Polyphenolic Compounds from Lespedeza Bicolor Root Bark Inhibit Progression of Human Prostate Cancer Cells via Induction of Apoptosis and Cell Cycle Arrest.

From a root bark of Lespedeza bicolor Turch we isolated two new (7 and 8) and six previously known compounds (1–6) belonging to the group of prenylated polyphenols. Their structures were elucidated using mass spectrometry, nuclear magnetic resonance and circular dichroism spectroscopy. These natural compounds selectively inhibited human drug-resistant prostate cancer in vitro. Prenylated pterocarpans 1–3 prevented the cell cycle progression of human cancer cells in S-phase. This was accompanied by a reduced expression of mRNA corresponding to several human cyclin-dependent kinases (CDKs). In contrast, compounds 4–8 induced a G1-phase cell cycle arrest without any pronounced effect on CDKs mRNA expression. Interestingly, a non-substituted hydroxy group at C-8 of ring D of the pterocarpan skeleton of compounds 1–3 seems to be important for the CDKs inhibitory activity.

The data were analyzed with the LabSolutions software (v. 1.0, Shimadzu, Kyoto, Japan).

HR-ESI-MS
HR-ESI-MS experiments were carried out using a Shimadzu hybrid ion trap-time of flight mass spectrometer (Shimadzu, Kyoto, Japan). The operating settings of the instrument were as follows: electrospray ionization (ESI) source potential, −3.8 and 4.5 kV for negative and positive polarity ionization, respectively, drying gas (N 2 ) pressure-200 kPa, nebulizer gas (N 2 ) flow-1.5 L/min, temperature for the curved desolvation line (CDL) and heat block-200 • C, detector voltage-1.5 kV and the range of detection-100-900 m/z. The mass accuracy was below 4 ppm. The data were acquired and processed using Shimadzu LCMS Solution software (v.3.60.361, Shimadzu, Kyoto, Japan).

MTT Assay
In vitro drug sensitivity MTT assay was performed as previously described [15]. In total, 6000 cells/well were seeded in 96-well plates, incubated overnight, and treated with the investigated drugs for 48 h.

Flow Cytometry Analysis
The effect of the compounds on apoptosis induction and cell cycle progression was analyzed by flow cytometry technique using PI staining as reported before [15]. Cells (0.2 × 10 6 cells/well) were seeded in six-well plates, incubated overnight, and treated with the investigated drugs for 48 h. The cells were further harvested, fixed with 70% EtOH/H 2 O, stained with propidium iodide (PI), and analyzed using a FACS Calibur machine (BD Bioscience, San Jose, CA, USA). The results were quantified using BD Bioscience Cell Quest Pro software (v.5.2.1., BD Bioscience).

Quantitative Real-Time PCR (qPCR)
The CDKs gene expression was measured using qPCR technique. PC-3 cells were seeded in Petri dishes (1 × 10 6 cells per ø 6 cm dish in 5 mL) in standard culture media, incubated overnight and treated with the investigated drugs in the fresh culture media for 24 h. Cells were harvested using the cell scraper, washed with PBS and homogenized using QIAshredder (QIAGEN, Hilden, Germany). The total RNA was isolated using PureLink ® RNA Mini Kit (Invitrogen, Carlsbad, CA, USA) and the on-column DNA digestion using PureLink™ DNase (Invitrogen). Correspondent RNA (2 µg in 30 µL) and was transcribed into cDNA using Maxima First Strand cDNA Synthesis Kit for RT-qPCR (Thermo Scientific, Vilnius, Lithuania). The following qPCR was performed using 2X KAPA SYBR FAST qPCR Master Mix Optimized for Roche LightCycler 480 (KAPA biosystems, Worburn, MA, USA) according to the manufacturer's protocol. In total, 20 ng of the template cDNA and 2 pmol of primers were used for each reaction. The PCR conditions were 30 s at 95 • C, followed by 40 cycles of 15 s at 95 • C, 5 s at melting temperature (Tm), and 26 s at 72 • C (fluorescence measurement). Melting curve analysis was performed under the following conditions directly after the PCR run: 10 s at 95 • C, 60 s at 65 • C and 1 s at 97 • C. Relative gene expression was calculated using the 2 -∆∆CT method. The expression of CDKs was normalized to GAPDH gene expression. The analysis was performed using primers, purchased from Eurofins MWG-Biotech AG (Ebersberg, Germany). Primers sequences and melting temperatures (Tm) is presented in Table 3.

