Establishment of a lipid metabolism disorder model in ApoEb mutant zebra�sh

Background and aims: ApoEb is a zebra�sh homologous to mammalian ApoE, whose de�ciency would lead to lipid metabolism disorders (LMDs) like atherosclerosis. We attempted to knock out the zebra�sh ApoEb, then establish a zebra�sh model with LMD. Methods: ApoEb was knocked out using CRISPR/Cas9 system, and the accumulation of lipids were con�rmed by Oil Red O staining, confocal imaging, and lipid measurements. The lipid-lowering effects of simvastatin (SIM), ezetimibe (EZE) and Xuezhikang (XZK), an extract derived from red yeast rice, were evaluated through in vivo imaging in zebra�sh larvae. Results: In ApoEb mutant, signi�cant vascular lipid deposition occurred, and lipid measurement performed in whole-body homogenate of larvae and adult plasma showed signi�cantly increased lipid levels. SIM, EZE and XZK apparently relieved the hyperlipidemia in ApoEb mutants, and XZK had a signi�cant inhibitory effect on the recruitment of neutrophils and macrophages. Conclusions: In this study, a LMD model has been established in ApoEb mutant zebra�sh. We suggest that this versatile model could be applied in studying hypercholesterolemia and related vascular pathology in the context of early atherosclerosis as well as the physiological function of ApoE.


Introduction
Apolipoprotein E (ApoE) is a kind of apolipoprotein mainly expressed in mammal livers [1].It is crucial in lipid metabolism, especially in the metabolism of very low-density lipoproteins (VLDLs), cholesterol and high-density lipoproteins (HDLs) [2].The relationship between ApoE and risk of cardiovascular disease (CVD) such as hyperlipidemia and ischemic heart disease has been discovered through numerous epidemiological observational and genetics studies as well as interventional clinical trials [3].
Approximately half of plasma ApoE is associated with triglyceride-rich lipoproteins, where ApoE serves as the main ligand for the low-density lipoprotein receptor (LDLR) and the LDLR related protein (LRP) [3].
Lipid-lowering drugs targeting various pathways have successfully helped human manage hyperlipidemia.Simvastatin (SIM), one of the principal drugs targeting Ldlr expression, is able to e ciently lower LDL-C and reduce the incidence of CVDs in human.Xuezhikang (XZK) is a commonly used traditional Chinese medicine extracted from red yeast rice [7].It has shown powerful lipid-lowering effect in clinic both in China and abroad [7,8].Despite its common use, the pharmacological and toxicological effects of XZK are still not totally clari ed.
Although various lipid-lowering drugs were invented, familial hypercholesterolemia (FH) patients are not sensitive to them, hence new lipid-lowering drugs remain to be developed, and e cient animal models are essential.But the procedures of the drug screening and hyperlipidemia-related research using ApoE -/- mice might be expensive, time-consuming, and labor-intensive.As a novel model animal, zebra sh have the advantages like body transparency, large progeny numbers and low maintenance costs.What's more, genes between human and zebra sh have a relative high conservation, including the genes encoding apolipoprotein family members like the ApoA family, ApoB, ApoC2 and Ldlr [9][10][11].In order to study lipid metabolism disorders (LMDs) like FH in depth, zebra sh lines with corresponding gene mutations have been established in recent years.The known main causes of FH are pathogenic variants in Ldlr, ApoB (the gene encodes the apolipoprotein B) and Pcsk9, as well as ApoE [12].Among them, Ldlr [11], ApoB [13] and Pcsk9 [14] mutant zebra sh lines have been established recently, yet the pathophysiology of FH remains elusive.To provide a different perspective for the study of the occurrence and development of FH and related CVDs, in this study, we used the CRISPR/Cas9 system to interrupt the expression of the zebra sh ApoEb gene, following with a short high-fat diet (HFD) feeding, established a novel zebra sh LMD model.And using this disease model, we have made a preliminary study on the mechanism of XZK in preventing LMDs.Our study paved another way for future pathophysiology studies of LMDs like hyperlipidemia and FH, as well as pharmacological and toxicological studies of lipid-lowering drugs.
Zebra sh larvae were fed twice a day, starting at the 5th day post fertilization (dpf), with either a normal diet (ND) (paramecium, from China Zebra sh Resource Center) as described [16] or a high-fat diet (HFD) (Yuanye, Shanghai, China), which contains 4% weight per weight (w/w) cholesterol and 20% w/w triglycerides.For drug treatment experiments, simvastatin (Sigma-Aldrich, Munich, Germany), ezetimibe (Selleckchem, Shanghai, China) and XZK (Beijing Peking University WBL Biotech Co., Ltd., Beijing, China) were dissolved in dimethyl sulphoxide (DMSO) (Sigma-Aldrich, Munich, Germany) and thoroughly mixed with HFD.According to the experiment, drugs were dissolved into different concentrations when used.The nal DMSO concentration was ≤0.5% volume/volume (v/v) in all treatment groups.Adult zebra sh of the Tuebingen strain were maintained in the same environment and fed brine shrimp (Bohai, Binzhou, China) twice a day as described [16].The HFD feed (containing 5% w/w cholesterol and 24% w/w triglycerides) for adult sh were bought from Trophic Corporation (Nantong, China).All animal studies were approved by the Experimental Animal Ethics Committee of Shanghai Changzheng Hospital.

