Introduction

Drugs are absorbed into the body as parent compounds or their metabolites1. Metabolites produced during drug metabolism can have pharmacological effects or toxicity. Therefore, it is essential to study them2. In particular, if activity is expected to result from metabolic conversion of the drug, understanding the pharmacokinetic properties of the biotransformation process of the drug helps ensure its safety and utility in clinical therapy3.

Korean red ginseng (KRG; Panax ginseng C.A. Meyer) is an extensively used in traditional oriental medicine for both preventive and therapeutic purposes4,5. It is valued as an adaptogen due to its ability to normalize body functions and strengthen systems damaged by stress6,7. Previous studies have reported its immune-enhancing effects, rejuvenating and anti-fatigue properties, and positive impact on blood flow as KRG’s major biological activities8,9,10. Major active components of KRG are ginsenosides, a diverse group of steroidal saponins. They are divided into the following two categories according to their hydroxylation position on the core triterpene saponin structure: 20(S)-protopanaxadiol (20(S)-PPD) (ginsenosides Rb1, Rb2, Rc, Rd, compound K, and PPD) and 20(S)-protopanaxatriol (20(S)-PPT) (ginsenosides Re, Rg1, Rf, Rh1, and PPT)11. In particular, PPD-type ginsenosides exhibit an array of pharmacological effects, including anti-inflammatory, anti-oxidant, and anti-cancer activities12.

Most natural products are metabolized by intestinal bacteria or structurally transformed by liver enzymes to produce metabolites13. PPD-type ginsenosides are partially metabolized to compound K by intestinal bacteria and absorbed into the blood of rats and humans14,15. In humans, ginsenosides Rb1, Rd, Rg1, and Rc are excreted in the urine within 3 h and do not remain in the blood, but compound K, which is produced by the intestinal microflora, appears in plasma 8 h after consuming ginseng lacking compound K16. Plasma compound K is biliary excreted and metabolized into PPD in the intestine, and the intestinal PPD metabolite is reabsorbed into systemic circulation17.

The aim of this study was to evaluate the actual in vivo exposure and accumulation of each active ingredient following single and repeated administration of KRG and its active ingredient, compound K, to experimental animals. Blood was collected at each time point after administration, and the concentrations of ginsenosides in the plasma were quantified using liquid chromatography tandem mass spectrometry (LC-MS/MS). In addition, to further expand the pharmacokinetic relevance of the study, exploratory in silico analyses were conducted to investigate potential molecular mechanisms underlying the anti-inflammatory effects of compound K.

Results

Analytical method validation

The 20(S)-PPD-type ginsenoside Rb1 undergoes metabolic conversion to Rd → F2 → compound K → PPD via the hydrolytic pathway (Fig. S1). Under the established analytical conditions, no endogenous interference was observed at the peaks of Rb1 (RT 4.6 min), Rd (RT 5.2 min), or berberine (internal standard, IS; RT 1.3 min) in blank plasma, regardless of whether spiked analytes were present or in authentic plasma samples. Similarly, no endogenous interference was detected for compound K (RT 2.3 min) or digoxin (IS, RT 1.1 min) (Fig. S2). Ginsenosides Rb1, Rd, and compound K exhibited linearity over the concentration range of 0 − 200 ng/mL, with correlation coefficients (R2) exceeding than 0.99 (Table S1). The limits of quantification (LOQ) for ginsenosides Rb1, Rd, and compound K was validated at 0.5 ng/mL using LC-MS/MS analysis. Intra- and inter-day accuracy and precision were evaluated at three concentrations: 1, 5, and 25 ng/mL. Intra-day accuracy ranged from − 1.65 to 13.10%, with precision between 0.29% and 14.05%. Inter-day accuracy ranged from − 2.57 to 4.04%, with precision ranging from 0.28 to 12.34%, indicating that the method’s suitability for analysis. Detailed parameters are presented in Table S2. Additionally, as summarized in Table S3, the recoveries of the target analytes using liquid-liquid extraction ranged from 82.1 to 102.0%. Matrix effects ranged from 90.9 to 105.0%, indicating no significant matrix effect. The stability of the target analytes after 24 h at room temperature ranged from 87.5 to 108.2%, confirming that all target anlaytes in rat plasma were stable under the experimental conditions.

Pharmacokinetic study 1: Pharmacokinetic study after single or repeated oral administration of KRG in rats

For both Rb1 and Rd (Fig. 1; Table 1), the maximum plasma concentration (Cmax) and area under the curve (AUC), measures of body exposure, significantly increased with repeated dosing. Additionally, the elimination half-life (T1/2) of the repeated administration group (8 wks; Rb1, 26.9 \(\:\pm\:\) 22.4 h; Rd, 42.4 \(\:\pm\:\) 18.6 h) was similar to that of the single administration group (Rb1, 25.1 \(\:\pm\:\) 12.9 h; Rd, 26.9 \(\:\pm\:\) 22.4 h).

