroserpin deficient zebrafish to study its role in CoCl2 chemically induced hypoxic injury. CoCl2 exposure was beginning at the embryonic stage. Development defects, neuronal loss, and vascular malformation was assessed by imaging microscopy. Neuroserpin deficient zebrafish showed more development defects, neuronal loss and vascular malformation compared to wide-type. Apoptosis and oxidative stress were evaluated to further identify the possible mechanisms. These findings indicate that neuroserpin could protective against CoCl2 induced hypoxic injury by alleviating oxidative stress.

In our study, hypoxia stress was induced by CoCl 2 in zebrafish. CoCl 2 mimics hypoxia by preventing prolyl and asparaginyl hydroxylase activity and proteasome degradation of the hypoxia inducible factor-1α (HIF-1α) under normoxic conditions, which is the most commonly used method in cells (Kamei and Duan, 2018;Munoz Sanchez and Chanez Cardenas, 2019). The most common methods to study effects of hypoxia in zebrafish during development is the use of CoCl 2 in the growth medium or incubation of the embryos in a hypoxic chamber with less than 5% O 2 concentration. However, zebrafish embryos are relatively tolerant to low oxygen (Mendelsohn et al., 2008;Elks et al., 2015). So, CoCl 2 induced chemical hypoxia injury is more reliable and economic than physical hypoxia chamber in this study.
Zebrafish have several advantages as a chemically induced hypoxia model. Firstly, hypoxia could be induced as early as embryonic stage due to its external development, which makes it possible to observe the phenotypic changes in the developing stage. External embryonic development in zebrafish also enable to study the effects of hypoxia in vivo easily and directly. Secondly, CoCl 2 hypoxia induction is more material and time economical compared to the hypoxia chamber method by addition of drugs direct to the embryo media (Elks et al., 2015). Thirdly, the small size of zebrafish embryos allows high-throughput behavioral screening at a time, minimizing deviation due to temporal and spatial differences.
To verify whether neuroserpin show protective effect on CoCl 2 induced injury in the zebrafish model, developmental morphological defects, behavioral change, neural lesion, and vascular damage were evaluated in a neuroserpin-deficient zebrafish model. Apoptosis and oxidative stress were assessed to identify related mechanisms. We demonstrate for the first time that neuroserpin exerts a protective role in CoCl 2 induced hypoxic injury in zebrafish by influencing oxidative stress induced apoptosis.

Zebrafish Maintenance
Zebrafish (Danio rerio) including wild type (WT), Tg (Huc: RFP) and Tg (Fli1: EGFP) were from the key laboratory of metabolism and molecular medicine, basic medical sciences, Fudan University. Zebrafish were raised and maintained in 14 h light/ 10 h dark cycle in a standard circulating laboratory environment (28.5°C). All zebrafish related procedures were approved by the Fudan University Shanghai Medical School Animal Care and Use Committee and were conducted in conformity with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals.

Real-Time PCR
Total RNA was extracted from zebrafish larvae at 7 days postfertilization (dpf) (n 30 per sample, three samples per group) using TRIZOL. Primescript RT reagent kit (TAKARA, RR037A, Japan) was used to synthesize cDNA. Real-time PCR was conducted using SYBR Premix Ex Taq II (Takara, RR 420D, Japan) in the StepONEPlus Real-time PCR system (Applied Biosystems, United States). The mRNA levels were normalized to the actin level. The primers for serpini1 are forward primer-TGACGGCTCAGATGACAGAC; reverse primer-TCCTGG TCAATGCGATCATA.

