Lycopene and Apo-10#-lycopenoic Acid Have Differential Mechanisms of Protection against Hepatic Steatosis in b-Carotene-9#,10#- oxygenase Knockout Male Mice

Background: Nonalcoholic fatty liver disease is positively associated with obesity and cardiovascular disease risk. Apo-10#lycopenoic acid (APO10LA), a potential oxidation product of apo-10#-lycopenal that is generated endogenously by b-carotene9#,10#-oxygenase (BCO2) cleavage of lycopene, inhibited hepatic steatosis in BCO2-expressing mice. Objective: The present study evaluated lycopene and APO10LA effects on hepatic steatosis inmicewithout BCO2 expression. Methods: Male and female BCO2-knockout (BCO2-KO) mice were fed a high saturated fat diet (HSFD) with or without APO10LA (10 mg/kg diet) or lycopene (100 mg/kg diet) for 12 wk. Results: Lycopene or APO10LA supplementation reduced hepatic steatosis incidence (78%and 72%, respectively) and severity in BCO2-KO male mice. Female mice did not develop steatosis, had greater hepatic total cholesterol (3.06 vs. 2.31 mg/g tissue) and cholesteryl ester (1.58 vs. 0.86mg/g tissue), but had lower plasma triglyceride (TG) (229 vs. 282mg/dL) and cholesterol (97.1 vs. 119 mg/dL) than male mice. APO10LA-mitigated steatosis in males was associated with reduced hepatic total cholesterol (18%) and activated sirtuin 1 signaling, which resulted in reduced fatty acids (FAs) and TG synthesis markers [stearoyl-coenzyme A (CoA) desaturase protein, 71%; acetyl-CoA carboxylase phosphorylation, 79%; AMP-activated protein kinase phosphorylation, 67%], and elevated cholesterol efflux genes (cytochrome P450 family 7A1, 65%; ATP-binding cassette transporter G5/8, 11%). These APO10LA-mediated effectswere notmimicked by lycopene supplementation. Intriguingly, steatosis inhibition by lycopene induced peroxisome proliferator–activated receptor (PPAR)aand PPARg-related genes in mesenteric adipose tissue (MAT) that increases mitochondrial uncoupling [cell death–inducing DNA fragmentation factor, a subunit-like effector a, 55%; PR domain-containing 16, 47%; uncoupling protein 3 (Ucp3), 55%], FAb-oxidation (PPARa, 53%; very long chain acyl-CoA dehydrogenase, 38%), and uptake (FA transport protein 4, 29%; lipoprotein lipase 43%). Expressions of 10 MAT PPAR-related genes were inversely correlated with steatosis score, suggesting that lycopene reduced steatosis by increasing MAT FA utilization. Conclusions: Our data suggest that lycopene and APO10LA inhibit HSFD-induced steatosis in BCO2-KO male mice through differential mechanisms. Sex disparity of BCO2-KO mice was observed in the outcomes of HSFD-induced liver steatosis and plasma lipids. J Nutr 2015;145:268–76.


Introduction
Nonalcoholic fatty liver disease (NAFLD) 8 is a chronic liver disease that is observed in 75-100% of overweight and obese adults and children (1). NAFLD describes a series of related disorders that can progress in stages from simple steatosis to fibrosis and cirrhosis (2). Patients with NAFLD are often associated with metabolic syndrome including insulin resistance and hypertriglyceridemia, as well as with increased cardiovascular disease (CVD) and related risk factors (3), which include an atherogenic lipid profile (high in serum TGs and LDL cholesterol; low in HDL cholesterol) and systemic inflammation (4,5). Liver has the capacity to synthesize lipids via de novo lipogenesis and secrete lipoprotein particles. Thus, liver dysfunction can modify risk factors of CVD (3,4). In light of the current obesity epidemic and increasing prevalence of NAFLD, the prevention of NAFLD-associated metabolic disorders through dietary means represents an important strategy.
