Comparison of the Cholesterol-Lowering Effect of Scallop Oil Prepared from the Internal Organs of the Japanese Giant Scallop ( Patinopecten yessoensis ), Fish Oil, and Krill Oil in Obese Type II Diabetic KK- A y Mice

: Due to the growing demand of n- 3 polyunsaturated fatty acids (PUFA) as supplements and pharmaceutical products worldwide, there are concerns about the exhaustion of n- 3 PUFA supply sources. We have successfully prepared high-quality scallop oil (SCO), containing high eicosapentaenoic acid and phospholipids contents, from the internal organs of the Japanese giant scallop ( Patinopecten yessoensis ), which is the largest unutilized marine resource in Japan. This study compared the cholesterol-lowering effect of SCO with fish oil (menhaden oil, MO) and krill oil (KO) in obese type II diabetic KK- A y mice. Four-week-old male KK- A y mice were divided into four groups; the control group was fed the AIN93G-modified high-fat (3 wt% soybean oil + 17 wt% lard) diet, and the other three groups (SCO, MO, and KO groups) were fed a high-fat diet, in which 7 wt% of the lard in the control diet was replaced with SCO, MO, or KO, respectively. After the mice were fed the experimental diet for 42 days, their serum, liver, and fecal lipid contents as well as their liver mRNA expression levels were evaluated. The SCO group had significantly decreased cholesterol levels in the serum and liver; this decrease was not observed in the MO and KO groups. The cholesterol-lowering effect of SCO was partly mediated by the enhancement of fecal total sterol excretion and expression of liver cholesterol 7α-hydroxylase, a rate-limiting enzyme for bile acid synthesis. These results indicate that dietary SCO exhibits serum and liver cholesterol-lowering effects that are not found in dietary MO and KO and can help prevent lifestyle-related diseases.


Lipid analysis of the experimental oils
The fatty acid FA composition was analyzed using a gas chromatography GC system GC-2014; Shimadzu Co., Kyoto, Japan equipped with Omegawax ® 250 column 30 m 0.25 mm i.d. 0.25 μm d.f., Merck KGaA following methylation with a boron trifluoride methanol complex solution 20 . The cholesterol content was analyzed using a GC system equipped with an SH-Rtx-5MS column 30 m 0.25 mm i.d. 0.25 μm d.f., Shimadzu GLC Ltd., Tokyo, Japan after saponification with sodium hydroxide, and 5α-cholestane was utilized as an internal standard 21 . The PL content was determined using a phosphorus assay 22 . The PL class composition was analyzed by thin-layer chromatography TLC with silica gel 60 glass plates Merck KGaA and chloroform/methanol/acetic acid/water 50:40:3:4, v/v/v/v as the solvent mixture. PL spots were then detected with a 50 v/v sulfuric acid solution and identified using authentic PL standards CAEP, PC, and PE . The PL class composition was calculated based on spot intensity using the JustTLC software version 4.0.3, Lund, Sweden . The lipid composition of the experimental oils is listed in Table 1.

Animal diet and care
The experimental protocol followed the Guide for the Care and Use of Experimental Animals issued by the Prime Minister s Office of Japan and was reviewed and approved by the Animal Ethics Committee of Kansai University Approval No. 1917 .
Four-week-old male KK-A y mice were purchased from CLEA Japan, Inc. Tokyo, Japan . Following an acclimatization period of seven days, the mice were divided into four groups of eight mice each with a similar mean body weight BW . The mice were kept in an air-conditioned room temperature, 20-22 ; light on, 08:00-20:00 and had free access to drinking water. Each mouse was given an experimental diet every day for 42 days by pair-feeding. The ingredient compositions based on the American Institute of Nutrition 93G formula 23 and FA composition of the experimental diets are shown in Table S1 and Table 2, respectively. Fresh diet and water were provided daily, and BW was measured each time. Before the mice were sacrificed, feces of each mouse were collected daily for two days, weighed, froze, and ground using a conventional mill. After the mice had been fed the experimental diets for 42 days, the mice that did not fast were anesthetized with isoflurane Fujifilm Wako Pure Chemical Co. and then weighed from 9:00 to 12:00 . Blood was collected without use of an anticoagulant, and serum was obtained by centrifugation at 1,500 g for 15 min. The liver and white adipose tissue WAT , including the epididymal, mesenteric, perirenal, and inguinal WAT, were removed, rinsed with cold saline, and weighed. A portion of the liver was stored in RNAlater ® solution Merck KGaA to measure mRNA levels. The collected serum and organs were frozen in liquid nitrogen and stored at 80 until analysis.

