*3.3. Metabolic Stability*

The results for the metabolic stability of HSG4112(S) and HSG4112(R) in rat (RLM), mouse (MLM), dog (DLM), and human liver microsomes (HLM) is shown in Figure 4 (left). The remaining amounts of HSG4112(S) at 120 min were 9.1 ± 3.2%, 6.9 ± 1.5%, 84.1 ± 6%, and 78.4 ± 0.5% in the RLM, MLM, DLM, and HLM, respectively. The half-life of HSG4112(S) was 15.5, 14.2, 664.8, and 424.6 min in the RLM, MLM, DLM, and HLM, respectively. The remaining amounts of HSG4112(R) at 120 min were 4.9 ± 3%,

**Figure 4.** Metabolic stability of HSG4112 in (**A**) rat, (**B**) mouse, (**C**) dog, and (**D**) human liver microsomes (*n* = 3). UDPGA, uridine 5′ -diphosphoglucuronic acid.

To investigate the effects of the phase II metabolism by glucuronidation, metabolic stability was tested after adding the glucuronidation cofactor, UDPGA. The results for the metabolic stability of HSG4112(S) and HSG4112(R) in rat, mouse, dog, and human liver microsomes with UDPGA are shown in Figure 4 (right). The remaining amounts of HSG4112(S) at 120 min were 11.9 ± 2.5%, 39.6 ± 1.5%, 12.8 ± 2%, and 78.9 ± 3.8% in the RLM, MLM, DLM, and HLM, respectively. The half-life of HSG4112(S) was 25.6, 64.3, 40.0, and 419.6 min in the RLM, MLM, DLM, and HLM, respectively. The remaining amounts of HSG4112(R) at 120 min were 3.2 ± 1.1%, 39.7 ± 0.9%, 66.5 ± 6.6%, and 52.2 ± 4% in the RLM, MLM, DLM, and HLM, respectively. The half-life values of HSG4112(R) were 9.1, 52.4, 226.8, and 150.5 min in the RLM, MLM, DLM, and HLM, respectively.

′

As glucuronidases are also predominantly distributed in the intestine, HSG4112 was tested in intestinal microsomes to confirm the metabolic stability results while considering glucuronidation. The results for the metabolic stability of HSG4112(S) and HSG4112(R) in rat (RIM), mouse (MIM), dog (DIM), and human intestinal microsomes (HIM) with UDPGA are shown in Figure 5. The remaining amounts of HSG4112(S) at 120 min were 87.9 ± 2.3%, 70.1 ± 6.1%, 37.9 ± 1.8%, and 76.7 ± 2.1% in RIM, MIM, DIM, and HIM, respectively. The half-life values of HSG4112(S) were 733.9, 247.9, 78.1, and 368.1 min in the RIM, MIM, DIM, and HIM, respectively. The remaining amounts of HSG4112(R) at 120 min were 95.8 ± 5.6%, 59.4 ± 6.3%, 65.9 ± 1.8%, and 75.5 ± 3.6% in RIM, MIM, DIM, and HIM, respectively. The half-life of HSG4112(R) was 1943.9, 136.7, 215.3, and 351.1 min in RIM, MIM, DIM, and HIM, respectively.

**Figure 5.** Metabolic stability of HSG4112 in (**A**) rat, (**B**) mouse, (**C**) dog, and (**D**) human intestinal microsomes (*n* = 3).

#### *3.4. Metabolite Profile*

To elucidate the main factor responsible for the stereospecific pharmacokinetic properties in rats, the metabolite profiles of HSG4112(S) and HSG4112(R) were investigated. When HSG4112(S) and HSG4112(R) were incubated in the rat liver microsomes for 60 min, the remaining amounts of HSG4112(S) and HSG4112(R) were 36.2% and 17.6%, respectively (Figure 6A). HSG4112(S) and HSG4112(R) were metabolized to yield metabolites by oxidation (M1a ~ M2b). M1c-1 and M1c-2 were

detected as one peak on a general C18 column but they were separated on a chiral column. Thus, M1c-1 and M1c-2 were supposed to be stereoisomers. The types of metabolites produced from the two isomers were similar, but the amounts produced were different. The predominant metabolites of HSG4112(S) were M1a and M2b, whereas those of HSG4112(R) were M1c-2 and M1a (Figure 6B). Subsequently, the metabolite profile was investigated in the rat plasma samples (on the 28th day of an oral dose of 100 mg/kg/day). A total of seven metabolites were detected (Figure 7); a metabolite with hydroxyl and carbonyl groups (M3a) and glucuronide conjugate (M4) were additionally detected besides the metabolites observed in liver microsomes. Their peak area values were plotted according to the time (Figure 8A), and the concentration of the parent (as racemate), M1c-1, and M1c-2 were quantitated (Figure 8B). M1c-2 was shown to be the predominant metabolite (Figure 8C). The accurate mass data for each postulated metabolite is tabulated in Table 5. The tentative structures of HSG4112 metabolites are provided as Supplementary data (Figure S2).