Observational data indicates that the intake of lycopene-rich foods is inversely associated with CVD risk (6,7) and with CVD risk indicators such as dyslipidemia and systemic inflammation (6,7). In vivo and in vitro studies have suggested that lycopene may use multifaceted mechanisms to reduce CVD and NAFLD risks (8)(9)(10), mechanisms that include modulating cholesterol metabolism, serving as an antioxidant, and inhibiting inflammation (8)(9)(10). Of note, NAFLD patients have substantially reduced plasma lycopene concentration (11), suggesting potential interactions between low lycopene status and CVD risk.
We and others have demonstrated that the enzyme bcarotene-9#,10#-oxygenase (BCO2) plays a critical role in the endogenous metabolism of nonprovitamin A carotenoids including lycopene (12,13). BCO2 metabolizes lycopene through asymmetric cleavage at the 9#,10# double bond and generates metabolites, which include apo-10#-lycopenal, apo-10#-lycopenol, and apo-10#-lycopenoic acid (APO10LA; chemical structure shown in Supplemental Figure 1) (12,13). Evidence suggests that these metabolites may exhibit more important biological roles than lycopene itself (9,14,15), and BCO2 ablation in mice altered lycopene metabolism (16,17). In particular, APO10LA supplementation was effective in inhibiting hepatic steatosis in genetically induced obese (ob/ob) mice (15) and attenuating high saturated fat diet (HSFD)-induced liver inflammation as well as tumor number and volume in C57Bl/6J mice (18). An important question remains as to whether the biological activity of lycopene could be different from its metabolite APO10LA in the absence of BCO2 expression. This information is critically needed because 19 single nucleotide polymorphisms of BCO2 have been found in humans (19), and BCO2 gene variants can account for the differential expression and activity of the BCO2 enzyme among individuals. BCO2 single nucleotide polymorphisms have been associated with increased circulatory proinflammatory IL-18 expression (19) and with decreased HDL cholesterol concentrations (19), suggesting a gene-diet interaction between the BCO2 enzyme and dietary lycopene on human health outcomes. We hypothesize that the lycopene biological effects would be different from APO10LA in mice when BCO2 enzyme expression is absent. Using a murine model that lacks BCO2 protein expression, we investigated the separate effects of lycopene or APO10LA supplementation on HSFD-induced hepatic steatosis and determined the underlying mechanisms thereof.

Methods
Animal experimental design. The animal experimental protocol was approved by the Institutional Animal Care and Use Committee at the Jean Mayer-USDA Human Nutrition Research Center on Aging at Tufts University. BCO2-knockout (BCO2-KO) mice with BCO2 protein ablation were generated as previously described (20). The schematic for the animal experimental design is shown in Supplemental Figure 2. Mice were fed the standard laboratory chow until the start of experiments (7012 Teklad LM-485; Harlan Laboratories), maintained on a 12-h light/dark cycle in a controlled temperature and humidity room, and consumed water ad libitum. Two separate experiments were conducted concurrently to evaluate APO10LA and lycopene effects. Sixwk-old male and female BCO2-KO mice were randomly assigned to 1 of the 4 diets: 1) an obesogenic HSFD [Bio-Serv; information cited in Ip et al. (21)], in which 60% of energy is lard derived; 2) an HSFD supplemented with APO10LA (HSFD+APO10LA; 10 mg/kg diet); 3) an HSFD supplemented with lycopene placebo beadlet (HSFD+LyP; 1000 mg beadlet/kg diet); and 4) an HSFD supplemented with lycopene (HSFD+Ly; 100 mg lycopene/kg diet) for 12 wk. We included both male and female mice in our experiments because there are no previously published studies to our knowledge that examined the biological effects of APO10LA in BCO2-KO mice. All mice in the experiments were given fresh diets every 2-3 d and maintained on their respective diets until the experiment was completed. Mice were weighed weekly and killed at 18 wk of age by exsanguinations under deep isoflurane (Isothesia; Butler Schein) anesthesia without food deprivation.
