CRISPR/Cas9-mediated Angptl8 knockout suppresses plasma triglyceride concentrations and adiposity in rats[S]

Angiopoietin-like protein (ANGPTL)8 is a liver- and adipocyte-derived protein that controls plasma triglyceride (TG) levels. Most animal studies have used mouse models. Here, we generated an Angptl8 KO rat model using a clustered regulatory interspaced short palindromic repeat (CRISPR)/CRISPR-associated protein 9 (Cas9) (CRISPR/Cas9) system to clarify the roles of ANGPTL8 in glucose and lipid metabolism. Compared with WT rats, Angptl8 KO rats had lower body weight and fat content, associated with impaired lipogenesis in adipocytes; no differences existed between the groups in food intake or rectal temperature. Plasma TG levels in both the fasted and refed states were significantly lower in KO than in WT rats, and an oral fat tolerance test showed decreased plasma TG excursion in Angptl8 KO rats. Higher levels of lipase activity in the heart and greater expression of genes related to β-oxidation in heart and skeletal muscle were observed in Angptl8 KO rats. However, there were no significant differences between KO and WT rats in glucose metabolism or the histology of pancreatic β-cells on both standard and high-fat diets. In conclusion, we demonstrated that Angptl8 KO in rats resulted in lower body weight and plasma TG levels without affecting glucose metabolism. ANGPTL8 might be an important therapeutic target for obesity and dyslipidemia.

In a subsequently retracted 2013 article, ANGPTL8 was renamed "betatrophin" to indicate that it regulated the replication of -cells in response to insulin resistance (8).
The authors reported that betatrophin/ANGPTL8 overexpression in mice induced a 17-fold increase in -cell proliferation. However, subsequent studies in mice contested that claim, with findings that Angptl8 KO mice exhibited normal glucose homeostasis, -cell area, and compensatory -cell expansion even if fed a high-fat diet (11,12) or if given an insulin receptor antagonist (12). In addition, ANGPTL8 overexpression in multiple cohorts of mice of various ages and genetic strains had no effect on -cell proliferation (13). The original article was finally retracted by the authors when they were unable to replicate their own findings in a larger number of mice (14). However, in a rat model, overexpression of ANGPTL8 by delivery of human ANGPTL8 gene plasmids to the pancreas, liver, and skeletal muscle reportedly increased -cell proliferation (15). It may be, therefore, that there are species-specific effects accounting for the differences in the mouse and rat models.
Regardless of whether ANGPTL8 affects -cells in certain species, the protein remains of great interest, and many investigators have tried to understand its role in metabolic disorders in humans, such as insulin resistance, obesity, dyslipidemia, and type 2 diabetes (16)(17)(18)(19). As noted above, most studies of the function of ANGPTL8 have been performed in mouse models, but if there are indeed interspecies differences, it is important to use other animal models to yield a better understanding of the role of ANGPTL8 in metabolic homeostasis. Rats have long been used in biomedical research, including in the development of drugs, but they have fallen out of favor with the advent of genebased research, as manipulating the rat genome has proven more difficult than working with the mouse genome (20). However, the development of the clustered regulatory interspaced short palindromic repeat (CRISPR) system using CRISPR-associated protein 9 (Cas9) has enabled editing of the rat genome (20,21). In this study, we generated an Angptl8 KO rat using the CRISPR/Cas9 system to investigate the protein's roles in glucose and lipid metabolism.
The rats were kept at a constant temperature of 23°C ± 1°C under a 14 h light and 10 h dark cycle (lights on 0700, lights off 2100) with free access to water and a standard diet (3.73 kcal/g, with 12% of calories from fat, F-2; Funabashi Farm, Chiba, Japan) unless otherwise stated. All animal care and experiments were conducted in accordance with the Guidelines for Animal Experiments of Kyoto University and were approved by the Animal Research Committee of Kyoto University.

Body weight, food intake, rectal temperature
Body weight was measured every week between 1500 and 1700 hours. Food intake was measured for 7 days and rectal temperature was determined with a digital thermometer (BAT7001H; Physitemp, Clifton, NJ).