ApoEb knockout using CRIPR/Cas9 system
The gRNA plasmid and Cas9 mRNA plasmid used in this study were kind gifts from Dr. Jing-Wei Xiong from Peking University.In vitro transcription of gRNAs and Cas9 mRNA were performed using T7 RNA polymerase (Ambion, TX, USA) according to the manufacturer's instructions and puri ed by phenol chloroform extraction.Two ApoEb genomic target sequences were selected: 5'-gcccagatgggaggagatggTGG-3' and 5'-gatgacgtgaagaaccgtgtCGG-3' in which the last 3 nt were the protospacer adjacent motif (PAM) required for CRISPR/Cas9 function.One hundred pg of gRNAs and 200 pg of Cas9 mRNA were co-injected into 1 to 2-cell stage embryos.For genotyping, the caudal ns of adult sh were partly cut down, and whole embryos were homogenized.The genomic DNA was extracted using the TransDirect Animal Tissue PCR kit (Transgen, Beijing, China).The genomic DNA fragment containing the target site was ampli ed using Ex Taq DNA polymerase (Takara, Tokyo, Japan).The primers used for PCR ampli cation of ApoEb cDNA anking exon 3 to exon 4 were listed in Supplementary Table 1.All primers and the Sanger sequencing of the PCR ampli cation products to con rm the genotypes in detail were provided by Biosune Ltd, Shanghai, China.

2.3
Quantitative RT-PCR (rt-qPCR) Total RNA was isolated from 15 larvae using Trizol reagent (Thermo Fisher, MA, USA), and reverse transcribed using FastQuant RT Kit (Tiangen, Beijing, China).Quantitative RT-PCR assays were performed using SYBR Green Realtime PCR Master Mix (Toyobo, Osaka, Japan).The primers used for mRNA detection were all listed in Supplementary Table 1.

Plasma lipase activity assay
Adult zebra sh were anesthetized in an ice-water mixture, and then tail amputation was performed.Blood collected from 5 adults were pooled and diluted 1:200 in natural saline and plasma was obtained by centrifugation at 3,000 g for 10 minutes.Two hundred μl of the diluted plasma was used to measure lipase activity with a Lipase activity assay kit (Solarbio, Beijing, China), following the manufacturer's manual.The assay buffer was used as a negative control.The reaction was conducted at 37 ℃.

Oil O Red staining
For Oil Red O (ORO) staining, the lipid in zebra sh body were stained with Oil Red O dye (Sigma-Aldrich, Munich, Germany).The zebra sh larvae were xed in 4% paraformaldehyde (Sangon, Shanghai, China) for 12 h, washed twice with phosphate buffer saline (PBS) (Sangon, Shanghai, China), and then incubated in 0.3% ORO solution for 3 h.Stained larvae were rinsed with PBS before imaging.

Imaging and counting
For in vivo visualization of vascular lipid deposits, the HFD was supplemented with 1 μg/g of a uorescent cholesteryl ester analog CHOLESTERYL BODIPY 576/589 C11 (Invitrogen, Carlsbad, USA), as previously described [18].After fed with this uorescently labeled diet for a speci c period according to the procedures of different experiments, Tg ( i1:EGFP) larvae were mounted in low melting point agarose (Sigma-Aldrich) in Danieu's solution with 0.04% Tricaine to keep them immobilized.Then, the larvae were scanned on a Leica TCS SP5 confocal microscope with 1.5 μm step size, 1,024 × 1,024 pixel at 400 Hz.
The in vivo imaging of angiogenesis of Tg ( i1: EGFP) sh line was performed using either the confocal microscope with the same parameters, or a Leica M205 uorescent microscope, which was also used to image the Tg ( k1:dsRed; lyz: EGFP) sh line.The amount of the neutrophils in 10 consecutive segments of the trunk was counted using Image-Pro Plus 6.0 software (Media Cybernetics, MD, USA).The heart beating images and movies of the Tg (cmcl2: mCherry) sh line were also taken using the Leica M205 uorescent microscope.The consequent measurements on the images were performed using Image-Pro Plus 6.0 software.Bright eld imaging for ORO staining and neutral red staining was taken using a stereomicroscope (Zeiss, Oberkochen, Germany).The integrated optical density (IOD) value of 5 consecutive segments of the trunk in all ORO staining or neutral red staining images were measured using Image-Pro Plus 6.0 software to relatively quantify the vascular lipid accumulation or recruitment of the macrophages [19].All the images are lateral views, anterior to the left, and dorsal is up unless speci cally noted.

