3.1 Determination of chemical composition
In the detection range, the peak area of each ingredient showed a linear relationship with the concentration (r > 0.9991); the repeatability RSD (n = 6) was < 2.67%; the precision RSD (n = 6) was < 2.22%; the stability was good within 24 h, and the stability RSD (n = 6) was < 2.56%; the recoveries of spiked samples were in the range of 95.3%~103.2%, and the RSD (n = 6) was less than 7.27%. The results showed that the established quantitative method is accurate, reliable, and meets the relevant requirements, so it can be used for the determination of the ingredients of black ginseng (Zhang JQ et al 2015; Xu YY et al 2021; Li M et al 2021; Zhu LL et al 2019; Li HY et al 2019; Rogério da Silva Moraes E et al 2022; Zhang YH et al 2020).
HPLC results are shown in Fig. 1. P1 ~ P6 were determined to be D-malic acid, malonic acid, acetic acid, citric acid, succinic acid, and L-malic acid, respectively. P7 was 5-HMF. P8 was maltol. P9 ~ P16 were identified as ginsenosides Rg1, Re, Rf, Rb1, Rc, Rb2, Rb3, and Rd, respectively, and were primary ginsenosides. P17 ~ P25 were identified as ginsenosides F4, Rk3, Rh4, 20 (S)-Rg3, 20 (R)-Rg3, Rk1, Rg5, 20 (S)-Rh2, and 20 (R)-Rh2, respectively, and were secondary ginsenosides. The content determination results are shown in Fig. 2. The raw ginseng S0D0, contained only primary ginsenoside, 5-HMF was not detected; a small amount of maltol was detected; and the organic acid and L-malic acid were high at 48.49 mg/g (Fig. 2(A)~(F)). S1D1 began to produce secondary ginsenoside, and the content of L-malic acid was drastically reduced to 1.40 mg/g. S2D2 had the highest overall ginsenoside content of 31.23 mg/g, with a higher content of acetic acid and a detectable 5-HMF of 0.008 mg/g. Accordingly, S2D2 exhibited a significant chemical difference from S0D0. The content of primary ginsenosides in the samples from S3D3 to S6D6 was gradually decreasing, and the content of secondary ginsenosides, 5-HMF, and malonate was gradually increasing. There was a significant increase in the content of secondary ginsenosides, 5-HMF, and malonate in the S7D7 as compared to the S6D6, and some primary ginsenosides could still be detected. The content of secondary ginsenosides in the S7D7 was the highest (17.41 mg/g). The primary ginsenosides were almost completely converted in the samples of S8D8 and S9D9, and the content of secondary ginsenosides was reduced significantly compared to that of S7D7. The content of overall ginsenosides was the lowest in S9D9 (3.29 mg/g), which was possibly caused by excessive processing. 5-HMF was also found to be the most important in S7D8 and S9D9. And gradually increased with the increase in the number of processing, with the highest content of 5-HMF in S9D9 (0.0291 mg/g).
According to Fig. 3, P1 ~ P4 were determined to be fructose, glucose, sucrose, and Maltose, respectively. Fructose and glucose gradually increased with the number of cycles of steaming and sun-drying (Fig. 2(G)). Sucrose content was highest in S0D0 (247.82 mg/g). Maltose content was highest in S1D1 (535.53 mg/g), and sucrose content was reduced drastically as compared to S0D0 (Jin Y et al 2015). Afterwards, the content of both maltose and sucrose gradually decreased.
As shown in Fig. 4, P1 ~ P18 were determined to be 18 protein amino acids, respectively. P19 and P20 were determined to be GABA and dencichin. The total amount of amino acids in the samples showed a gradual decrease with the number of cycles of steaming and sun-drying (Fig. 2(H)~(J)). The total amount of essential amino acids tends to stabilize after S2D2. In S9D9, except for aspartic acid, alanine, and histidine, they were no longer detectable. Cysteine, methionine, and tryptophan were not detected in this experiment. The contents of both dencichin and GABA showed a decreasing trend with the increase in the number of cycles of steaming and sun-drying; the content of dencichin decreased to 0 in S5D5, and the content of GABA decreased to 0 in S8D8(Fig. 2(K)).