Fig. 1
figure 1

Mean plasma concentration vs. time profile of ginsenoside Rb1 (−●−) and Rd (−■−) following oral administration of Korean red ginseng at 2 g/kg to rats over different dosing periods; (A) single, (B) 2 wks, (C) 4 wks, and (D) 8 wks. Data expressed as mean ± S.D. of five samples for each time point.

Full size image
Table 1 Pharmacokinetic profiles of ginsenoside Rb1 and Rd in plasma of rats (n = 5 per group) according to the duration of oral administration at 2 g/kg of Korean red ginseng.
Full size table

For compound K (Fig. 2; Table 2), plasma concentrations were observed at the LOQ from the 4 wks repeated dose group, albeit limited to a few samples. The Cmax and AUC of compound K were not significantly different between the 4 and 8 wks dosing groups.

Fig. 2
figure 2

Mean plasma concentration vs. time profile of compound K following oral administration of Korean red ginseng at 2 g/kg to rats for 4 wks (−●−) and 8 wks (−■−). Data expressed as mean \(\:\pm\:\) S.D. of five samples for each time point.

Full size image
Table 2 Pharmacokinetic profiles of compound K in plasma of rats (n = 5 per group) according to the duration of oral administration at 2 g/kg of Korean red ginseng.
Full size table

Pharmacokinetic study 2: Pharmacokinetic study of compound K in rats

For comparative evaluation, a pharmacokinetic study of compound K was performed in rats using the available serial blood draws. To compare changes in drug exposure parameters (Cmax and AUC) according to the dosing duration at two different dose levels (Fig. 3; Table 3), it was observed that in the single-dose group, systemic exposure increased with increasing dose (Cmax: 5 mg/kg, 6.2 \(\:\pm\:\) 4.1 ng/mL; 10 mg/kg, 16.6 \(\:\pm\:\) 15.3 ng/mL; AUC: 5 mg/kg, 25.8 \(\:\pm\:\) 13.6 ng*h/mL; 10 mg/kg, 62.1 \(\:\pm\:\) 49.9 ng*h/mL). However, due to the small sample size and high inter-individual variability, statistical significance was not observed. In contrast, after 4 wks of dosing, the difference in systemic exposure between dose levels was greater than that seen in the single-dose group (Cmax: 5 mg/kg, 31.5 \(\:\pm\:\) 4.9 ng/mL; 10 mg/kg, 93.5 \(\:\pm\:\) 31.0 ng/mL; AUC: 5 mg/kg, 106.5 \(\:\pm\:\) 13.5 ng*h/mL; 10 mg/kg, 438.4 \(\:\pm\:\) 170.9 ng*h/mL), and this was statistically significant. Additionally, at the same dose, an increase in Cmax and AUC was observed with increasing duration of administration. The accumulation factor (the degree to which a drug accumulates in the body with repeated administration), calculated by dividing the AUC value obtained after repeated dosing by the AUC value of a single dose, was confirmed to be 4 in the 5 mg/kg group and 7 in the 10 mg/kg group.

Fig. 3
figure 3

Mean plasma concentration vs. time profile of compound K following oral administration of two doses (5 and 10 mg/kg) of compound K to rats for single or 4 wks; 5 mg/kg for single (−●−), 10 mg/kg for single (−■−), 5 mg/kg for 4 weeks (−▲−), and 10 mg/kg for 4 weeks (−▼−). Data expressed as mean \(\:\pm\:\) S.D. of five samples for each time point.

Full size image
Table 3 Pharmacokinetic profiles of compound K in plasma of rats (n = 5 per group) according to the dose and duration of oral administration of compound K.
Full size table

We also confirmed the metabolic conversion of compound K to PPD. However, PPD concentrations above the LOQ (4 ng/mL) were detected in only a few samples (2 to 4 out of 5 rats per group), making pharmacokinetic analysis unreliable. Additionally, despite prolonged administration, plasma concentrations of PPD remained relatively similar within each dose group. At the 8–12 h time points, the concentrations were as follows: single, 5 mg/kg – 6.61 \(\:\pm\:\) 0.20 ng/mL; 4 wks, 5 mg/kg ‒ 5.04 \(\:\pm\:\) 0.58 ng/mL; single, 10 mg/kg ‒ 11.84 \(\:\pm\:\) 3.49 ng/mL; 4 wks, 10 mg/kg ‒ 15.51 \(\:\pm\:\) 3.46 ng/mL.

