Hepatoprotective Bioactivity of the Glycoprotein, Antrodan, Isolated from Antrodia cinnamomea Mycelia

Antrodan, a protein-bound polysaccharide isolated from Antrodia cinnamomea mycelia, was demonstrated to exhibit significant anti-inflammatory bioactivity in vitro. However, its role in hepatic injury in vivo still remains unclear. We hypothesized that antrodan may have beneficial hepatoprotective effects. To verify this, a lipopolysaccharide (LPS)-Sprague-Dawley rat model was used. Antrodan protected against liver damage by suppressing LPS-stimulated serum glutamine-oxaloacetic transaminase (GOT), glutamic-pyruvic transaminase (GPT), interleukin (IL)-6, hepatic thiobarbituric acid reactive substances (TBARS), nitric oxide (NO), inducible NO synthase (iNOS) and nuclear factor (NF)-κB, and by effectively alleviating the downregulated hepatic superoxide dismutase (SOD), catalase, and glutathione peroxidase (GSH-Px). Hematoxylin-eosin staining revealed that antrodan at a dosage of 40 mg/kg was able to alleviate LPS-induced liver damage to a normal status. In addition, we identified the partial main architectural backbone of antrodan to have a 1→3 linear β-glycosidic backbone of mannan linked by β-1→3 glucosidic branches. Conclusively, antrodan can potentially ameliorate liver damage in vivo by suppressing oxidative stress induced by LPS.


Introduction
Beta-glucans are commonly found at high levels in fungi, yeast, oats, barley, and bacteria [1,2]. Their potential bioactivities were reported to involve immunomodulation [3,4]. anti-inflammation [5], anticarcinogenicity [6,7], enhancement of natural killer cells [8], resistance against bacterial and parasitic infections [9], antidiabetes, and reduction of associated cardiovascular risks [10]. Recently, the protein-bound polysaccharide K (trade name Krestin, or PSK) was approved in a phase 1 clinical trial of breast cancer by the US Food and Drug Administration in 2007 [11]. One of the important action mechanisms of PSK was described as being related to activation of antitumor immune responses [12]. Combining PSK with chemotherapy prolonged the survival rate of patients with gastric cancer, colorectal cancer, and small-cell lung carcinoma [13]. Generally terpenoids predominantly occur in Antrodia cinnamomea (syn. A. camphorata) (AC) fruiting bodies, in which 78 compounds have been identified [14]. AC fruiting bodies are effective against numerous diseases including hepatitis, diarrhea, abdominal pain, hypertension, and tumorigenic diseases [15]. To the present, the cited biological effects of AC are mostly limited to its triterpene content [15,14], while the hepatoprotective bioactivity of AC polysaccharides has not been examined.
b-Glucans can act on several immune receptors including Dectin-1, complement receptor (CR)3 and Toll-like receptor (TLR)-2/6, and can also trigger groups of immune cells including macrophages, neutrophils, monocytes, natural killer (NK) cells, and dendritic cells and enhance phagocytosis [4]. The literature demonstrates that even a slight difference among glucans in terms of the glycosidic linkages, higher-order structure, molecular weight, solubility, protein and lipid pendants, and/or higherorder aggregates can result in great differences in innate immune activities [16,17].
Decoking is thought to produce an extract comprising both triterpenes and a diversity of soluble polysaccharides, and we hypothesized that AC polysaccharides could contribute comparable, if not more-prominent, immunobioactivity that ultimately may be associated with certain promising hepatoprotective effects. To verify this, we carried out the present experiments.

