Targeting hepatic heparin-binding EGF-like growth factor (HB-EGF) induces anti-hyperlipidemia leading to reduction of angiotensin II-induced aneurysm development

Objective The upregulated expression of heparin binding EGF-like growth factor (HB-EGF) in the vessel and circulation is associated with risk of cardiovascular disease. In this study, we tested the effects of HB-EGF targeting using HB-EGF-specific antisense oligonucleotide (ASO) on the development of aortic aneurysm in a mouse aneurysm model. Approach and results Low-density lipoprotein receptor (LDLR) deficient mice (male, 16 weeks of age) were injected with control and HB-EGF ASOs for 10 weeks. To induce aneurysm, the mice were fed a high fat diet (22% fat, 0.2% cholesterol; w/w) at 5 week point of ASO administration and infused with angiotensin II (AngII, 1,000ng/kg/min) for the last 4 weeks of ASO administration. We confirmed that the HB-EGF ASO administration significantly downregulated HB-EGF expression in multiple tissues including the liver. Importantly, the HB-EGF ASO administration significantly suppressed development of aortic aneurysms including thoracic and abdominal types. Interestingly, the HB-EGF ASO administration induced a remarkable anti-hyperlipidemic effect by suppressing very low density lipoprotein (VLDL) level in the blood. Mechanistically, the HB-EGF targeting suppressed hepatic VLDL secretion rate without changing heparin-releasable plasma triglyceride (TG) hydrolytic activity or fecal neutral cholesterol excretion rate. Conclusion This result suggested that the HB-EGF targeting induced protection against aneurysm development through anti-hyperlipidemic effects. Suppression of hepatic VLDL production process appears to be a key mechanism for the anti-hyperlipidemic effects by the HB-EGF targeting.


Introduction
Heparin binding EGF-like growth factor (HB-EGF), which is a member of epidermal growth factor (EGF) family member and a ligand for EGF-receptor (EGFR) [1], is involved in various pathophysiological processes including atherosclerosis and cancer development [2][3][4][5][6]. Among EGF family members, HB-EGF is a representative mediator for the integral EGFR transactivation by various stress conditions [7,8]. HB-EGF regulates proliferation of vascular smooth muscle cell (VSMC) [9,10] and inflammatory gene expression in the aortic endothelium under hyperlipidemic environment [11]. In addition, recent reports indicate that HB-EGF concentration in blood circulation correlates with circulatory cholesterol concentration [12] and risk of coronary artery disease in humans [13].
Hyperlipidemia is a key risk factor for the development of vascular diseases including aneurysm and atherosclerosis [14,15]. For lipid or lipoprotein homeostasis, the balance of production and clearance of VLDL in the liver and capillary endothelium of peripheral tissues is critical [16]. For the production of VLDL in the liver cells, the expression and stability of apolipoprotein B (apoB) and lipid transferring protein microsomal triglyceride transfer protein (MTP) are key determinants [17]. The clearance of VLDL in circulation is mainly regulated by vascular endothelial lipoprotein lipase (LPL) [18].
Infusion of angiotensin II (AngII) into hyperlipidemic mouse models (e.g., LDLR deficient mice under high fat diet or ApoE deficient mice under chow diet) have been widely used as aneurysm models for the last decade [19,20]. AngII infusion or hyperlipidemia alone can induce aortic aneurysm but the intensities and frequencies of aneurysm development was quite limited [14,21]. There was a significant gender difference on aneurysm development in the model mice as male mice showed greater incidence and severity of aneurysm development [14,22,23].
In this study, we targeted HB-EGF gene transcription using HB-EGF-specific antisense oligonucleotide (ASO) administration to determine the targeting effects on aortic aneurysm development. In summary, we observed that the HB-EGF ASO administration induced an efficient protection against aneurysm developments in ascending and abdominal aorta. The HB-EGF targeting induced a remarkable anti-hyperlipidemic effect by suppressing hepatic VLDL secretion, which appears to be a key mechanism for the protection.