Dietary supplementation. The APO10LA used in the experiments of this study was provided by Dr. Hansgeorg Ernst (BASF, Ludwigshafen, Germany) with 99% purity, analyzed by HPLC as previously described (Supplemental Figure 3) (13,14). The lycopene supplement was in the form of a 10% lycopene beadlet (BASF). The composition of the lycopene placebo beadlet is listed in Supplemental Table 1. APO10LA, lycopene, or placebo beadlet without lycopene was incorporated directly into the HSFD to achieve a homogenous diet mixture and stored as previously described (18). We selected the same APO10LA dosage (10 mg/kg diet) as in our published study (18), where APO10LA supplementation for 24 wk reduced HSFD-promoted hepatic inflammation in C57Bl/6J male mice. This APO10LA-supplemented dose (10 mg/kg diet) is approximately equivalent to 0.36 mg/d APO10LA in a 60-kg adult man (22,23). A/J mice supplemented with the same APO10LA dose for 14 wk had a plasma APO10LA concentration of ;1.0 nmol/L (14). This plasma concentration is comparable to the sum of apo-lycopenals (1.9 nmol/L) found in human plasma of individuals who had consumed tomato juice (with 21.8 mg/d lycopene) for 8 wk (24). Based on the assumption that BCO2-KO mice have similar carotenoid absorption as other strains of mice on a vitaminsupplemented, semipurified diet (;1/10 of human absorption) (14,15,25), our lycopene-supplemented dose (100 mg/kg diet) is approximately equivalent to 8.1 mg/d lycopene in a 60-kg adult man (22,23). The mean human dietary lycopene was ;8 mg/d (26,27). In addition, our recently published study (28) showed that BCO2-KO mice supplemented with the identical lycopene dose for 24 wk had mean hepatic lycopene concentrations that were within human ranges (0.1-20.7 nmol/g tissue) (29).
Liver and adipose tissue collection and processing. After the mice were killed, whole livers and mesenteric adipose tissue (MAT) were removed from experimental mice and processed for biochemical and histologic analyses as described previously (18).
Plasma lipid and lipoprotein profile. Blood from cardiac puncture was collected into EDTA-coated tubes, and plasma was separated from RBCs by centrifugation at 1500 3 g for 20 min at 4°C. Plasma total cholesterol (TC), HDL cholesterol, and TG concentrations were determined on a Cobas Mira automated analyzer with use of enzymatic reagents (Roche Diagnostics). Non-HDL cholesterol was calculated as the difference between TC and HDL cholesterol.
Hepatic cholesterol and TG contents. Hepatic TC and free cholesterol (FC) contents of extracted liver lipids were analyzed by GC as described previously (30). Cholesteryl ester (CE) was computed as the difference between TC and FC (30). Liver tissue (25 mg of wet weight) was used to determine hepatic TG content as previously described (15). Hepatic cholesterol and TG were corrected for the wet liver weight used in the analysis.
RNA extraction and qRT-PCR. Total RNA extraction from frozen liver and MAT tissue and subsequent qRT-PCR were performed as described previously (18). Primer sequences are listed in Supplemental Table 2.
Protein isolation and Western blotting. Protein isolation from frozen liver tissue and subsequent Western blotting analysis of specific proteins were executed as described previously (18). The following antibodies were used for western blotting: acetyl-CoA carboxylase (ACC), AMPactivated protein kinase (AMPK) a, phosphorylated-ACC (Ser79), phosphorylated-AMPKa (Thr172), stearoyl-CoA desaturase (SCD1) (Cell Signaling), IL-6 (R&D), acetylated-Foxhead box protein O1 (FoxO1), heme-oxygenase (HO)-1, and sirtuin 1 (SIRT1). Horseradish peroxidase-conjugated secondary antibody (Bio-Rad) was used to detect specific proteins, and the specific bands were visualized by a SuperSignal West Pico Chemiluminescent Substrate Kit (Pierce) according to the manufacturerÕs instructions. Dilution series and calibration curve were performed for each of the antibodies used to quantify protein. b-Actin protein was used for loading normalization of some proteins unless specified otherwise, which was detected by an anti-actin antibody (Sigma-Aldrich). A GS-710 Calibrated Imaging Densitometer (Bio-Rad) was used to quantify the intensities of protein bands.
