Characterization of hyperglycemia due to sub-chronic administration of red ginseng extract via comparative global proteomic analysis

Ginseng (Panax ginseng Meyer) is commonly used as an herbal remedy worldwide. Few studies have explored the possible physiological changes in the liver although patients often self-medicate with ginseng preparations, which may lead to exceeding the recommended dose for long-term administration. Here, we analyzed changes in the hepatic proteins of mouse livers using quantitative proteomics after sub-chronic administration of Korean red ginseng (KRG) extract (control group and 0.5, 1.0, and 2.0 g/kg KRG) using tandem mass tag (TMT) 6‐plex technology. The 1.0 and 2.0 g/kg KRG groups exhibited signs of liver injury, including increased levels of aspartate transaminase (AST) and alanine aminotransferase (ALT) in the serum. Furthermore, serum glucose levels were significantly higher following KRG administration compared with the control group. Based on the upregulated proteins found in the proteomic analysis, we found that increased cystathionine beta-synthase (CBS) and cystathionine gamma-lyase (CSE) levels promoted greater hydrogen sulfide (H2S) synthesis in the liver. This investigation provides novel evidence that sub-chronic administration of KRG can elevate H2S production by increasing protein expression of CBS and CSE in the liver.

www.nature.com/scientificreports/ to nitric oxide (NO) and carbon monoxide, which plays a role in the regulation of inflammatory responses, apoptosis, oxidative stress, and angiogenesis [14][15][16][17] . Additionally, H 2 S is reportedly related to the regulation of glucose metabolism 18,19 . H 2 S is now recognized as an important cellular signaling molecule due to its important functions in several aspects of human health and disease 20 . Despite the beneficial effects of KRG in various physiological states, the way KRG affects H 2 S levels has not previously been explored. Considering the importance of the liver for H 2 S production, it is important to explore how H 2 S affects glucose metabolism and liver injury after long-term overexposed KRG administration. Thus, in this investigation, we explored the protein dynamics in the livers of mice after sub-chronic KRG administration using proteomic analysis.

Results
Changes in clinical characteristics. In general, the daily recommended oral doses of ginseng for rodents is 500 mg/kg. These doses were calculated based on actual doses of red ginseng calculated for human beings (1.5-3.0 g/person/day) 21,22 . In this study, we administered KRG extract to mice orally every day for 28 days at a maximum of 4 times the recommended dose (0.5, 1.0, and 2.0 g/kg) to investigate the clinical effects of highdose KRG (Fig. 1A). We quantified the ginsenoside contained in the KRG extract used in the experiment in our previous study 23 . These dosages have been used previously in general toxicity assessments in rodent models 24,25 . After 4 weeks of administration, we measured several serum parameters to characterize liver function. It was determined that glucose levels increased significantly, and serum AST and ALT levels were also significantly higher in the 1.0 and 2.0 g/kg KRG groups, respectively, which indicated liver injury (Fig. 1B). Furthermore, the hepatic pathologic change was confirmed in H&E stained tissues (Fig. 1C). Histopathological results were dis-   Table S1. There were no other significant differences in terms of body weight, liver weight, or other serum parameters between groups.
Global protein profiling in mice administered with KRG. The process used for protein profiling in this investigation is presented in Fig. 2A. In the 3 group, we identified a combined total of 1005 proteins, where 881 of them were quantified based on comparative proteomic analysis (Table S2). All data contained technical duplicates and the mass error for all identified peptides was assessed, with the highest mass error being > 0.02 Da (Fig. S1A). We conducted a Pearson correlation analysis of the reporter ion intensities to determine the quantitative accuracy of MS-based proteomics, with a resulting of 0.8775 (Fig. S1B).