For estimation, the dosing interval was calculated based on 80% of the T \(\:>\) IC50 (time above half maximal inhibitory concentration) value (Table 4). For prostaglandin E2 (PGE2), the estimated dosing intervals were approximately 18 and 30 h following single administrations of 5 and 10 mg/kg, respectively. For nitric oxide (NO), the corresponding intervals were 7 and 13 h, respectively.

Table 4 Estimation of dosing intervals for compound K based on IC50 (half maximal inhibitory concentration) values for prostaglandin E2 (PGE2) and nitric oxide (NO) inhibition, using Pharmacokinetic parameters from a single oral administration in rats.
Full size table

Pharmacokinetic study 3: Pharmacokinetic study of compound K in mice

To evaluate the oral bioavailability of compound K, a pharmacokinetic study was conducted in mice via two routes of administration. As shown in Fig. 4; Table 5, the T1/2 of intravenous injection (IV) and oral administration (PO) derived from this experiment were 6 and 14 h, respectively. The oral bioavailability of compound K in mice was approximately 12%. Based on this result, the equilibrium concentration was calculated for repeated oral administration to mice at a dose of 10 mg/kg, resulting in a mean concentration in plasma of 6 ng/mL.

Fig. 4
figure 4

Mean plasma concentration vs. time profile of compound K in mice following (A) intravenous injection (IV) and (B) oral administration (PO) of compound K at a 2 mg/kg and 10 mg/kg, respectively. Data expressed as mean \(\:\pm\:\) S.D. of three samples for each time point.

Full size image
Table 5 Pharmacokinetic profiles of compound K in plasma of mice (n = 3 per group) according to administration route of compound K.
Full size table

Network pharmacology and molecular docking study for compound K

A total of 154 targets for compound K were predicted, among which 143 genes overlapped with inflammation-associated genes (Fig. S3A). Gene ontology (GO) enrichment analysis revealed strong associations with the inflammatory response (adjusted p = 1.05e-18), regulation of immune system process (adjusted p = 1.74e-11), and regulation of cytokine production (adjusted p = 9.6e-5) (Fig. 5A). Among the five binding cavities predicted by CB-Dock2 for signal transducer and activator of transcription 3 (STAT3) (Fig. 5B, S3B, and Table 6), pocket 1 exhibited the strongest binding affinity with compound K (‒8.8 kcal/mol), suggesting a stable interaction. Pocket 2 contained critical residues such as LYS591, ARG595, TRP623, GLU625, TYR640, CYS718, and THR721, which are located within or near the SH2 domain (residues ~ 580–690) of STAT3 (Table S4).

Fig. 5
figure 5

Network pharmacology and molecular docking analysis results of compound K. (A) The Top 10 gene ontology (GO): Biological process terms related to inflammation that are significantly enriched among compound K-associated targets (adjusted p-value \(\:<\) 0.05). (B) Two major binding pockets of STAT3 identified for compound K: pocket 1 (binding affinity ‒8.8 kcal/mol) and pocket 2 (binding affinity ‒8.0 kcal/mol).

Full size image
Table 6 Docking analysis results of compound K with signal transducer and activator of transcription 3.
Full size table

Discussion

KRG contains diverse active ingredients that undergo metabolic conversion in vivo18. For active substance undergoing complex understanding their pharmacokinetic properties is of critical importance19. All the active metabolite of KRG exhibit distinct bioactivity, and studies have been conducted to regulate their content using various methods to enhance their bioactivity20,21. Due to this metabolic conversion, determining the active ingredient concentrations of KRG in vivo based on the initial dose alone is difficult. This complexity necessitates further research into the pharmacokinetic properties of KRG.