Isolation and Structure Elucidation of Compounds 1-8
We used a column chromatography on polyamide sorbent to isolate the fractions of the bioactive prenylated polyphenolic compounds. The fractions were tested for the presence of polyphenolics using TLC plates treated with FeCl 3 and an HPLC-PDA-MS technique. The fractions evaluated positively in a FeCl 3 test were subsequently separated using silica gel column. Individual compounds were further purified using preparative HPLC. Compounds 1-6 were identified by comparison of their HPLC-PDA-MS and NMR spectra with previously published data [13]. Apart from prenylated polyphenolic compounds 1-6, we isolated and identified two new prenylated pterocarpan derivatives 7 and 8 ( Figure 1). prenylated polyphenolic compounds. The fractions were tested for the presence of polyphenolics using TLC plates treated with FeCl3 and an HPLC-PDA-MS technique.  The structures of the isolated compounds were elucidated using mass spectrometry, NMR and CD spectroscopy (for the detailed experimental spectral data of the new compounds please see Supplementary data). Optically active prenylated polyphenolic compound 7 was isolated from L. bicolor root bark as a white amorphous powder. Its molecular formula was determined as  The structures of the isolated compounds were elucidated using mass spectrometry, NMR and CD spectroscopy (for the detailed experimental spectral data of the new compounds please see Supplementary data). Optically active prenylated polyphenolic compound 7 was isolated from L. bicolor root bark as a white amorphous powder. Its molecular formula was determined as C 26 (Table 1). 1 H NMR spectrum also revealed characteristic signals of protons of the pterocarpan skeleton at δ H 3.48 (1H, m), 3.66 (1H, t, J = 11.0), 4.22 (1H, dd, J = 5.0, 11.0) and 5.40 (1H, d, J = 7.0) assigned to H-6a, two H-6, and H-11a protons, respectively. The signals at δ H 7.40 (1H, d, J = 8.4), 6.55 (1H, dd, J = 2.6, 8.4), and 6.40 (1H, d, J = 2.6) belonged to H-1, H-2, and H-4 protons of ring A, respectively, and formed a spin system. The singlet signal at δ H 6.66 (1H, s) was assigned to the aromatic proton at C-7 of ring D. The location of the geranyl side chain at C-10 of the pterocarpan skeleton was determined based on the correlation between the proton signal of 2H-1' at δ H 3.32 and carbon signals of C-9, C-10 иC-10a at δ C 144.5, 111.8, and 152.4, respectively, in the HMBC spectrum of 7. The position of the methoxy group was determined as C-8 based on the observed correlation fbetween the signal of its protons at δ H 3.85 and the signal of C-8 at δ C 141.0 in the HMBC spectrum of 7 (Table 1). The signals in the 1 H and 13 C spectra were completely assigned using HSQC, ROESY, and HMBC data ( Table 1) [18].
We observed a correlation between the proton signal of H-7 at δ H 7.34 (1H, s) of the pterocarpan fragments and the carbon signal of C-4 of the carbonyl group at δ C 194.6 in the HMBC spectrum of 8. Thus, we concluded that the carbonyl group was attached to C-8 of a pterocarpan fragment. The signal of the hydroxy group proton at C-9 of the pterocarpan fragment had a δ H of 12.83, which also confirmed that the carbonyl group was located at C-8 and formed a hydrogen bond with this hydroxy group. The signals in 1 H and 13 C NMR spectra of 8 were assigned using HSQC, HMBC, and ROESY experiments. Thus, compound 8, named lespebicolin A, was assumed to be a dimeric flavonoid consisting of the 2-arylbenzofuran and the pterocarpan fragments linked via carbonyl group.
The absolute configurations of the asymmetric centers at C-6a and C-11a in the pterocarpan fragment of 8 were determined based on the negative optical rotation value ([α] D 24 − 95º) and CD spectral data similar to those of lespecyrtins H 1 -H 4 [18]. Thus, in addition to the six polyphenolic compounds 1-6 previously isolated from L. bicolor stem bark [13], we isolated two new pterocarpans 7 and 8 from L. bicolor root bark. A comparison of the HPLC profiles of L. bicolor stem bark and root bark showed that compounds 7 and 8 are also present in L. bicolor stem bark but in smaller amounts than in root bark (Figure 2).
We observed a correlation between the proton signal of H-7 at δН 7.34 (1H, s) of the pterocarpan fragments and the carbon signal of C-4 of the carbonyl group at δC 194.6 in the HMBC spectrum of 8. Thus, we concluded that the carbonyl group was attached to C-8 of a pterocarpan fragment. The signal of the hydroxy group proton at C-9 of the pterocarpan fragment had a δН of 12.83, which also confirmed that the carbonyl group was located at C-8 and formed a hydrogen bond with this hydroxy group. The signals in 1 H and 13 C NMR spectra of 8 were assigned using HSQC, HMBC, and ROESY experiments. Thus, compound 8, named lespebicolin A, was assumed to be a dimeric flavonoid consisting of the 2-arylbenzofuran and the pterocarpan fragments linked via carbonyl group.
The absolute configurations of the asymmetric centers at C-6a and C-11a in the pterocarpan fragment of 8 were determined based on the negative optical rotation value ([α]D 24 − 95º) and CD spectral data similar to those of lespecyrtins H1-H4 [18].
Thus, in addition to the six polyphenolic compounds 1-6 previously isolated from L. bicolor stem bark [13], we isolated two new pterocarpans 7 and 8 from L. bicolor root bark. A comparison of the HPLC profiles of L. bicolor stem bark and root bark showed that compounds 7 and 8 are also present in L. bicolor stem bark but in smaller amounts than in root bark (Figure 2). Recently, Korean colleagues isolated five pterocarpans, two new coumestans, and two new arylbenzofurans with prenyl and geranyl substituents in their structures [14]. These isolated compounds contain a methoxy group at C-1 and exhibited antiproliferative effects on human leukemia cells. Of note, we could not find these compounds in the stem bark and root bark of L. bicolor harvested in the South of the Primorskiy region of the Russian Far East. The polyphenolic compounds isolated by us from the stem bark and root bark of L. bicolor also contained the pterocarpan skeleton and geranyl side chains in their structures. However, they did not have a methoxy group at C-1. Moreover, for the first time, we have isolated a new dimeric flavonoid lespebicolin A (8) harboring two geranyl side chains. Previously, the related dimeric flavonoids had been found in L. homoloba, L. floribunda and L. cyrtobotria [18,19]. Thus, the chemical compositions of polyphenolic compounds of L. bicolor growing in the Primorskiy region and South Korea differed significantly. This may be because L. bicolor samples collected in the Primorye Region (the South of the Russian Far East) and South Korea belong to different variants of this species [20].