Lipid analysis of serum, liver, and feces
The serum TAG, PL, total cholesterol, HDL-C, and non-HDL-C contents were measured with an automatic chemistry analyzer Olympus AU5431; Olympus Co., Tokyo, Japan .
Liver lipids were extracted using the method described by Bligh and Dyer 24 . The total lipids of the liver were dissolved in 2-propanol, and liver TAG content was then measured using the Triglyceride E Test Fujifilm Wako Pure Chemical Co. , following the manufacturer s instructions. Liver PL and cholesterol contents were determined using the same methods as described above, respectively. The FA composition of the total liver lipids was analyzed by GC as described above 20 .
The fecal neutral sterols, including cholesterol and coprostanol, were analyzed by GC, as described above 21 . The fecal total bile acid BA content was analyzed using the total bile acid test kit Fujifilm Wako Pure Chemical Co. , following the manufacturer s instructions. The fecal total sterol content was calculated by adding neutral sterols and total BA contents. The fecal BA composition was measured using GC-mass spectrometry MS , as described in our previous report 25 , with some modifications. Briefly, the fecal sterols were extracted twice using ethanol containing 1 M NaOH and an internal standard 23-nordeoxycholic acid, NDCA at 80 for 1 h. The neutral sterols were removed using diethyl ether, and BA was extracted with diethyl ether after adjusting the pH to 1. The obtained BA was redissolved in methanol and then methylated and trimethylsilylated using trimethylsilyldiazomethane and N-trimethylsilylimidazole with trimethylchlorosilane, respectively. The derivatized BA was then analyzed through GC-MS GCMS-QP2010 PARVUM 2; Shimadzu Co. equipped with SPB-1 column 30 m 0. 25  using helium at a flow rate of 1 mL/min as the carrier gas. The column temperature program was initialized at 230 , ramped at 2 /min to 310 , and held at 310 for 5 min. Identification of each derivatized BA was carried out using an authentic BA standard, and the content of each derivatized BA in feces was quantified using a correction factor obtained using NDCA as an internal standard. The fecal short-chain fatty acid SCFA content was analyzed using GC equipped with a Nukol TM capillary GC column 30 m 0.25 mm i.d. 0.25 μm d.f., Merck KGaA after extracting SCFA using diethyl ether as described in our previous report 25 . Each SCFA was identified using a mixture of SCFA Volatile Free Acid Mix certified reference material, Merck KGaA .

