Enhancement of Anticancer Potential of Pterostilbene Derivative by Chalcone Hybridization

Pterostilbene, a natural metabolite of resveratrol, has been indicated as a potent anticancer molecule. Recently, several pterostilbene derivatives have been reported to exhibit better anticancer activities than that of the parent pterostilbene molecule. In the present study, a series of pterostilbene derivatives were designed and synthesized by the hybridization of pterostilbene, chalcone, and cinnamic acid. The cytotoxic effect of these hybrid molecules was determined using two oral cancer cell lines, HSC-3 and OECM-1. (E)-3-(2-((E)-4-Hydroxystyryl)-4,6-dimethoxyphenyl)-1-(2-methoxyphenyl)prop-2-en-1-one (4d), with IC50 of 16.38 and 18.06 μM against OECM-1 and HSC-3, respectively, was selected for further anticancer mechanism studies. Results indicated that compound 4d effectively inhibited cell proliferation and induced G2/M cell cycle arrest via modulating p21, cyclin B1, and cyclin A2. Compound 4d ultimately induced cell apoptosis by reducing the expression of Bcl-2 and surviving. In addition, cleavage of PARP and caspase-3 were enhanced following the treatment of compound 4d with increased dose. To conclude, a number of pterostilbene derivatives were discovered to possess potent anticancer potentials. Among them, compound 4d was the most active, more active than the parent pterostilbene.


Introduction
Pterostilbene, a natural metabolite from dimethylation of resveratrol, is isolated from the Pterocarpus santalinus and is also found in blueberries and grapes [1]. The antioxidant activity of pterostilbene has been implicated in numerous potential preventive and therapeutic properties in neurological, carcinogenic, inflammation, metabolic, and bacterial disorders [2,3]. Similar to resveratrol, pterostilbene has been proved to be a potent anticancer agent [1] with anticancer activities through different mechanisms, such as modulation of signal transduction pathways [4], cell cycle regulation [5], cell differentiation [6], and oncogenes suppression [7]. Pterostilbene is generally safe for human use and the dose can be up to 250 mg/day in a clinical study [8]. However, more than five times the equivalent mean human intake dose of pterostilbene was required to significantly inhibit tumor growth in several in vivo studies [9][10][11][12].
The half-maximal inhibitory concentration (IC 50 ) profile of pterostilbene in some cancer cell lines has been investigated, such as cervical cancer cell line (IC 50 = 108.7 µM for HeLa; 44.45 µM for CaSki; 91.15 µM for SiHa) [13], human colon cell line (IC 50 = 47.1 µM for HCT-116, 80.6 µM for HT-29) [14], breast cancer cell line (IC 50 = 65 µM for MCF-7) [15], and cisplatin-resistance human oral cancer cell (IC 50 = 98.29 µM for CAL27) [16]. Recently, several derivatives of pterostilbene were synthesized and reported with improved activity [17]. Hsieh et al. reported a series of ester derivatives from modification of pterostilbene. One of these derivatives, , exhibited potent activity against CAL27 and improved the IC 50 from 98.29 to 32.6 µM, which showed a promising therapeutic potential [16,18]. Pterostilbene carboxaldehyde thiosemicarbazone (PTERC-T), a pterostilbene derivative, was reported to effectively decrease the IC 50 of MCF-7 from 65 to 25 µM and reducing the tumor volumes in Ehrlich ascitic cell xenograft [15]. 3 -Hydroxypterostilbene (HPSB), a more potent pterostilbene analogue, effectively inhibited the growth of human colon cancer cells with IC 50 = 40.2 µM for HCT-116 and 70.9 µM for HT-29 by inducing apoptosis and autophagy [14]. These results strongly suggest that the modification of pterostilbene may generate analogues with more enhanced anticancer activity. However, for therapeutic use, the efficacy of these analogues still needs to be improved.
Chalcone has the structure of 1,3-diaryl-2-propen-1-one, using three carbon units with α, β-unsaturated ketones to connect two sides of the aryl group. Chalcone is classified as a flavonoid and its derivatives can be widely found in natural fruits and vegetables. Chalcone derivatives also existed in many traditional therapeutical plants. Modern research of chalcone derivatives has revealed that its structure has diverse pharmacological effects, such as anticardiovascular and metabolic diseases [19,20], anticancer [21,22], antiretrovirus [23], anti-inflammation [24,25], and antibacteria activities [26]. In particular, the chalcone structure has the potential ability of autophagy-inducing, vascular-inhibition, and kinase-inhibition, which play an important role in anticancer drug development [27][28][29]. Cinnamic acid containing an aryl ring with α, β-unsaturated ketone, was reported to have anticancer activity. Therefore, both chalcone and cinnamic acid have been adapted as versatile structural moieties in many cases of cancer drug development [30].
In a previous study, the combination of resveratrol and chalcone have been synthesized and tested for their anticancer activity. Resveratrol-chalcone derivatives were first published by Ruan et al. and found to be a potential antitubulin agent with IC 50 values 6.07-350.20 µM [31]. Similar derivatives, published by Shin et al. in 2019, was found to against HCT-116 with IC 50 values 3.42-91.41 µM [32]. In this study, pterostilbene, chalcone, and cinnamic acid were hybridized to generate two series of pterostilbene derivatives ( Figure 1). Totally, 19 compounds were obtained and IC 50 of each compound was tested on two human oral squamous carcinoma cell lines (OECM-1, HSC-3). According to the IC 50 profile of all pterostilbene derivatives, (E)-3-(2-((E)-4-hydroxystyryl)-4,6dimethoxyphenyl)-1-(2-methoxyphenyl)prop-2-en-1-one (4d), with the highest antiproliferation ability was selected for subsequent serial evaluations of anticancer performance as well as the mechanism.
For compound 2j, esterification of pterostilbene and caffeic acid gave complicated results, probably due to the interference of both hydroxyl groups in caffeic acid. We circumvented this dilemma by using malonic acid and acetic anhydride as started materials to obtain compound 6. Then, pterostilbene was reacted with compound 6 and 3,4-dihydroxybenzaldehyde through two-step reactions to obtain compound 2j (Scheme 2). All of the synthesized compounds were examined by NMR, LC-Mass for structural identifications, and the data were shown in the section of materials and methods (NMR spectra could be found in Supplementary Materials). Scheme 2. Synthesis of compound 2j.

