Epidemiological studies have consistently shown that regular consumption of fruits and vegetables as well as whole grains is strongly associated with a reduced risk of developing chronic diseases, such as cancer and CVD, which are the top two causes of death in most industrialised countries(Reference Pomerleau, Lock and Knai1). These studies highlighted the unique chemical features of fruits, vegetables, spices and grains. Plant-based whole foods provide thousands of primary and secondary metabolites to the human diet that are absorbed into the body due to their low-molecular lipophilic nature. These phytochemicals are bioactive non-nutrient plant compounds that can be classified into various families, such as alkaloids, flavonoids, isoflavones, isothiocyanates, organo-sulphur compounds, capsaicinoids, phytosterols, carotenoids and terpenoids(Reference Liu2).
Particularly, terpenoids are perhaps the most diverse family of natural products synthesised from plants, serving a range of important physiological functions. Terpenoids are the major volatile constituents of essential oils that generate the ‘flavour fingerprint’ of vegetables, fruits and spices(Reference Goff and Klee3). The essential oils of plants, fruits and vegetables are natural, complex, multi-component systems composed mainly of terpenes in addition to some other non-terpene components. Formed via the isoprene pathway, these compounds are grouped into hydrocarbons (limonene), alcohols (menthol), aldehydes (citral), ketones (carvone), acids and esters (geranyl acetate). Since this pathway operates to a different extent in plants, many of the same chemical constituents are present in a wide variety of essential oils. A wide range of terpenoids has demonstrated pharmaceutical activity against human ailments such as inflammatory diseases(Reference Aggarwal and Rangari4), CVD(Reference Tirapelli, Ambrosio and da Costa5) and cancer(Reference Sethi, Ahn and Pandey6).
Our group previously reported the hepatoprotective activity of an ethanol extract of Rosmarinus tomentosus Huberth-Morat and Maire (Lamiaceae) in CCl4- and thioacetamide-induced cirrhotic rats(Reference Galisteo, Suárez and del Pilar Montilla7, Reference Galisteo, Suárez and Montilla8). Because reactive oxygen species are generated during CCl4-induced liver toxicity, we explored the free-radical-scavenging capacity of the most frequent terpenes in Rosmarinus spp. essential oil. The ethanol extract in Rosmarinus officinalis contains a variable mixture of phenolic diterpenes (carnosic acid, carnosol and 12-O-methylcarnosic acid), caffeoyl derivatives (rosmarinic acid) and flavones with other terpenes in lower quantities(Reference del Baño, Lorente and Castillo9). The present study contributes extensive data on the identification of β-caryophyllene (CAR) as a free-radical-scavenging agent and on its capacity to inhibit lipid peroxidation in vitro. CAR protected rat livers from CCl4 toxicity and reduced fibrosis and the gene expression of type I collagen α1 (Col1a1), transforming growth factor-β1 (Tgfb1) and tissue inhibitor of metalloproteinase-1 (Timp1). Ex vivo, CAR protected CFCS-2G cells, a rat hepatic stellate cell (HSC) line, from free radical damage and effectively down-regulated the expression of the Col1a1 and Timp1 genes. CAR is a volatile sesquiterpene found in large amounts in the essential oils of many herbal drugs, spice and food plants, including oregano, cinnamon, black pepper, carrot, bay laurel, grapes, melon, red blush grapefruit, strawberry, guava, ginger, among others. The present results demonstrated that CAR is an antifibrogenic candidate for the prevention and treatment of hepatic fibrosis. Since patients with liver fibrosis have an associated condition of malnutrition, our data support the need to increase the consumption of fruits, vegetables and whole grains as a source of bioactive non-nutrient natural compounds to maintain a healthy liver status.
Experimental methods
Chemicals
All chemicals (unless specified) were obtained from Sigma-Aldrich Química S.A., at the highest available grades.
In vitro cell culture
The CFSC-2G cell line has a similar phenotype to that of early-passage primary HSC(Reference Greenwel, Rubin and Schartz10). CFSC-2G cells were maintained in culture with minimum essential medium supplemented with 10 % fetal bovine serum, 0·1 % non-essential amino acids and 1·1 % penicillin/streptomycin. For each experiment, cells were incubated overnight with no additions and then treated with appropriate concentrations of CAR at 24 h before treatment with an Fe2+–ascorbate solution (final concentration 50–200 μm). Culture medium and cells were collected 6 h later.