Percent body fat and adipocyte size
Epididymal fat for calculation of percent body fat was dissected and weighed and immediately fixed in 10% neutral buffered formalin. For histology, 5 m-thick sections were stained with hematoxylin and eosin. Adipocyte size in the epididymal fat was evaluated on micrographs taken with a fluorescent microscope and analyzed (BioRevo BZ-9000; KEYENCE, Osaka, Japan), assessing the cross-sectional adipocyte area in 10 random fields (magnification ×200).

Adipocyte differentiation
For culture and differentiation of preadipocytes from subcutaneous fat, we applied a modification of a procedure described previously (23). Briefly, after isolation of subcutaneous inguinal fat from 15-week-old rats, the fat tissues were minced and digested with 0.2 mg/ml collagenase type IV at 37°C for 30-60 min. The suspension was passed through a 250 m pore mesh, and the effluent was centrifuged at 300 g for 5 min. After decantation, 20 ml of DMEM (Thermo Fisher Scientific, Waltham, MA) were added, and the suspension was passed through a 40 m pore mesh again to remove undigested clumps and debris. The effluent was centrifuged at 300 g for 5 min and the pellet was washed twice in 5 ml of DMEM. The resulting isolated cells were plated in a 24-well tissue culture dish and cultured in DMEM supplemented with FBS (Thermo Fisher Scientific). After growth to confluence, DMEM with 0.17 M insulin, 0.5 mM isobutyl-1-methylxanthine, 0.25 M dexamethasone, and 1 M pioglitazone was used to induce differentiation for 3 days. The medium was replaced with DMEM with 0.17 M insulin for two additional days. This procedure was performed twice, after which the differentiated adipocytes were analyzed.

Biochemical analysis
After an overnight fast and then again 2 h after refeeding, blood was collected from 12-week-old rats, and the plasma was isolated by centrifugation (1,150 g) for 10 min. Glucose levels were determined by the glucose oxidase method using a reflectance glucometer (LIFE CHECK; GUNZE, Kyoto, Japan). Plasma TG, total cholesterol, and NEFA were measured using enzymatic kits (Triglyceride E-test, Cholesterol E-test, and NEFA C-test; Wako Pure Chemical Industries, Osaka, Japan). Plasma insulin levels were measured using an enzyme-linked immunosorbent assay kit (rat insulin kit; Morinaga Institute of Biological Science, Yokohama, Japan). Plasma TG and cholesterol levels in chylomicron (CM), VLDL, LDL, and HDL fractions were determined using high-performance liquid chromatography (LipoSEARCH; Skylight Biotech, Akita, Japan) as previously described (24). The upper limit of the size of CMs that pass through the column is about 500 nm.

Glucose, insulin, and fat tolerance tests
A glucose tolerance test (GTT) was performed by injecting rats intraperitoneally with 2.0 mg/g glucose after an overnight fast. Blood samples were obtained from the tail vein before and 15, 30, 60, and 120 min after the glucose load for measurement of blood glucose and plasma insulin levels. For an insulin tolerance test (ITT), rats were injected intraperitoneally with 1.0 U/kg human regular insulin (Humulin R; Eli Lilly Japan, Kobe, Japan) after an overnight fast, and blood glucose levels were measured before and 30, 60, 120, and 180 min after the insulin load. For a fat tolerance test (FTT), rats were given 10 l/g corn oil orally after an overnight fast, and blood samples were obtained before and 1, 2, 3, and 4 h after the fat load to measure plasma TG levels. Incremental area under the curve (iAUC) relative to the fasting value for TG during the FTT was calculated.

TG content in tissues
Rats aged between 30 and 48 weeks were euthanized between 1500 and 1700 hours, and individual organs were dissected to obtain tissue samples. For measurement of TG content in tissues, we applied a modification of a procedure described previously (25). Briefly, TGs in heart, skeletal muscle, and liver were extracted with 2-propanol (Nacalai Tesque, Kyoto, Japan) and heptane (Nacalai Tesque). After evaporation of the solvent, ethanol was added to solubilize the lipid and the TG content was measured using an enzymatic kit.