Statistical analysis. SAS 9.3 software was used to perform the statistical analysis. Values in the text are means 6 SEMs or medians (IQRs). Two-factor ANOVA was used to examine the overall, diet, sex, or diet 3 sex interaction effects on experimental outcomes modified by APO10LA or lycopene supplementation. StudentÕs two-sample unequal variance t test, Wilcoxon signed-rank test, and x-square test were used to examine the differences for the following comparisons: 1) HSFD and HSFD+APO10LA, and 2) HSFD+LyP and HSFD+Ly. SpearmanÕs rank correlation was used to perform nonparametric measures of statistical dependence between 2 variables. The P value was set at 0.05 for the comparisons to reach statistical significance.

Results
APO10LA and lycopene supplementation ameliorated HSFD-induced hepatic steatosis in BCO2-KO male mice. Final mean body and liver weights did not differ significantly between supplemented and nonsupplemented BCO2-KO mice (Tables 1 and 2). There was a marked difference in the response of female and male mice to the HSFD for liver steatosis, cholesterol accumulation, and plasma lipid concentrations.  Male mice weighed significantly more than female mice at the start and end of the experiments (Tables 1 and 2). The percentage weight gain was 63.4 (10.3 g) in the female mice and 83.3 (16.2 g) in the male mice. In male mice, APO10LA and lycopene supplementation significantly decreased HSFD-induced liver steatosis incidence and score ( Figure 1, Tables 1 and 2). APO10LA or lycopene supplementation tended to decrease hepatic TGs (Tables 1 and 2). In female mice, the HSFD did not induce steatosis, and APO10LA or lycopene supplementation did not alter hepatic steatosis outcome or TGs (Tables 1 and 2). In both male and female mice, APO10LA, but not lycopene supplementation, significantly reduced plasma TG and hepatic FC and CE concentrations (Tables 1 and 2). Final body weight (BW), BW gain, and hepatic FC and TC were correlated with steatosis score in both female and male mice (R 2 = 0.52, 0.52, 0.33, 0.35; P = 0.001, 0.001, 0.030, 0.020, respectively). Female mice had significantly greater hepatic TC and CE than males, but FC was not significantly different between males and females (Tables 1 and 2). Female mice had less weight gain, and plasma TGs, TC, and non-HDL cholesterol than males (Tables  1 and 2). Female mice had lower plasma TGs than males only within the HSFD and HSFD+APO10LA groups (Table 1). These results suggest that lycopene and APO10LA had biological effects in the absence of BCO2. Sex disparity exists on diet-induced hepatic steatosis and lipid metabolism in BCO2-KO mice.
APO10LA, but not lycopene, decreased HSFD-induced hepatic inflammation in BCO2-KO male but not female mice. APO10LA, but not lycopene supplementation, significantly reduced hepatic inflammatory foci incidence in male but not female BCO2-KO mice (Table 1). This APO10LA-mediated inflammation suppression in male mice was associated with reduced hepatic IL-6 protein (94%, P = 0.06; Figure 2A) and with elevated anti-inflammatory HO-1 protein (74%; Figure 2B) expression in male mice.

Discussion
The present study evaluated the effects of lycopene and APO10LA supplementation on hepatic steatosis in male and female mice that were deficient in BCO2 expression. BCO2-KO male mice developed hepatic steatosis as examined by histology after 12 wk of being fed an HSFD, suggesting that this in vivo model can be useful in evaluating dietary effects on liver steatosis. Supplementation with either lycopene or APO10LA inhibited high-saturated-fatinduced hepatic steatosis incidence and severity in BCO2-KO male mice. Dietary lycopene supplementation mitigated steatosis despite the absence of BCO2 enzyme expression, suggesting that lycopene itself exhibits biological functions. It should be noted that APO10LA or lycopene supplementation tended to decrease hepatic TGs in BCO2-KO male mice, but the effects were not significant (P = 0.07 and 0.09, respectively). This discrepancy between our histopathologic steatosis grading and hepatic TGs was likely because of the regional distribution of steatosis in the liver and the small sample size that led to high variability within each experimental group. Therefore, we believe that the histopathologic steatosis grading of left liver lobes was more valid in evaluating hepatic steatosis.