Analysis of liver mRNA expression
The liver stored in RNAlater ® solution was crushed with a bead beater-type homogenizer MicroSmash MS-100R, Tomy Seiko Co., Ltd., Tokyo, Japan . Liver total RNA was isolated and purified using the TRIzol ® reagent Thermo Fisher Scientific Inc., Waltham, MA, USA according to the manufacturer s protocol. The content and purity of total RNA were determined by measuring the absorbance at wavelengths of 260 and 280 nm by ultraviolet spectroscopy UV-1800, Shimadzu Co. using Hellma ® TrayCell TM Hellma GmbH & Co. KG, Müllheim, Germany . cDNA was synthesized using the GoScript TM Reverse Transcription System Promega Co., Fitchburg, WI, USA . The mRNA expression levels were measured using a Thermal Cycler Dice ® Real Time System Takara Bio Inc., Kusatsu, Japan using GoTaq ® qPCR Master Mix Promega Co. . The primer sequences used for the detection of the ATP-binding cassette Abc a1, Abcg5, Abcg8, acetyl-Coenzyme A acetyltransferase 1 Acat1 , cytochrome P450 family 2 subfamily c polypeptide 70 Cyp2c70 , cholesterol 7α-hydroxylase Cyp7a1 , farnesoid X receptor Fxr , 3-hydroxy-3-methylglutaryl coenzyme A reductase Hmgcr , low density lipoprotein receptor Ldlr , liver X receptor Lxr , small heterodimer partner 1 Shp1 , scavenger receptor class B type 1 Srb1 , sterol regulatory element binding factor 2 Srebf2 , and glyceraldehyde 3-phosphate dehydrogenase Gapdh , all of which were designed using Primer3Plus http://primer3plus.com/ , are listed in Table S2. Results were normalization to Gapdh expression level and expressed as the fold-change in mRNA expression relative to that in the control group.
2.6 16S rRNA amplicon sequence and bioinformatic analysis Four samples from the experimental groups were randomly selected for the experiment. Genomic DNA from the feces was extracted using ISOSPIN Fecal DNA Nippon Gene Co., Ltd., Tokyo, Japan according to the manufacturer s protocol. The 16S rRNA amplicon sequence analysis was performed using a next-generation sequencing system Ion PGM TM , Thermo Fisher Scientific Inc. with an Ion 16S Metagenomics Kit according to the manufacturer s protocol and our previous report 25 .
The obtained sequence data were analyzed using Ion Reporter Software 16S Metagenomics Workflow ver. 5.12 Thermo Fisher Scientific Inc. automatically accessible to MicroSEQ 16S Reference Library v2013.1 Thermo Fisher Scientific Inc. and Greengenes v13.5 The Greengenes Database Consortium, http://greengenes.secondgenome. com/ . α-Diversity indices Chao-1 and Simpson as well as bacteria and their sequence numbers identified at the phylum and genus levels for each sample were obtained from Ion Reporter Software. The obtained sequence numbers were used to determine the relative abundance of bacteria in each taxonomic class. Hierarchical clustering was performed using R 26 based on each group s proximate bacterial composition at the phylum and genus levels.

Statistical analysis
Data are expressed as the mean standard error of the mean SEM . The differences between multiple groups were evaluated using analysis of variance ANOVA and Tukey s multiple comparison test. Statistical significance was set at p 0.05. All analyses were performed using GraphPad Prism version 7.0; GraphPad Software, San Diego, CA, USA .

Growth parameters and relative organ weights
Growth parameters and relative organ weights are listed in Table 3. The initial BW, final BW, BW gain, food intake, and food efficiency were not significantly different among the groups. The relative liver weights in the MO and KO groups were significantly higher than those in the control and other groups, respectively. There were no significant differences in the relative WAT weights epididymal, mesenteric, perirenal, and inguinal WAT among the groups.

Lipid contents in the serum and liver
The lipid contents of the serum and liver are listed in Table 4. The serum TAG contents in the SCO and KO  groups were significantly lower than those in the control group. The SCO, MO, and KO diets significantly decreased the serum PL content compared to the control diet. The total serum cholesterol and non-HDL-C contents in the SCO group were significantly lower than those in the control, MO, and KO groups. In addition, mice fed the KO diet had significantly higher serum HDL-C content than mice fed the SCO diet. The liver cholesterol content in the SCO group was significantly lower than that in the control, MO, and KO groups. The KO diet significantly increased the liver TAG and cholesterol contents and decreased the liver PL content compared to the other diets.
The FA compositions of the total liver lipids are listed in Table 5. The SCO diet increased the C18:3n-3 and EPA ratio, and decreased the C18:1n-9 and C20:1n-9 ratio compared with the control group. In addition, the SCO group was increased the C16:0, C16:1n-7, C18:1n-7, C18:3n-3, C20:4n-6, and EPA ratio and decreased the C18:1n-9, C20:1n-9, and DHA ratio compared to the KO group. There were no significant differences in FA compositions of the total liver lipids between the SCO and MO groups.