Animals and treatments
Female Wistar rats (150–170 g) were housed in cages at constant temperature (22°C) and humidity (55 %) with a 12 h light–12 h dark cycle (06.00–18.00 hours). The animals were handled in accordance with recommendations of the American Physiological Society (Council of Europe, 1982). The animals were randomly divided into five groups of six individuals with free access to water and rodent chow. To produce acute liver damage, the experimental design consisted in a single CCl4 (0·15 ml/100 g body weight) intraperitoneal injection. Dose selection for CAR treatment was based on previously published studies(Reference Cho, Chang and Lee11). Similarly to epigallocathechin-3-gallate treatment, CAR administration was started in animals subjected to acute CCl4 on the day before CCl4 poisoning(Reference Zhen, Wang and Huang12). Body weight was monitored throughout the experiment. Rats were deprived of food for 12 h, anaesthetised and blood was collected from the abdominal aorta. The liver was removed, washed with cold saline, weighed and stored at − 80°C.
Liver histology and collagen content
A portion of the liver was fixed in 10 % formaldehyde, embedded in paraffin and cut into 5 mm-thick sections. The samples were stained with periodic acid-Schiff and fast green–neutral red. A semi-quantitative score for hepatic injury ranging from 0 to 5 was assigned to each rat according to the presence or absence of hepatocyte necrosis, nodular transformation, fibrous septae and biliary hyperplasia. Fibrosis severity was scored as normal (0), minimal (1), mild (2), moderate (3), marked (4) or severe fibrosis (5). Collagen content was measured in 15 μm-thick liver sections using a colorimetric method based on the selective capacity of two dyes (Sirius red F3B and fast green FCF) to bind to collagen and non-collagenous proteins, respectively, in the presence of picric acid(Reference Pérez, Suárez and Gómez-Capilla13).
Measurement of free radical-scavenging activity
Radical-scavenging activity against the 1,1-diphenyl-2-picrylhydrazyl (DPPH) radical was determined by using the method of Mensor et al. (Reference Mensor, Menezes and Leitão14). Scavenging activity against the hydroxyl radical was determined by the inhibition of chemiluminescence during linoleic acid peroxidation by the Fenton reaction, as described by Xun et al. (Reference Xun, Jingdong and Zhaonan15). Scavenging activity against the superoxide anion was assayed in non-enzymatic and enzymatic systems. First, the compounds were tested to inhibit the photochemical oxidation of o-dianisidine by riboflavin, as described by Bridges et al. (Reference Bridges, Scott and Belch16). Second, the enzymatic production of the superoxide anion was assayed by monitoring the reduction in cytochrome c (Reference Yin and Chan17). The inhibition percentage of superoxide anion generation was calculated according to the formula: percentage inhibition = (A control–A sample)/A control× 100. The inhibition of the formation of lipid peroxides by 5-lipoxygenase was performed as described by Liu & Pan(Reference Liu and Pan18).
Inhibition of lipid peroxidation
Liver microsomes were obtained as described by Suárez et al. (Reference Suárez, Ramírez-Tortosa and Gil19). Lipid peroxidation was induced in vitro in liver microsomes with a Fe2+–ascorbate solution (50–200 μm). The test compounds were added to liver microsomes 30 min before initiating lipid peroxidation. Thiobarbituric acid-reactive substances (TBARS) were assayed in the medium at 30 min after induction and expressed as nmol malondialdehyde/mg protein(Reference Suárez, Ramírez-Tortosa and Gil19).
Analytic determinations
Total bilirubin levels were determined with the standard 2,5-dichlorophenyldiazonium tetrafluoroborate diazo reagent, according to the manufacturer's instructions (Roche Applied Science). Plasma protein content was determined according to the method of Bradford(Reference Bradford20). Lactate dehydrogenase (LDH) was spectrophotometrically assayed to test cell viability (Sigma-Aldrich Química S.A.). Aspartate aminotransferase (ASAT), alanine aminotransferase (ALAT), alkaline phosphatase and γ-glutamyltranspeptidase activities were measured as described previously(Reference Galisteo, Suárez and del Pilar Montilla7, Reference Galisteo, Suárez and Montilla8). Enzyme units were expressed per mg of plasma protein.