**Table 5.** Accurate mass data for HSG4112 and its postulated metabolites in rats. **−**

**Δ <sup>−</sup>**


−

**Figure 6.** Metabolic profiles of HSG4112(S) and HSG4112(R) in rat liver microsomes. (**A**) Percent remaining amounts of HSG4112(S) and HSG4112(R). (**B**) Relative peak areas of the metabolites generated from HSG4112(S) and HSG4112(R). HSG4112(S) and HSG4112(R) were incubated separately in rat liver microsomes for 60 min and the resulting amount of the parent drug and its metabolites was measured based on peak area.

**Figure 7.** Representative extracted ion chromatograms of HSG4112 and its metabolites in rat plasma.

**Figure 8.** *Cont*.

**Figure 8.** Metabolic profiles of HSG4112 in rat plasma after oral administration (100 mg/kg/day, 28th day, *n* = 3). (**A**) Plot for time-peak area of HSG4112 and its metabolites. (**B**) Plot for time-plasma concentration of HSG4112, M1c-1 and M1c-2. (**C**) Area under curve values of HSG4112 and its metabolites calculated from (**A**).

Meanwhile, the dog plasma sample showed only two metabolites (M1d and M4); M4, the glucuronide metabolite, was the predominant metabolite (Figure 9). Due to the unavailability of the sample and reference standards, the time–plasma concentration profile could not be obtained.

**Figure 9.** Representative extracted ion chromatograms of HSG4112 and its metabolites in dog plasma.

#### **4. Discussion**

In this study, concentration profiles of HSG4112(S) and HSG4112(R) after intravenous and/or oral administration of HSG4112 were investigated in rats and dogs, and the stereoselectivity in the metabolism of HSG4112 was investigated in vitro and in vivo. The resulting data showed a characteristic stereoselective pharmacokinetic pattern depending on the species.

When HSG4112(S) and HSG4112(R) were quantified in rat plasma, the concentration of HSG4112(S) generally measured higher, and, consequently, the AUC of HSG4112(S) was 4.9–7.8 times higher than that of HSG4112(R). This trend was also observed in the plasma samples from rats administered intravenously; the concentration ratio (S/R) over time was 1.4–4.8, and the AUC of HSG4112(S) was 2.7 times higher than that of HSG4112(R). Therefore, the systemic exposure of HSG4112(S) was significantly higher than that of HSG4112(R) in rats, indicating stereospecificity in the pharmacokinetics. This stereoselective pharmacokinetic property in rats is supposedly due to the stereoselective metabolism of HSG4112. The possible metabolic pathways of HSG4112 isomers in rats are presented in Figure S3. The metabolic stability data and the metabolite profiles in rat plasma showed that HSG4112(R) is metabolized more extensively than HSG4112(S), and their main metabolic pathways are different. The metabolic profile data suggested that the formation of M1c is predominantly responsible for the stereoselective pharmacokinetics of HSG4112.

When HSG4112 was orally administered to beagle dogs, the concentration ratio (S/R) over time was 0.06–0.24, and the concentration of HSG4112(R) measured at a much higher level. The AUC ratio of HSG4112(S) and HSG4112(R) was 0.12–0.13 in the oral administration group. In addition, when HSG4112 was intravenously administered to beagle dogs, the concentration ratio (S/R) over time was 0.3–0.9, and the AUC ratio of HSG4112(S) and HSG4112(R) was 0.5. Therefore, the systemic exposure of HSG4112(R) was significantly higher than HSG4112(S) in beagle dogs, indicating stereospecificity in the pharmacokinetics. Interestingly, the predominant isomer form in dogs was opposite to that of rats. The metabolic stability data for phase I metabolism did not exhibit any difference between HSG4112(S) and HSG4112(R). However, the metabolic stability data including glucuronide formation showed significant differences between HSG4112(S) and HSG4112(R); the metabolic rate of HSG4112(S) was much higher than that of HSG4112(R). This metabolism pattern was also obviously observed in dog intestinal microsomes, which also have high glucuronidase activities. This revealed that the stereoselective glucuronide formation is mainly responsible for the stereoselective pharmacokinetics of HSG4112 in dogs. This was also supported by the in vivo metabolism profile data of dog plasma; the glucuronide metabolite (M4) was found to be the predominant metabolite in dog plasma. Notably, the longer terminal half-life was observed after oral administration compared with IV injection in dogs. Thus, it is supposed that HSG4112 has a general-case pharmacokinetic property in rats but a flip-flop pharmacokinetic property in dogs. This phenomenon occurs when the absorption rate of drugs is much slower than the elimination rate [11,12]. In this case, bioavailability factors such as the absorption rate and extent mainly affect the terminal slope of oral administration than clearance and volume of distribution.