In Study 1, the doses-dependent changes in the pharmacokinetics of representative active ingredients of KRG have been characterized in experimental rodents. Increased maximum plasma concentrations and body exposure were observed for both Rb1 and Rd with longer dosing periods compared with those for a single dose, which was assessed for up to 8 wks. For Rb1, an accumulation factor of approximately 3 was observed in the 2 wks group, similar to that for the 8 wks dosing group. Contrarily, Rd showed an accumulation factor of 2.2 in the 2 wks group, nearly doubling to a value of 5 in the 8 wks group. This suggested that, in addition to repeated dose-induced accumulation, Rd production via the metabolic conversion of Rb1 may also contribute to its accumulation22. The accumulation of a drug reflects the balance between dosing frequency and drug elimination rate23. In practice, the drug accumulation factor helps assess whether dosing regimens need to be adjusted to optimize efficacy while minimizing toxicity24. The accumulation factor of Rb1 and Rd observed in the present study indicate that even with the same dosing interval, the degree of accumulation of each ginsenoside may vary due to metabolic conversion. Therefore, it may be necessary to adjust the dosing interval according to the target ginsenoside, such as by extending the interval to account for differences in accumulation. Additionally, the repeated administration of KRG extract to humans for 15 days did not result in any increase in the Cmax and AUC values of compound K and less accumulation of compound K in the plasma compared with that of other ginsenosides25. Similar behavior was observed in the present study, where detectable levels of compound K were observed in the blood from 4 wks of repeated dosing, but no significant accumulation compared to 8 wks of repeated dosing. As in humans, the ability of KRG to be converted to compound K is low in rats, suggesting that metabolic conversion is less likely to result in accumulation. Another reason is likely due to the relatively short half-life of compound K26. However, in this study, even in rats, which have relatively low metabolic conversion rates, detectable levels of compound K were observed after 4 wks of administration. Given that the content of compound K in the administered KRG was below the LOQ according to our analytical method, the compound K detected in the bloodstream after repeated KRG administration is presumed to have been generated via metabolic transformation of major KRG constituents within the body. This suggests that the parent compounds of compound K had accumulated sufficiently to allow for its formation.

After assessing the extent of conversion to compound K following KRG administration, in Study 2, which characterized the pharmacokinetic properties of compound K itself following repeated dosing, Cmax and AUC, indicators of drug exposure, exhibited significant dose- and duration-dependent differences. This was supported by the dose-related accumulation factor (5 mg/kg, 4; 10 mg/kg, 7) with repeated dosing. When examining the conversion to PPD in relation to the increase in AUC of compound K, concentrations above the LOQ were observed in only a few samples after 8 h. Regardless of the duration of administration, conversion to PPD increased only in proportion to the compound K dose, suggesting that metabolic saturation is unlikely with prolonged administration. Compound K is absorbed into the liver via transport, and a dose-dependent increase in bioavailability is predicted because of the saturation of the uptake transport in the liver with increasing doses27. Cytochrome P450, located in the liver and responsible for the oxidative metabolism of most drugs, is inhibited by the intestinal metabolites of ginsenosides, including compound K and PPD, which may alter the in vivo drug interactions28. This inhibition may have contributed to the increased AUC of compound K with repeated dosing observed in our experiments. Although the AUC increased because of the low in vivo metabolic conversion efficiency, there was no significant difference in the half-life, indicating that the elimination pattern remained unchanged. Half-life is determined by two independent factors, volume of distribution and clearance, regardless of the dose of the drug administered29. Therefore, further research is needed in this regard via IV injection of compound K in rats.

In Study 3, the pharmacokinetics of compound K after a single IV injection was evaluated in mice, given the difficulty and economic considerations of repeated IV administration because of the amount of compound K required. In this study, the oral bioavailability of compound K in mice was approximately 12%. This is higher than the previously reported 4% in rats at the same dose level30. Differences in animal species and the time of blood sampling might explain this variation despite clinically speaking these values fall in the same range of low F%. An equilibrium concentration is achieved when the rate of drug elimination equals the rate of drug absorption into systemic circulation31. Most therapeutic drugs are developed and evaluated using dosing regimens designed to reach a predetermined steady-state concentration32. Consequently, therapeutic drug monitoring is typically conducted at steady state, as subtherapeutic levels may result in insufficient efficacy, while supratherapeutic levels may lead to toxicity. According to well-established pharmacokinetic principles, steady-state concentrations are generally achieved after approximately five elimination half-lives33. Based on pharmacokinetic parameters obtained from Study 3 and reference equation34, repeated oral administration of compound K at a dose of 10 mg/kg in mice, over a period exceeding five half-lives, is predicted to result in an average steady-state plasma concentration of approximately 6 ng/mL. Although directly correlating plasma drug concentrations with in vitro IC₅₀ values has limitations, a comparison may provide a general reference for evaluating the dose level. In this context, the estimated plasma concentration is substantially lower than the in vitro IC₅₀ values required for antitumor activity in lung cancer cell lines, such as 16.11 µg/mL for A549 and 14.46 µg/mL for PC-9 cells35,36. This discrepancy is likely attributable to the low oral bioavailability of compound K in mice. In preclinical studies, the initial dosing regimen is typically derived from human equivalent doses or effective concentrations observed in in vitro experiments. The low oral bioavailability observed in our study suggests that the current dosing regimen may be insufficient to reach pharmacologically effective plasma concentrations. Therefore, higher doses or more frequent dosing may be necessary to attain therapeutic levels. These findings highlight the inherent challenges in interspecies extrapolation of pharmacokinetic data, particularly with respect to differences in systemic bioavailability and tissue distribution, which may compromise the accurate prediction of pharmacological efficacy across species.