In the present study, the biochemical and molecular analyses of liver and MAT suggest that lycopene and APO10LA have differential mechanisms of biological action in the BCO2-KO mice. APO10LA-mitigated steatosis and inflammation in BCO2-KO male mice coincided with induced SIRT1 protein, deacetylation of SIRT1 direct downstream target, and FoxO1, as well as with reduced SIRT1 protein cleavage. These results suggest that APO10LA modulated SIRT1 protein in a post-translational manner, which is similar to our previous findings (18). APO10LAmediated steatosis reduction and SIRT1-signaling activation in BCO2-KO mice were associated with decreased lipogenic SCD1 gene and protein as well as with increased AMPK activation and ACC phosphorylation. Importantly, Scd1 gene expression in APO10LA-supplemented BCO2-KO mice correlated with hepatic steatosis score and negatively correlated with SIRT1 protein. SIRT1-signaling activation protected mice from HSFD-induced steatosis and inflammation through inhibiting FA biosynthesis (32,33), promoting FA b-oxidation (32,33), inhibiting Scd gene induction (34), and activating AMPK (35). AMPK is an enzyme that inhibits ACC by phosphorylation to the inactive form of the enzyme (36). Hepatic ACC activation promotes FA biosynthesis and steatosis while inhibiting FA b-oxidation (34)(35)(36). Therefore, these data suggest that APO10LA reduced steatosis through SIRT1-signaling activation. Moreover, the APO10LA biological effects in the present study are in agreement with our previous findings using mouse strains with BCO2 expression (15,18), indicating that APO10LA may not be further metabolized by the BCO2 enzyme despite the presence of a 9#,10# double bond in its structure.
Hepatic FC accumulation was shown to sensitize rats to cytokine-mediated steatohepatitis (37). In the present study, the decreased hepatic FC in APO10LA-fed BCO2-KO male mice coincided with reduced hepatic inflammation and IL-6 expression (P = 0.06) as well as elevated HO-1 expression. APO10LA also induced hepatic cholesterol efflux and bile synthesis genes, without altering other genes involved in hepatic cholesterol metabolism. APO10LA supplementation stimulated HO-1 in the lungs of A/J mice (14). The SIRT1-activator resveratrol induced hepatic HO-1 (38), cholesterol efflux, and bile synthesis genes (39). In contrast, hepatocyte-specific deletion of SIRT1 in mice suppressed hepatic Cyp7a1 and promoted hepatic cholesterol accumulation, steatosis, and inflammation (33). Notably, APO10LA-induced hepatic Abcg5/8 in the present study was correlated with SIRT1 protein. Abcg5/8 and Cyp7a1 are also targets of the transcription factor farnesoid X receptor (40), but APO10LA did not modulate 2 other farnesoid X receptor-modifiable genes, Abcb11 and Cyp8b1. Therefore, these results suggest that APO10LA modulated Abcg5/8 and Cyp7a1 genes through SIRT1 activation.
In the present study, lycopene protected BCO2-KO male mice against HSFD-induced hepatic steatosis despite the absence of hepatic SIRT1-signaling modulations. Lycopene can accumulate in adipose tissue (10) and modulate adipose lipid/inflammationassociated signaling cascades (41,42). Evidence showed that the portal vein directly connects the liver with the visceral MAT (43,44), and visceral fat content is highly correlated with the degree of liver steatosis in humans (45,46). In the present study, the hepatic steatosis reduction by lycopene was associated and inversely correlated with induced PPARa-and PPARg-associated genes in MAT. PPARa activation can result in the net effect of reducing FA synthesis and secretion from cells (47). PPARg activation in adipose tissue increases adipocyte differentiation, FA sequestration, and FA utilization (48,49). Therefore, lycopene-mediated upregulation in MAT PPARa-and PPARg-associated genes may have promoted MAT FA utilization, and subsequently reduced lipid delivery from MAT to the liver. Nevertheless, it should be noted that lycopene effects on PPARa-and PPARg-mediated signaling were relatively modest; thus, this might explain why lycopene did not reduce final BW. Lycopene modulations in PPAR-associated signaling were observed in the adrenal glands, liver, and kidney of rodents in previous studies (16,50,51). During the preparation of this manuscript, a recent publication by Tan et al. (50) showed that 3wk-old BCO2-KO male mice on a lycopene-supplemented (250 mg/ kg diet) AIN-93G semipurified diet for 3 wk modulated hepatic PPAR-associated signaling that could potentially affect hepatic responses to metabolic, infectious, or chemical stress. However, the present study showed that lycopene in HSFD-fed BCO2-KO mice only modified PPARa-associated genes in MAT and not in the liver. This discrepancy between the present study and the findings published by Tan et al. may be the result of the differential experimental designs, diet composition, and/or lycopenesupplemented dosage/duration. Future research is needed to examine whether BCO2 expression can modify tissue specificity of lycopene modulations in PPARa-associated signaling.