Lipase activity
To assess the plasma lipase activities, postheparin plasma was collected from each ad libitum-fed rat genotype 10 min after a bolus injection of heparin (0.2 U/g) into the tail vein. Heparinreleasable LPL activity was measured using the LPL/HTGL activity assay kit (Immuno-Biological Laboratories Co., Ltd., Gunma, Japan) according to the manufacturer's instructions. Total lipase activity in heart and skeletal muscle was determined using a LPL activity fluorometric assay kit (BioVision, Milpitas, CA) according to the manufacturer's instructions.

Total RNA extraction and quantitative real-time PCR
Total RNA was extracted from the heart, skeletal muscle, and liver with TRIzol (Life Technologies, Carlsbad, CA) according to the manufacturer's instructions. Total RNA (1 g) was reversetranscribed with an iScriptTMcDNA synthesis kit (Bio-Rad Laboratories, Hercules, CA). Quantitative real-time PCR was performed with the StepOnePlus real-time PCR system using TaqMan (Applied Biosystems, Foster City, CA). Relative amounts of mRNAs were normalized with ribosomal 18 S RNA. The sets of primers and probes used are listed in supplemental Table S1.

Immunostaining
Pancreases were isolated and immediately fixed in 10% neutral buffered formalin. After proper fixation, the organs were routinely dehydrated through a graded series of ethanol concentrations, embedded in paraffin, sectioned, stained, and examined microscopically. Shortly after deparaffinization and hydrophilization, endogenous peroxidase activity was blocked with 3% hydrogen peroxidase. For immunohistochemical staining for insulin, sections were incubated with guinea pig polyclonal anti-porcine insulin antibody (Dako, Carpinteria, CA). Immunolocalization was demonstrated using the EnVision+ system (anti-rabbit HRP-labeled polymer) (Dako) and DAB+ substrate kit (Dako).

Statistics
Data are expressed as mean ± SEM. Comparisons were made using a t-test or one-way factorial ANOVA followed by the Tukey-Kramer test, where appropriate. Statistical analysis of changes in body weight and the results of the GTT, ITT, and FTT were performed using two-way factorial ANOVA with repeated measurement followed by the Tukey test. A P value <0.05 was considered statistically significant.

RESULTS
Angptl8 KO rats were generated using the CRISPR/Cas9 system Thirty-one pups (21 males and 10 females) were born. PCR amplification of the targeted loci showed that two male rats (lines #1 and #2) had Angptl8 deletions. Further sequencing confirmed that the two rats had different frameshift mutations in exon 2 (Fig. 1A). Line #1 carried a 7 bp deletion and line #2 a 980 bp deletion in the ANGPTL8 gene, resulting in a premature termination codon. The expression of ANGPTL8 in the liver, white adipose tissue, and brown adipose tissue was significantly lower in Angptl8 KO rats compared with WT rats (Fig. 1B). Rats Het for Angptl8 had gene expression levels intermediate between those of the WT and Angptl8 KO rats (Fig. 1B). Line #1 was chosen to establish a colony of Angptl8 KO rats (designated as KO), as the phenotypes of lines #1 and #2 were identical.

Angptl8 KO rats were resistant to diet-induced obesity
Body weights of Het and Angptl8 KO rats were lower than those of WT rats throughout the study ( Fig. 2A). However, food intake and rectal temperature measured at the age of 25 weeks to assess energy balance did not differ among the three groups (Fig. 2B, C). There was also no significant difference in weight between Het and Angptl8 KO rats. As indicated by epididymal fat weight, the percent body fat of Angptl8 KO rats was significantly lower than that of WT and Het rats (Fig. 2D). In addition, the adipocytes in epididymal fat were significantly smaller in Angptl8 KO rats than in WT rats (Fig. 2E).
To confirm the influence of Angptl8 KO on body weight, we measured weights when the rats were on a high-fat diet (5.24 kcal/g, with 60% of calories from fat, D12492; Research Diets Inc., New Brunswick, NJ), which increases the expression of ANGPTL8 in liver and adipose tissue (7,26,27). Rats were fed a high-fat diet from the age of 8 weeks, and they all gained weight gradually, but from the age of 13 weeks on, Angptl8 KO rats gained significantly less weight than WT rats (Fig. 2F). There were, however, no differences in food intake or body temperature between Angptl8 KO and WT rats (food intake, 35