To link the pharmacokinetics of compound K with its anti-inflammatory activity, we estimated the optimal dosing interval using in vitro IC₅₀ values for PGE₂ and NO inhibition, applying a conservative reference threshold of 80% T \(\:>\) IC₅₀ [37−39]. Based on these values, a single 5 mg/kg dose of compound K would need to be administered approximately every 18 h to achieve optimal anti-inflammatory effects through PGE2 inhibition, and every 7 h through NO inhibition, with dosing intervals extending proportionally as the dose increases. Given the limited data available for compound K, this approach served as an exploratory means of assessing effective exposure duration. However, we acknowledge the limitations of translating in vitro potency into in vivo efficacy without a formal pharmacokinetic–pharmacodynamic (PK/PD) or in vitro–in vivo dose extrapolation (IVIVE) model. Thus, the estimated T \(\:>\) IC₅₀ values are presented for hypothesis generation only, not as a basis for dose recommendation.

While the main objective was to investigate the pharmacokinetics of compound K, supplementary in silico analyses were undertaken to expand the biological relevance of the pharmacokinetic data by identifying potential anti-inflammatory targets and pathways. The anti-inflammatory effect of compound K was confirmed by potential target analysis and subsequent GO: Biological process enrichment, which included terms related with inflammatory process. Additionally, to investigate the anti-inflammatory mechanism of compound K, a molecular docking simulation was conducted against STAT3, an inflammation-related target closely associated with PGE2 and identified with high probability among the predicted targets. The SH2 domain of STAT3 plays a key role in mediating phosphorylation-dependent dimerization and nuclear translocation, which are essential for the transcription of inflammation-related genes40. The present docking simulation suggests that compound K binds to the SH2 domain (pocket 2) with moderate to high affinity (–8.0 kcal/mol), particularly interacting with residues ARG595 and TYR640. The SH2 domain of STAT3, essential for its phosphorylation-dependent dimerization and activation, has emerged as a promising therapeutic target, as binding to key residues within this domain can effectively inhibit STAT3 activity41,42. These findings support the hypothesis that compound K may exert anti-inflammatory activity by targeting the SH2 domain of STAT3 and preventing its activation, highlighting its potential as a pharmacokinetically significant active component of KRG.

Although this study has examined various pharmacokinetic characteristics of compound K, several limitations should be considered for future research. While the sample size of five animals per group was validated to be statistically sufficient to determine significant differences in key pharmacokinetic parameters, it may have contributed to large standard deviations due to inter-individual variability. This could limit the reliability and generalizability of the results, and should be taken into account when interpreting the findings. Furthermore, due to the limited availability of previous studies on the bioavailability (F value) of compound K in experimental animals and constraints in securing the pure compound, the F value was derived from the mice study and used in estimation analysis, despite possible interspecies differences. This method may fail to fully reflect interspecies pharmacokinetic differences, which should be considered another limitation of the study. Lastly, while the present pharmacokinetic study focused on evaluating plasma concentrations, it remains a limitation that tissue specific distribution was not assessed. Evaluating the concentrations of individual ginsenosides and their metabolites in specific tissues could help clarify the relationship between active compounds, tissue-specific efficacy, and dosing.

Conclusion

In this study, we evaluated the impact of repeated KRG administration on the pharmacokinetic properties of ginsenosides and included supplementary mechanistic analyses to support the relevance of compound K as a potential pharmacokinetic subject of interest. The in vivo metabolic conversion of KRG is affected by various factors, and continuous exposure to the drug, such as through repeated administration, can alter metabolite formation because of the effects of metabolic enzymes and the intestinal environment. Additionally, the pharmacokinetics of ginsenosides are highly variable among individuals and can differ significantly among species of experimental animals, especially after long-term administration. Therefore, the pharmacokinetic properties of ginsenosides to be studied should be considered to achieve optimal animal study results.