Gene expression
Total RNA was extracted from the rat liver tissue or CFSC-2G cells. Briefly, 10 μg of total RNA were electrophoresed in agarose gels, transferred to nylon filters, fixed with a UV crosslinker and hybridised with an α-32P-deoxycytidine triphosphate-labelled complementary DNA probe specific for Col1a1 or Timp1 mRNA and, sequentially, with an α-32P-deoxycytidine triphosphate-labelled oligonucleotide complementary to glyceraldehyde-3-phosphate dehydrogenase (Gapdh) mRNA. Semi-quantitative PCR was used to measure Col1a1, Tgfb1 and Timp1 gene expression in the rat liver as described by Pérez et al. (Reference Pérez, Suárez and Gómez-Capilla13).
Statistical analysis
All data are presented as mean values with their standard errors. All variables were checked for normality and homogeneous variance by using the Kolmogorov–Smirnov and Levene tests, respectively. A two-way ANOVA was used to evaluate differences according to compound and concentration, and a posteriori Bonferroni tests were carried out for specific mean comparisons. One-way ANOVA and a post hoc mean comparison using the Bonferroni test were used to analyse the effects between CAR, vitamin E and control groups, and to assess concentration effects for specific compounds. A value of P <0·05 was considered significant. Data analyses were performed using SPSS for Windows 15.0 (SPSS, Inc.).
Results
β-Caryophyllene is a free-radical-scavenging constituent of Rosmarinus spp. essential oil
Based on published data on essential oil constituents in Rosmarinus spp.(Reference del Baño, Lorente and Castillo9), we first tested the inhibitory capacity of α-pinene, camphene, α-humulene, camphor, borneol, 1,8-cineole and CAR against liver microsome lipid peroxidation. Only α-humulene and CAR showed inhibitory capacity (data not shown), which was further assayed in a concentration-dependent test. At all concentrations, CAR had a significantly higher inhibitory action against lipid peroxidation than did α-humulene, probucol and vitamin E (Fig. 1). Probucol and vitamin E were included as positive controls for synthetic and natural antioxidant compounds. CAR showed scavenging activity against the hydroxyl radical (Fenton reaction) and superoxide anion (Riboflavin assay) in a concentration-dependent manner (Table 1). CAR also inhibited superoxide anion-dependent cytochrome c reduction in the hypoxanthine/xanthine oxidase system (Table 1). CAR inhibited lipoxygenase-dependent oxidation of linoleic acid in a similar manner to α-tocopherol. CAR showed greater free-radical-scavenging activity in comparison with probucol and greater superoxide anion- and similar hydroxyl radical-scavenging activities in comparison with α-tocopherol.
Antioxidant activity of probucol (
), β-caryophyllene (□), α-tocopherol (■) and α-humulene (
) on the Fe2+/ascorbate-induced lipid peroxidation of liver microsomes, expressed as the percentage inhibition of the production of thiobarbituric acid-reactive substances. Values are means, with their standard errors represented by vertical bars (n 6). a,b,c,dMean values with unlike letters were significantly different (P< 0·05).
Scavenging effect of probucol, α-tocopherol and β-caryophyllene (CAR) at different concentrations (Mean values with their standard errors, n 6)
IC50, half-maximal inhibitory concentration; DPPH, 1,1-diphenyl-2-picrylhydrazyl radical; XO, xanthine oxidase; LPO, 5-lipoxygenase.
a,b,cMean values within a row with unlike superscript letters were significantly different (P< 0·05).