Investigation of Anticancer In Vitro Activity of the Compounds 1-8
Next, we examined the cytotoxic activity of the isolated compounds in prostate cancer cell lines as well as in non-cancer cells. To get a first impression on the impact of these novel compounds for the treatment of advanced prostate cancer, human drug-resistant cell lines PC-3 and 22Rv1 were used. Both cell lines are known to be castration-resistant (androgen-independent). In addition, the cell lines are known to exhibit resistance to novel second-generation androgen receptor (AR) targeting drugs, e.g., abiraterone and enzalutamide, due to the loss or alternative splicing of the AR, respectively [21,22]. Moreover, PC-3 cells have been reported to be rather resistant to docetaxel, and therefore, are known as one of the most aggressive prostate cancer cell lines used as in vitro and in vivo models. Notably, all eight isolated compounds exhibited a cytotoxic activity in both prostate cancer cell lines in the micromolar range, while non-malignant MRC-9 cells were less affected under Recently, Korean colleagues isolated five pterocarpans, two new coumestans, and two new arylbenzofurans with prenyl and geranyl substituents in their structures [14]. These isolated compounds contain a methoxy group at C-1 and exhibited antiproliferative effects on human leukemia cells. Of note, we could not find these compounds in the stem bark and root bark of L. bicolor harvested in the South of the Primorskiy region of the Russian Far East. The polyphenolic compounds isolated by us from the stem bark and root bark of L. bicolor also contained the pterocarpan skeleton and geranyl side chains in their structures. However, they did not have a methoxy group at C-1. Moreover, for the first time, we have isolated a new dimeric flavonoid lespebicolin A (8) harboring two geranyl side chains. Previously, the related dimeric flavonoids had been found in L. homoloba, L. floribunda and L. cyrtobotria [18,19]. Thus, the chemical compositions of polyphenolic compounds of L. bicolor growing in the Primorskiy region and South Korea differed significantly. This may be because L. bicolor samples collected in the Primorye Region (the South of the Russian Far East) and South Korea belong to different variants of this species [20].