Methods

Materials and chemicals

KRG [Hong Sam Jung; lot no. H2006(2)1043] was provided by the Korea Ginseng Corporation (Daejeon, Korea). Previously, major ginsenosides Rb1 (6.0 mg/g) and Rd (1.0 mg/g) were estimated using high-performance liquid chromatography analysis43. However, the compound K content in KRG was below the LOQ (0.5 ng/mL) and could not be measured. This was similar to the ginsenoside content profile of commercial KRG extracts available on the market21. Compound K (lot no. C2203255) was purchased from Shanghai Aladdin Biochemical Technology Co. (Shanghai, China). The standard compounds, including ginsenosides Rb1, Rd, compound K, and PPD, were purchased from Sigma-Aldrich (St. Louis, MO, USA). Berberine and digoxin (Sigma-Aldrich) were used as IS. All other chemicals and solvents were of reagent with a purity of \(\:>\) 90%.

Standard preparations

Ginsenosides Rb1, Rd, compound K, and PPD were accurately weighed and dissolved in methanol to obtain a concentration of 1000 µg/mL each. Stock solutions were divided and mixed according to the sample preparation method. Ginsenosides Rb1 and Rd were mixed and diluted with methanol to a final concentration of 1000 ng/mL. Stock solutions of ginsenosides Rb1 and Rd were serially diluted with methanol to obtain working solutions of 5, 10, 20, 50, 200, and 500 ng/mL for calibration standards and quality control (QC) samples. Compound K and PPD were diluted with methanol to a concentration of 5000 ng/mL each. Stock solutions of compound K and PPD were serially diluted with methanol to obtain working solutions of 50, 100, 200, 500, 2000, and 5000 ng/mL for the calibration standards and QC samples. All the solutions were stored at – 20\({}^ \circ C\).

Sample preparation

Plasma samples (200 µL) were collected at each time point and processed according to the sample extraction method. For the extraction of ginsenosides Rb1 and Rd, calibration standards and QC samples were prepared by adding 5 µL working solution into 45 µL of blank plasma to achieve final concentrations of 0.5, 1, 2, 5, 20, and 50 ng/mL. To each 50 µL of plasma sample, 600 µL of an IS solution (0.05 ng/mL berberine in methanol) was added. The mixture was vortexed for 5 min and then centrifuged at 13,700 \(\:\text{x}\) g for 15 min. Subsequently, 500 µL of the supernatant was transferred to a clean tube and allowed to evaporate to dryness under a nitrogen stream at 40\({}^ \circ C\). The residue was reconstituted in 50 µL of 70% methanol containing 0.1% formic acid.

For compound K and PPD extraction, calibration standards and QC samples were prepared by spiking 95 µL of blank plasma with 5 µL of working solution to achieve final concentrations of 0.5, 1, 2, 5, 20, and 50 ng/mL. To 100 µL of plasma samples, 5 µL of an IS solution (10 µg/mL digoxin in methanol) and 800 µL of methyl tert-butyl ether was added. The mixture was vortexed for 20 min and then centrifuged at 13,700 \(\:\text{x}\) g for 15 min. After centrifugation, 800 µL of the supernatant was transferred to a clean tube and allowed to evaporate to dryness under a nitrogen stream at 40\({}^ \circ C\). The residue was resuspended in 50 µL of 80% methanol containing 0.1% formic acid.

LC-MS/MS analysis

The concentrations of ginsenosides Rb1, Rd, compound K, and PPD were analyzed using an Agilent 6470 triple quadrupole LC-MS/MS system (Agilent Technologies Inc., Santa Clara, CA, USA). Ginsenosides Rb1 and Rd were separated using a Hypersil GOLD™ C18 column (3 μm, 2.1 \(\:\times\:\) 100 mm, Thermo Fisher Scientific, Leicester, UK). The mobile phase comprised 0.1% formic acid in distilled water (A) and 0.1% formic acid in methanol (B), with gradient conditions as follows: 0–2 min (37 − 37% A), 2–4 min (37 − 15% A), 4–7 min (15 − 10% A), and 7–7.5 min (10 − 37% A) at a flow rate of 0.270 mL/min. Quantification of the separated analyte peak was performed at m/z 1131.8 → 789.6 for Rb1, m/z 969.7 → 789.4 for Rd, and m/z 336.2 → 320.1 for berberine (IS) in the positive ionization mode, with collision energy (CE) of 50, 50, and 30 eV, respectively. Sample injection volume was 10 µL.

Compound K and PPD were separated using an Eclipse Plus C18 column (1.8 μm, 2.1 \(\:\times\:\) 50 mm, Agilent Technologies Inc.). The mobile phase consisted of distilled water (5%) and 0.1% formic acid in methanol (95%) under isocratic elution at a flow rate of 0.120 mL/min. Quantification was performed at m/z 645.6 → 203.1 for compound K, m/z 425 → 109.2 for PPD, and m/z 803.4 → 673.1 for digoxin (IS) in the positive ionization mode with CE of 35, 25, and 50 eV, respectively. Sample injection volume was 5 µL. Throughout the analytical run, the autosampler temperature was maintained at 4 °C. Analytical data were quantified using the MassHunter software (Version 10.1, Agilent).

