*3.3. Intestinal Metabolism of Compound K*

β To understand the metabolism of compound K, we measured PPD levels in the plasma, urine, and feces samples from rats and mice. Previous studies have demonstrated that compound K is metabolized to PPD via β-glucosidase in intestinal microbacterium (Figure 5) [1,13]. As shown in Figure 6, PPD was detected in the plasma and fecal samples from both mice and rats following intravenous injection of compound K (2 mg/kg). However, PPD was not detected in the rat and mouse urine samples. These results suggest that compound K was metabolized to PPD in both rats and mice. β

**Figure 5.** Structure and reported metabolic pathway of compound K to PPD. Glc: glucose; PPD: 20(S)-protopanaxadiol.

**Figure 6.** Representative multiple reaction monitoring chromatogram of compound K and PPD in plasma and feces samples collected from rats (**A**,**B**) and mice (**C**,**D**) following intravenous injection of compound K at a dose of 2 mg/kg.

The concentrations of compound K and PPD in plasma and feces samples are shown in Figure 7. Following intravenous injection of compound K, the PPD peak was detected in the mouse plasma samples between 4 and 24 h and in the rat plasma samples between 8 and 24 h (Figure 7A,B), suggesting that the metabolism of compound K to PPD might be a slow process. The PPD concentration in the plasma samples at 48 h was below the lower limit of quantification (LLOQ; 5 ng/mL). The metabolic ratio calculated from the PPD to compound K ratio was much greater in rat plasma than in mouse plasma (0.02–0.08 in mouse vs. 0.2–1.6 in rat; Figure 7A,B). Similarly, the ratio of PPD to compound K in the feces samples was also significantly greater in rats than in mice (Figure 7C,D). The recovery of PPD was 3.4-fold higher than that of compound K in rat feces, whereas the recovery of PPD was 1.2-fold higher than that of compound K in mouse feces. As shown in Table 4, compound K and its metabolite PPD were exclusively recovered from feces but not from urine. The higher percentage of PPD recovery in rats compared to mice suggest the higher metabolism of compound K to PPD in rats compared to mice (Figure 1 and Table 3).

**Figure 7.** Plasma concentration vs. time profile of compound K and PPD (a metabolite of compound K) following intravenous injection of compound K at a single dose of 2 mg/kg in (**A**) rats and (**B**) mice. Fecal excretion of compound K and PPD following intravenous injection of compound K at a single dose of 2 mg/kg in (**C**) rats and (**D**) mice. Data expressed as mean± standard deviation of four samples for each time point.



ND: Not detected. Data expressed as mean ± standard deviation from four rats and mice per group.

Next, we compared the pharmacokinetic features of compound K between the bile-cannulated rats and the non-bile-cannulated rats. In the bile-cannulated rats, biliary excretion of compound K was very fast and most of the excreted compound K was collected during the 0–2 h period (Figure 8B), which could be due to high and fast distribution of compound K to the liver (Figure 2). Because

of the fast distribution to the liver and biliary excretion, the plasma concentration of compound K disappeared rapidly and was best fitted to the one-compartment model. The elimination constant was estimated using 5 points and yielded 0.90 h−<sup>1</sup> with an *r* <sup>2</sup> value of 0.92 (Figure 8A). The T1/<sup>2</sup> was calculated as 0.78 ± 0.11 h in the bile-cannulated rats. In contrast, the AUC value for compound K in the non-bile-cannulated rats was significantly greater than that of the non-bile cannulated control rats (Table 5). The plasma concentration profile of compound K in the control group showed 2-exponential decay. The elimination constant was estimated using 3 points and yielded 0.29 h−<sup>1</sup> with *r* <sup>2</sup> value of 0.94 (Figure 9A). As results, *T*1/<sup>2</sup> was calculated as 2.40 ± 0.61 h in the non-bile cannulated rats, which is significantly greater than that in bile cannulated rats. The results suggest that the compound K enters into the systemic circulation from other compartments. − −

● ○ **Figure 8.** (**A**) Plasma concentration and (**B**) biliary excretion of compound K following intravenous injection at a single dose of 2 mg/kg in the bile-cannulated rats (•) and in the non-bile-cannulated rats (#). Dotted lines represent the regression line of the elimination constant from the plasma concentrations of compound K. Data expressed as mean ± standard deviation from four rats per group.

**Table 5.** Pharmacokinetic parameters of compound K in rats of bile cannulation and non-bile cannulation.


∞ *C*0: initial plasma concentration; AUC12 h or AUC∞: area under the plasma concentration-time curve from 0 to last sampling time or infinity; *T*1/2: half-life. Data expressed as mean ± standard deviation from four rats. \*: *p* < 0.05 compared with non-bile cannulated rat group.