Extraction and purification of antrodan and analysis of its glycosidic linkages
Antrodan, a kind of glycoprotein, was prepared from mycelia of A. cinnamomea according to our previous report [18]. In brief, defatted mycelial powder (1 kg) obtained by supercritical fluid extraction was refluxed with 20 L of double-distilled water (DDW) at 90uC for 2 h with constant stirring at 400 rpm. The extraction was repeated three times to deplete the water-soluble material. The residue was desiccated under vaccuo followed by extraction at a ratio of 1:10 w/v with a hot alkaline solution (pH 9.0) at 80uC for 1 h. This extraction was repeated three times. After being cooled, the extracts were combined and filtered with the aid of aspiration. To the filtrate was added 1 N HCl to pH 4.0. A 2-fold volume of ethanol (95%) was then added. The precipitated polysaccharides were collected and re-purified in 4 L of hot water (90uC). In the final step, the soluble polysaccharides were reprecipitated by the addition of a 3-fold volume of ethanol (95%). The solution was left to stand for 4 h. The precipitate was collected, lyophilized, and pulverized (this was called AC-II). AC-II was further subjected to gel permeation chromatographic separation according to Carbonero et al. [19]. Briefly, 1 mL 0.05 N NaOH was added to AC-II (10 mg) and heated to 50uC while being stirred. After being cooled, the solution was centrifuged at 13,0006g for 5 min to eliminate any undissolved residue. The supernatant was subjected to a Sepharose CL-6B column (,6id = 8263 cm) and eluted with DDW (adjusted to pH 11.0 with NaOH) at a flow rate 0.5 mL/min. The eluent was collected with a fractionator (Advantec SF-2120, Toyo Kaisha, Tokyo, Japan) at a velocity 1 tube/10 min. In total, 100 tubes were collected. The total sugar content of the eluent was determined, and the optical density (OD) at 280 nm was monitored. By referring to the OD at 280 nm, the main fractions (16 th to 40 th ) were selected and combined. The combined solution was lyophilized to yield a pure glucan, called ''antrodan'' in the text hereafter. To analyze the glycosidic linkage of the polysaccharide moiety (2 mg), partially methylated alditol acetates were prepared from permethyl derivatives by hydrolysis (with 2 M trifluoroacetic acid at 121uC for 2 h), reduction (with 10 mg NaBD 4 /mL for 2 h at room temperature), and acetylation (with acetic anhydride at 100uC for 1 h). A gas chromatographic-mass spectrometric (GC-MS) analysis was performed on an Agilent Gas Chromatograph 6890 (Santa Clara, CA, USA) connected to an Agilent 5973 Mass Selective Detector. Samples were dissolved in hexane and injected into the HP-5MS fused silica capillary column (,6id = 30 m60.25 mm, 625 mm, Agilent Technologies) at an injector temperature of 250uC. The helium carrier gas was operated at a flow rate of 1 ml/min. The oven, initially held at 100uC for 1 min, was programmed at a rate of 10uC/min to 200uC and held for 1 min at 200uC, and finally increased to 300uC at a rate of 15uC/min [20].

Animal experiment
The animal experiment was approved by the Institutional Animal Care and Ethics Committee of Hungkuang University (Taichung, Taiwan). Thirty male Sprague-Dawley rats, aged 6 weeks and weighing 265,287 g, were purchased from the BioLasCo Animal Center (Taipei, Taiwan). These rats were housed in stainless cages in an animal room maintained at 2062uC, and a relative humidity of 65%66%, with a 12/12-h light/dark cycle. These rats were fed ad libitum on basic granular chow (Fu-So Feed Stocks, Taichung, Taiwan) and water and acclimated in the animal room for the first week. The rats were randomly divided into five groups, with six in each group. The grouping involved the control, LPS control, antrodan control, LPS + low-dose antrodan (40 mg/kg) (antrodan L+LPS), and LPS + high-dose antrodan (80 mg/kg) (antrodan H+LPS) groups. Animals were separated and caged, with two or three rats in each cage. Antrodan was administered by gavage. The placebo group was fed basic chow only. Normal saline was used instead of antrodan in the LPS control. Access to chow and water was ad libitum. The entire experiment lasted 7 days. The initial and final body weights (BWs) were recorded. One day before being euthanized, LPS (5 mg/kg) (in saline) was intraperitoneally (i.p.) injected to induce acute hepatic injury. Then animals were anesthetized with CO 2 , and blood was collected from the hepatic portal vein and stored at 4uC. The heart, kidneys, and liver were excised, rinsed twice with phosphate-buffered saline (PBS), dewatered of soft skin tissues, weighed, and immediately immersed in liquid nitrogen for storage at 280uC for further use. The collected blood was centrifuged at 30006g for 15 min to separate the serum, which was used in the following biochemical analyses.
Analysis for serum glutamine-oxaloacetic transaminase (GOT) and glutamic-pyruvic transaminase (GPT) Serum levels of GOT and GPT were assayed with Colorimetrics Slides (Fuji Dri-Chem Slide GOT/AST-PIII and Fuji Dri-Chem Slide GPT/ALT-PIII; Fuji, Japan) according to instructions given by the manufacturer.