Carbon tetrachloride-induced liver fibrosis was markedly suppressed by β-caryophyllene
A fourth experiment was designed to determine the protective effect of CAR on CCl4-induced liver injury. Pretreatment with CAR significantly reduced plasma ASAT, ALAT and LDH activities and total bilirubin levels in a dose-dependent manner (Table 2). Representative images of liver microscopic morphology are shown in Fig. 2. Histological analysis of the liver of CCl4-treated rats revealed marked fibrosis with hepatocyte necrosis, inflammatory cell infiltration, fibrosis and microvesicular steatosis. In contrast, CAR significantly reduced the liver fibrosis score. Image analysis showed that CAR reduced the liver collagen content (Table 2) and significantly ameliorated the histological structure of the liver compared with non-treated fibrotic rats, showing no signs of steatosis, fibrosis or hepatocyte necrosis (Fig. 2). Timp1, Tgfb1 and Col1a1 gene expressions were higher in the liver of CCl4-treated rats than in the liver of normal rats (Fig. 3). Furthermore, the CAR treatment significantly attenuated the CCl4-induced up-regulation of both genes. Hence, it was demonstrated that CAR protects the rat liver from CCl4-induced injury by reducing fibrosis and down-regulating Timp1, Tgfb1 and Col1a1 gene expressions.
Serum biochemistry of healthy rats (control), carbon tetrachloride-treated rats (CCl4) and carbon tetrachloride-treated rats receiving β-caryophyllene (CAR) at different concentrations (Mean values with their standard errors, n 6)
ASAT, aspartate aminotransferase; ALAT, alanine aminotransferase; LDH, lactate dehydrogenase; AP, alkaline phosphatase; BRB, total bilirubin.
a,b,cMean values within a row with unlike superscript letters were significantly different (P< 0·05).
Protective effect of β-caryophyllene on carbon tetrachloride-induced liver damage. (A) Sections of the liver of healthy rats, (B) carbon tetrachloride-treated rats and (C) carbon tetrachloride-treated rats that received 2 mg/kg of β-caryophyllene. Original magnification, 200 × .
Effect of β-caryophyllene (CAR) on the hepatic mRNA expression of the type I collagen α1 (Col1a1, □), transforming growth factor-β1 (Tgfb1, 
) and tissue inhibitor of metalloproteinase-1 (Timp1, ■) genes in carbon tetrachloride-induced liver fibrosis. Rats received different concentrations of CAR. Data are values of mRNA expression relative to glyceraldehyde-3-phosphate dehydrogenase (Gapdh). Values are means, with their standard errors represented by vertical bars (n 3). a,bMean values with unlike letters were significantly different (P< 0·05).
β-Caryophyllene protects CFSC-2G hepatic stellate cells against free radical damage
The next step was to assess whether CAR protected hepatic CFSC-2G cells against oxidative damage induced by a Fe2+–H2O2 solution. The degree of cell damage was measured with Trypan blue exclusion assay, intracellular TBARS products and enzymatic markers, present in cell media due to cytoplasmic leakage. Oxidative damage reduced cell viability to 93·3 % and increased intracellular TBARS products to 3·72 nmol/mg protein (Table 3). LDH, alkaline phosphatase, γ-glutamyltranspeptidase, ALAT and ASAT activities were increased in cell media due to free radical-induced cell damage. The CAR treatment increased cell viability and reduced enzyme activities and TBARS production. The present results indicate that CAR effectively protected CFSC-2G cells from free radical-induced cell damage.
Protection of β-caryophyllene (CAR) against oxidative stress-induced damage to CFS-2G hepatic stellate cells (Mean values with their standard errors, n 6)
TBARS, thiobarbituric acid-reactive substances; LDH, lactate dehydrogenase; ASAT, aspartate aminotransferase; ALAT, alanine aminotransferase; AP, alkaline phosphatase; γ-GT, γ-glutamyl transpeptidase.
a,b,c,dMean values within a row with unlike superscript letters were significantly different (P< 0·05).
* Control cells received vehicle alone (ethanol).
† Oxidative damage was induced in cells by addition of an Fe2+/ascorbate 50 μm/200 μm solution. Cells were incubated with α-tocopherol or CAR before oxidative damage.
β-Caryophyllene reduces oxidative stress-induced activation of CFSC-2G hepatic stellate cells
HSC play a key role in response to liver injury. Oxidative stress activates HSC that become the main source of the post-injury extracellular matrix deposit that results in liver fibrosis. An experiment was designed to confirm that CAR reduces HSC activation by reactive oxygen species and down-regulates the mRNA of the Timp1 and Col1a1 genes. Exposure of CFSC-2G cells to oxidative damage up-regulated 2·06-fold the expression of the Col1a1 gene and 3·12-fold that of the Timp1 gene in comparison with untreated cells (Fig. 4). The addition of CAR significantly reduced the expression of Col1a1 and Timp1 in a dose-dependent manner.