Animals
Male LDLR deficient mice (colony bred from original stock from The Jackson Laboratory; Stock No. 002207; 16 weeks of age; The strain has been backcrossed to C57BL/6J mice for 10 Noninvasive tail cuff method to measure blood pressure Blood pressure was measured using the Kent blood pressure machine as described in the previous report [28]. Refer to S1 File Materials and Procedure-Extended for details of the procedure. http://dx.doi.org/10.17504/protocols.io.iygcftw

Lipoprotein-associated cholesterol distribution analysis by FPLC
Blood was collected from mice in EDTA-coated tubes by cardiac puncture, and plasma was isolated by centrifugation. The cholesterol distribution among lipoprotein classes was determined after separation of plasma by gel filtration chromatography based upon the method described previously [29]. Refer to Supplementary Procedure for procedure details. http://dx. doi.org/10.17504/protocols.io.izhcf36

Quantification of liver tissue lipid content
We followed the procedure described by Temel RE et al. [30]. Refer to S1 File Materials and Procedure-Extended for details. http://dx.doi.org/10.17504/protocols.io.iy7cfzn No 005061) were used for the assays. Refer to S1 File Materials and Procedure-Extended for details. http://dx.doi.org/10.17504/protocols.io.iy3cfyn

GC analysis of fecal neutral sterols
We followed the procedure described by Temel RE et al. [30]. Refer to S1 File Materials and Procedure-Extended for the procedure details. http://dx.doi.org/10.17504/protocols.io.iy9cfz6 Quantitative RT-PCR Total RNAs were extracted from mouse tissues using an RNA isolation kit (Qiagen, Cat No. 74106). cDNAs were prepared using cDNA synthesis kit (Bio-Rad, Cat No. 1708891). For amplification and quantification of PCR products, SYBR Green Master Mix (ThermoFisher Scientific; Cat No. K0223) and a PCR amplification apparatus (Bio-Rad, CFX96 Touch™ Real-Time PCR System) were used. Expression level of GAPDH mRNA was used for normalization. More than two sets of primers were tested for PCR amplification for each gene. Primer sequence information is described in S1 Table. Western blotting Plasma samples or freshly isolated tissue samples were grinded and lysed in RIPA buffer (Cell Signaling; Cat No. 9806) containing PMSF (1mM) and protease and phosphatase inhibitor cocktails (Sigma-Aldrich; Cat No. P8340 and P5726). Protein samples were separated by SDS-PAGE using 4-20% Bio-Rad Mini-PROTEAN TGX gels following a standard procedure. Separated proteins were transferred onto polyvinylidene difluoride (PVDF) membranes using a Trans-Blot Turbo™ Transfer System (Bio Rad, Cat No. 1704155). After incubation with an anti-apoB antibody (Meridian Life Science, Cat No. K34005G-1; HRP-conjugated) in 1% BSA or 5% fat free milk containing TBST buffer solution, apoB bands were detected using enhanced chemiluminescent solution (Amersham, Cat No. RPN2232). ApoB bands were captured and quantified using image analysis myECL imager (Thermo Fisher, Cat No. 62236).

Statistical analysis
Results are presented as mean ± standard deviation (SD) unless mentioned otherwise. Test group samples are compared to control group samples by Student's T-test or Two-way ANOVA. If required, multiple comparison correction by Sidak-Bonferroni method was applied. A significant p value less than 0.05 was considered statistically different.