Based on the results of HPLC, we found that during the processing, C3 and C20 sugars in ginsenosides Ra, Rb1, Rb2, Rb3, Rc, and Rd were hydrolyzed to generate Rg3 and Rh2; Rg3 generates Rk1 and Rg5 through dehydration reactions; and Rg2 was hydrolyzed and dehydrated to produce Rg6, F4, Rk3, and Rh4 (A.M. Metwaly et al 2019). Consequentially, this caused a steady decrease in the amount of primary ginsenosides in black ginseng, an increase in the amount of secondary ginsenosides, and the primary ginsenosides progressive conversion into secondary ginsenosides. As the number of steaming and sun-drying cycle periods increased, sucrose and maltose were continually hydrolyzed into one glucose molecule, one fructose molecule, and two glucose molecules, respectively. Therefore, their content gradually declined while fructose and glucose content steadily grew. The Maillard reaction, a non-enzymatic thermal reaction in the processing of TCM (Han Z et al 2022), alters the flavor and color of ginseng. One of the byproducts of the Maillard process is the short-molecule aldehyde compound 5-HMF, which possesses antioxidant and neuroprotective properties (Ya B et al 2017). Polyphenols are the secondary metabolites of ginseng, and maltol has a high concentration and antioxidant activity (Han Y et al 2015). The amounts of 5-HMF and maltol gradually rose as processing cycles increased. In samples with various steaming and drying periods, eighteen protein amino acids, as well as two non-protein amino acids, dencichin and GABA, were found. Amino acids could regulate pH values, and glutamic acid aids in the synthesis of secondary ginsenosides (Liu Z et al 2020). The amount of protein amino acids and non-protein amino acids regularly reduced as processing cycles increased.
3.2 Principal component analysis
According to Fig. 5(A), which depicts the principal component analysis of the major chemical ingredients from S0D0 to S9D9, PC1 and PC2 contributed 58.0% and 38.8%, respectively, for a total of 96.8%. It was found that the two principal components were responsible for the majority of the chemical information in the black ginseng samples. S0D0 had the largest projection on PC1 and PC2 and was located in the first quadrant of the coordinates of PC1 and PC2, while S7D7, S8D8, and S9D9 were in the second quadrant; S1D1 was in the third quadrant; S2D2 was in the fourth; and S3D3, S4D4, S5D5, and S6D6 were relatively close to the coordinate axes, which indicated that there was a considerable difference between chemical ingredients of samples in various steaming and sun-drying cycle periods. As shown in Fig. 5(B), the primary ginsenosides mostly contributed positively to PC1, whereas the secondary ginsenosides mostly contributed negatively. The content of each secondary ginsenoside and primary ginsenoside of black ginseng altered dramatically during the processing, which was the main reason for the different positions of various black ginseng samples in the coordinate system.
3.3 Orthogonal partial least squares discriminant analysis
The differential ingredients of each sample could be effectively differentiated utilizing the OPLS-DA analysis with SIMCA software. This analysis showed a fit index of 0.978 for the independent variable (R2x), 0.982 for the dependent variable (R2y), and 0.862 for the model prediction index (Q2), with R2 and Q2 exceeding 0.5 to indicate that the results of the model fit were acceptable. An intersection between the Q2 regression line and the vertical axis was smaller than zero after 200 permutation tests, suggesting that the model was not overfitted and that it had been verified (Musumarra G et al 2011). It indicated that they might be applied to the analysis of the differential ingredients of black ginseng samples in various steaming and sun-drying cycle periods, and the results are shown in Fig. 6. The active ingredients' VIP values are displayed in Table. 1, and the top 10 differential ingredients—fructose, glucose, ginsenoside 20 (S)-Rg3, 20 (R)-Rg3, 20 (S)-Rh2, dencichin, glutamic acid, ginsenoside Rg1, Re, and Rc—are filtered out based on the requirement that their VIP values be more than one.
Table.1 VIP value of OPLS-DA of each ingredient
ingredient
|
VIP value
|
ingredient
|
VIP value
|
ingredient
|
VIP value
|
fructose
|
1.38
|
citric acid
|
1.09
|
isoleucine
|
0.95
|
glucose
|
1.34
|
L-malic acid
|
1.07
|
maltol
|
0.93
|
20 (S)-Rg3
|
1.34
|
tyrosine
|
1.05
|
GABA
|
0.92
|
20 (R)-Rg3
|
1.29
|
Rd
|
1.04
|
glycine
|
0.91
|
20 (S)-Rh2
|
1.24
|
acetic acid
|
1.03
|
5-HMF
|
0.91
|
dencichin
|
1.23
|
Rb2
|
1.02
|
Rb3
|
0.90
|
glutamic acid
|
1.16
|
20 (R)-Rh2
|
1.00
|
serine
|
0.89
|
Rg1
|
1.13
|
aspartic acid
|
0.98
|
Cysteine
|
0.89
|
Re
|
1.13
|
Rf
|
0.97
|
Malonic acid
|
0.89
|
Rc
|
1.13
|
proline
|
0.97
|
Maltose
|
0.88
|
arginine
|
1.12
|
sucrose
|
0.97
|
Rg5
|
0.86
|
methionine
|
1.11
|
D-malic acid
|
0.95
|
Rk1
|
0.86
|
3.4 Network pharmacological analysis
The KEGG pathway annotation analysis of key targets of S0D0 ~ S9D9 against prostate cancer showed that they were enriched with 115, 122, 122, 122, 122, 118, 118, 118, 31, 31 signaling pathways, respectively (Fig. 7). The enrichment results showed that the anti-prostate cancer effects of black ginseng samples during the "nine steaming and nine sun-drying" process mainly involved the neuroactive ligand-receptor interaction, glutamatergic synapse, GABAergic synapse, notch signaling pathway, PI3K-Akt signaling pathway, etc.