Analysis of serum IL-6 and nitrite/nitrate
Levels of serum IL-6 were determined by using a ELISA kit purchased from R&D Systems (Minneapolis, MN) according to the protocol provided by the manufacture. Briefly to a 96-well plate, 100 mL of capture antibody was added. The plate was tightly sealed and stored at room temperature (RT) overnight. After washes, the plates were blocked with Block Buffer at room temperature for 1 h. Authentic IL-6 or serum (100 mL) was added and left to react for 2 h at RT. Dilute detection antibody was loaded to each well, and the plate was incubated for an additional 2 h at RT. Avidin peroxidase (100 mL) was added and allowed to react for 30 min at ambient temperature. A substrate solution (ABTS liquid substrate solution) (100 mL) was added, and the mixture was allowed to react for 10 min while avoiding direct sunlight. The absorbance was read at 405 nm, with wavelength correction set at 650 nm using the ELISA reader (VersaMax, Molecular Devices, Sunnyvale, CA, USA). A calibration curve was similarly established using authentic IL-6. Serum nitrite (including nitrate) concentration was measured by a modified method of the Griess assay, described by Miranda et al. [21]. In brief, one hundred microliters of deproteinized serum samples were mixed with 100 ml of VCl 3 , rapidly followed by the addition of the Griess reagents (1% sulfanilamide dissolved in 5% phosphoric acid + 0.1% naphthylethylenediamine dihydrochloride, 1:1, v/v). The absorbance was immediately measured at 540 nm after the mixture being reacted for 15 min in dark at RT. Nitrite concentration was calculated using a NaNO 2 standard curve and expressed as micromoles per liter.

Protein extraction and quantification
The Bradford protein binding assay [22] was followed to determine the protein content in tissue. To 0.5 g of hepatic tissue, a 10-fold volume of phosphate-EDTA buffer (0.1 mM, pH 7.0) was added, homogenized for 20 min, and centrifuged at 30006g for 15 min. The supernatant (1 mL) was transferred to a microcentrifuge tube and centrifuged at 4uC and 12,0006g for 10 min. The protein content of the supernatant was determined (called Hgn hereafter).

Tissue superoxide dismutase (SOD) activity
The method of Marklund and Marklund [23] was followed to determine the tissue SOD activity. Briefly, to 40 mL of Hgn phosphate-EDTA buffer (50 mM, pH 7.0, 56 mL), a Triton X-100 solution (2%, pH 8.2, 96 mL) was added, mixed well, and centrifuged at 4uC and 12,0006g for 5 min. The supernatant was decanted and transferred to a 5-mL reaction vessel. Tris-HCl (3 mL, pH 8.2, 50 mM) was added and agitated for 5 min. To 10 mL of the mixture, methanolic pyrogallol (50 mM) was added, immediately mixed, and read at 325 nm. Readings were successively made every 15 s for a total period of 3 min. Water was used as the blank, and a similar protocol was conducted. The activity is expressed in U/mg protein. The following equation was used to calculate SOD activity: where DA is the change in the OD/min after adding water, DA S is the change in the OD/min after adding the sample solution, V t is the total volume of the reaction mixture (mL), 4.8 is the dilution extent, V s is the total sample volume (mL), P is the total protein content (mg), A 1 is the OD at 0 s after adding water, A 2 is the OD at 120 s after adding water, and t is the time (min).