Classification of differentially expressed proteins after the administration of high-dose KRG. Proteins identified from the liver tissue of the 3 KRG groups were assigned a reporter ion ratio over the control group for tandem mass tag (TMT) analysis to identify the differentially expressed proteins (DEPs) associated with KRG-induced liver injury. We identified a total of 58 DEPs after comparing the 3 KRG groups and the control group (Table S3). A total of 19, 40, and 31 DEPs were identified at the 0.5, 1.0, and 2.0 g/kg KRG groups, respectively (log 2 ratio ≥ 0.585 (upregulated) or log 2 ratio ≤ − 0.585 (down-regulated)), for each KRG group over the control group (Fig. S1C). In addition, we employed one-way t-tests with a p value cutoff of 0.05 to identify proteins whose levels changed significantly after KRG administration. A Volcano Plot of DEP was illustrated by fold change (log 2 Difference) versus significance (-Log10 p value) using a threshold value of 0.05. Protein IDs in red were considered significantly up-or down-regulated using the Perseus software package (Fig. 2B). Based on these criteria, 9 upregulated proteins showed statistical significance ( Table 1).

Verification of cystathionine β-synthase (CBS) and cystathionine gamma-lyase (CSE) protein expression and H 2 S levels.
After exploring the role of each protein and its association with hyperglycemia after KRG administration for 4 weeks, we focused on the most upregulated protein, CBS, and validated the accuracy of the TMT-labeled quantitative proteomics results using Western Blot (Fig. S2A). Furthermore, we determined that CSE is also an important protein in this study even though the proteomics results include no upregulation of CSE in the KRG groups. This is explained by the abundance of CBS and CSE in the liver and their involvement in the endogenous production of H 2 S and its metabolism 26 . CBS protein expression levels were found to be significantly higher in all KRG groups (0.5 g/kg; KRG relative level to control 2.07 (p < 0.05), 1 g/kg KRG; relative level to control 2.49 (p < 0.01), 2 g/kg KRG; relative level to control 2.28 (p < 0.01)), supporting the quantitative proteomics analysis. Also, the relative ratio of CSE protein expression was 1.6-fold (p < 0.01) which was increased in the only 2.0 g/kg KRG group (relative level to control, 1.0) (Fig. 3A). These changes in protein levels support the belief that dominant H 2 S-producing enzymes are highly expressed in the liver after high-dose KRG administration. When we measured the H 2 S levels in the liver samples using ELISA assays to further characterize CBS and CSE proteins, a significant dose-dependent increase in H 2 S concentration was seen in 1.0 and 2.0 g/kg KRG mouse liver tissues (3.1 umol/L, (p < 0.001); 3.5 umol/L, (p < 0.001)) ( Fig. 3B). www.nature.com/scientificreports/ Protein expression levels and the association with oxidative stress and hyperglycemia through H 2 S. Phospho-NF-κB p65, NF-κB, TNF-α, and IL-6 expression levels were measured using Western Blot analysis to assess the inflammatory response to KRG over-dose (Fig. S2B). A significant increase in NF-κB p65 (relative level to control 1.88, (p < 0.01)), TNF-α (relative level to control 2.87, (p < 0.05)), and IL-6 (relative level to control 1.99, (p < 0.01)) was seen in the 2.0 g/kg KRG group compared to the control group (relative level to control 1.0) (Fig. 3C). Furthermore, to explore the role of H 2 S in the liver in terms of gluconeogenesis, we assessed the protein levels of phospho-5' AMP-activated protein kinase alpha (pAMPKα), AMPKα, and PEPCK, which are involved in the gluconeogenesis metabolic pathway (Fig. S2C). High levels of H 2 S activity in the liver  . Changes in protein levels with increased H 2 S and the effects on liver injury (C) and the gluconeogenesis pathway (D) in mouse liver following the administration of high-dose concentrations of KRG. Bands were quantified using a densitometer in the ImageJ program, and the H 2 S results are expressed as mean ± SE (n = 5-6). Statistical significance was set at * p < 0.05, ** p < 0.01, and ***p < 0.001 compared with the control based on an ANOVA (SPSS Statistics 23 software package, https:// www. ibm. com/ analy tics/ spss-stati stics-softw are).  3D). Interestingly, high-dose KRG increased PEPCK protein levels (2 g/kg KRG; relative level to control 2.81, (p < 0.05)) by increasing H 2 S levels. In summary, we demonstrated that KRG over-dose contributes to H 2 S synthesis, resulting in increased oxidative stress in the liver and hyperglycemia (Fig. 4).