Angptl8 KO impaired adipogenesis
To investigate the influence of Angptl8 KO on adipocytes themselves, we induced adipocyte differentiation to evaluate adipogenesis of preadipocytes from the stromal vascular fraction of the inguinal fat pads. Microscopic inspection of the culture dishes revealed lower numbers of Oil Red O-stained cells in KO cultures (Fig. 3A). In addition, we assessed the expression of adipocyte marker genes using real-time PCR and found no differences between WT and Angptl8 KO cultures in the expression of PPAR-2, the master regulator for the development of adipocytes. However, SREBP-1c and stearoyl-CoA desaturase (SCD)-1 expression was significantly lower in Angptl8 KO cultures, and leptin expression appeared to be lower, although not significantly so (P = 0.058) (Fig. 3B).

Angptl8 KO rats had reduced plasma TG levels
To evaluate the influence of Angptl8 KO on lipid metabolism in rats, we measured plasma lipid levels of each rat genotype fed a standard diet. Plasma TG levels in both the fasted and refed states were more than 50% lower in Ang-ptl8 KO rats compared with WT rats, and Het rats had intermediate plasma TG levels (Fig. 4A). There were no significant differences in total cholesterol levels in the fasted and refed states among WT, Het, and Angptl8 KO rats (Fig. 4B). Although NEFA levels in the refed state were not determined, they did not differ in the fasted state among WT, Het, and Angptl8 KO rats (Fig. 4C). Plasma TG levels in the fed state were significantly lower in the CM, VLDL, and HDL fractions in Angptl8 KO compared with WT rats (Fig. 4D).
As plasma TG levels of Angptl8 KO rats were low in both the fasted and refed states, we performed oral FTTs to investigate plasma TG excursion. Plasma TG levels of Angptl8 KO rats were significantly lower not only in the fasted state but also after a fat load compared with WT rats (Fig. 4E), and the iAUC for Angptl8 KO rats was 51% lower than that for WT rats (205.3 ± 31.6 mg/dl/h vs. 419.8 ± 54.5 mg/ dl/h, P < 0.05). Although CM production was not measured, this suggests that Angptl8 KO rats had increased TG clearance. Finally, we determined postheparin LPL activity in each rat genotype. Postheparin plasma LPL activity appeared to be higher by Angptl8 KO in a dose-dependent manner, although not significantly so (Fig. 4F).

Angptl8 KO rats revealed accelerated -oxidation in heart and skeletal muscle
Plasma TG levels of Angptl8 KO rats were low in both the fasted and refed states even on a high-fat diet (Fig. 5A). Total lipase activity in heart tissue of Angptl8 KO rats was higher than in WT rats, but not significantly so (25.4 ± 2.8 vs. 15.6 ± 4.3, P = 0.081) (Fig. 5B). LPL expression levels in heart (1.00 ± 0.07 vs. 1.37 ± 0.21, P = 0.20) and skeletal muscle (1.00 ± 0.15 vs. 1.02 ± 0.14, P = 0.91) did not differ between WT and Angptl8 KO rats (Fig. 5C). LPL catalyzes the hydrolytic cleavage of TGs into fatty acids for subsequent oxidation and/or storage. Therefore, we evaluated gene expressions related to -oxidation and lipid synthesis in heart and skeletal muscle. The -oxidation markers, carnitine palmitoyltransferase (CPT)-1b and acyl-CoA oxidase (ACOX), of Angptl8 KO rats were significantly elevated in the heart (Fig. 5D), and a tendency toward this was observed in the skeletal muscle (Fig. 5E). On the other hand, the lipid synthesis markers, SREBP-1c, SCD-1, and fatty acid synthase, did not differ between WT and Angptl8 KO rats (supplemental Fig. S2A, B). Finally, we checked lipid accumulation in heart, skeletal muscle, and liver tissues and found that TG content in all those tissues did not differ between WT and Angptl8 KO rats (Fig. 5F).