3.5 Entropy weighting method assignment
First, the findings are data normalized with an assumption of n samples and k evaluation indices for each sample, thus forming a sequence of raw data units {Xij} (i = 1, 2, 3...n; j = 1, 2, 3...k; i = 10, j = 10 in this study). Eq. (1) was used to standardize the raw data. Yij is the standardized data, Xij is the jth indicator value of the ith sample, and {Xj} is the data series of the jth indicator (Ardila JA et al 2015).
$${Y}_{ij}=\frac{{X}_{ij}-min\left\{{X}_{j}\right\}}{max\left\{{X}_{j}\right\}-min\left\{{X}_{j}\right\}}·················\left(1\right)$$
Calculate the information entropy (E) of each indicator using equations (2) and (3), according to the definition of information entropy in information theory. Ej is the information entropy of the jth indicator of the sample; if Pij = 0, then lnPij = 0 is defined.
\({P}_{ij}=\frac{{Y}_{ij}}{\sum _{i=1}^{n}{Y}_{ij}}···································\left(2\right)\) \({E}_{j}=\frac{1}{{ln}n}\sum _{i=1}^{n}{P}_{ij}{ln}({P}_{ij})·····················(3)\)
Finally, the weight of each indicator (Wj) was calculated according to Eq. (4).
$${W}_{j}=\frac{1-{E}_{j}}{\sum _{i=1}^{n}{(1-E}_{j})}························\left(4\right)$$
The Wj of the 10 differential ingredients in S0D0 ~ S9D9 were weighted with their contents in the samples in various steaming and sun-drying cycle periods. The final assigned value (ρ) of the ingredient in SnDn was obtained according to equation. (5), and the results are shown in Table 2.
$$\rho =Wj\times samples{\prime }content···························\left(5\right)$$
Table.2 Weights of the main differentiating ingredients ( ρ )
ingredient
|
S0D0
|
S1D1
|
S2D2
|
S3D3
|
S4D4
|
S5D5
|
S6D6
|
S7D7
|
S8D8
|
S9D9
|
fructose
|
3.95
|
2.99
|
5.72
|
9.37
|
7.63
|
6.61
|
19.43
|
22.73
|
23.48
|
21.11
|
glucose
|
0.97
|
0.62
|
1.12
|
2.20
|
1.97
|
2.07
|
4.34
|
3.70
|
4.68
|
3.62
|
20 (S)-Rg3
|
0
|
0.03
|
0.04
|
0.10
|
0.07
|
0.12
|
0.10
|
0.28
|
0.14
|
0.05
|
20 (R)-Rg3
|
0
|
0.06
|
0.07
|
0.11
|
0.10
|
0.11
|
0.11
|
0.30
|
0.14
|
0.07
|
20 (S)-Rh2
|
0
|
0.01
|
0.02
|
0.03
|
0.03
|
0.04
|
0.04
|
0.07
|
0.04
|
0.02
|
dencichin
|
0.50
|
0.36
|
0.39
|
0.21
|
0.11
|
0.09
|
0.15
|
0.04
|
0
|
0
|
glutamic acid
|
0.42
|
0.49
|
0.49
|
0.45
|
0.17
|
0.12
|
0.29
|
0.09
|
0
|
0
|
Rg1
|
1.72
|
1.47
|
0.87
|
0.67
|
0.46
|
0
|
0
|
0
|
0
|
0
|
Re
|
0.15
|
0.07
|
0.04
|
0.03
|
0.03
|
0.02
|
0.02
|
0.01
|
0
|
0
|
Rc
|
0.39
|
0.19
|
0.17
|
0.16
|
0.11
|
0.22
|
0.10
|
0.08
|
0
|
0
|
By weighting each target in accordance with the assignment (ρ) of each ingredient and the Degree value of the target in the PPI network, Eq. (6) was utilized to generate the final assignment (σ) of each target in SnDn.