Tissue catalase (CAT) activity
The method of Aebi [24] was adopted with slight modifications. Briefly, to 40 mL Hgn phosphate-EDTA buffer (50 mM, pH 7.0, 56 mL), a Triton X-100 solution (2%, pH 8.2, 96 mL) was added, mixed well, and centrifuged at 4uC and 12,0006g for 5 min. The supernatant (5 mL) was decanted and transferred to a 5-mL reaction vessel and diluted with phosphate-EDTA buffer (50 mM, pH 7.0, 5 mL). To the diluted solution (2 mL), H 2 O 2 (1 mL, 30 mM) was added and immediately mixed, and the OD was read at 240 nm. Readings were successively taken every 15 s for a total reading period of 3 min. CAT activity is expressed as U/mg protein.
The following equation was used to calculate CAT activity: where A 2 is the OD at 180 s, A 1 is the OD at 0 s, t is the time (min), V t is the total volume of the reaction mixture (mL), 4.8 is the dilution extent, , is the light path (cm), e 240 is the extinction coefficient of 0.0395 mM 21 cm 21 , V s is the total sample volume (mL), and P is the total protein content (mg).

Tissue glutathione peroxidase (GSH-Px) activity
The method of Lawerence and Burk [25] was followed with slight modifications. Briefly, to 100 mL of Hgn, 800 mL potassium phosphate buffer (1 mM EDTA, 1 mM NaN 3 , 0.2 mM NADPH, 1 U/mL GSH reductase, and 1 mM GSH) was added and mixed well. The mixture was left at ambient temperature for 5 min to facilitate the reaction. To this reaction mixture, 100 mL of 2.5 mM H 2 O 2 was added and mixed well. The OD was read at 340 nm. Readings were successively taken every 5 s for a period 3 min. The GSH-Px activity is expressed as U/mg protein.
The following equation was used to calculate GSH-Px activity: where A 2 is the OD at 180 s, A 1 is the OD at 0 s, t is the time (min), V t is the total volume of the reaction mixture (mL), 4.8 is the extent of dilution, , is the light path (cm), e 340 is the extinction coefficient of NADPH of 6.22 mM 21 cm 21 , V s is the total sample volume (mL), and P is the total protein content (mg).

Tissue thiobarbituric acid reactive substances (TBARS) level
The method of Buege and Aust [26] with slight modifications was used to determine the hepatic tissue TBARS level. Briefly, to 0.5 g of liver tissue, 5 mL of normal saline was added and homogenized. To determine the TBARS level, 2 mL of homogenate was transferred to a spiral tube, and successively 1.5 mL 0.01 N HCl and 0.5 mL 15% TCA +0.375% TBA +0.25 N HCl were added and agitated to mix it well. The mixture was heated to 100uC for 15 min, left to cool at ambient temperature, and centrifuged at 30006g for 10 min. The supernatant was separated and read at 535 nm. A calibration curve was established using trimethylolpropane as the reference compound and treated similarly. The hepatic tissue TBARS level was calculated using this curve.

Hematoxylin-eosin (H&E) staining
Liver tissue samples were collected at 24 h after LPS administration. After excision, the liver was cut into slices at a thickness of 0.5,1.0 cm and immersed in a 10% neutral formalin solution for 3 days. Specimens were embedded in paraffin and frozen at 2,8uC. The frozen slices were sliced with a rotary microtome to yield slices with a thickness of 4 mM. The microtomed slices were mounted onto microscopic glass slides, and treated with 50,60uC water. After being tempered overnight in an oven held at 37uC, the slices were stained with H&E, sealed with fat-soluble gel, and examined microscopically (magnification, 2006).

Western blot analysis of iNOS and NF-kB p65 in cytoplasm
Liver samples were homogenized and cytoplasmic proteins were isolated as following. One gram liver tissue was homogenized in 10 mL of ice-cold buffer A (10 mM HEPES, pH 7.9, 1.5 mM MgCl 2 , 10 mM KCl, 1 mM dithiothreitol, and 1 mM phenylmethylsulfonylfluoride) and incubated on ice for 10 min. The samples were then centrifuged at 8506g for 10 min at 4uC. The supernatants were discarded and the pellets were resuspended in 1 mL of buffer A with 0.1% Triton X-100 per mg of tissue, incubated for 10 min on ice and centrifuged as described above. The supernatant was removed and saved as the cytoplasmic fraction. The cytoplasmic protein was used for detection of inducible nitric oxide synthase (iNOS) and NF-kB p65 in Western blot analysis. Sample (30 mg of protein) was subjected to SDSpolyacrylamide gel electrophoresis (8% gel for iNOS and 12% for NF-kB p65). Proteins in the gel were transferred to a polyvinylidene difluoride membrane by electroblotting, and the membrane was incubated sequentially with primary antibodies that recognize iNOS (BD Biosciences, San Diego, CA; 1:2000), NF-kB p65 (Millipore, Billerica, MA; 1:500) and b-actin (Cell Signaling Technology, Danvers, MA; 1:5000) for 2 h, and horseradish peroxidase-conjugated secondary antibody for 1 h at room temperature. After washing with phosphate-buffered saline (PBST), goat anti-rabbit IgG horseradish peroxidase (HRP)conjugated secondary antibodies (Santa Cruz) (1:1000 dilution in PBST) were incubated for1 h at roomtemperature. Immunoreactive proteins were detected using an enhanced chemiluminescence kit (ECL, PerkinElmer, Waltham, MA), and the relative expression of the protein bands was quantified by densitometry with ImageQuant TL software (GE Healthcare).