Lipid level measurement
Fifteen larvae in the same group were euthanized in icy PBS and homogenized by ultrasonic fragmentation, supernatants were collected and used as 1 homogenate sample.Lipid levels in diluted pooled adult plasma or larvae homogenate were measured using total cholesterol (TC), triglyceride (TG), low-density lipoprotein cholesterol (LDL-C), very-low-density lipoprotein (VLDL) and high-density lipoprotein cholesterol (HDL-C) quanti cation kits (Jiancheng, Nanjing, China) according to the manufacturer's protocols, respectively.

Neutral red staining
To evaluate recruitment of the macrophages, 15 live larvae from each group were respectively cultured in a 12-well plate containing 2.5 mg/L of neutral red dye solution (Solarbio, Beijing, China), which can aggregate in the macrophages, forming red dots due to the recruitment of macrophages.After incubation for 8 h, the larvae were anesthetized and imaged under a microscope.

Statistical analysis
Normality was evaluated by Shapiro-Wilk test.The data t normal distribution were presented as the mean ± standard error of mean (SEM), while boxplots were used to present skew distribution data.GraphPad Prism software was used for data analysis.For the normal distribution data, Levene's test of homogeneity of variance was further performed.When the data tted the homogeneity of variance, two-tailed Student's t test or one-way ANOVA was applied.For the data that did not t the homogeneity of variance, Mann-Whitney test, Brown-Forsythe ANOVA test, or Kolmogorov-Smirnov test was performed.Mann-Whitney test or Kruskal-Wallis test was performed for skew distribution data.All Pvalues were two-sided and P < 0.05 was considered statistically signi cant.

Generation of the ApoEb mutant zebra sh
In zebra sh, the ApoE gene has two paralogs -ApoEa (NCBI gene ID: 553587) and ApoEb (NCBI gene ID: 30314) [20][21][22].Despite of its less than 30% identity with mouse or human ApoE (Fig. S1A), ApoEb, but not ApoEa, contains a conserved sequences encoding amino acid sequence similar to the lipoprotein receptor-binding region (LRR) of human ApoE [23], and its protein sequences satisfy the common structural features depicted for potential amphipathic helices characteristic of plasma apolipoprotein [24].Previous study also has found that ApoEa expression gradually reduced during sh development, while ApoEb increased [25].All these indicated that with conserved functional motifs like mammalian ApoE (Fig. S1B), ApoEb may play an important role in lipid metabolism of zebra sh.In order to effectively interrupt the expression of ApoEb gene using CRISPR/Cas9, we designed 2 gRNAs targeting its exon 3 and 4 respectively (Fig. 1A).These could cause a 378 nt deletion in ApoEb mRNA, and consequently lead to a 126 amino acid (aa) loss and a 1 aa mistranslation in the corresponding ApoEb protein (Fig. 1B).The truncated protein translated might lost parts of the apolipoprotein domain and the LRR required for ApoEb function (Fig. 1C).Hence, the mutant might only have loss-of-function ApoEb protein.The gRNAs and Cas9 mRNAs were synthesized and injected immediately into the cytoplasm of the fertilized egg cells as soon as possible, to obtain mutant founders (F0).Sanger sequencing of the F0 genomic DNA showed there were 542 nucleotide deletions (Δ542) occurred in several mutants, as shown in Fig. 1D.In order to carry this mutation to other zebra sh lines with various transgenic backgrounds, successfully mutagenized F0 animals were raised to adulthood and crossed to other sh lines to obtain the F1 generations.After the sexual maturity, the F1 male were crossed with F1 female to obtain F2 embryos carrying the same mutation.Through 2 generations, homozygous ApoEb mutated Tg ( 1:EGFP), Tg ( k1:dsRed; lyz: EGFP) and Tg (cmcl2:mCherry) lines were then screened out.Rt-qPCR detection for the normal ApoEb mRNA with the integrated sequence showed ApoEb mRNA level was almost undetectable in 5 dpf ApoEb -/-mutants, as well as in the livers of 2-month-old adults (Fig. 1E).This part of results indicated that there existed not normally functionable ApoEb protein in homozygous.