$$\sigma =\sum _{i=1}^{n}\rho \times Degree··························\left(6\right)$$
The score (ω) of each route was then thoroughly derived using equations (7) and (8) based on the assignment results of each target and the enrichment of the major anti-prostate cancer pathways. GeneInTerm is the number of targets in the pathway in which the drug treats the disease; count is the total number of targets in which the drug treats the disease.
$$GeneRatio=\frac{GeneInTerm}{count}··············\left(7\right)$$
$$\omega =\sum _{i=1}^{n}\sigma \times GeneRatio····················\left(8\right)$$
The result of S0D0 ~ S9D9 anti-prostate cancer pathway assignment is shown in Fig. 8, in which the scores of the neuroactive ligand-receptor interaction pathway (A), glutamatergic synapse pathway (B), PI3K-Akt signaling pathway (C), Chemical carcinogenesis-receptor activation pathway (D), and Notch signaling pathway (E) were changed regularly with the increase of the various steaming and sun-drying cycle periods. As the number of steaming and sun-drying cycles increased, the scores for pathways A and B steadily dropped, showing that the benefits of black ginseng on these two pathways faded as the number of processing cycles rose. The scores of pathways C, D, and E showed a tendency to increase and then decrease, and the scores of pathways C and D in S7D7 were the highest.
According to the HPLC data, dencichin, glutamate, Rg1, Re, and Rc content trended to decline in S0D0–S9D9 and then reached zero in S8D8 and S9D9. Fructose and glucose concentrations rise gradually and steadily, but ginsenosides 20 (S)-Rg3, 20 (R)-Rg3, and 20 (S)-Rg2 concentrations fall after peaking in S7D7. Fructose and glucose had a certain promoting effect on prostate cancer, with 20 (S)-Rg3, 20 (R)-Rg3, and 20 (S)-Rh2 having a certain inhibitory effect that is greater than the promoting effect of fructose and glucose. Based on the results of HPLC analysis and vector space network pharmacology, we found that a decrease in the content of decicin, glutamate, Rg1, Re, and Rc, as well as an increase in the content of fructose, glucose, 20 (S)-Rg3, 20 (R)-Rg3, and 20 (S)-Rh2, weakened the regulation of black ginseng on the neuroactive ligand-receptor interaction pathway and glutamatergic synapse pathway, and enhanced its regulatory effects on the PI3K-Akt signaling pathway, chemical carcinogenesis-receptor activation pathway, and Notch signaling pathway. The variation in the content of different components in black ginseng during processing leads to its different regulatory effects on prostate cancer-related pathways, which is the main reason for the change in the anti-prostate cancer effect of black ginseng in various steaming and sun-drying cycle periods.
Carbohydrates are crucial components of cellular metabolism and may provide prostate cancer cells with more energy. The expression of glucose transporter 1 (Glut-1) in prostate cancer cells rises with the grade of the malignant tumor (Effert P et al 2004), while the fructose transporters Glut5 and Glut9 are markedly elevated in prostate cancer patients (Carreño DV et al 2021). Ginsenoside is the main active ingredient in ginseng and black ginseng. Ginsenoside Rg3 inhibits the expression of AQP1 in highly metastatic PC-3M prostate cancer cells to reduce their migration level (Pan XY et al 2012). Rg3 may also inhibit the cell cycle by increasing the amount of ROS produced by prostate cancer cells (Peng Y et al 2019). Ginsenoside Rh2 can upregulate ROS, superoxide, and PPAR-δ; downregulation of p-STAT3 induces apoptosis in prostate cancer cells (Tong-Lin Wu T et al 2018). Many patients with prostate cancer are associated with the transmission of neural active ligand-receptor interaction signals (Nagaya N et al 2021). The transmission of various glutamate receptors, such as mGluRs and iGluRs, is involved in the process of cell proliferation and migration in cancer (Stepulak A et al 2014). The activation of PI3K/Akt in the PI3K/Akt pathway enhances the levels of Bcl-2 and XIAP, thereby improving the survival rate of prostate cancer cells (Tewari D et al 2022). PI3K/Akt may also work together with other pathways to worsen the condition of prostate cancer (Yan Y et al 2019). The activation of Erk/MAPK signals in the chemo-carcinogenic receptor activation pathway is the mechanism of cadmium-induced prostate cancer (Dasgupta P et al 2020). Notch signaling is highly expressed in prostate cancer, and its expression increases with the cancer grade (Deng G et al 2016; Su Q et al 2016). These accounts concur with the findings of our experiment. Therefore, vector space network pharmacology methods can accurately and effectively analyze the anti-prostate cancer mechanism of black ginseng.