CoCl 2 Induced Hypoxia Injury
Cobalt (II) chloride hexahydrate (CoCl 2 6H 2 O, C8661, Sigma, United States) was dissolved in a 100 mM stock solution with sterilized water and stored at −80°C freezer. To find out the preferred concentration for hypoxic injury, different concentrations of CoCl 2 solution (0, 1, 10, 20, and 50 mM) were applied to zebrafish embryos every 12 h until 96 hpf. Survival rate was calculated under different concentration working solution every 12 h. WT and serpini1 −/− zebrafish were exposed to the preferred concentration for phenotype observation (survival rate≥50% at 96 hpf) according to the survival curve. Morphology phenotypes including hatch rate at 48hpf, survival rate at 96 hpf, teratogenic effects including pericardial edema, spine deformation, and abnormality of the head, eye (n 200 per group) were analyzed by visual assessment under a dissection microscope (Olympus, DP73, Japan).

Behavior
To study the spontaneous activities of WT and serpini1 −/− , a 40minute behavioral test (10-minute acclimation period with illumination and 30-minute continuous illumination period) was performed at 5 dpf. The detection and recording of zebrafish larvae locomotor activities were achieved by using Frontiers in Pharmacology | www.frontiersin.org March 2021 | Volume 12 | Article 632662 the tracking mode of ZebraLab software (ViewPoint Life Sciences, France). Videos of zebrafish larvae were taken at the rate of 25 fps (frames per second) and were pooled into 1-minute time bins. The detection threshold was set at 25, a level that allowed the software to accurately detect the movement of the larvae. The threshold for inactive (no locomotion activity) was set to 0 cm/s. General locomotor activities were recorded and analyzed according to the distance moved by the zebrafish larvae in the wells of the 24-well plate (n 96 per group).

Imaging Acquisition and Processing
For imaging, live zebrafish larvae anesthetized with 200 mg/L tricaine (Sigma-Aldrich, A5040, United States) were mounted in 3% methylcellulose (Sigma-Aldrich, M7027, United States). Brains were dissected under a dissection microscope (Olympus, DP73, Japan) at the time point of 4 dpf. Fixed samples were balanced and located in 80% glycerol before imaging. Images of neurons in the diencephalon area of Tg (Huc: RFP) and Tg (serpini1 −/− -HuC: RFP +/+ ), and vascular of Tg (Fli1: EGFP) and Tg (serpini1 −/− -Fli1: EGFP +/+ ) were acquired from a confocal microscope system (Olympus, FV3000, Japan). Using the Z stack strategy, top and bottom of the structure of interest were verified and a 2 um interval was selected to obtain the high-resolution images. Image procession and intensity measurements were done using ImageJ software (n 10 for each group).

Acridine Orange Staining
To compare the apoptotic cell numbers of each group, zebrafish larvae at 4 dpf were incubated in 10 μg/ml of AO (Thermo Fisher, A1301, United States) in the culture medium. Phenylthiourea (PTU) treatment can be applied to reduce pigmentation after 24 hpf. After 30 min of staining, larvae were washed three times in culture medium. Larvae were then transferred to 48-well plates and anesthetized with tricaine for imaging (Corning, catalog #4520, Corning, NY). Images were taken in GFP channels (n 25 per group).