Method validation

The validation parameters, including specificity, linearity, recovery, and matrix effects, were assessed according to the “Guidelines for Validation of Bioanalytical Methods” established by the Korean Ministry of Food and Drug Safety44. Specificity was evaluated by comparing the chromatograms from six lots of the plasma to ensure the absence of interfering peaks at the respective retention times of the ginsenoside analytes and IS at the LOQ. Each calibration curve was constructed by plotting the peak area ratio (y-axis) of the analyte relative to the IS area against the spiked concentration (x-axis) of each ginsenoside analyte. Linearity was assessed using a weighted (1/x2) least squares regression analysis. The LOQ was based on the standard deviation of the response and the slope of the calibration curve, with a signal-to-noise (S/N) ratio of 10. Extraction recovery was evaluated for three different QC samples with three replicates each by comparing the peak areas of the extracted samples (spiked before extraction) with those of the unextracted samples (spiked after blank extraction). Matrix effects were investigated by comparing the peak areas of the analytes dissolved in the pretreated blank plasma with the corresponding concentrations prepared in the resuspension solution, with three replicates.

Animal study

Sprague Dawley (SD) rats and ICR mice for the pharmacokinetics study were purchased from Orient Bio Inc. (Seungnam, Korea). The animals were provided ad libitum access to water and food and acclimated for one week prior to the experiments. All animal care and experimental procedures were conducted following the guidelines of the Animal Care and Use Committee of the Chungnam National University (Approval No. 202112 A-CNU-216). In addition, all experimental protocols were approved by an Animal Care and Use Committee of the Chungnam National University. Animals were euthanized using CO2 at the end of the experiments.

In pharmacokinetic study 1, for the pharmacokinetic study of KRG, male SD rats of varying ages (7-week-old for single administration, 6-week-old for 2 wks administration, and 4-week-old for 4 and 8 wks administration, n = 5 rats per group) received KRG at a single dose of 2 g/kg/day via PO administration. KRG was dissolved in phosphate-buffered saline (PBS). This dosage, which was slightly higher than usual, was chosen based on human pharmacokinetic studies, low compound K blood levels, and repeated administration experiments, to ensure accurate ginsenosides evaluation. Blood samples (200 µL) were collected from the tail vein after 0, 1, 2, 4, 6, 8, 10, 12, 24, 48, 72, and 96 h from the last administration. The blood samples were centrifuged at 13,700 \(\:\text{x}\) g for 15 min, and 100 µL aliquots of plasma were stored at − 80\({}^ \circ C\) until the analysis of ginsenosides Rb1, Rd, and compound K.

In pharmacokinetic study 2, for the pharmacokinetic study of compound K, aimed to evaluate its absorption rate based on dosage and administration period, male SD rats (7-week-old for single administration and 4-week-old for 4 wks administration, n = 5 rats per group) were divided into four groups and treated with compound K, dissolved in PBS, at doses of 5 and 10 mg/kg/day via oral gavage after single or multiple (4 wks) administration. Blood samples were collected from the tail vein after 0, 1, 2, 4, 6, 8, 10, 12, and 24 h of administration.

In pharmacokinetic study 3, to confirm the oral bioavailability of compound K, 7-week-old male ICR mice (n = 3 mice at each time point) were randomly divided into two groups: an IV group and a PO group. The IV dose was determined based on a previous study17, and the PO dose was set at five times the IV dose to consider absorption rates. The IV group received compound K dissolved in normal saline containing 20% dimethyl sulfoxide at a single dose of 2 mg/kg via tail vein injection. The PO group was administered compound K dissolved in PBS at a single dose of 10 mg/kg. Subsequently, the mice were euthanized and blood samples were collected from the abdominal artery after 0, 1, 2, 4, 6, 8, 10, 12, 24, and 48 h of treatment with compound K. Blood samples were centrifuged at 13,700 \(\:\text{x}\) g for 15 min, and 100 µL aliquots of plasma were stored at – 80\({}^ \circ C\) until compound K analysis.

Pharmacokinetic data and statistical analysis

Pharmacokinetic parameters were calculated using the ThothPro software (version 4.3.0; ThothPro LLC, Gdańsk, Poland). The pharmacokinetic parameters for KRG were determined using a non-compartmental model. For compound K, the IV group parameters were estimated using a non-compartment model, and those of PO group were estimated using a one-compartment model to derive the absorption rate constants necessary for inferring the equilibrium concentrations after repeated oral administration. For mice, a naïve pooled analysis was applied to the plasma data from each individual to calculate pharmacokinetic indices using samples obtained from non-contiguous blood draws at each time point.