Sterol contents and short-chain fatty acid composition in feces
The fecal neutral sterol, total BA, and total sterol con-  tents are shown in Table 6. The fecal cholesterol and coprostanol contents in the SCO and KO groups were significantly higher than those in the control and MO groups. The fecal cholesterol content in the KO group was higher than that in the SCO group. The fecal total BA content in the SCO group was higher than that in the control group, whereas the fecal total BA content in the KO group was higher than that in all other groups. The fecal BA content is shown in Fig. 1. The SCO and KO diets significantly increased the fecal α-muricholic acid MCA , βMCA, and deoxycholic acid DCA contents and decreased the fecal hyodeoxycholic acid HDCA and ursodeoxycholic acid UDCA contents compared with the control diet. The fecal ωMCA content in the SCO group was significantly higher than that in the MO and KO groups. In contrast, no significant differences were observed in fecal SCFA contents, including acetic acid, propionic acid, isobutyric acid, butyric acid, isovaleric acid, and valeric acid, among the groups Fig. S1 .

Relative liver mRNA expression levels
The relative mRNA expression levels related to cholesterol metabolism are shown in Fig. 2. The liver Hmgcr and Ldlr expression in the KO group was significantly decreased compared to that in the control group. The KO group exhibited a significant increase in the expression of Abcg5 compared to the SCO group. In addition, the expression of Cyp7a1 was significantly higher in the SCO group than in mice fed the control diet.

The relative abundance of fecal microbiota
There was no significant difference in the number of good reads among the groups after reading length and read abundance filtering using Ion Reporter. α-Diversity indices Chao-1 and Simpson were not different among the groups data not shown . Hierarchical clustering and relative abundance of main fecal bacteria at the phylum level are shown in Figs. 3A and 3B. In the hierarchical clustering at the phylum level Fig. 3A , the MO, SCO, and KO groups were separated from the control group, and the MO group was separated from the SCO and KO groups. There was no significant difference in the relative abundance of main fecal bacteria at the group s phylum level Fig. 3B . The relative fecal abundance and main bacteria composition at the genus level are shown in Figs. 3C and 3D. The relative abundance of Bacteroides in the SCO group was significantly higher than that in the control group. The KO diet was significantly decreased the relative abundance of Lactococcus compared to the control diet. Besides, mice fed the KO diet were increased the close relative abundance of Lactobacillus genus considerably compared with mice fed the other diets.