Oxidative Stress Level Measurements
The content of lipid peroxidation (MDA) was measured by lipid peroxidation assay (Beyotime Institute of Biotechnology, Changsha, China S0131) according to the protocol from the manufacture. 100 larvae from each group were homogenized in ice-cold physiological saline by 1 ml springe and homogenates were centrifuged at 3,000 × g at 4°C for 15 min. Then the supernatants were collected and the assay was carried out immediately. For MDA detection, the supernatants from samples and standard substances were reacted with thiobarbituric acid contained in the kit, and the reaction products were measured spectrophotometrically at 532 nm. Finally, the MDA levels of the samples was calculated according to the standard curve. And the MDA level of samples were normalized against total protein levels determined by the BCA Protein Assay Kit (Beyotime Institute of Biotechnology). Antioxidant enzyme activity assay: glutathione peroxidase (GPX) (Glutathione peroxidase activity assay: Beyotime Institute of Biotechnology, Changsha, China S0056), catalase (CAT) (Catalase assay: Beyotime Institute of Biotechnology, Changsha, China S0051), and superoxide dismutase (SOD) (Superoxide dismutase assay: Beyotime Institute of Biotechnology, Changsha, China S0101) activity were measured using commercial kits according to the manufacturer's instructions. Samples from 100 larvae for each group were homogenized and collected. The determination of GPx activity was by detecting reduced NADPH in absorbance at 340 nm. For CAT, different concentrations of hydrogen peroxide (H 2 O 2 ) are catalyzed by peroxidase in the kit to produce the red product, N-4-antipyry l-3-chloro-5-sulfonate-p-benzoquinonemonoimine in absorbance at 340 nm. Then, a standard curve for H 2 O 2 concentration and absorption value was generated. Samples were treated with 250 mM hydrogen peroxide (H 2 O 2 ) for 1-5 min. And the remaining H 2 O 2 was measured at 520 nm. CAT could be calculated according to the standard curve. SOD inhibited the process of superoxide transforming WST-8 to a stable water-soluble WST-8 formazan, which could be tested in the optical density at 450 nm. So, the level of SOD in samples could be calculated according to the A450 absorption. Protein concentration for each sample was quantified by a protein assay kit (Beyotime Institute of Biotechnology). All enzyme activities were normalized as U (unit) per mg of protein.

Statistical Analysis
Statistical analyses were performed using GraphPad Prism 6 software. Data are expressed as mean ± SEM. Survival rates difference in WT zebrafish under different concentration of CoCl 2 at the same timepoint were evaluated by one-way ANOVA followed by Dunnett's multiple comparisons. Statistical analysis for serpini1 mRNA level, developmental defects, locomotion, neuron loss, and vessel malformation, and oxidative stress level in WT and serpini1 −/− group under CoCl 2 induced injury were assessed using unpaired t-test. Differences were considered as statistically significant as p-value < 0.05.

Establishment of Neuroserpin-Deficient Zebrafish Model
The serpini1 −/− mutant zebrafish were successfully generated via CRISPR/Cas9 technique targeting in the fourth exon of the serpini1 sequence ( Figure 1A). The DNA sequencing of target-specific PCR products confirmed that the serpini1 targeted allele caused a frameshift mutation with deletion of five bases ( Figure 1B). The mutated variant resulted in a truncated protein with 271 amino acids, which disrupted protein secondary structure ( Figure 1C). serpini1 mRNA expression was significantly reduced in the serpini1 −/− zebrafish compared with the WT ( Figure 1D). Neuroserpin protein could not be detected due to lack of anti-neuroserpin antibody for zebrafish in the present.

Increased Locomotor Impairment and More Neuron Loss in Neuroserpin Deficient Zebrafish under CoCl 2 Induced Hypoxic Injury
To further examine the behavioral change under CoCl 2 induced hypoxia injury, 5 dpf zebrafish from WT and serpini1 −/− were incubated in E3 medium in 24-plate and general locomotor movement was recorded. CoCl 2 induced hypoxic injury caused decreased locomotor activities in both groups as shown in Figures  3A,B Figure 3D).

Apoptosis and Oxidative Stress Level in Neuroserpin Deficient Zebrafish under CoCl 2 Induced Hypoxic Injury
To further evaluate the CoCl 2 induced injury, AO staining was used to identify the apoptotic cells. 10 mM CoCl 2 induced gross apoptotic damage in the brain, heart, and spinal cord ( Figure 5A). serpini1 −/− group displayed increased apoptosis compared to WT ( Figure 5A). To identify the mechanism, oxidative stress levels were assessed in both groups under 10 mM CoCl 2 induced injury. As shown in Figures 5B-E), serpini1 −/− group significantly increased peroxidation product MDA ( Figure 5B)