Glucose metabolism was not perturbed in ANGPLT8 KO rats
Fasting blood glucose and plasma insulin levels did not differ between WT and Angptl8 KO rats fed a standard diet (Fig. 6A, B). To evaluate the influence of Angptl8 KO on glucose tolerance and insulin sensitivity in rats, we performed GTTs and ITTs. In the GTTs, there were no significant differences in blood glucose or plasma insulin levels in the fasted state or after the glucose load between WT and Angptl8 KO rats (Fig. 6C, D). The results of ITTs were similar, with no significant differences in blood glucose levels between WT and Angptl8 KO rats (Fig. 6E), suggesting that Angptl8 KO did not affect insulin sensitivity. Feeding a highfat diet for 4 weeks induced insulin resistance and increased Fig. 2. Body weight and adipose tissue in Angptl8 KO rats. A: Body weight monitored weekly in WT, Het, and KO rats fed a standard diet (n = 5-14 per group). Data were generated from four independent experiments. Data were analyzed by two-way factorial ANOVA with repeated measurement followed by the Tukey test. **P < 0.01 KO versus WT, † †P < 0.01 Het versus WT. Food intake (B) and rectal temperature (C) measured in WT, Het, and KO rats (n = 4-14 per group, age 25 weeks). Data were generated from four independent experiments. Data were analyzed by one-way factorial ANOVA followed by the Tukey-Kramer test. D: Epididymal fat weight of WT, Het, and KO rats (n = 7-8 per group, age 35-48 weeks). Data were analyzed by one-way factorial ANOVA followed by the Tukey-Kramer test. *P < 0.05, **P < 0.01. E: Adipocyte areas in epididymal fat of WT and KO rats (n = 10 per group, age 41 weeks). Data were analyzed by t-test. **P < 0.01. F: Body weight change monitored weekly in WT, Het, and KO rats fed a high-fat diet (n = 4-9 per group). Data were generated from three independent experiments. Data were analyzed by two-way factorial ANOVA with repeated measurement followed by the Tukey test. **P < 0.01 KO versus WT, †P < 0.05 Het versus WT, † †P < 0.01 Het versus WT.
plasma insulin levels in all rats, but there were again no differences in fasting blood glucose and plasma insulin levels between WT and Angptl8 KO rats (Fig. 6A, B). Finally, we evaluated histologic changes in the pancreas. The increase in -cell area of rats fed a high-fat diet for 34 weeks did not differ between WT and Angptl8 KO rats (Fig. 6F).