Hyperlipidemia in adult ApoEb mutants
No apparent physical differences were found between 3-month-old male wild-type siblings and ApoEb mutant zebra sh (Fig. 2A-B).To test whether ApoEb de ciency resulted in hypertriglyceridemia, we drew blood from adult male zebra sh fed with an 8-week normal diet (ND) or high-fat diet (HFD) to measure the lipid pro le in plasma.The results showed that ND fed ApoEb mutants had higher plasma TC, LDL-C and VLDL level.Although there was no statistical difference, plasma TG level in ApoEb mutant tended to increase.After HFD feeding, TC, TG, LDL-C and VLDL levels in ApoEb mutant's plasma signi cantly increased, while HDL-C remained unchanged compared with siblings (Fig. 2C).Meanwhile, plasma lipase activity was signi cantly decreased in normally fed ApoEb mutants compared with siblings (Fig. 2D).Due to ApoEb probably was a ligand for zebra sh LDLR, which is required for extracellular cholesterol uptake and regulation of the Srebp pathways in the liver to regulate plasma cholesterol [25], we next tested whether ApoEb knockout had in uence on the Srebp pathways, which was also a target of many lipid-lowering drugs, in the livers of normally fed 2-month-old ApoEb mutants (Fig. 2E).The detection of the expression level of Hmgcr and Ldlr, two target genes of Srebp-2, showed an interesting upregulation without that of Srebp-2 in ApoEb mutant.An increase in mRNA expression of Srebp-1 and Fasn, one of its target genes, could be observed in ApoEb mutant.

3.3
Early angiogenesis defects in ApoEb mutant larvae Three dpf ApoEb mutant larvae showed no signi cant defects in development and yolk utilization and growth (Fig. 3A).The body length, yolk area and yolk extension did not show differences between 3 dpf siblings and ApoEb mutants (Fig. 3B).However, in about 40% of ApoEb mutants at 3 dpf, cerebral hemorrhage could be observed, mainly in midbrain and hindbrain (Fig. 3C-D, indicated by red triangles).
Confocal imaging of 3 dpf ApoEb mutant Tg ( i1: EGFP) larvae showed the integrity of the cerebral vascular network was signi cantly interfered.For instance, in Fig. 3E, the red triangles indicated the narrowed posterior mesencephalic central arteries (PMCtAs), while the white triangles pointed out the absence of the metencephalic arteries (MtAs) in ApoEb mutants.We then observed the development of the mutant's intersegmental vessels (ISVs).Fig. 3F showed the development of wild-type and ApoEb mutated larvae ISVs within 48 hpf, with signi cant narrowed ISVs at 30, 36 and 48 hpf in the mutants.However, within 6 dpf, the time and length of ISV spreading were not statistically different (Fig. 3F), though the diameter of ISVs shortened in 48 hpf ApoEb mutants.These results demonstrated that angiogenesis defects occurred at early developmental stage of ApoEb mutants.