**Figure 9.** Representative multiple reaction monitoring chromatogram of compound K and PPD in (**A**) bile collected for 0–2 h and 2–12 h from bile cannulated rats, (**B**) intestinal feces samples collected from rats of non-bile cannulation at 2 or 12 h, (**C**) plasma samples taken at 2 and 8 h from bile cannulated rats, and (**D**) plasma samples taken at 2 and 8 h from rats of non-bile cannulation following intravenous injection of compound K (2 mg/kg).

We compared the compound K and PPD levels in the plasma and bile or intestinal feces samples in the bile duct cannulated rats and the control rats (without bile duct cannulation) to determine whether biotransformation from compound K to PPD occurred in the intestine or the plasma. For this, bile samples were collected for 0–2 h and 2–12 h period after administration of compound K from the rats with bile duct cannulation. Since these bile samples were collected directly from the bile duct, the compound K in the sample was distributed to the liver and excreted through the biliary route and, therefore, did not reach the intestinal microbiota. In control group, however, the compound K in the bile sample was excreted via the bile duct, so it reached the intestine and was subjected to further metabolism to PPD. As shown in Figure 8, all of the bile samples collected over a 12 h period from the bile-cannulated rats contained only compound K without PPD (Figure 9A), suggesting no further metabolism of compound K to PPD occurred in the rat plasma, liver, and bile. However, the intestinal fecal samples collected from rats without bile cannulation at 2 and 12 h following intravenous injection of compound K showed peaks for both compound K and PPD, and the amount of PPD was greater in the 12 h samples compared to the 2 h samples (Figure 9B). In addition, PPD was detected in the plasma samples at 8 h in the control rats (Figure 9D), suggesting that the PPD detected in the 8 h sample (which was not present in the 2 h plasma sample) was absorbed from the intestine after metabolism of

compound K occurred. These results were consistent with the results from Figure 7A. Contrary to the control group, PPD was not detected in any of the plasma samples collected from bile-cannulated rats (Figure 9C). Taken together, the results suggest slow biotransformation of compound K to PPD in the rat intestine followed by the reabsorption of PPD into the systemic circulation.

#### **4. Discussion**

Although much is known about the pharmacological effects of compound K from in vitro studies and in vivo disease models, research on the pharmacokinetics as well as the relationship between the pharmacokinetics and drug response of compound K has been limited. This study aimed to understand the pharmacokinetic features of compound K and to compare its pharmacokinetic behavior in rats and mice, which are often used for disease models. Our study found that biliary excretion of compound K is a major elimination pathway, and fast and extensive liver distribution of compound K was demonstrated in both rats and mice. Oatp transporter-mediated hepatic uptake could be a possible mechanism for the dominant liver distribution compared to other tissues such as kidney, brain, spleen, and testis in both species (Figures 2 and 3), as the hepatic uptake of compound K was significantly inhibited by the pretreatment of rifampin, an Oatp inhibitor (Figure 4). Our results indicate that biotransformation of compound K to PPD occurs in the intestine rather than in the plasma or the liver, based on the comparison between the non-bile-cannulated rats and the bile-cannulated rats (Figure 9). The plasma PPD in rats and mice was then be reabsorbed from the intestine after the metabolism of the excreted compound K (Figures 1 and 9). In addition, the higher plasma AUC and *T*1/<sup>2</sup> of compound K in the non-bile cannulated rats compared with those in bile cannulated rats (Table 5) suggests that the excreted compound K was reabsorbed from the intestinal lumen. The *T*1/<sup>2</sup> of compound K in non-bile cannulated rats was significantly greater (2.40 ± 0.61 h) than in the bile-cannulated rats (0.78 ± 0.11 h). In addition, the *T*1/<sup>2</sup> of compound K in the control rats calculated from the plasma concentration profiles for 48 h was 7.3 ± 0.4 h (Table 3), which is much longer than the T1/<sup>2</sup> calculated from the plasma concentration profiles for 12 h. Also, the compound K plasma concentrations in the bile-cannulated rats could be fitted to a 1-compartment model while the plasma concentrations in the non-bile cannulated control rats showed 2-exponential decay (Figure 8A). Moreover, the compound K plasma concentrations increased or maintained at 2–4 h after sharply decreasing in the 0–2 h period, and then showed a slow decrease over the 4–48 h time period in both rats and mice (Figure 7A,B). This pattern could be attributable to the continuous reabsorption of compound K. The lipophilicity (LogP value 3.85 for compound K; 5.53 for PPD) and moderate permeability (0.5–2 × 10−<sup>6</sup> cm/s for compound K; 1.15 × 10−<sup>6</sup> cm/s for PPD) of compound K and PPD in Caco-2 cells also support the possibility of compound K and PPD reabsorption [3,7,14,15].