HB-EGF targeting using ASO administration reduced aneurysm development
The HB-EGF ASO administration significantly downregulated HB-EGF mRNA expression level in liver tissue ( Fig 1A). The expressions of the other EGF family members or EGFR were not affected by HB-EGF ASO administration (S1A and S1B Fig). We used male LDL deficient mice fed high fat diet (23% fat and 0.2% cholesterol; w/w) and AngII infusion for the induction of aortic aneurysm (S2A Fig for animal treatment scheme). There was no difference of body weight gains between control and HB-EGF ASO treatment groups (S2B Fig). Two mice from each group died of aortic rupture between 7 to 10 days of AngII infusion (Fig 1B).  Targeting of HB-EGF against aneurysm ASO administration also decreased the number of mice with dissections in the ascending aortic arch area ( Fig 1D).
In addition, HB-EGF ASO administration significantly suppressed abdominal aortic aneurysm (AAA) development. Representative intact aorta images with and without AAA were shown in Fig 1E. Images for all aortas of control and HB-EGF ASO treatment groups were listed in S3A and S3B Fig. There was a significant reduction of the number of mice with abdominal aortic dilation incidence by HB-EGF targeting ( Fig 1E and S3C Fig). We also confirmed that HB-EGF ASO administration effectively suppressed atherosclerotic lesion formation in the aortic arch lumen area as shown in S3D Fig. Other groups previously reported that prenatal whole body-or vascular endothelial-specific HB-EGF gene deletions induced cardiac hypertrophy with gross enlargement of heart ventricular chambers [32-34]; however, postnatal induction of HB-EGF gene deletion did not induce the phenotype change [35]. In a control experiment using LDLR deficient mice under normal diet, we tested the effects of HB-EGF ASO administration on the cardiac structure. First, we confirmed downregulation of HB-EGF mRNA level by the HB-EGF ASO administration in multiple tissues including liver, aorta, and kidney (S4A Separately, we tested the effects of HB-EGF ASO administration in wild type C57BL/6 mice under normal diet. The C57BL/6 was the genetic background of the LDLR deficient mice that we adopted for aneurysm study. As expected, the HB-EGF ASO administration did not induce any vascular defects like aneurysm or atherosclerosis (data not shown). There were no changes of heart size or chamber structures by the ASO administration (data not shown). However, still we observed a significant reduction of circulatory lipid levels by the HB-EGF ASO administration in the C57BL/6 mice (S6A and S6B Fig).
HB-EGF targeting induced anti-hyperlipidemic effects by suppressing the hepatic VLDL secretion HB-EGF ASO induced a remarkable suppression of systemic total cholesterol and TG concentrations in the LDLR deficient mice under HFD (Fig 2A and 2B). FPLC fractionation of lipoprotein-associated cholesterol of plasmas collected at termination step showed remarkable and moderate downregulations of VLDL and LDL-cholesterols in circulation, respectively ( Fig  2C); however, there was no change for HDL-associated cholesterol concentration by the HB-EGF ASO administration. In correspondence, there was a downregulation of apoB protein concentration in blood circulation ( Fig 2D).
We observed a significant increase of liver weight ( Fig 3A) and elevation of hepatic neutral lipid contents (i.e., TG and cholesterol ester) by the HB-EGF ASO treatment (Fig 3B and 3C). There was no difference of free cholesterol content in the liver (Fig 3D).
The circulatory VLDL level was suppressed but there was simultaneous elevation of neutral lipid contents in the liver by the HB-EGF administration. Thus, we hypothesized that HB-EGF positively regulates VLDL secretion from the liver. To test this hypothesis, we directly measured hepatic VLDL secretion rate using C57BL/6 mice [31]. As shown in Fig 4A, the HB-EGF ASO administration induced a significant suppression of hepatic VLDL secretion rate; in contrast, injection of recombinant HB-EGF (human, active form) significantly increased hepatic   Targeting of HB-EGF against aneurysm