Statistical analysis
Data obtained were statistically treated with a one-way analysis of variance (ANOVA). Tukey's test or a least significant different (LSD) test was used to analyze differences in significance. P,0.05 was considered to indicate a significant difference between groups.

Characteristics of antrodan
The original yield of the crude AC-II fraction was 12.0%60.3%, and the yield of the purified product was 9.2%60.4%, which is hereafter called ''antrodan''. The yield of antrodan was slightly lower than that previously reported [18]. Recently, we identified antrodan to be a glycoprotein with a molecular weight of 442611 kDa which contains 14.10% carbohydrates, 71.0% protein, and a profound content of uronic acid (152.660.8 mg/g) [18]. In order to determine the nature of the linkages among different monosaccharides in the polysaccha-ride moiety of antrodan, the reduced polymers were permethylated and subjected to a GC-MS analysis (Fig. 1A). In the analysis of sugar linkages, the compound 2,4,6-Me 3 -Manp was found to be the most abundant among all monosaccharide derivatives. The second most abundant was 2,4,6-Me 3 -Glcp (Table 1). From the overall relative amounts of these derived monosaccharides, we suggest that antrodan exhibited mainly an architecture of a mannose backbone connected to glucose branches by 1,3glycosidic linkages (Fig. 1B).

Variations of BWs and LWs and the LW/BW ratio
After 7 days of feeding, the apparent BWs did not significantly vary among all rat groups ( Table 2). The absolute LW was found to have been highly increased in the antrodan L+LPS group, but was substantially suppressed in the antrodan H+LPS group to a level comparable to that of the antrodan control (Table 2). However, the LW/BW ratio was found to have been significantly reduced by LPS treatment. Antrodan alone or the antrodan+LPS combination revealed no effect on the LW/BW ratio (Table 2).

Antrodan alleviated the in vivo LPS-induced suppression of hepatic antioxidant enzymes activities
LPS treatment suppressed the hepatic SOD level to 0.5660.17 U/mg protein after 24 h. In rats treated with antrodan alone, the level was 0.7460.06 U/mg protein. Thus, compared to control (1.5460.12 U/mg protein), LPS reduced SOD by about 77% (p,0.05) and antrodan by 52% (p,0.05) (Fig. 3A). The low dose of antrodan did not have any improving effect, while the higher dose apparently showed a higher recovery efficiency ( Fig. 2A).
After treatment for 24 h, LPS reduced the CAT activity by 71% compared to the control (46.1760.24 vs. 158.26614.29 U/mg protein). Antrodan alone suppressed the CAT level to 68.54615.74 U/mg protein, a 57% decrease (p,0.05). It is worth noting that antrodan L+LPS and antrodan H+LPS showed no better effect, and respective CAT activities were still 61.5564.68 and 46.2367.18 U/mg protein (Fig. 3B). GSH-Px activity was not affected by either LPS or antrodan. GSH-Px activities were 5.7660. 26, 4.8360.4, 5.1560.03, 6.1360.72, and 5.6460.15 U/ mg protein in the control, LPS control, antrodan control, antrodan L+LPS, and antrodan H+LPS groups, respectively. LPS inhibited hepatic GSH-Px. Antrodan alone seemed to have little effect on rescuing hepatic GSH-Px. However in counteracting LPS-induced GSH-Px inhibition, the lower dose of antrodan (40 mg/mL) seemed to be more effective than the higher dose (Fig. 3C).