Discussion
The presence of n-3 PUFA was characterized by SCO, MO, and KO containing EPA and DHA Table 1 . The EPA contents in SCO, MO, and KO were 30.5, 16.8, and 29.6 mg/ g, respectively. The DHA content was almost the same in SCO, MO, and KO 10.9, 12.0, and 12.6 mg/g, respectively . PL contents in SCO and KO were 16.3 and 35.2 mg/g, respectively, while MO did not contain PL. The PL class composition of SCO was 63.4 wt of PC and 21.1 wt of PE, and that of KO was 61.3 wt of PC and 29.0 wt of PE. Furthermore, SCO contained 8.7 wt of CAEP, which is a sphingolipid possessing a C-P bond and is absent in KO.
The KO diet significantly increased the relative liver weight compared to the control diet Table 3 . Additionally, the liver TAG content in the KO group was significantly   higher than that in the other groups Table 4 . Wang et al.
reported that a cholesterol diet 1.0 , w/w increased liver weight and TAG content because the β-oxidation of FA in the liver was suppressed in rats 27 . In addition, our previous study demonstrated that the intake of a cholesterol diet 0.5 , w/w approximately doubled the liver weight compared to a cholesterol-free diet in rats 28 . Since KO contained high levels of cholesterol 45.1 mg/g , the KO diet contained 0.32 w/w cholesterol. The increase in liver weight and liver TAG content in the KO group was probably due to the high cholesterol content in the KO diet. However, this phenomenon cannot occur in humans, because cholesterol metabolism in humans is different from that of mice. A previous study reported that KO intake decreased serum total cholesterol content in humans 29 . Moreover, liver expression levels of Hmgcr, encoding the rate-limiting enzyme for the synthesis of cholesterol, and Ldlr, which encodes the enzyme responsible for cholesterol uptake from the blood, were significantly lower in the KO group than in the control group Fig. 2 . A previous study showed that KO intake lowered liver Hmgcr and Ldlr expression levels in mice 30 . On the other hand, Zou et al. demonstrated that a high-cholesterol diet decreased liver Hmgcr and Ldlr expression 31 . In this study, it was impossible to determine whether the reduction in liver Hmgcr and Ldlr expression in the KO group was due to KO or cholesterol intake. To clarify the cause of the reduction in liver Hmgcr and Ldlr expression in the KO group, it is necessary to evaluate the cholesterol content of experimental diets at the same level. In addition, the total fecal sterols, including neutral sterols and total BA excretions, were significantly increased in the KO group compared to those in other groups Table 6 . The high excretion of neutral sterols in the feces was attributed to the excretion of cholesterol obtained from the diet. Cholesterol in the liver is metabolized into BA and then secreted into the bile 32 . The increased fecal total BA excretion in the KO group was attributed to the conversion of a large amount of cholesterol taken up to the liver. Therefore, the increase in liver TAG content and fecal total sterol excretion in the KO group was partly due to the high cholesterol content in KO.
The serum TAG content in the SCO group was significantly lower than that in the control group Table 4 . The serum TAG content is partly due to the alteration of liver FA metabolism, and liver FA metabolism is mainly regulated by transcriptional factors, including SREBP-1 and nuclear receptors, and peroxisome proliferator-activated receptor alpha PPARα 33 . EPA and DHA suppress FA synthesis in the liver through SREBP-1c 34 and are natural ligands of PPARα, which regulates the stimulated β-oxidation of FA 35 . In this study, the FA composition of SCO including EPA and DHA was reflected in the total liver lipid s FA composition Table 5 . Our previous study showed that the activity of acyl-coenzyme A oxidase-1, a key enzyme of peroxisomal FA β-oxidation, was significantly increased by glycerophospholipid structure compared to dietary TAG structure 20 . Therefore, the reduction in serum TAG content in the SCO group suggested that the intake of EPA, DHA, and PL contained in SCO could partly lower serum TAG content. On the other hand, the DHA ratio in the liver total lipid of the KO group was higher than other three groups. In previous reports, the DHA ratio in the liver of rodents fed fish oil was as high as about 10 20, 25 , suggesting that DHA binds not only to PL but also to TAG. Therefore, the increase in the DHA ratio of the KO group may be related to increase the DHA-containing TAG content.