Discussion
In this study, we demonstrated that the HB-EGF targeting significantly suppressed aneurysm and atherosclerotic lesion development in a mouse disease model. Mechanistically, an efficient lipid-lowering by the HB-EGF targeting appears to be a primary mechanism for the protection against the aneurysm and atherosclerosis. This study also indicates that the HB-EGF via EGFR signaling could be an important positive regulator for the production of VLDL in liver. The injection of recombinant HB-EGF enhanced hepatic VLDL production rate, but the HB-EGF ASO administration suppressed circulatory lipid levels in both normolipidemic and hyperlipidemic conditions, which suggested that HB-EGF would be a general regulator for the regulation of hepatic VLDL production. This new information could be of potential importance in understanding lipoprotein metabolism and lipid homeostasis in the liver and blood circulation [37,39].
The molecular and cellular mechanism for the regulation by HB-EGF signaling on hepatic VLDL production is unclear yet. The process of VLDL production in the hepatocytes is highly complicated process, which is controlled by multi-factors including lipid substrate availability [17, 40] and expression and function of apoB and microsomal triglyceride transfer protein (MTP) proteins in the ER lumen space [41,42]. Recently, Lee RG et al. showed that suppressions of apoB or MTP expressions in the liver tissue using ASO administrations induced efficient reductions of circulating lipid levels but with significant increases of neutral lipid levels in the liver tissue [25]. Because the HB-EGF ASO administration did not induce significant changes of apoB mRNA or MTP expression levels (mRNA and protein) in the liver (data not shown), the apoB-or MTP-independent mechanism might be involved in suppressing hepatic VLDL production by the ASO administration. The maturation of primordial VLDL initially synthesized in endoplasmic reticulum (ER) lumen space and subsequent secretory process in the Golgi apparatus are another important factors [43]. Because there was a significant elevation of neutral lipid contents in the liver and increase of lipid droplet formation in the liver cell, the HB-EGF targeting appears to cause delay of mobilization of lipid substrates from the cytosolic lipid droplets in the hepatocytes [40]. Other groups showed that the combination of suppression of apoB expression and enhanced de novo lipogenesis, which happens under insulin resistance or metabolic syndrome conditions, induced enlargement of VLDL particles in circulation [44]. Because we observed that both TG and apoB levels were reduced by the HB-EGF ASO administration, the size of the VLDL particle appears to be constant. Thus, not the deficiency of apoB or MTP but the insufficiency of the lipid substrate for VLDL assembly appears to be a key mechanism for the suppression of VLDL production in the liver by the HB-EGF ASO administration.
Previous reports delineated that local expression of HB-EGF in the aortic vessel was associated with burden of atherosclerosis [2,4,45]. Though the liver tissue is one of the top organs with efficient ASO distribution [46], the HB-EGF ASO administration significantly suppressed HB-EGF expression in the aorta, which may contribute to the protection against development of aneurysm and atherosclerosis. We showed that the HB-EGF-EGFR signaling mediates inflammatory gene transcription in the vascular endothelial cells by oxidized phospholipids [11] and the signaling also regulates the proliferation and migration of the vascular smooth muscle cells [10]. An additional complication is that HB-EGF mediates angiotensin II signaling in the vascular smooth muscle cells [10,[47][48][49]. Thus, local suppression of HB-EGF expression may directly inhibit angiotensin II signaling in the vessel wall. It would be difficult to study tissue-specific effects of the HB-EGF targeting using ASO administration. Further use of inducible VSMC or vascular endothelial-specific HB-EGF gene knockout systems would be helpful to define the local effects of the HB-EGF knockdown.
Two separate groups developed HB-EGF floxed mouse system for the induction of tissuespecific HB-EGF knockout using loxP-Cre system [32,33]. Prenatal whole body-, vascular endothelium-, or vascular smooth muscle cell-specific knockouts of HB-EGF induced cardiac hypertrophy with severe defects of valvulogenesis causing gross enlargement of ventricular chambers [32-34]. Similar phenotype was shown by genetic knockouts of EGFR or TACE/ ADAM17, which is a downstream receptor of HB-EGF and the metalloproteinase that activates HB-EGF on cell surface, respectively. Mice with defects of HB-EGF ectodomain shedding fragment also showed similar cardiac hypertrophy [50]. In contrast, post-natal induction of HB-EGF gene deletion in the vascular endothelium or in the liver did not show structural defects of heart [35, 51,52]. We also show that the HB-EGF ASO administration does not cause structural problems in the heart. Compared with genetically modified model systems, the suppression of HB-EG expression by HB-EGF ASO was moderate, about 50% mRNA level reduction. In control experiments using LDLR deficient mice under normal diet or C57BL/6 mice under normal diet, we confirmed absence of deleterious effects of HB-EGF ASO on the heart structure and size of ventricular chambers.