β-Caryophyllene (CAR) reduces oxidative stress-induced activation of CFS-2G hepatic stellate cells. Data of mRNA expression of the type I collagen α1 (Col1a1, □) and tissue inhibitor of metalloproteinase-1 (Timp1, ■) genes relative to glyceraldehyde-3-phosphate dehydrogenase (Gapdh) are presented. Oxidative damage was induced with a Fe2+/ascorbate 50 μm/200 μm solution. Cells were treated with CAR at the indicated concentrations before oxidative damage. Values are means, with their standard errors represented by vertical bars (n 3). a,b,cMean values with unlike letters were significantly different (P< 0·05).
Discussion
CAR is a natural volatile bicyclic sesquiterpene that is present in numerous herbs, spices and foods. CAR can be isolated from cinnamon, clove oil and copaiba balsam, all of which have been used as natural remedies. Because of its weak aromatic taste, CAR is commercially used as a food additive and in cosmetics. The Flavour and Extract Manufacturers Association granted CAR the status of ‘generally recognised as safe’, and it has been approved by the US Food and Drug Administration for food use due to its low toxicity(Reference Adams, Gavin and McGowen21). CAR has been shown to have anti-microbial(Reference Sabulal, Dan and J22), anti-inflammatory(Reference Aggarwal and Rangari4), cannabinoid(Reference Gertsch, Leonti and Raduner23), anti-carcinogenic(Reference Legault and Pichette24) and skin penetration-enhancing properties(Reference Cornwell and Barry25). Our group previously reported the hepatoprotective activity of an ethanol extract of R. tomentosus Huberth-Morat and Maire (Lamiaceae) in CCl4- and thioacetamide-treated cirrhotic rats(Reference Galisteo, Suárez and del Pilar Montilla7, Reference Galisteo, Suárez and Montilla8). The composition of the ethanol extract was unclear, but it contained most of the volatile terpenes present in its essential oil, of which CAR is an important component(Reference del Baño, Lorente and Castillo9).
Lipid peroxidation of cell membranes is the main damage caused by reactive oxygen species in cells. It is initiated by hydroxyl radicals and superoxide anions, leading to the formation of peroxyl radicals that ultimately propagate the chain reaction in lipids. In a first step, we explored the antioxidant activities of several constituents of Rosmarinus spp. essential oil. CAR was the only terpene to show a strong inhibitory effect on in vitro lipid peroxidation (Fig. 1). This inhibition was significantly higher than that of α-humulene, α-tocopherol or probucol. Given that different radical species initiate lipid peroxidation, we further tested the free radical-scavenging activity of CAR (Table 1). According to the present results, CAR showed scavenging activity against hydroxyl radical (Fenton reaction) and superoxide anion (riboflavin assay) in a concentration-dependent manner (Table 1). Its scavenging capacity was significantly higher than that of probucol. The scavenging properties of α-tocopherol and CAR were similar. However, both sesquiterpenes, CAR and α-humulene failed to scavenge DPPH, a molecule that contains a stable free radical. This result may explain why CAR has not been previously reported as an antioxidant, since the DPPH test is the most widely used technique to characterise the anti-radical capacity of compounds. Interestingly, CAR was able to markedly reduce the in vitro formation of lipid peroxides by 5-lipoxygenase. Taken together, our data indicated that CAR is a highly effective chain-breaking antioxidant in in vitro lipid peroxidation and possesses greater scavenging activities against reactive oxygen species than against stable organic radicals.
Based on these results, we hypothesised that CAR may protect from oxidative stress-induced injury. Hepatic fibrosis caused by CCl4 has been extensively used in experimental rat models. Hepatic responses in rats to acute CCl4 administration are shown to be similar to human hepatic fibrosis(Reference Pérez Tamayo26). CCl4 metabolism in the liver stimulates lipid peroxidation and the production of free radicals(Reference Basu27), causing necrosis of hepatocytes, inducing inflammation and further promoting the progression of hepatic fibrogenesis. In this mechanism, activation of Kupffer cells, where 5-lipoxygenase is basically expressed in the liver(Reference Titos, Clària and Planagumà28), and the subsequent release of pro-inflammatory mediators are an early step in the pathogenesis of fibrosis. Since CAR inhibited in vitro lipid peroxidation and 5-lipoxygenase activity, we tested its effect on CCl4 liver toxicity. Acute poisoning with CCl4 increased bilirubin concentration and ALAT, ASAT, alkaline phosphatase and LDH activities in rat plasma, decreased in plasma total protein values, and produced severe fibrosis of hepatocyte necrosis, inflammatory cell infiltration, fibrosis and microvesicular steatosis (Table 2; Fig. 2). CAR demonstrated hepatoprotective activity, markedly restoring the values of plasma markers of liver damage and reducing the degree of steatosis and the extent of liver fibrosis.