Antrodan alleviated the in vivo LPS-induced suppression of hepatic TBARS
LPS significantly stimulated hepatic TBARS to 127.166 4.10 nmole/mg protein compared to 92.1160.70 nmole/mg protein in the control, yielding an increase of 38% (p,0.05) (Fig. 3D). To our surprise, antrodan alone further raised the hepatic TBARS to 150.1468.64 nmole/mg protein for a 63% increase (p, 0.05). In contrast, in the antrodan L+LPS and antrodan H+LPS groups, TBARS levels were significantly ameliorated to 107.48614.37 and 107.95619.84 nmole/mg protein, respectively. It seemed that a low dose of antrodan (40 mg/kg) had a full-strength alleviating effect (Fig. 3D).

Antrodan alleviated LPS-induced suppression of serum NO in vivo
LPS highly stimulated serum NO to 55.566.0 mM (Fig. 3E). Antrodan moderately alleviated this adverse effect at a dose of as low as 40 mg/kg and completely alleviated the effect at a dose of

H&E staining of hepatic tissues
The histopathological examination revealed no morphological alterations in either the control (Fig. 4A) or antrodan group (Fig. 4C). In contrast, LPS damaged hepatic tissues and increased neutrophils and lymphocytes, inducing hepatic necrosis and cell lysis, releasing a tremendous number of nuclei and eliciting degenerative death of hepatic cells (Fig. 4B) (indicated by arrows). Oral administration of antrodan daily for 7 days significantly improved the state of hepatic necrosis while the antrodan control group shown as non-toxic injury on liver (Fig. 4D-E). Antrodan was found to be dose-responsively beneficial in alleviating cytotoxicity. At the lower dose, antrodan-treated liver leucocytes aggregated around blood vessels, and leucocyte infiltration was found in hepatic cells (Fig. 4D). However, the phenomenon was much highly improved over that found in the LPS-treated group. Interestingly, the high dose of antrodan aggravated LPS-induced hepatic damage. Inflammation with neutrophil aggregation was very common (indicated by arrows in Fig. 4E).

Western blot of cytosolic NF-kB and iNOS
LPS highly upregulated cytosolic NF-kB and substantially downregulated iNOS activity (Fig. 4). Antrodan alone showed comparable levels of NF-kB and iNOS as the control. Both antrodan L+LPS and antrodan H+LPS moderately suppressed iNOS upregulation (Fig. 5). Amazingly, the higher dose of antrodan was found to be less effective than treatment with the lower dose (Fig. 5).