Gene ontology and Kyoto encyclopedia of genes and genomes pathway analysis of the DEPs based on bioinformatics. We analyzed the gene ontology (GO) and Kyoto Encyclopedia of Genes and
Genomes (KEGG) pathways by calculating the Fisher's exact test p value for 47 upregulated DEPs in order to increase our understanding of the mechanisms by which the sub-chronic administration of KRG in 3 groups, at different doses, led to liver injury. Fig. S3 summarizes the top five annotations from the Fisher exact tests for each category. The GO-based distribution of upregulated proteins was also evaluated to characterize the liver injury caused by sub-chronic high-dose KRG administration in mouse liver tissue. The KEGG pathway analysis showed that steroid hormone biosynthesis and sulfur metabolism were significantly enhanced with KRG administration. The domain annotation enrichment from the InterPro results indicated that cytochrome P450 was present at significantly higher concentrations in mouse liver after receiving high doses of KRG.

Discussion
It has long been believed that ginseng is a nontoxic herbal medicine, but investigations have been conducted to determine its safety. Most of the previous ginseng research focused only on its protective effects as a medicinal herb for the treatment of a variety of medical conditions 27,28 . In this investigation, we showed for the first time that high levels of H 2 S may play an important role in regulating oxidative stress and hyperglycemia levels in the liver by sub-chronic high-dose KRG administration. To explore the mechanisms underlying the negative effects of high-dose KRG, an in vivo model was used for proteomic analysis with TMT labeling technology. We demonstrated that sub-chronic administration of high doses of KRG caused high serum AST and ALT levels and elevated blood glucose (Fig. 1B). Although we couldn't find reference studies be comparable to subchronic administration of high dose KRG to normal mice, previous studies on liver function decline following long-term administration or high-dose administration for 90 days were confirmed. For examples, subchronic administration for 90 days of the main KRG metabolite of ginseng saponin, compound K, in beagle dogs showed increased ALT level in groups with compound K 29 . And a single case study reported KRG 6 g/day dose for 12 weeks, the liver function test showed increased AST and ALT level compared to prior liver enzymes (6 month earlier) 30 . These previous results could be suggest possibilities of negative effects on liver enzymes due to sub chronic high-dose KRG administration. Although rare, as can be seen from the case of acute hepatotoxicity after consumption of ginseng-related products, the need for mechanistic research on hepatotoxicity derived from ginseng used as a health functional food is proposed. We listed 47 upregulated proteins to determine the biological mechanisms of KRG-induced hepatotoxicity. Interestingly, CBS was the most upregulated protein in all the KRG groups (Table 1) and H 2 S-generating reactions were catalyzed by CBS and CSE in the trans-sulfuration pathway 31 . Thus, the hyperglycemia caused by sub-chronic administration in the high-dose KRG groups may be associated with increased H 2 S synthesis via CBS and CSE in the liver, given that CSE is an important H 2 S-producing enzyme that can be upregulated by NO 26 . Several earlier studies that explored KRG treatment and its association with H 2 S showed, for example, that KRG treatment inhibited H 2 S in vivo and in vitro 32,33 . Choi et al. reported that KRG at doses of 50-100 μg/ ml decreased both CBS and CSE expression in human umbilical vein endothelial cells (HUVECs) 34 . In this study, the protein expression levels of CBS and CSE determined by immunoblotting (Fig. 3a) were increased by administration of high-dose KRG. That difference is considered because the KRG dose within the recommended range for pharmacological effects significantly lowered the expression of inflammatory mediators.