It is the exuberant deposition of matrix proteins after injury that causes liver fibrosis and eventually cirrhosis. The mechanism of hepatic fibrogenesis is related not only to an enhanced matrix biosynthesis but also to an inhibition of matrix breakdown(Reference Friedman29). In human subjects, chronic alcohol-induced damage to the liver significantly diminishes collagenolytic activity in parallel with a dramatic rise in TIMP-1 and TGF-β levels(Reference Iredale, Benyon and Arthur30). In the present experiment, CCl4 poisoning markedly increased the expression of the Col1a1, Tgfb1 and Timp1 genes. Protection by CAR against CCl4-induced liver fibrosis was associated with a reduction in liver collagen content, the histology score and the expression of the Col1a1, Tgfb1 and Timp1 genes (Fig. 3). These results indicate that CAR is a hepatoprotective compound that down-regulates the expression of profibrotic genes and reduces hepatic histological lesions.
Oxidative stress promotes the activation of HSC that are the main source of the post-injury extracellular matrix deposit resulting in liver fibrosis. Oxidative stress promotes HSC proliferation and up-regulates Col1a1, Tgfb1 and Timp1 gene expressions in HSC(Reference Galli, Svegliati-Baroni and Ceni31), and this process is modulated by antioxidants(Reference Orr, Leel and Cameron32, Reference Martucci, Ziulkoski and Fortuna33). Based on our previous results, we hypothesised that CAR may prevent liver fibrosis by reducing HSC activation. An experiment was designed to test whether CAR inhibited oxidative stress-mediated HSC activation. Reactive oxygen species have been shown to disrupt cell membrane integrity due to the oxidation of PUFA, which causes the spillage of cytoplasmic enzymes. As expected, exposure of CFCS-2G cells to a pro-oxidant solution reduced cell viability and increased LDH, ASAT, ALAT, alkaline phosphatase and γ-glutamyltranspeptidase leakage, and malonaldehyde concentration in culture media (Table 3). Oxidative damage up-regulated the expression of the Col1a1 and Timp1 genes in CFSC-2G cells (Fig. 4). Pretreatment of CFSC-2G cells with CAR significantly reduced oxidative stress-induced damage in a dose-dependent manner. The main result was the down-regulation of Col1a1 and Timp1 gene transcription exerted by CAR. In keeping with our hypothesis, the present results confirmed that CAR is an effective inhibitor of cell damage induced by reactive oxygen species and can prevent oxidative stress-mediated up-regulation of the Col1a1 and Timp1 genes in HSC.
In summary, we provide the first demonstration that CAR, a compound present in numerous plants and foods, is an effective inhibitor of lipid peroxidation, probably due to its free radical-scavenging activity against hydroxyl radicals, superoxide anions and lipid peroxides. Our data provide novel insights into the mechanisms of CAR in the prevention of liver fibrosis that may be related at least in part to its antioxidant activity. Since CAR is present in numerous plants, spices and foods, further experiments are needed to consolidate the potential application of CAR as a bioactive natural compound that reduces oxidative stress-mediated injury in human diseases.
Acknowledgements
None of the authors had a commercial interest, financial interest and/or other relationships with pharmaceutical, laboratory supplies or medical companies. The authors are grateful to the Spanish Ministry of Public Health for providing funds during the tenure of research. A. S. was involved in the data collection, data analysis, data interpretation, literature search and manuscript preparation. M. A. C., J. M. V., T. M. and M. I. T. were involved in the data collection, data analysis and data interpretation. A. S., M. J. F. and A. G. were involved in the study design and data interpretation.