The features that differed most significantly between the rats and the mice were the higher plasma concentration (C<sup>0</sup> and AUC) of compound K in mice and the greater fecal recovery of PPD in rats. The percent recovery of the parent form (compound K) in mouse feces was much higher than in rat feces (13.8% in rats vs. 28.4% in mice), while the total fecal recovery (sum of compound K and PPD) was similar for both mice and rats (60.4% in rats vs. 62.8% in mice), suggesting that the elimination process of compound K could differ between rats and mice in addition to the difference in Vd between rats and mice. Multiple previous studies have shown that the tri- or four-glycosylated PPD-type ginsenosides (major components in red ginseng; Rb1, Rb2, Rc, and Rd) have been metabolized to compound K (monoglycosylated PPD-type ginsenoside) and further hydrolyzed to PPD, the final metabolite of the PPD-type ginsenosides, in the presence of lactic acid bacteria and gut microbiota [13,16]. Previous studies have reported that Bacteroides sp., Eubacterium sp., and Bifidobacterium sp. could potentially be involved in compound K metabolism and that subjects who have a higher composition of Bacteroides sp., Eubacterium sp., and Bifidobacterium sp. strains showed higher metabolism of Rb1 to compound K [1,10]. Similarly, differences in the composition of the intestinal microbiota in rats and mice could lead to the different rates of metabolism of compound K to PPD. Wang et al. identified the predominant bacterium in human and animal fecal samples [17]. In human fecal samples, 55% of colonies were

identified as Bacteroides sp. but Eubacterium sp. and Bifidobacterium sp. were also present in smaller proportions. In mice and rat fecal samples, Bacteroides sp. showed relatively low expression compared to the human samples. Eubacterium sp. and Bifidobacterium sp. also showed lower expression than other species (Clostridium, Fusobacterium, and Peptosreptococcus sp.). The data suggests that there are species-dependent factors that play a role in the gut metabolism of ginsenosides in mice, rats, and humans. Kim et al. [10] reported that human subjects who have a higher proportion Bacteroides sp. in their fecal microbiota showed 6-fold higher metabolic activity of compound K than the subject group that had a smaller proportion of Bacteroides sp. Choi et al. [7] reported that inter-subject variability in gut metabolism of compound K rather than the intestinal absorption of compound K may contribute to the large inter-individual variations in plasma compound K concentrations. Therefore, differences in the gut metabolism of compound K could also explain the variability of the compound K pharmacokinetics between species. In addition, cytochrome P450 3A-mediated metabolism of PPD has been reported in human plasma and urine samples. Multiple oxidized PPD metabolites were identified from human plasma and urine samples, and cytochrome P450 3A is thought to be involved in this process [18,19]. This metabolism of PPD in the liver microsomes could explain the unrecovered portion of compound K and PPD at 48 h following intravenous administration of compound K in this study.

Collectively, the proposed enterohepatic circulation of compound K and PPD (Figure 10) could explain how compound K shows efficacy in vivo despite its fast and exclusive biliary excretion. The distribution of compound K into the liver could be a possible link to the hepatoprotective effect of compound K. However, the therapeutic use of compound K in other tissues and the oral administration of compound K and PPD may be limited because of poor aqueous solubility (33 µg/mL for compound K; <50 ng/mL for PPD) and P-gp-mediated efflux [4,14,15]. The use of nanocrystals for PPD formulation improved oral bioavailability and brain delivery [15]. The use of the metabolism inhibitor piperine [20] in the formulation of PPD and the use of P-gp inhibitor, α-Tocopheryl polyethylene glycol 1000 succinate (TPGS) [21], in the formulation of PPD and compound K enhanced oral absorption and anticancer efficacy of these PPD and compound K [22,23]. These approaches may provide a strategy for developing formulations for compound K and PPD by modulating their pharmacokinetic features. α

**Figure 10.** Proposed pharmacokinetic pathway of compound K following intravenous injection. Compound K underwent several steps: (i) intravenous entry of compound K into systemic circulation, (ii) Oatp-mediated hepatic uptake of compound K, (iii) biliary excretion of compound K into the intestinal lumen, (iv) metabolism of compound K into PPD in intestine, and (v) the absorption of compound K and PPD from intestine in blood.