DISCUSSION
To the best of our knowledge, this is the first report that CRISPR/Cas9-mediated Angptl8 KO rats had reduced body weight and plasma TG levels without any changes in glucose metabolism. In addition, we demonstrated that deletion of Angptl8 led to decreased fasting and postprandial plasma TG concentrations and increased expression of -oxidation genes in heart and muscle, without any extra lipid accumulation in heart, skeletal muscle, and liver. Rats are considered to be a better model than mice in terms of their physiologic and behavioral characteristics, which are in some respects more relevant to humans (20,28,29). As we have reported previously, for example, while Het Pparg KO mice have increased insulin sensitivity, Het Pparg KO rats show insulin resistance like patients with Het Pparg dominant-negative mutation (30). In this context, we have thus far generated KO rats for Leptin (31) and Bscl2/Seipin (32) besides Pparg (30) and investigated the roles of each gene. Therefore, we generated a rat model to analyze the functions of ANGPTL8. In this study, our data using Angptl8 KO rats is similar to that found with Angptl8 KO mice (11,12), confirming physiologic roles of ANGPTL8 in energy, glucose, and lipid metabolism in two different species. CRISPR/Cas9-mediated cleavage is achieved using the RNA-guided DNA nuclease, Cas9. Following generation of a double-stranded break, the DNA can be repaired by nonhomologous end joining, which can generate KO mutations by the introduction of random deletions of a few to hundreds of base pairs (20). We obtained two lines of Ang-ptl8 KO rats with different deletion lengths, 7 bp and 980 bp, in the ANGPTL8 gene (Fig. 1A). Decreased ANGPTL8 transcript levels in KO rats were revealed both in lines #1 and #2 by real-time PCR (Fig. 1B). Unfortunately, because there are no good antibodies for rat ANGPTL8, we could not evaluate protein levels. Off-target mutations are a major concern with the CRISPR/Cas9 system (20)(21)(22). Therefore, we analyzed and carefully compared two lines of Ang-ptl8 KO rats to assess the adequacy of Angptl8 KO and off-target mutations in our Angptl8 KO rats. Both lines #1 and #2 exhibited similar phenotypes in terms of energy, glucose, and lipid metabolism (supplemental Fig. S3A-F), suggesting that the ANGPTL8 gene was definitely knocked out and that off-target mutations were avoided in our Angptl8 KO rats. Line #1 female Angptl8 KO rats exhibited similar phenotypes to male Angptl8 KO rats (supplemental Fig. S4A-F), indicating that the effect of Angptl8 KO on energy, glucose, and lipid metabolism was similar regardless of the rats' sex.
Angptl8 KO suppressed weight gain in a dose-dependent manner, which was more evident when the rats were fed a high-fat diet ( Fig. 2A, F). Suppression of weight gain in Angptl8 KO rats was associated with reduced accumulation of adipose tissue (Fig. 2D, E), which is consistent with the phenotype of Angptl8 KO mice (11). We speculate that there are two mechanisms underlying decreased adiposity. One of the mechanisms may be increased energy expenditure, Fig. 4. Lipid metabolism in Angptl8 KO rats. A-C: Plasma TG, total cholesterol, and NEFA levels in WT, Het, and KO rats fed a standard diet in the fasted and refed state (n = 4-12 per group, age 12 weeks). Data were generated from four independent experiments. Data were analyzed by one-way factorial ANOVA followed by the Tukey-Kramer test. D: Plasma TG and cholesterol levels in lipoprotein fractions, including CMs, VLDL, LDL, and HDL in WT and KO rats measured by high-performance liquid chromatography (n = 3-4 per group, age 12 weeks). Data were analyzed by t-test. E: Plasma TG levels during a FTT in WT and KO rats (n = 3 per group, age 20 weeks). Data were generated from two independent experiments. Data were analyzed by two-way factorial ANOVA with repeated measurement followed by the Tukey test. F: Postheparin plasma LPL activity of WT, Het, and KO rats fed with a standard diet (n = 3-6 per group). Data were generated from two independent experiments. Data were analyzed by one-way factorial ANOVA followed by the Tukey-Kramer test. *P < 0.05, **P < 0.01. because ANGPTL8 blockade with a monoclonal antibody reduced body fat content and weight secondary to increased energy expenditure in mice (33). Although we did not measure whole-body expenditure in the present study, there were no differences in rectal temperature or food intake (Fig. 2B, C). In addition, there