The sex difference in BW gain and hepatic steatosis outcomes have previously been demonstrated in a number of mouse and rat models (52)(53)(54)(55). Indeed, the present study showed that HSFD feeding induced hepatic steatosis in male but not in female BCO2-KO mice. BCO2-KO females gained less BW and exhibited differential hepatic lipid metabolism compared to males. NAFLD patients are more likely to be males and postmenopausal women (56), and the difference in NAFLD prevalence may be attributed to the putative protective effects of estrogen. Estrogen treatment inhibited hepatic lipid storage in estrogen-deficient mice (57,58) and decreased BW in HSFD-fed C57Bl/6J female mice (59). Therefore, the sex disparity on HSFD-induced steatosis and BW in the present study may be the difference in circulatory estrogen between male and female BCO2-KO mice. Other factors including the elevated Abcg1 gene in BCO2-KO females compared to males may also protect females from hepatic lipid accumulation. This is because targeted disruption of Abcg1 in mice resulted in accumulation of neutral lipids and phospholipid in hepatocytes (60).
Female BCO2-KO mice in the present study had greater hepatic TC and CE and lower plasma TGs, TC, and non-HDL cholesterol but no differences in HDL cholesterol compared to male mice. These observations in BCO2-KO females were associated with lower Abca1, ApoA1, and ApoB gene expression, but greater Abcg1 gene expression than male mice. ApoB is required for the assembly and secretion of intestinal and hepatic apolipoproteins that transport CE and TG (61). Therapeutic inhibition of ApoB synthesis reduced plasma TC and TG in humans (61). Therefore, it is plausible that the reduced ApoB expression in BCO2-KO females compared to males may contribute to the sex disparity in plasma lipids. Hepatic cholesterol efflux to HDL requires ABCG1, whereas efflux to the HDL precursor ApoA1 requires ABCA1 (60). Because female BCO2-KO mice had elevated Abcg1 but reduced Abca1 and ApoA1 genes compared to males, these results may explain why male and female BCO2-KO mice had similar HDL cholesterol.
Because we aimed to isolate the differential effects of lycopene and APO10LA by using mice without BCO2 expression, we did not include wild-type mice in the present study. However, we acknowledge that the scope of the present study would be broadened with the inclusion of wild-type mice. In addition, BCO2 is believed to be the key enzyme involved in lycopene eccentric metabolism (12,13), and BCO2-KO mice had elevated lycopene accumulation in tissue and serum compared to wild-type mice (16,17). Lycopene-supplemented mice without carotene-15,15#-oxygenase (BCO1) expression did not have elevated lycopene accumulation in tissue and serum compared to wildtype mice (16,17). BCO1 also did not cleave lycopene in vitro in previous studies (10,62). However, recent studies suggested that lycopene could be metabolized to acyclo-retinal by purified recombinant human BCO1 (63) and by recombinant murine BCO1 (64). Further investigations with BCO1/BCO2-double KO mice are currently ongoing in our laboratory to examine if lycopenemediated functions are the effects of lycopene metabolites generated by BCO1 cleavage.
Taken together, the results from the present study suggest that both lycopene and APO10LA supplementation were effective in reducing high saturated fat-induced liver steatosis in mice where BCO2 expression was absent. The APO10LA-mediated ability to prevent hepatic steatosis and inflammation in male mice was associated with SIRT1-signaling activation in the liver, whereas the lycopene-mediated ability to prevent hepatic steatosis was associated with modulations in MAT lipid metabolism. Female BCO2-KO mice were less prone to HSFD-induced liver steatosis, weight gain, and elevated plasma TGs and TC than male mice.