An investigation of the process implicated in H 2 S-induced inflammation and ROS, determined that high H 2 S concentrations over a short period of time triggered the toxicity of H 2 S via the inhibition of mitochondrial www.nature.com/scientificreports/ cytochrome c oxidase and mitochondrial respiration 35 . Another study reported that administration of 500 μM NaHS could increase ROS formation through the inhibition of cytochrome c oxidase and the depletion of GSH in rat primary hepatocytes, which could lead to hepatotoxicity 36 . H 2 S exposure activated the NF-κB pathway, resulting in an increase in the protein levels of NF-κB, TNF-α, and IL-10. These results collectively indicate that H 2 S can induce oxidative stress via redox homeostasis disorders in the liver. Tan et al. (2017) reported H 2 S levels in wild-type mice liver tissue of approximately 1 μM/L 37 , which is similar to our results for the control group (1.43 μM/L) (Fig. 3b). Another study reported that 5-week HFD in mice induced a significant increase in hepatic H 2 S production, which was associated with elevated levels of CBS and CSE expression 31 . H 2 S also reportedly stimulates gluconeogenesis and glycogenolysis, but inhibits glycogenesis and glycolysis, contributing to increased levels of glucose in the liver 18 . Overexpressed CSE in HepG2 cells stimulates H 2 S generation, resulting in attenuated glycogen storage. The signaling pathways for H 2 S are closely associated with gluconeogenesis and glucose production. These findings were supported by our immunoblotting results for glucose metabolism, including pAMPK and PEPCK (Fig. 3c). H 2 S activates PEPCK by strengthening the glucocorticoid receptors and blocking AMPK activity 18 and stimulates G6Pase and FDP expression through key gluconeogenic transcription factor S-sulfhydrating PGC-1α 19 . Furthermore, other studies have reported that lower levels of H 2 S are related to liver dysfunction and result in hepatic fibrosis and cirrhosis, whereas higher levels of H 2 S strengthen insulin resistance and diabetes 38,39 .
In conclusion, we demonstrated that high H 2 S levels may be central to liver injury and elevated glucose levels may be due to sub-chronic administration of high-dose KRG (Fig. 4). This toxic mechanism is also associated with increased protein expression of CBS and CSE in the liver. In this study, KRG-induced liver toxicity was observed for the first time, and the results help increase our understanding of the biological mechanisms underlying KRG toxicity. Nonetheless, additional research under different physiological conditions is required to further delineate the mechanisms involved in ginseng toxicity.

Methods
Study design. The male C57BL/6 N mice were obtained from Orient Co. (Seongnam, Korea) and randomly housed at 4 mice per cage. The mice were acclimated for 1 week under controlled laboratory conditions (temperature of 22 ± 2 °C, humidity of 55 ± 5%, and 12-h light/dark cycle) before the experiments, and fed standard rodent chow and tap water ad libitum. The mice (mean body weight 20.3 ± 0.6 g) were randomly divided into 4 groups (6 mice per group) before administration of KRG. Daily oral administration of KRG in the treatment groups (0.5, 1.0, and 2.0 g/kg) took place over 4 weeks. The KRG extract used in this study was obtained from Punggi Ginseng Cooperative Association (Punggi, Korea) and was prepared using a traditional process that involves repeatedly steaming and drying the roots, hot-water extraction, and concentration 27 . The 13 ginsenosides Rb1, Rb2, Rc, Rd, Re, Rf, 20(S)-Rh1, 20(S)-Rh2, Rg1, 20(S)-Rg3, F1, F2, and compound K were absolutely quantified by LC-MS/MS 23 . The mice fasted for 12 h with free access to water before sacrifice to obtain blood and liver tissue. All animal experiments and methods were conducted in accordance with the guidelines of the Institutional Animal Care and Use Committee of Kyungpook National University and carried out in compliance with the ARRIVE guidelines.