Glabridin, the motive compound of HSG4112, has the R configuration at carbon C-3. According to the previous report, the systemic bioavailability of glabridin was very low (about 7.5% in rats). Glabridin is mainly metabolized by glucuronidases in the intestine and liver, and the first-pass effects of glucuronidation are one of the main factors responsible for the low oral bioavailability of glabridin. Guo et al. investigated the tissue and species differences in the glucuronidation of glabridin. In their study, glabridin was metabolized to yield two glucuronide metabolites: M1 and M2. M2 formation was predominant whereas the formation of M1 was negligible in most species tested. The authors could not determine the exact position of glucuronidation for M1 and M2, but they suggested that M2 might have a glucuronide moiety at the C-4 hydroxyl group in the B ring based on the earlier reports. The C-4 hydroxyl group of HSG4112 is masked by an alkyl chain. This is the reason why the glucuronidation was not the main metabolic pathway in humans and rats in our study with HSG4112. It is notable that dog liver microsomes showed the highest value of the intrinsic clearance by M1 formation (glucuronidation at the C-2 hydroxyl group) in glabridin. This result agrees with our finding that the formation of the glucuronide metabolite predominantly occurred in dog liver and intestinal microsomes. The results by Guo et al. showed that the metabolic clearance by glucuronidation both for M1 and M2 was negligible in humans. This is also consistent with our data that the contribution of glucuronidation to the metabolic clearance of HSG4112 was minimal.

Stereoselectivity for HSG 4112 metabolism was also shown in human liver microsomes. The metabolism pattern in human liver microsomes is somewhat close to that in rat liver microsomes. Thus, this is supposedly due to CYP-mediated metabolism rather than glucuronidation. Although the metabolic rate in humans is lower than that in other species, the present data suggests that stereoselective pharmacokinetics by stereoselective metabolism (i.e., higher exposure of the (S)-isomer) could be shown in the clinical trial of HSG4112. Another factor that should be considered regarding

stereoselective pharmacokinetics is the possibility of interconversion between the isomers. To address this issue, HSG4112 isomers were analyzed after oral administration of each isomer to rats. When one isomer was administered, the other isomer was not detected (data not shown). This suggests that interconversion between isomers does not occur at least in rats. Further investigation should be followed in humans.

In conclusion, the present study demonstrates that HSG4112 showed species-specific stereoselective pharmacokinetics. Notably, the pharmacokinetic stereoselectivity of HSG4112 isomers showed opposing patterns between rats and dogs. This is because the major metabolic pathways involved in the clearance of HSG4112 are different in rats and dogs. Each metabolic enzyme may have different stereoselectivity. We presented the possible mechanisms for these stereoselective pharmacokinetic patterns by the metabolite profiling data in vitro and in vivo. These results will provide helpful information for understanding the pharmacological and toxicological effects of HSG4112, depending on its configurations. In addition, caution should be taken in extrapolating preclinical data with different experimental animal models to clinical data for humans.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/1999-4923/12/2/127/s1, Figure S1: Representative chromatograms of HSG4112(S) and HSG4112(R) in (A) rat (oral, 100 mg/kg, 6h) and (B) dog (oral, 100 mg/kg, 6h) plasma, Figure S2: Proposed chemical structures of HSG4112 metabolites, Figure S3: Postulated metabolic pathways of HSG4112 isomers, Table S1: Blood drain time points for oral pharmacokinetic analysis in rats (repeated dose, 100 mg/kg/day, 28th day), Table S2: Linearity of HSG4112(S) and HSG4112(R) in rat plasma (*n* = 3), Table S3: Intra-day and inter-day accuracy and coefficient of variation for determination of HSG4112(S) and HSG4112(R) in rat plasma, Table S4: Matrix effect, recovery and process efficiency data for HSG4112(S) and HSG4112(R) in rat plasma (*n* = 3), Table S5: Stability of HSG4112(S) and HSG4112(R) in rat plasma, Table S6: Linearity of HSG4112(S) and HSG4112(R) in dog plasma, Table S7: Intra-day and inter-day accuracy and coefficient of variation for determination of HSG4112(S) and HSG4112(R) in dog plasma, Table S8: Matrix effect, recovery and process efficiency data for HSG4112(S) and HSG4112(R) in dog plasma (*n* = 3), Table S9: Stability of HSG4112(S) and HSG4112(R) in dog plasma.

**Author Contributions:** Contributed to the research design: H.H.Y., S.-K.Y., K.K.; Conducted experiments: I.Y.B, Y.S.J.; Contributed to data analysis and interpretation: M.S.C.; Contributed to the preparation of the manuscript: I.Y.B., M.S.C., H.H.Y. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by the National Research Foundation of Korea funded by the Korea government (NRF-2017R1A2B4001814).

**Conflicts of Interest:** The authors declare no conflict of interest. S.K.Y. and K.K. are from Glaceum Inc., the company had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, and in the decision to publish the results.