Investigation of Anticancer In Vitro Activity of the Compounds 1-8
Next, we examined the cytotoxic activity of the isolated compounds in prostate cancer cell lines as well as in non-cancer cells. To get a first impression on the impact of these novel compounds for the treatment of advanced prostate cancer, human drug-resistant cell lines PC-3 and 22Rv1 were used. Both cell lines are known to be castration-resistant (androgen-independent). In addition, the cell lines are known to exhibit resistance to novel second-generation androgen receptor (AR) targeting drugs, e.g., abiraterone and enzalutamide, due to the loss or alternative splicing of the AR, respectively [21,22]. Moreover, PC-3 cells have been reported to be rather resistant to docetaxel, and therefore, are known as one of the most aggressive prostate cancer cell lines used as in vitro and in vivo models. Notably, all eight isolated compounds exhibited a cytotoxic activity in both prostate cancer cell lines in the micromolar range, while non-malignant MRC-9 cells were less affected under the treatment ( Table 4). The selectivity index evaluation (SI; MRC-9 vs. PC-3 cells) has revealed compounds 1, 2, 3 as well as 6 and 8 to be the most promising and selective in human drug-resistant prostate cancer cells (SI = 2.5~8) ( Table 4). Remarkably, a well-established cytotoxic chemotherapeutic drug cisplatin applied for the treatment of different cancer types and used in the current research as a positive control was less selective (SI = 0.72) ( Table 4). For further investigations on the mechanisms of action, we have chosen the aggressive and drug-resistant prostate cancer PC-3 cell line. To investigate the mechanism of action contributing to the cytotoxicity in PC-3 cells, we investigated the effects on DNA fragmentation, a well-established marker of cellular apoptosis. Therefore, we evaluated the effects of the isolated compounds at four different concentrations of 0.5 µM, 1 µM, 5 µM, and 10 µM by flow cytometry (Figure 3). All the substances apart from structurally different lespebicolin A (8) revealed a significant apoptosis induction at the tested concentrations ( Figure 3). Therefore, prenylated dimeric flavonoid 8 may either reveal a non-apoptotic character of cell death or a predominantly antiproliferative rather than cytotoxic effect. Note, compound 3 exhibited the most pronounced apoptosis-inducing effect.  To investigate the mechanism of action contributing to the cytotoxicity in PC-3 cells, we investigated the effects on DNA fragmentation, a well-established marker of cellular apoptosis. Therefore, we evaluated the effects of the isolated compounds at four different concentrations of 0.5 μM, 1 μM, 5 μM, and 10 μM by flow cytometry (Figure 3). All the substances apart from structurally different lespebicolin A (8) revealed a significant apoptosis induction at the tested concentrations ( Figure 3). Therefore, prenylated dimeric flavonoid 8 may either reveal a non-apoptotic character of cell death or a predominantly antiproliferative rather than cytotoxic effect. Note, compound 3 exhibited the most pronounced apoptosis-inducing effect. Next, we examined the effects of the isolated compounds on the cell cycle progression of PC-3 cells. For compounds 1-3, a significant accumulation of the cells in S-phase as well as slight accumulation in G2/M-phase was observed ( Figure 4). Thus, these compounds mainly induced a Sphase arrest, whereas the compounds 4-8 led to a pronounced G1-phase cell cycle arrest (Figure 4). Due to the significant alterations of the cell cycle, we further investigated the effects of the isolated compounds on the cyclin-dependent kinases (CDKs). These proteins are known to control proliferation as well as some apoptotic processes. We have examined the expression of mRNAs corresponding to the nine most studied CDKs known to be involved in the cell cycle progression of cancer cells, namely, CDK1-CDK9. Indeed, we detected the inhibitory effects of compounds 1, 2, and 3 on the expression of mRNA corresponding to several essential CDKs ( Figure 5). The most pronounced effects were observed for CDK1, CDK2, CDK4, and CDK5. Remarkably, these effects correlated well with the cytotoxicity of the isolated compounds. Thus, the strongest inhibitory effect on the corresponding mRNA expression was observed with the most cytotoxic compound 3 (IC50 = 2.1 μM in PC-3 cells, Table 3).
CDKs are known to play an important role in the development and progression of several cancer types, including human prostate cancer. These kinases are involved in the cell cycle control and are often overexpressed or mutated in cancer cells [23]. Therefore, CDKs are an attractive target in anticancer therapy and are currently actively explored [23]. In particular, CDK1 and CDK6 are known to phosphorylate AR and activate its transcriptional activity at different phosphorylation sites finally leading to a poorer clinical prognosis [24][25][26]. Expression and activity of CDK2 are associated with prostate cancer relapse in patients and with cancer cell invasion [27]. In addition, the particular Due to the significant alterations of the cell cycle, we further investigated the effects of the isolated compounds on the cyclin-dependent kinases (CDKs). These proteins are known to control proliferation as well as some apoptotic processes. We have examined the expression of mRNAs corresponding to the nine most studied CDKs known to be involved in the cell cycle progression of cancer cells, namely, CDK1-CDK9. Indeed, we detected the inhibitory effects of compounds 1, 2, and 3 on the expression of mRNA corresponding to several essential CDKs ( Figure 5). The most pronounced effects were observed for CDK1, CDK2, CDK4, and CDK5. Remarkably, these effects correlated well with the cytotoxicity of the isolated compounds. Thus, the strongest inhibitory effect on the corresponding mRNA expression was observed with the most cytotoxic compound 3 (IC 50 = 2.1 µM in PC-3 cells, Table 3). Biomolecules 2020, 10, x 14 of 16