Cmax and time to peak concentration (Tmax) were calculated based on the plasma concentration of each ginsenoside analyte. The AUC, representing drug exposure in the body, was calculated by summing the areas between each adjacent observed concentration as if they formed a trapezoid. The accumulation factor was calculated as the ratio of the AUC after the administration of a single dose to the AUC during the dosing interval under steady-state for the same dose45:

\(\:\text{A}\text{c}\text{c}\text{u}\text{m}\text{u}\text{l}\text{a}\text{t}\text{i}\text{o}\text{n}\:\text{f}\text{a}\text{c}\text{t}\text{o}\text{r}=\frac{{AUC}_{\left(0-t\right)}\:of\:repeated\:administraiton\:group\:}{{AUC}_{\left(0-t\right)}\:of\:single\:administraiton\:group\:}\) (at the same dose).

The equilibrium concentration at steady-state was derived using the following equation34:

$$\:\text{P}\text{l}\text{a}\text{s}\text{m}\text{a}\:\text{c}\text{o}\text{n}\text{c}\text{e}\text{n}\text{t}\text{r}\text{a}\text{t}\text{i}\text{o}\text{n}=\:\frac{F\cdot\:D\cdot\:{k}_{a}}{Vd({k}_{a}-{k}_{e})}\cdot\:\left(\frac{\text{e}\text{x}\text{p}(-{k}_{e}\cdot\:\tau\:)}{(1-\text{e}\text{x}\text{p}(-{k}_{e}\cdot\:\tau\:\left)\right)}-\frac{\text{e}\text{x}\text{p}(-{k}_{a}\cdot\:\tau\:)}{(1-\text{e}\text{x}\text{p}(-{k}_{a}\cdot\:\tau\:\left)\right)}\right)$$

where F, oral bioavailability; D, dose; Vd, volume of distribution;

ka, absorption rate constant; ke, elimination rate constant; \(\:\tau\:\), dosing interval.

The following equation, adapted from models used in antibiotic evaluation, was applied to estimate the dosing interval based on the T \(\:>\) IC50 target46. The IC50 values of compound K for PGE2 and NO were set at 4 µM and 12 µM, respectively, based on the study by Park et al.47, and the F value was obtained from current mouse data.

$$\:\text{D}\text{o}\text{s}\text{e}\:\text{I}\text{n}\text{t}\text{e}\text{r}\text{v}\text{a}\text{l}=\:\frac{\text{ln}\left(\frac{Dose}{Vd\cdot\:BW\cdot\:{IC}_{50}}\right)\cdot\:{T}_{1/2}\cdot\:100}{\text{ln}2\cdot\:\%T>{IC}_{50}}$$

where Dose, administered dose; Vd, volume of distribution; BW, body weight;

IC50, half maximal inhibitory concentration; T1/2, elimination half-life;

%T \(\:>\) IC50, percentage of time that the drug concentration remains above the IC50.

All pharmacokinetic parameters except Tmax are presented as mean \(\:\pm\:\) standard deviation. Tmax is the categorical pharmacokinetic parameter and expressed as median and range48. Statistical significance was determined as p-value \(\:<\) 0.05 or 0.01 using a two-tailed t-test and GraphPad Prism (version 9, GraphPad Inc., La Jolla, CA, USA).

In Silico mechanistic exploration of compound K’s anti-inflammatory activity

The potential molecular targets for compound K (CID: 9852086) were predicted using the SuperPred tool49. These targets were cross-referenced with inflammation-associated genes retrieved from the GeneCards database using the keyword, inflammation, resulting in a set of overlapping genes presumed to mediate anti-inflammatory effects. The overlapping genes were then subjected to functional enrichment analysis via g: Profiler. GO: Biological Process terms were evaluated with a significance threshold of adjusted p \(\:<\) 0.05, and inflammation-relevant categories were selected using keyword filters (inflammation, immune, cytokine, and interleukin).

To further investigate the interaction between compound K and a representative inflammatory target, molecular docking simulations were conducted using the CB-Dock2 platform50. The 3D structure of compound K was generated from its SMILES using Open Babel51, and the STAT3 protein structure (AF-Q14765-F1) was obtained from the AlphaFold Protein Structure Database. Both ligand and receptor structures were submitted to CB-Dock2 for cavity prediction and blind docking based on AutoDock Vina. Docking outputs included predicted binding sites, binding affinities (Vina scores), pocket volumes, and interacting residues.