KRG preparation. We were provided with ginseng products by Punggi Ginseng Cooperative Association (Punggi, Korea). Thirteen ginsenosides were prepared according to a traditional routine process 40,41 . KRG is processed by placing the cleaned and sorted ginseng on bamboo or wooden shelves in a closed steam chamber and applying steam slowly. Steaming lasted approximately 50-90 min, depending on the size of the ginseng, until it was thoroughly prepared. Then the ginseng was transferred outdoors to cool and remove the moisture. Finally, the ginseng is moved to the baking room and baked at a low temperature until proper dryness is achieved. Serum chemistry. Blood samples were collected from the inferior vena cava and maintained at room temperature. The blood was centrifuged for 15 min at 4000 × g to obtain the serum samples. These samples were stored at − 80 o C until they were tested. Serum parameters including liver function levels and lipid parameters were analyzed at the Hoseo Toxicological Research Center (Asan, Chungcheongnam-do, Korea) using a Hitachi 7020 Chemistry Analyzer (Hitachi, Tokyo, Japan). Liver sample preparation for proteomic analysis. Liver tissues (n = 3) of mice from each group were washed twice with cold phosphate-buffered saline and homogenized to extract the protein with 1% SDS buffer (1% SDS, 2 mM EDTA, 10 mM Tris-HCl, pH 7.5) containing a protease inhibitor cocktail (Thermo Fisher Scientific Inc., Rockford, IL). The liver homogenates were centrifuged at 12,000 × g for 10 min at 4 °C and the supernatants were transferred to new tubes. All samples from each group were pooled to minimize sample biological variation. Protein was reduced with 15 mM dithiothreitol and incubated at 56 °C for 30 min and then alkylated with 15 mM iodoacetamide at room temperature for 30 min in the dark. To purify the protein, 10% trichloroacetic acid was added to the protein samples and incubated for 4 h at 4 °C. The protein pellets were then washed twice with ice-cold acetone. The protein samples were solubilized in 100 mM triethylammonium bicarbonate buffer; then, Trypsin (Promega, Madison, WI) was added and the samples were incubated overnight at 37 °C. After trypsin digestion, the final concentration of 1% TFA was added to stop enzyme activation. The samples were then centrifuged at 12,000 × g for 10 min, the digested peptides were collected, and the 100 μg samples were obtained. The samples were kept in − 80 °C until use.
Histopathology. Sliced  Identification of differentially expressed proteins. MaxQuant 1.5 integrated with the Andromeda search engine was used for searching MS/MS data 42 . Tandem mass spectra were analyzed against a total of 51,444 mouse sequences in the UniProtKB/Swiss-Prot database concatenated with the reverse decoy database and common contaminants. Trypsin was specified as a cleavage enzyme, allowing for two missed cleavages. Carbamidomethylation on cysteine was set as a fixed modification, whereas methionine oxidation and protein N-term acetylation were set as variable modifications. For protein quantification through TMT labeling, we calculated the ratio of reporter ions using the TMT 6-plex method at MaxQaunt 1.5. All other parameters in MaxQuant were set to default values. We excluded contamination to obtain a high-quality protein list. The differentially expressed peptides include peptides with FDR ≤ 0.01, score > 40, and absolute log2-fold-changes 0.58 (1.5-fold), which are calculated from the TMT reagent ion reporter ratio of different-dose KRG groups, compared to the control group. DEPs are defined as proteins with a fold change greater than 1.5 or less than 0.666 in terms of relative abundance.
Bioinformatics. The list of GO analysis, KEGG 43 , and InterPro were classified according to the top 5 enrichment annotations among the upregulated proteins. The basis for calculating enrichment values for the mouse proteins was achieved using DAVID 6.7 online software (https:// david. ncifc rf. gov/) 44,45 . The high enrichment was categorized by the EASE score and a modified Fisher's exact test p value below 0.05. Visualizations, such as the volcano plot, for proteomics results were drawn using Perseus 1.6 (http:// www. perse us-frame work. org) 46 .
H 2 S assays. A total of 20 mg of homogenated liver samples were centrifuged for 15 min at 1500 × g with cold PBS. The supernatant was then transferred into new tubes and assayed immediately or aliquoted and stored at − 80 °C. H 2 S (Kamiya biomedical company, Seattle, WA, USA) was measured as per the manufacturer's protocols. Briefly, the samples and standards were incubated at 37 °C for 1 h after mixing with 50 μL of conjugate solution in antibody pre-coated plates. Substrate solutions were then reacted in the plate for 10-15 min after 5 washes with wash buffer. The stop solution was added and absorbance was measured at 450 nm. All samples were quantified and standardized to the same protein concentration using the BCA kit and calculated per mg of pro-