Conclusion
In conclusion, we isolated two new (7 and 8) and six recently reported (1-6) polyphenolic compounds from the root bark of L. bicolor. Structures were established using mass spectrometry and CDKs are known to play an important role in the development and progression of several cancer types, including human prostate cancer. These kinases are involved in the cell cycle control and are often overexpressed or mutated in cancer cells [23]. Therefore, CDKs are an attractive target in anticancer therapy and are currently actively explored [23]. In particular, CDK1 and CDK6 are known to phosphorylate AR and activate its transcriptional activity at different phosphorylation sites finally leading to a poorer clinical prognosis [24][25][26]. Expression and activity of CDK2 are associated with prostate cancer relapse in patients and with cancer cell invasion [27]. In addition, the particular importance and clinical relevance of CDK5 for prostate cancer growth have been highlighted in different publications. Thus, it was reported to phosphorylate and stabilize AR leading to the promotion of its transcriptional activity [28,29]. Moreover, CDK5 stimulates the growth of AR-negative prostate cancer cells via Akt activation [30]. Of note, the CDK4/6 inhibitor is currently undergoing clinical trials in metastatic castration-resistant prostate cancer. They can promote cell senescence as well as disruption of cancer cells in vivo by the induction of cytotoxic T cell-mediated cell death. Moreover, inhibition of CDK4/6 may cause S-phase cell cycle arrest, which has been observed for compounds 1-3 ( Figure 4) [31,32].
Of note, the other compounds 4-8 did not exhibit pronounced effects on CDKs expression ( Figure 5) but still inhibited cell cycle progression, however, in the G1 phase ( Figure 4). This effect may result from the other processes related either to apoptosis induction or inhibition of different CDKs at the post-transcriptional (inhibition of CDKs protein expression) or even post-translational levels (inhibition of CDKs activity). Indeed, a free hydroxyl group at C-8 (e.g., pterocarpans 1-3) seems to be important for the inhibition of CDKs mRNA expression by these natural compounds. Hence, compound 2, containing 8-OH was fairly active in this experiment, however, very structurally similar new compound 7 containing a methylated hydroxyl group (8-OCH 3 ) did not exhibit this activity; moreover, it was less cytotoxic and had a different effect on cell cycle progression (i.e., G1-phase arrest instead of S-phase arrest). In addition, new compound 8 has a pterocarpan fragment related to compounds 1, 2 and 3. However, this compound is substituted at C-8 and therefore similar to 7 did not exhibit any pronounced effect on CDKs mRNA expression and could not induce an S-phase cell cycle arrest.

Conclusions
In conclusion, we isolated two new (7 and 8) and six recently reported (1-6) polyphenolic compounds from the root bark of L. bicolor. Structures were established using mass spectrometry and NMR spectroscopy. The natural compounds were active and selective in human drug-resistant prostate cancer cells in vitro. Prenylated pterocarpans 1-3 effectively inhibited cell cycle progression of human cancer cells in S-phase. This was associated with the inhibition of mRNA expression corresponding to several human CDKs. Compounds 4-8 induced G1-phase cell cycle arrest without any pronounced effect on CDK mRNA expression. The non-substituted hydroxy group at C-8 in ring D of the pterocarpan skeleton seems to be important for the CDKs inhibitory activity.

Conflicts of Interest:
The authors declare no conflict of interest.