Discussion
Recently, we showed antrodan characteristically to be a proteinbound polysaccharide in which b-glucans occupied a higher percentage (14.20%) than a-glucans (1.45%). In addition, a large amount of uronic acid (152.6 mg/g) coexists in the complex [18]. The high content of uronic acid may contribute to antrodan's bioactivities. A high uronic acid content in polysaccharide conjugates was shown to exhibit stronger reactive oxygen species-scavenging activities [27]. Certain polysaccharide-protein complexes obtained from mushrooms are capable of stimulating the non-specific immune system, exerting antitumor activities through stimulating a host's defense mechanism [28]. A similar structural type glucoxylan-protein complex prepared from Hericium erinaceus was been found to have immunomodulatory activity [29].
Nonetheless, the b-glucan content of antrodan is slightly lower compared to yeast which has a b-(1R3) glucan content of 45.47% [18]. Importantly, the animal experiment revealed that antrodan was nontoxic and safe in nature, hence feasible for use in nutraceutical therapy (Table 2). Liu et al. [30] demonstrated that proteins in polysaccharides directly contribute to free radicalscavenging activities. Protein-bound polysaccharides extracted from fruiting bodies of Ganoderma atrum also revealed multiple strong antioxidant activities [31]. As mentioned, we showed antrodan to be a glycoprotein [18]. Now we found that linkages of the polysaccharide moieties in antrodan exhibited mostly backbones of 1R3 linear b-glycosidic chains of mannose linked by b-1R3 glucose branches (Table 1, Fig. 1B). These might contribute to certain specific biological functions of antrodan. Animal studies demonstrated that the 1R3 linear b-glycosidic backbone of bglucans in fact cannot be readily digested in vivo [4]. Most bglucans enter the proximal small intestine, and some are captured by macrophages. First, they are internalized and fragmented within cells and then transported by macrophages to the marrow and endothelial reticular system. There the reduced b-glucan fragments are eventually released by macrophages and taken up by other immune cells leading to a variety of immune responses [4].
IL-6 has been found both in local and systemic acute inflammatory responses. However, the main sources of IL-6 in vivo are stimulated monocytes/macrophages, fibroblasts, and vascular endothelial cells [32]. In liver tissues, different sources of IL-6 present different functions. The IL-6 in hepatocytes induces STAT3 activation and plays important roles in hepatoprotection and liver regeneration. In Kupffer cells, IL-6 induces transient STAT3 activation and contributes to the pro-inflammatory response. In sinusoidal endothelial cells or hepatic stellate cells, IL-6 induces STAT3 activation and subsequently promotes cell survival [33]. Therefore, IL-6 could have anti and proinflammatory functions and the circulating levels of IL-6 are directly and closely correlated with the severity of disease [34]. IL-6 in serum was determined in this study would be more corresponding to reality of LPS-induced inflammatory responses of rats. Direct contact of b-glucans present in fungal cell walls is sufficient to trigger the rapid synthesis and secretion of the IL-6 protein, a post-transcriptional regulator of IL-6 in response to fungal extracts mediated by the p38 mitogen-activated protein kinase pathway [35]. b-Glucans are also beneficial against diabetes and associated cardiovascular risks [10].
LPS exerts oxidative stress and damages rat livers [35]. Activities of alanine aminotransferase (ALT or GPT), aspartate aminotransferase (AST or GOT), alkaline phosphatase (ALP), lactate dehydrogenase (LDH), and levels of serum total bilirubin, total protein, TNF-a, and IL-1b all increased [34]. Consistently, we showed that LPS stimulated activities of serum GOT, GPT, IL-6, and NO ( Fig. 2A-D). Antrodan was found to be beneficial to the LPS-damaged liver ( Table 2) as evidenced by suppressed serum levels of GOT and GPT (Figs. 1B, 2A).
b-Glucan stimulates lymphocyte formation in the injured body, producing the cytokine, IL-1, and an immunoglobulin M (IgM) antibody and adjusting the immune function [4]. Hepatic levels of TBARS, protein carbonyl content (PCC), CAT, glutathione peroxidase (GSH-Px), and myeloperoxidase (MPO) were stimulated by LPS. Conversely, reduced glutathione (GSH) and SOD were suppressed by LPS [36]. Yokoyama et al. [37] demonstrated that LPS administration increased the conversion of oxidized cytochrome c into reduced cytochrome c in the perfusate, indicating that superoxide anions had formed in the hepatic sinusoids, suggesting that superoxide anions in hepatic sinusoids may be one of the pathogenic factors behind damage to epithelial cells of hepatic sinusoids caused by LPS [37]. LPS administration induced lipoperoxidation and depletion of antioxidant enzyme activities such as SOD, CAT, and GSH-Px activities [38]. We found that LPS downregulated hepatic SOD, CAT, and GSH-Px ( Fig. 3A-C), resulting in substantial elevation of the hepatic TBARS level (Fig. 2D). Astonishingly, although the higher dose (80 mg/kg) of antrodan showed a better effect for alleviating hepatic SOD (Fig. 3A), it was less effective for serum IL-6 ( Fig. 1C), hepatic CAT (Fig. 3B), and serum NO (Fig. 3E). In contrast, hepatic GSH-Px was almost completely ameliorated by antrodan (Fig. 3C).