Takemura et al. showed that the induction of liver-specific HB-EGF knockdown enhanced liver injuries induced by thioacetamide (TAA), CCl 4 , or bile duct ligation procedure [51][52][53]. The mechanism for the increase of sensitivity to the liver damages is not clear yet. Possibly, the upregulation of neutral lipid contents in the liver might be one cause for the increased liver damage. The lipid accumulation in the liver by HB-EGF ASO administration suggests possible limitation in applying the ASO for the subjects with hyperlipidemia or hyperlipidemia-associated vascular diseases. Lee RG et al. showed that the administration of apoB and MTP ASOs in the liver induced increases of neutral lipid contents in the liver in the LDLR deficient mice under HFD [25]. However, the apoB ASO induced less accumulation of lipids with formation of smaller size of lipid droplets in the liver. Recently, Conlon et al. showed the mechanism for the differences of lipid accumulation by apoB and MTP ASOs [54]. The apoB ASO induced an autophagic pathway that helps removal of neutral lipids in the liver cells. Because the lipid disposing autophagic pathway requires MTP function, MTP targeting by MTP ASO causes severe fat accumulation in the liver. The HB-EGF ASO administration keeps MTP expression intact in the liver, which may be a favorable sign for the induction of the lipid disposing autophagic pathway.
Interestingly, we observed that the HB-EGF ASO administration induced a significant downregulation of basal blood pressure in control experiments using LDLR deficient mice as shown in S5C Fig. The mechanism for the anti-hypertensive effects is unclear yet. Previous reports on the effects of whole-body knockout of HB-EGF on the blood pressure were inconsistent [32, 33, 55]; two groups showed that the HB-EGF knockouts did not change basal blood pressure [32, 55], but one group showed a significant lowering of blood pressure [33]. We presume that the reduction of blood pressure was not by direct effects of heart dysfunction because the ASO administration did not induce apparent changes of heart structure. As one clue, we detected that the HB-EGF ASO administration significantly suppressed HB-EGF expression in the kidney. Previously, other groups reported that targeting of HB-EGF induced protections against vasospatic response by endothelin-1 administration [55] or against renal injuries induced by chronic infusion of AngII [35] and ischemic reperfusion [56]. Vascular endothelial-specific HB-EGF knockout also leads to protection against development of progressive crescentic glomerulonephritis [57]. For determination of the mechanism of the antihypertensive effects shown by the HB-EGF ASO administration, further study of the effects on the kidney function would be required.
Previous reports showed that an EGFR blocker gefitinib was shown to downregulate circulatory lipid levels in a mouse model [58] and another EGFR blocker AG1478 also induced a significant protection against atherosclerotic lesion development in a hyperlipidemic mouse model [59]. Multiple EGFR ligands are expressed in the liver as shown in S1A and S1B Fig. Previous reports suggested that each EGF family members induces both overlapped and distinctive intracellular EGFR signaling depending on the type of receptor it interacts with [60,61]. Although this study was focused on HB-EGF signaling, the results of this study suggests a possibility that the other EGF family members than HB-EGF might also contribute to the pathophysiological regulation of hepatic VLDL production, which would be an important subject in understanding homeostasis of hepatic lipoprotein metabolism.
We observed that the HB-EGF ASO administration protects against both thoracic and abdominal aortic aneurysms (TAA and AAA). Multiple groups demonstrated that TAA and AAA have differential pathological mechanisms [14,15]; however, hyperlipidemia appears to be a common risk factor for both types of aneurysms. Collectively, in this study, we demonstrated HB-EGF targeting significantly reduced aortic aneurysm developments in the aorta. The HB-EGF targeting remarkably suppressed circulating lipid concentration by delaying hepatic VLDL production. Further evaluation of liver-or brain-associated side effects would be required to evaluate the usefulness of HB-EGF targeting against aneurysm and possibly against atherosclerosis.

S1 Fig. HB-EGF ASO administration downregulated hepatic HB-EGF mRNA levels. (A)
LDLR deficient mice (LDLR KO) fed normal diet (ND) were injected weekly with control and HB-EGF ASOs for 12 weeks (40 and 20 mg/kg/wk by 6 week interval). Relative mRNA levels of the EGF family members, EGFR, and ERBB4 in the liver tissues were determined by RT-PCR. The number in () indicates PCR cycle no. The cycle number for each gene was optimized to detect differences of template amounts by standard reactions using serial dilution of pooled RNA samples. PCR reactions for amphiregulin, which is a member of EGF family, with 2 different sets of primers showed no products. Housekeeping gene GAPDH product bands were used for normalization. Similar results were reproduced by more than 2 times of repeated PCR reactions using the same total RNA samples.