LMD occurred in ApoEb mutant fed with HFD
We performed ORO staining, confocal imaging, and lipid measurements to test whether the ApoEb mutant developed LMD.We found that both the yolk sac and blood vessels of ApoEb mutants had lipid content (Supplementary Fig. S2A).Lipid deposition gradually decreased with development, and existed until 6 dpf, indicating that the lipids derived from the yolk sac had been consumed completely at this stage (Supplementary Fig. S2A).Compared with ND feeding for 1 day, 7 dpf ApoEb mutant larvae fed with ND for 2 days showed slightly ORO staining in the blood vessels (Supplementary Fig. S2B), indicating moderately elevated neutral lipid levels in the circulation.The IOD value, which could be used to relatively quantify the intensity of intravascular ORO staining con rmed these ndings (Supplementary Fig. S2C).When fed for 5 days, ORO staining of 10 dpf larvae fed with ND showed signi cant intravascular lipid deposition in ApoEb mutants (Supplementary Fig. S2D).However, uorescently labeled cholesterol esters (576/589-CE, the numbers indicate excitation/emission wavelengths) accumulated in vessels did not show signi cant change in 5-day ND fed 10 dpf ApoEb mutants (Supplementary Fig. S2D-E).Interestingly, the angiogenesis defects in ISVs at early stage seemed to be corrected at this stage.In other words, the vasculature of 7 dpf mutants was normal (Supplementary Fig. S2D).We then tested whether a 2-day HFD feeding was able to induce LMD in ApoEb mutant larvae, as previously reported in wild-type [26].The results showed that the phenotype in ApoEb mutants fed with a 2-day HFD was more robust (Supplementary Fig. S2F-G).In 10 dpf ApoEb mutants fed with a 5-day HFD, much more intravascular lipid accumulation also could be observed through ORO staining (Fig. 4A-B).A HFD feeding, even for just 2 days, might lead to signi cantly increased 576/589-CE deposition in 7 dpf ApoEb mutant larvae (Fig. 4A).The calculation of average lipid deposits per segment also told this (Fig. 4C).
To gure out the primary effects of ApoEb knockout on the Srebp pathways, we performed detections in 5 dpf fasted sibling and ApoEb mutant larvae (Supplementary Fig. S2H), which showed an interesting upregulation of Ldlr.Though a 2-day ND did not alter the expressions of the Srebp pathways much, a 2day HFD feeding signi cantly activated both Srebp pathways in ApoEb mutant (Supplementary Fig. S2I).This was consistent with the expression of Srebp-1 and -2 pathways in 10 dpf ApoEb mutant fed with a 5day diet, which showed that ND might induce the upregulation of Hmgcr, while HFD upregulated the expressions of all these genes (Fig. 4D).So as the results from the livers of 8-week HFD feeding ApoEb mutants (Supplementary Fig. S2J).Furthermore, lipid measurement of whole-body homogenates of 7 dpf larvae showed signi cant higher TC, TG, LDL-C and HDL-C levels in ApoEb mutants fed with HFD for 2 days, while a 2-day ND also increased the cholesterol levels in ApoEb mutants (Supplementary Fig. S2K).However, a continuous 5-day ND did not elevate the lipid levels (Supplementary Fig. S2L), while a 5day HFD feeding increased the levels of TC, TG, VLDL and LDL-C in whole-body homogenate of 10 dpf ApoEb mutants, with unchanged HDL-C level (Fig. 4E).
We also investigated the effect of ApoEb de ciency on systemic in ammation.We found that the expression of IL-1β and TNF-α were signi cantly upregulated in 5-day HFD fed 10 dpf ApoEb mutants, while IL-6 expression remained unchanged (Supplementary Fig. S2M).Besides, we discovered that acute lipid burden like a 2-day HFD feeding might result in impaired cardiac function and arrhythmia in ApoEb mutant Tg (cmcl2: mCherry) larvae.Supplementary Fig. S2N showed that after 2-day HFD feeding, ApoEb mutants manifested higher heart rates, lower ventricular fractional shortening (FS), and smaller heart area as well as shorter venous sinus (SV) -artery ball (BA) distance which indicated restricted heart beats.Supplementary Movie 1&2 demonstrated signi cant diastolic prolongation and premature beats in 2-day HFD fed 7 dpf ApoEb mutant, compared to the sibling.The results above suggested that HFD feeding might lead to serious LMD in ApoEb mutant zebra sh.
3.5 Simvastatin, ezetimibe and Xuezhikang signi cantly relieved LMD in ApoEb mutant To evaluate whether ApoEb mutant zebra sh subjected to a short-term (2 days) HFD feeding could be useful for drug screening experiments, we need to nd proper positive controls.We tested the effect of simvastatin (SIM), an Hmgcr inhibitor as well as a widely applied lipid-lowering drug, on intravascular lipid accumulation.Firstly, 5 dpf larvae were fed with a 2-day HFD, supplemented with various concentrations (0, 0.1, 0.2, 0.3, 0.4, 0.5, and 1.0 μg/mL) of SIM at the last 24 hours.The survival test of 24-hour SIM treatment showed that when treated with 0.4 μg/mL SIM for 12 hours, the survival rate maintained at a high level (Supplementary Fig. S3A).However, a 24-hour SIM treatment has relatively higher fatality rate (Supplementary Fig. S3A), which was consistent to a previous report [27], as well as signi cant teratogenic effect to zebra sh larvae.When treated for 24 hours, the larvae demonstrated doubled rate of slow touch-evoked escape (a parameter used to measure the body movement ability and the ability to respond to external stimuli of zebra sh larvae, Supplementary Fig. S3B) as well as elevated rate of abnormal morphology (Supplementary Fig. S3C-D, red triangles indicate the bent tails).For longer drug treatment experiments, according to former studies [27], we tested ezetimibe (EZE), which was much safer than SIM for larvae when treated for more than 2 days.As a result, larvae treated with 5 or 10 mM EZE within 5 days showed higher survival rates (Supplementary Fig. S3E-F).We also found that larvae had an acceptable 12-hour survival rate when treated with Xuezhikang (XZK) ≤120 μg/mL (Supplementary Fig. S3E-F), and the optimal concentration (30 μg/mL) of XZK for 5-day treatment (Supplementary Fig. S3E).
Having this con rmed, the lipid lowering effect of SIM, EZE, and XZK had been tested.ORO staining demonstrated that a 12-hour treatment with 30 μg/mL XZK was effective to reduce intravascular lipid accumulation in 7 dpf ApoEb mutants fed with a 2-day HFD, while 10 mM ezetimibe or 0.4 μg/mL SIM treatment was more potent (Fig. 5A-B).We also examined the intravascular accumulation of lipid deposits directly through confocal imaging (Fig. 5C).Consistently, the vascular deposition of 576/589-CE in the XZK treatment group was signi cantly reduced, and further reduced in the EZE and SIM treatment groups (Fig. 5C-D).In 5-day treatment experiments, the lipid measurements of whole-body homogenates of 10 dpf ApoEb mutant larvae fed with 5-day HFD suggested that compared to the DMSO group (treated with 0.5% DMSO), both 10 mM EZE and 30 μg/mL XZK signi cantly reduced lipid levels in tissue, with an elevated HDL-C level (Fig. 5E).We also detected the effects of EZE or XZK treatments on the expressions of Srebp-1 and -2 pathways in 10 dpf ApoEb mutant larvae fed with 5-day HFD (Fig. 5F).
The results implied that EZE might activates both pathways while XZK mainly upregulated the Srebp-2 pathway, with increased Hmgcr and Ldlr expression (Fig. 5F).
The above results proved that the ApoEb mutant LMD model could be utilized in drug screening procedures.SIM and EZE could be applied as positive controls in 12-hour and 5-day treatments, respectively.Through a rapid relatively quanti cation method, we also found that XZK was able to relief LMD by lowering general TC, TG, VLDL and LDL-C levels while increasing HDL-C level in zebra sh.