One current hypothesis for the molecular mechanisms of septic shock is that enhanced NO production by mitochondrial (mt) NOS leads to excessive peroxynitrite (ONOO -) production and protein nitration in the mitochondrial matrix, resulting in mitochondrial dysfunction and contractile failure [39].
Antrodan showed a promising in vivo suppressive effect on LPSinduced production of serum NO. We speculated that the potent biological effect by antrodan of scavenging NO might be attributed to its high uronic acid content (Fig. 3E, Table 1). LPS increased (p,0.05) TBARS, 13,14-dihydro-15-keto-prostaglandin F(2a) and NO [40]. Surface chemiluminescence was suggested to be a useful assay to assess inflammation and oxidative stress in situ in the liver and skeletal muscles. Liver chemiluminescence in inflammatory processes and phagocyte chemiluminescence were found to spectrally differ from spontaneous liver chemiluminescence with increased emission at 440,600 nm [39]. Boveris et al. pointed out the possibility that NO 2 and ONOO 2 could participate in reactions leading to the formation of excited species [39].
While evidence for 'baseline' iNOS expression has been elusive, IRF1 and NF-kB-dependent activation of the iNOS promoter supports inflammation-mediated stimulation of this transcript. iNOS produces large quantities of NO upon stimulation, such as by proinflammatory cytokines (e.g., IL-1, TNF-aa, and interferonc) [41]. Induction of high-output iNOS usually occurs in an oxidative environment, and thus high levels of NO have the opportunity to react with superoxide leading to peroxynitrite formation and cell toxicity. These properties may define the roles of iNOS in host immunity, enabling its participation in antimicrobial and antitumor activities as part of oxidative bursts of macrophages [42].
The phenomena that antrodan alone actively raised the hepatic TBARS to a high level but was only moderately effective in counteracting LPS-induced TBARS elevation (Fig. 3D) imply dual bioactivities of antrodan to simultaneously act as both an oxidative stress-defensive (Fig. 3A, 3B) and oxidative stress-inducing agent (Fig. 2C, D). LPS induced hepatotoxicity in rats with subacute pretreatment, and part of its action mechanism was associated with iron sequestration from the plasma to the liver compartment [38]. H&E staining indicated that antrodan enhanced hepatic cell proliferation (Fig. 3C). Antrodan alleviated these cytotoxicities (Fig. 4A-D). Alternatively, the cause of serum NO upregulation was due to upregulation of iNOS, and the liver damage exerted by LPS was attributed to the upregulated iNOS and downregulated cytosolic NF-kB (Fig. 4). Antrodan moderately ameliorated these expressions (Fig. 5). Similar to the above-mentioned findings, a high dose of antrodan conversely aggravated LPS-induced hepatic damage where inflammation with neutrophil aggregation was very common (indicated by arrows in Fig. 4).
It is worth noting that the active polysaccharide fractionated from the fruiting body of Agaricus blazei exhibited main components of FI0, a b-(1R6)-; (1R3)-b-D-glucan; FA-1, an-a acidic (1R6)-; (1R4)-a-D-glucan; FA-1, an acidic b-(1R6)-; a-(1R3)-D-glucan; and FA-2, a b-acidic RNA-protein complex [43]. The purified fraction, HM3-G (MW 380 kDa) was shown to be associated with the highest tumoricidal activity. Its main sugar is glucose (90%). Structurally, its main component is (1R4)-alpha-D-glucan with (1R6)-beta branching, in a ratio of approximately 4:1 [43]. Extensive investigation of cell-surface carbohydrates in tumor cells revealed two opposing roles in tumor metastasis: depending on their structures: carbohydrates can either promote or suppress metastasis. N-Glycans (N-acetylglucosamine attached to asparagine) were shown to play several important roles in tumor metastasis [44][45][46]. On the other hand, Core2 O-glycans allow tumor cells to evade NK cells of the immune system and survive longer in the circulatory system, thereby promoting tumor metastasis. Core3 O-glycans or O-mannosyl glycans suppress tumor formation and metastasis by modulating integrin-mediated signaling [47].
In conclusion, a low dose of antrodan (40 mg/kg) can act as a promising hepatoprotective agent. In addition, we identified the partial main architectural backbone of antrodan to have a 1R3 linear b-glycosidic backbone of mannan linked by b-1R3 glucosan branches. The findings of certain adverse effects that might be caused by the higher dose of antrodan inspired us to conduct further studies on i) reducing the molecular weight (or molecular size); ii) eliminating most of the Core2 O-glycans, if any; and iii) eliminating the N-glycan content, if any. However, according to previous experiences [4], the strategy of reducing the molecular weight might be feasible to obtain a more potential candidate in treating LPS-induced inflammatory disease.