were no differences in core body temperature or energy expenditure, evaluated by oxygen consumption and respiratory exchange ratio, between Angptl8 KO mice and WT littermates (11). Taken together, the main cause of suppressed weight gain may be other than increased energy expenditure. The other mechanism may be due to the adipocytes themselves. The expression of ANGPTL8 is exclusive to mature adipocytes in the adipose tissue, and its level is markedly increased during adipogenesis (6). When we examined adipocyte differentiation using primary cultures, we found that lipid accumulation was impaired in Angptl8 KO cultures (Fig. 3A). Knockdown of Angptl8 in murine 3T3L1 cells with siRNA was reported to reduce lipid accumulation (6), which is compatible with our results in rats. Real-time PCR analyses showed that PPAR-2 transcript levels were similar, but lipogenic genes, such as SREBP-1c and SCD-1, were significantly lower in Angptl8 KO compared with WT rats (Fig. 3B), suggesting that lipogenesis in adipocytes was impaired by Angptl8 KO. The reason for suppression of weight gain in Angptl8 KO rats remains to be elucidated.
Plasma TG levels in both the fasted and refed states were almost half in Angptl8 KO rats compared with WT rats (Figs. 4A, 5B), and the iAUC during the FTT showed decreased plasma TG excursions in Angptl8 KO rats (Fig. 4E). TG clearance is regulated by the balance of production and consumption of TG. Lipid synthesis and fatty acid oxidation in the liver, the main TG synthetic organ, were examined by gene expressions, but no differences were observed (supplemental Fig. S5). On the other hand, tissues such as the heart, skeletal muscle, and white adipose tissue rely on plasma TGs as an important source of fatty acids for oxidation or to be stored (34). Utilization of plasma TGs is dependent on LPL, which is attached to the capillary endothelium and catalyzes the hydrolytic cleavage of TGs into fatty acids. LPL is minutely regulated to adjust fatty acid uptake by multiple mechanisms at the transcriptional and Effects of Angptl8 KO on lipase activity in oxidative tissues and ectopic fat accumulation. A: Plasma TG levels in WT and KO rats fed a high-fat diet in the fasted and refed state (n = 4-12 per group, age 12 weeks). B: Total lipase activity in heart and skeletal muscle of WT and KO rats fed a high-fat diet (n = 4-6 per group, age 30-42 weeks). C: Expression of LPL in heart and skeletal muscle of WT and KO rats fed a high-fat diet (n = 4-6 per group, age 30-42 weeks). Expression of genes related to -oxidation in heart (D) and skeletal muscle (E) of WT and KO rats fed a high-fat diet (n = 4-6 per group, age 30-42 weeks). F: TG content in heart, skeletal muscle, and liver of WT and KO rats fed a high-fat diet (n = 4-6 per group, age 30-42 weeks). Data were generated from three independent experiments. Data were analyzed by t-test. *P < 0.05. posttranslational levels to meet the requirements of the underlying tissue. In the fed state, LPL activity is increased in white adipose tissue and reduced in heart and skeletal muscle (34,35), suggesting that LPL is triggered to deliver TGs to adipose tissue rather than to oxidative tissues in the fed state. In this study, the postheparin plasma LPL activity appeared to be higher by Angptl8 KO in a dose-dependent manner, although not significantly so (Fig. 4F). In addition, total lipase activity in the heart of fed Angptl8 KO rats was slightly higher, though not significantly so, compared with that of WT rats (Fig. 5B). The transcriptional levels of LPL did not differ (Fig. 5C). LPL in the heart and the skeletal Fig. 6. Glucose metabolism in Angptl8 KO rats. Fasting blood glucose (A) and fasting plasma insulin (B) levels of WT and KO rats fed a standard (SD) or a high-fat diet (HFD) (n = 4-6 per group, age 12 weeks). Data were generated from four independent experiments. Data were analyzed by t-test. Blood glucose levels (C) and plasma insulin levels (D) during a GTT in WT and KO rats (n = 3-7 per group, age 26 weeks). Data were generated from four independent experiments. Data were analyzed by two-way factorial ANOVA with repeated measurement followed by the Tukey test. E: Blood glucose levels during an ITT in WT and KO rats (n = 3-7 per group, age 27 weeks). Data were generated from four independent experiments. Data were analyzed by two-way factorial ANOVA with repeated measurement followed by the Tukey test. F: Representative images of the pancreas