DISCUSSION
Research on protective therapy for cerebral ischemia revealed that neuroserpin is an essential therapeutic agent in the pathological condition and gives us a hint that it could protect against neonatal hypoxic ischemia. (Cinelli et al., 2001;Lebeurrier et al., 2005;Rodriguez Gonzalez et al., 2011b;Gelderblom et al., 2013). Due to its high efficacy in gene-manipulation, we establish a neuroserpin knockout zebrafish to study whether neuroserpin inserts protective role in CoCl 2 induced hypoxia injury in this study. Moderate to high concentration of CoCl 2 could cause embryotoxicity and development defects in zebrafish (Cai et al., 2012). In our study, CoCl 2 was exposed in the embryonic stage. Reduced hatchment, increased mortality and more teratogenic effects were observed in neuroserpin deficient zebrafish group. These developmental defects result in more severe behavioral impairment and neuronal loss in brain. Our study reflects the same phenotypes in neonatal hypoxia brain injury such as developmental retardation, spine deformation, microcephaly and microphthalmia, which causes long-term disability and mortality. Plasminogen activator (PA) system is critical for hypoxic-ischemic brain injury in newborns. Endogenous neuroserpin and tissue plasminogen activator (tPA) highly express at the early development stage in central nervous system and decline to a moderate level at the adult stage, which are involved in neuronal migration, axonal outgrowth, and synapse plasticity (Lee et al., 2017). Recent researches also detected that neuroserpin is expressed in the developing cortical plate and subplate, which is sensitive to neonatal ischemia hypoxia injury (Adorjan et al., 2019;Kondo et al., 2015). tPA induction was deleterious in hypoxic-ischemic Vannucci model of hypoxic-ischemic encephalopathy and in the adult ischemic stroke model (Adhami et al., 2008;Omouendze et al., 2013). So, serine-protease inhibitors could be a protective therapy strategy. It has been shown that plasminogen activator inhibitor-1(PAI-1) by intranasal delivery or ventricle injection protect against neonatal hypoxic-ischemia injury by reduced axonal degeneration and cortical neuron death (Yang et al., 2009;Yang et al., 2013;Yang and Kuan, 2015). As with PAI-1 and other serine-protease inhibitors, neuroserpin shares the essential conserved inhibitory center reactive center loop (RCL) to inhibit tPA activity. These evidences indicate that neuroserpin could be considered as a protective agent for neonatal hypoxic ischemia.
CoCl 2 provokes HIF-1α intracellular accumulation, which stimulates vascular endothelial growth factor (VEGF) secretion, angiogenesis, and hemorrhages. It has been verified that blood-brain barrier (BBB) is impaired in both ischemic stroke and neonatal hypoxic-ischemia encephalopathy (Abu Fanne et al., 2010;Tu et al., 2011). Excessive tPA is induced under hypoxia stress in microvascular endothelial cells  and elevated tPA could trigger the opening of BBB through plasmin-dependent pathway or matrix metallopeptidases activation (Su et al., 2008). Our previous data in an intracerebral hemorrhage mouse model identify that neuroserpin restores blood-brain barrier by re-establishing tight junction of vascular endothelial cells. In the present study, aggravated vascular malformation was observed in neuroserpin deficient group and these vessels are fragile and prone to rupture, which may due to dysfunction of vascular endothelial cells. It may highlight the complicated role of neuroserpin in endothelial cells. Neuroserpin has been found to express in non-neuronal cells such as the ependyma and epithelial cells of choroid plexus (Teesalu et al., 2004) and other tissues such as endocrine and immune systems (Fell et al., 2002;Hill et al., 2002). The impact of neuroserpin in non-neuronal cells and tissues may be involved in regulation of cell-cell adhesion or modification of extracellular matrix Lee et al., 2008;Gupta et al., 2017). More evidence is needed to be explored in the future study. Cobalt-induced hypoxia associated mechanisms of toxicity are not yet understood. Prolonged or high-dose treatment with