3.6
The effect of XZK on LMDs ORO staining for XZK treated Tg ( k1:dsRed; lyz:EGFP) larvae carrying the ApoEb mutation showed that along with the increase of XZK concentrations, the intravascular lipid deposition gradually decreased (Supplementary Fig. S4A-B, Fig. 6A showed the representative images), which could be con rmed by the quanti cation of intravascular staining intensity (Fig. 6B).This lipid-lowering effect might play an important role in the preventive effect of XZK on hyperlipidemia.In order to prove that the cells with green uorescence in this sh line were neutrophils rather than macrophages, we conducted immuno uorescence experiments to verify (Fig. 6C).As showed in Fig. 6C, EGFP (green) driven by the Lyz promoter had obvious colocalization (yellow) with neutrophil speci c marker Mpo (red), but no colocalization with macrophage speci c marker MFAP4.This indicated that the cells expressing EGFP in the sh line were indeed neutrophils.On the other hand, uorescence photography revealed that after HFD feeding for 2 days, EGFP-expressing neutrophils gathered in large numbers around the trunk large blood vessels (Fig. 6D).While in XZK and SIM treatment groups, neutrophils were mostly concentrated in the intestine rather than blood vessels (Supplementary Fig. S4C-D, Fig. 6D showed the representative images), and as the drug concentration increased, neutrophils accumulated in blood vessels decreased (Fig. 6E).These results indicated that XZK has an inhibitory effect on the aggregation of neutrophils, which might be another mechanism of XZK's LMD-preventing effect.To evaluate whether XZK treatment affected macrophage recruitment, Neutral Red staining was performed in 10 dpf ApoEb mutants fed with 5-day HFD supplemented with EZE (10 μM) and XZK (30 μg/mL).It was observed that both of them effectively reduced the recruitment of macrophages (Fig. 6F-G).Besides, the mRNA measurements of IL-1β, IL-6 and TNF-α in the pooled whole-body homogenates of 10 dpf larvae fed a 5-day HFD supplemented with 0.5% DMSO (DMSO), XZK or EZE showed that EZE signi cantly induced the downregulation of IL-1β and TNF-α.XZK treatment reduced mRNA expression of IL-1β either but did not affect that of TNF-α.IL-6 was not affected by both of XZK and EZE (Fig. 6H).