The SCO diet significantly decreased the total cholesterol and non-HDL-C contents in the serum and liver cholesterol content compared to the other diets Table 4 . Several mechanisms could explain the liver cholesterollowering effect of dietary oils. The first possibility is enhanced fecal sterol excretion 36 . The SCO diet significantly increased fecal neutral sterol excretions, including of cholesterol and coprostanol, compared to the control and MO diets Table 6 . Glycerophospholipid intake is generally known to inhibit cholesterol absorption in the small intestine 37 . A previous study reported that the glycerophospholipids PC and PE inhibited cholesterol absorption in the small intestine by inhibiting the hydrolysis of micellar PL by phospholipase A2 38,39 . In addition, sphingomyelin SM , which is one of sphingolipid, inhibited cholesterol absorption through hydrogen bonding with the hydroxyl group of cholesterol 40,41 . Furthermore, SM digestion products, ceramide and sphingoid bases, inhibited the intestinal absorption of cholesterol, because of the interactions between the hydroxyl group of cholesterol and the carboxylic acid group of FA 42 . Although no study has reported the inhibition of cholesterol absorption by CAEP intake, CAEP possesses the potential to inhibit cholesterol absorption as it has been hydrolyzed into free sphingoid bases during the digestion process 43 . In this study, SCO was found to be mainly composed of PC, PE, and CAEP Table  1 , enhancing the fecal neutral sterols excretion. Additionally, fecal total BA excretion was significantly higher in the SCO group compared to the control and MO groups Table  6 . Intestinal bile salt hydrolase BSH activity, which is involved in the removal of the N-acyl amidation glycine or taurine of BA, is associated with fecal BA excretion, because conjugated BA is more efficiently reabsorbed in the ileum 44,45 . The distribution of BSH has been recorded in Bacteroides, Bifidobacterium, Clostridium, Enterococcus, and Lactobacillus 46 . In this study, the relative abundance of the genus Bacteroides, which was the highest among the fecal microbiota, was significantly higher in the SCO group than that in the control group Fig.  3 . The increase in fecal total BA excretion in the SCO group could be associated with an increase in the relative abundance of Bacteroides. Therefore, the enhancement of fecal neutral sterols and total BA excretions by SCO intake was partly related to reduced serum and liver cholesterol contents.
The second possible cause of the lowered cholesterol content is the alteration of cholesterol metabolism in the liver. The expression of liver Cyp7a1, encoding the ratelimiting enzyme in the classical BA biosynthetic pathway, in the SCO group was significantly increased compared to that in the control group Fig. 2 . Our previous study also showed that SCO intake increased liver Cyp7a1 expression level 11 . Hayashi et al. reported that increasing taurinebinding βMCA TβMCA , an antagonist of FXR, in the ileum enhanced liver Cyp7a1 expression by suppressing the FXR/SHP1 pathway 47 . The SCO diet significantly increased fecal βMCA content compared to the control diet Fig. 1 . BA is usually conjugated in the liver 48 , and the most conjugated BA in mice is taurine-conjugated 49 . The higher fecal βMCA content in the SCO group was considered to be due to the higher TβMCA content in the ileum than that in the control group. Therefore, SCO enhanced liver Cyp7a1 expression, partly because of the increased TβMCA composition in the ileum. Moreover, αMCA and βMCA are biosynthesized from CDCA by liver CYP2C70, a 7α-hydroxylase 50 .
Since there was no difference in the Cyp2c70 expression level Fig. 2 , the increase in fecal αMCA and βMCA contents in the SCO group was associated with a decrease in the amount of αMCAand βMCA-metabolizing bacteria. However, since bacteria with enzymes that metabolize αMCA and βMCA have not been reported, the 16S rRNA amplicon sequence results could not be utilized.
Several limitations should be considered when interpreting the results of this study. Firstly, this study could only evaluate the mRNA expression of genes. Second, this study could not measure PL subclasses of SCO and KO. The plasmalogen, a type of PL subclass, has reported to change cholesterol metabolism in mice 49 . Further evaluations of gene expression and PL subclasses are necessary to strengthen this hypothesis in the future.

Conclusion
This study compared the cholesterol-lowering effect of SCO with that of MO and KO in obese type II diabetic KK-A y mice. SCO intake decreased the serum and liver cholesterol contents; however, this effect was not observed by the intake of MO and KO. The reductions in the serum and liver cholesterol contents by SCO intake were partly mediated by the enhancement of fecal total sterol excretion and liver Cyp7a1 expression levels. Therefore, SCO, containing n-3 PUFA and PL, is expected to be a healthpromoting food material with a cholesterol-lowering effect not found in MO and KO.