CoCl 2 switches cell metabolism from aerobic respiration to anaerobic glycolysis (Regazzetti et al., 2009). This induction of oxidative stress can provoke cell death through apoptosis (Stewart et al., 2017). In our study, we found augmented apoptosis and enhance oxidative stress level in serpini1 −/− zebrafish under cobalt-induced hypoxia. The generation of excessive oxidative damage is believed to be involved in the pathogenesis of neonatal hypoxic-ischemic injury. A study of extracellular application of recombinant neuroserpin to primary hippocampal culture confirmed that it could rescue hydrogen peroxide (H 2 O 2 )-induced oxidative stress neurotoxicity (Cheng et al., 2017). Our finding was consistent with the result and it could be an essential mechanism for protection role of neuroserpin.
Although we confirmed that neuroserpin protects against CoCl 2 induced hypoxia injury in the zebrafish model in present study, it should be noted that this study uses a hypoxia-only model without ischemia. Human hypoxicischemic injury is usually more complicated with insufficient blood supply due to vascular occlusion or rupture. So, this artificial hypoxia-only system is not ideal for further complex conclusion. In addition, there are many other effects of hypoxia-ischemia that are not mimicked by CoCl 2 treatments. Hypoxia downregulates HIF-1α during Statistical analyses were performed with Student's t-test, ***p < 0.001 compared with WT, ns: no significant. (C) The level of GSH-Px in WT and serpini1 −/− under 10 mM CoCl 2 induced injury. The data are presented as the mean ± SEM, n 100 per group. Three trails for the experiment. Statistical analyses were performed with Student's t-test, **p < 0.01 compared with WT, ns: no significant. (D) The level of CAT in WT and serpini1 −/− under 10 mM CoCl 2 induced injury. The data are presented as the mean ± SEM, n 100 per group. Three trails for the experiment. Statistical analyses were performed with Student's t-test, ***p < 0.001 compared with WT, ns: no significant. (E) The level of SOD in WT and serpini1 −/− under 10 mM CoCl 2 induced injury. The data are presented as the mean ± SEM, n 100 per group. Three trails for the experiment. Statistical analyses were performed with Student's t-test, ***p < 0.001 compared with WT, ns: no significant.
Frontiers in Pharmacology | www.frontiersin.org March 2021 | Volume 12 | Article 632662 8 prolonged oxygen reduction. The decrease in HIF-1α protein levels could be mediated by the increase in the levels of the prolyl hydroxylases (Berra et al., 2003) or by the induction of antisense HIF-1α (Rossignol et al., 2002), which can act in a negative manner. However, the negative feedback regulation is not established when HIF-1α is stabilized by CoCl 2 , which affect severely the proliferative capacity of Hela cell (Triantafyllou et al., 2006). So, the phenotypes we observed in our study maybe different from the pathological change in hypoxic-ischemic injury. Furthermore, CoCl 2 have alternative and more complex mechanisms of cobalt activity. High concentration of CoCl 2 used to induce hypoxia has been shown to influence hematological cells of fish (Saeedi Saravi et al., 2009). Also, CoCl 2 is known to cause mitochondrial damage (Mansfield et al., 2005). We may not exclude the off-target effects of CoCl 2 in our study. These problems deserve further investigation and exploration.
To the best of our knowledge, this is the first report that demonstrates the protection of neuroserpin in zebrafish under CoCl 2 induced hypoxia injury. This study contributes to a better understanding of the protective role of neuroserpin in hypoxic injury.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The animal study was reviewed and approved by the Fudan University Shanghai Medical School Animal Care and Use Committee.

AUTHOR CONTRIBUTIONS
DZ maintained the transgenic zebrafish; SH designed and constructed the transgenic zebrafish model; SH performed the behaviors and cofocal images; DZ analyzed the data with the assistance of SH; LW, QD, and XW carried out experimental designs and contributed toward the manuscript writing, together with SH.