Discussion
At present, both ApoE and Ldlr knockout mice supplemented with HFD have become routine animal models of LMDs such as atherosclerosis and hyperlipidemia.ApoE knockout mouse demonstrate a marked increase in blood lipids, a large amount of lipid deposits in the aortic wall -as a result, developed to atherosclerosis in a short period of time [28].Even though the phonotype of ApoE and Ldlr mutant mice are similar to each other, they differ in their dietary needs for developing atherosclerosis, prominent lipoproteins and effect of hepatic lipase de ciency [29].The advantage of the ApoE -/-mice model is that complex vascular lesions readily develop in animals fed the normal low-fat rodent chow, and atherogenesis can be notably accelerated by the feeding of a high-fat, high-cholesterol Western type diet [30], which is a necessary condition for establishing an atherosclerosis model in Ldlr knockout mice.In addition, apolipoprotein ApoE was reported to have several other roles in the pathological process of atherosclerosis besides directly participating in lipid metabolism, such as the regulation of monocyte activation [31] and smooth muscle cell (SMC) migration and proliferation [32], nitric oxide synthase (NOS) mediated platelet aggregation [33], and so on.In brief, different factors may show different functions in ApoE and Ldlr knockout mice, respectively.In other words, ApoE and LDLR mutants cannot be substituted for each other.
In a recent review, Vedder et al wrote, "The development of ApoE -/-zebra sh would create a new tool for atherosclerosis research that could be compared with mouse models" [34].In this study, we aimed to establish a model of LMD in zebra sh with loss-of-function of ApoEb, which is encoded by one of the 2 subfunctionalized paralogs of ApoE during the evolution of zebra sh [22].Zebra sh ApoEb gene expressed in a species-speci c manner [22,35,20], and its speci c function and mechanism research is still lacking.Several studies had chartered zebra sh ApoEb protein has similar functional domains to mammalian ApoE protein and their high homology [24,23].Hereinafter in this study, we supposed zebra sh ApoEb functioned as the ApoE gene in mammals, and the ApoEb mutant sh line adds to the set of zebra sh models of lipid abnormalities including ApoC2 mutant zebra sh and others [18].Although these mutated zebra sh lines have been established, there are various physiological and pathophysiological differences between ApoEb mutants and others.In the present study, we discovered that ApoEb de ciency in zebra sh led to cerebral hemorrhage, defected cerebral vascular network, and ISV stenosis.Under lipid burden, ApoEb mutants manifested early cardiac dysfunction and signi cant vascular lipid deposits.Previous studies reported a WT larvae zebra sh hypercholesterolemia model with accumulation of vascular lipid deposits whose establishment required a 2-week HCD feeding [36], and Ldlr loss-of-function mutant zebra sh model with a similar phenotype which needed HCD feeding for 5 days [11].However, the time consumption to induce robust vascular lipid accumulation by HFD feeding in our new ApoEb mutant could be reduced to 48 hours, according to speci c research purpose, for instance, the rapid reaction of neutrophils in acute lipid burden.What's more, the establishment of our ApoEb mutant zebra sh model is not only valuable for research on LMDs, but also can be used for nervous system development and disease research like Alzheimer's disease [37] as well as research on the innate immune system like bacterial antigen immunity [38], etc.
In our results, fasting 5 dpf ApoEb mutants showed increased Ldlr expression which indicated the activation of Srebp-2 pathway and decreased expression of Fasn which indicated the suppression of the Srebp-1 pathway.Through the measurements at various time points of feeding, we con rmed the activation of both Srebp-1 and -2 pathways in HFD feeding ApoEb mutants, and consistently, the upregulation of Ldlr.It is well-known that Ldlr is the mediator of hepatic cholesterol uptake [39], and predominantly regulated by SREBP-2 [40].When hepatocytes lack cholesterol relative to that required for their physiological needs, the upregulation of Ldlr would be mediated via a negative feedback mechanism that is tightly controlled by SREBP-2 [41].On the other hand, FASN is a multifunctional protein directly correlated with the fatty acids synthesis, playing a crucial role in de novo lipogenesis in mammals [42].Remarkably, the downregulation of Fasn in ApoEb mutant we found in the current study was in contrast to its upregulation in Ldlr mutant [11].However, other Srebp genes, including Hmgcr and Pcsk9, were not signi cantly changed in fasting larvae.It seems that different target genes are regulated by Srebp-1 or Srebp-2 in different manners, and the underlying mechanisms, which might provide new insights and suggest new approaches for differential targeting of genes regulating cholesterol homeostasis, are worth further exploring.