with staining for insulin in WT and KO rats (magnification ×20, SD, age 37 weeks; HFD, age 42 weeks). muscle catalyzes lipolysis of plasma lipoproteins, reducing plasma TGs and providing free fatty acids as substrates for oxidation to produce energy (36,37). Changes of LPL activity in heart and skeletal muscle, especially, can strongly affect plasma TG levels. For example, mice with cardiac muscle-specific overexpression of LPL reportedly have lower plasma TG levels (38). In addition, overexpression of CPT-1 in skeletal muscle reportedly increases fatty acid oxidation and decreases plasma TG levels in rats (39). In this study, expression levels of genes related to -oxidation were significantly higher both in the heart and the skeletal muscle of fed Angptl8 KO rats associated with increased LPL activity (Fig. 5D, E). These data suggest that greater LPL activity in Angptl8 KO rats accelerated -oxidation in oxidative tissues, leading to increased TG clearance and decreased plasma TG levels. However, the mechanism linking ANGPTL8 deficiency and decreased LPL activity with increased -oxidation is unknown.
Glucose metabolism was unchanged in Angptl8 KO rats (Fig. 6A-F), similar to findings in Angptl8 KO mice (11,12). Despite the findings of Chen et al. (15) that ANGPTL8 promoted -cell proliferation in rats, we found that deletion of Angptl8 did not affect -cells or glucose metabolism in rats. According to the previous rat study, deletion of Angptl8 was likely to exacerbate glucose metabolism. However, based on our data, suppression of ANGPTL8 for treating dyslipidemia in humans would not worsen the control of diabetes, which will provide new evidence for the translational research toward a therapy targeted to ANGPTL8 in humans. The discrepancy between studies of the same species may relate to differences in study design, the one using overexpression of ANGPTL8 and ours using deletion of ANGPTL8. Further investigation is needed to determine whether ANG-PTL8 regulates -cell replication in rats.
Increased plasma ANGPTL8 levels have been reported in several diseases, including obesity, dyslipidemia, and type 2 diabetes (16)(17)(18)(19), disorders that are all risk factors for atherosclerosis leading to CVD (40). In this regard, inhibition of ANGPTL8 is anticipated as a potential therapy to prevent CVD events. In particular, despite lowering of plasma LDL-cholesterol levels with a statin, elevated TG levels may remain a residual risk for CVD (41,42). The Japan Diabetes Complications Study, a prospective study to assess risk factors for coronary heart disease and stroke, showed that serum TG was a leading predictor of coronary heart disease (43). Therefore, decreasing ANGPTL8 levels, for example by a monoclonal antibody (33,44), might well prevent CVD both by preventing weight gain and by substantially reducing plasma TG levels.
ANGPTL3 and ANGPTL4, having high homology with ANGPTL8, have also been reported as regulators of LPL activity and are expected to be potential targets to treat dyslipidemia (45). Angptl3 KO mice showed increased postheparin plasma LPL activity and decreased plasma TG levels (46). A phase 1 study of antisense inhibitor of ANG-PTL3 was conducted, reporting that decreased TG levels were observed in the treatment group (47). Angptl4 KO mice were also reported to have increased postheparin plasma LPL activity (46), and treatment with monoclonal antibody for mouse ANGPTL4 demonstrated that inhibition of ANGPTL4 led to decreased TG levels in mice (48). Similar to Angptl8 KO rats, both Angptl3 and Angptl4 KO mice showed increased LPL activity in heart or skeletal muscle (49,50). However, neither Angptl3 nor Angptl4 KO mice showed a significant change in body weight compared with WT mice (46). Although the reason for this discrepancy is not clear, decreased adipocyte expression of lipogenic genes, suggestive of impaired lipogenesis, could contribute to the lower body weight in Angptl8 KO rats. With its unique effect on body weight, ANGPTL8 may be a more suitable target for treating dyslipidemia with obesity.
In conclusion, we generated for the first time Angptl8 KO rats using the CRISPR/Cas9 system and demonstrated that Angptl8 KO reduced body weight, adiposity, and plasma TG levels without affecting glucose metabolism. These data provide proof of concept for studying Angptl8 KO in the rat model and indicate that ANGPTL8 may be a novel target in developing drugs to treat obesity and dyslipidemia.