In the 12-hour drug treatment experiments, we used SIM to test if the established model could be applied in lipid-lowering drug screening.SIM prevented HFD-induced accumulation of vascular lipid deposits effectively, elevated the plasma HDL-C level, indicating a promoted reverse transportation of cholesterol back to the liver.These results were in contrast to a previous report that SIM did not show explicit lipidlowering effect in WT zebra sh [27].This discrepancy may be caused by the underlying coupling effect of Srebp pathway activation induced by ApoEb dysfunction and HMGCR inhibition caused by SIM treatment and partly by the different absorption of SIM due to different methods of administration.Notably, previous work reported the skeletal muscle toxicity of SIM to zebra sh larvae [43], as well as human patients [44].Consistently, our results con rmed in vivo the myotoxicity of SIM characterized by decreased body movement, slow response to external stimuli and increased abnormal morphology (bent tails and scoliosis).In the 5-day drug treatment experiments, EZE was employed as a safe and e cient positive control to be compared to XZK.Our results showed EZE signi cantly activated the Srebp-2 pathway, upregulated the expression of Srebp-2 and its downstream genes, Hmgcr, Pcsk9 and Ldlr, consistent to previous studies in mice [45], and XZK demonstrated similar effects on the Srebp-2 pathway.Which was inconsistent to mice studies [46], we found EZE also upregulated Srebp-1 expression, without affecting its downstream, Fasn.In the present study, it was hard to con rm whether and where the fatty acid synthesis was activated, which is worthy of further exploration.Therefore, we supposed that the ApoEb mutant larvae can be used for mechanistic studies, as well as for initial genetic or drug screening to identify new therapeutic targets, as well as their potential hazards.
In the early stages of atherosclerosis, neutrophils are recruited in large quantities under the vascular endothelium, releasing a variety of cytokines and in ammatory mediators, causing local in ammatory reactions.Neutrophils then induce monocytes and macrophages to enter and engulf oxidized LDL (oxLDL), forming foam cells, promoting the formation of lipid streaks, further activating macrophages, and aggravating atherosclerosis.A number of clinical studies have found that in advanced atherosclerotic plaques, the number and activity of neutrophils are positively correlated with the incidence of acute coronary syndrome (ACS), unstable plaque rupture, thrombosis and other cardiovascular events [47,48].Previous study reported that traditional Chinese medicines such as mustard seeds and their extracts can inhibit the recruitment of neutrophils and exert their anti-atherosclerotic effects [49].In our experiments, we have observed signi cant suppression in intravascular neutrophil recruitment after XZK treatment, which indicated that XZK may also have a similar inhibitory effect on neutrophils.In 5-day EZE treatment, reduced intravascular macrophage recruitment could be observed, which was consist to mice models [50].Like EZE, XZK also demonstrated an inhibitory effect on macrophage recruitment.What's more, both EZE and XZK signi cantly downregulated IL-1β expression, while EZE alone reduced TNF-α expression, either.
There were several limitations in our study.Firstly, adult zebra sh has a long experimental period and is not suitable for large-scale experiments like drug screening.The larvae sh have a short cycle and low cost, but the volume is tiny, and the development of organs are still in the early stage, making it di cult to conduct blood tests and slices.In fact, even for adult zebra sh, due to the small size and disease progression limitations, research on zebra sh atherosclerosis is heavily leaning on the lipid metabolism [34].Therefore, only the adult sh blood lipids were measured in this study, and it was not observed whether atherosclerotic plaque formation occurred early after feeding HFD.This could be a drawback, but the easy genetic manipulation and opacity of larvae offers an opportunity to study early atherosclerosis pathomechanisms.Maybe, with the progress of detection technology, the additional effect of ApoEb deletion on leukocyte accumulation and atherosclerotic plaque formation in adult zebra sh will be clari ed in the future.Secondly, it is impossible to accurately quantify the feeding, consumption, and excretion of each larva.In the experiment, we observed the development and feeding of larvae under the microscope regularly and abandoned the unhealthy or non-feeding larvae in time, but still could not completely eliminate the in uence of individual differences on the experimental results.
Thirdly, whether ApoEb knockout will lead to other pathological consequence related to lipid metabolism is still unclear, which requires further observation.

Conclusion
In summary, we introduced a new genetic model of LMD.ApoEb mutant zebra sh larvae subjected to a short HFD feeding resulted in robust and consistent accumulation of intravascular lipid deposits.We propose that this new model may pave another way for the pathophysiological and therapeutic target screening research on LMDs like hyperlipidemia and FH, as well as the pharmacological and toxicological studies of novel lipid-lowering drugs.

Supplementary Files
This is a list of supplementary les associated with this preprint.Click to download.
Financial support This work was supported in part by the National Key Research and Development Program of China (2019YFA0802700, 2017YFA0103700), the National Natural Science Foundation of China (91339205, 81130005).CRediT authorship contribution statement Yang-Xi Hu: Conceptualization, Methodology, Software, Formal analysis, Data curation, Visualization, Writing -Original draft preparation.Hong-Min You: Validation.Rong-Fang Zhu: Conceptualization, Methodology.Yu-Lai Liang and Fang-Fang Li: Methodology.Yong-Wen Qin and Xian-Xian Zhao: Supervision, Resources.Chun Liang and Qing Jing: Writing -Review & Editing, Project administration, Funding acquisition.

Figures Figure 1
Figures

Figure 2 Adult
Figure 2

Figure 3 Development
Figure 3