Screening of Compound 2 and 4 Series by Cytotoxic Activity
In the beginning, the effect of each compound on the viability of two human oral squamous carcinoma cell (OSCC) lines was studied using MTT (dimethylthiazol-dipheny ltetrazolium bromide) assay. Among these two cell lines, HSC-3 was one of the most highly aggressive OSCC line. The cell viability and IC 50 was assessed after 72 h treatment with each synthesized compound at different concentrations (Table 1). Pterostilbene, with IC 50 = 40.19 µM and > 50 µM in OECM-1 and HSC-3, respectively, was used as a positive control. Most of the pterostilbene-cinnamic acid derivatives 2a-2i showed no cytotoxic effect against HSC-3 with IC 50 of >50 µM with the exception of compound 2j, which exhibited an IC 50 of 18.53 µM. Compound 2g exhibited comparable cytotoxicity to the positive pterostilbene against OECM-1. The IC 50 of 2j was 22.26 µM against OECM-1, while compounds 2a-2f, 2h, and 2i were non-cytotoxic with IC 50 of >50 µM, respectively.
Another series of pterostilbene-chalcone derivatives 4 exhibited more potent cytotoxic activity. Except for compound 4h (IC 50 > 40 µM), all the other tested compounds inhibited the viability of OECM-1 cells and the IC 50 ranged from 9.62 to 22.8 µM. Among these compounds, compound 4i exhibited the most potent cytotoxic activity against OECM-1 with IC 50 9.62 µM. Similar results were observed in HSC-3 cells; all tested compounds except compound 4h (IC 50 > 50 µM) showed antitumor activities. Among these compounds, the IC 50 of 4c-d and 4g were less than 20 µM. Compound 4d with IC 50 18.06 µM exhibited the most potent antitumor activity against HSC-3.
This evidence strongly supported our design, in that the cytotoxic effect of pterostilbene in OSCC lines could be enhanced apparently by a combination with chalcone. As a result, compound 4d, one of the most potent anticancer compounds, was chosen for further mechanism studies.

Structure-Activity Relationship
From the results of preliminary screening, we figured out the structure-activity relationship ( Figure 2). Compound 2 series, the ester derivatives of pterostilbene and selected cinnamic acid, was anticipated to show the characteristics of the prodrug, which could be metabolized to regenerate pterostilbene and cinnamic acid and exhibit synergistic anticancer activity. However, most of the compounds in this series showed no activity against OECM-1 and HSC-3. Only the caffeic acid derivative 2j, 3 ,4 -dihydroxy substituted, had the anticancer activity with IC 50 18.53 µM against HSC-3. We thus inferred that the anticancer activity was produced by caffeic acid moiety. Compound 4 series were the derivatives of pterostilbene and chalcone. Except for compound 4h, all compounds in this series exhibited more potent inhibitory activity on cell viability than pterostilbene. Moreover, the anticancer effects of compound 4 series with chalcone modification were better than compound 2 series. We also figured out the relationship between the substituent's site and the anticancer activity in HSC-3. While comparing methoxy-substituted compounds (4b, 4c, 4d), the anticancer activities against HSC-3 in increasing order were as follows: 4b (IC 50 : 28.93 µM) < 4c (IC 50 : 18.22 µM) < 4d (IC 50 : 18.06 µM). The result indicated that 2-substituted derivative was more active than the 3-substituted counterpart, which in turn was more active than the 4-substituted counterpart. The similar order was also found in fluoro-substituted compounds, 4e (IC 50 : 19.13 µM) ∼ = 4f (IC 50 : 20.41 µM) < 4g (IC 50 : 18.43 µM), and hydroxy-substituted compounds, 4h (IC 50 > 50 µM) < 4i (IC 50 : 19.45 µM). There was no obvious difference on anticancer activity between different substituents in HSC-3. However, with substituent was more active than no substituent in most cases except for compound 4h.

Compound 4d Effectively Inhibited the Cell Proliferation through Modulation of Cell Cycle-Regulatory Proteins
To further investigate the anticancer performance of compound 4d, the cell proliferation and cell cycle of the two cell lines were evaluated after compound treatment. Cell proliferation of HSC-3 and OECM-1 cells was assayed by the incorporation of BrdU, an analogue of thymidine, which is taken up into the newly formed DNA by dividing cells. The result showed that the BrdU incorporation was significantly inhibited by compound 4d at the concentrations of 10, 25, and 50 µM, while pterostilbene treatment did not inhibit but slightly enhanced the incorporation with the same dose ( Figure 3). Specifically, as compared with the control group, treatment of 10, 25, and 50 µM of compound 4d resulted in 15, 68, 70% decrease and 58, 89, 93% decrease in BrdU incorporation of HSC-3 and OECM-1, respectively ( Figure 3). Cell cycle was analyzed in these cells after compound treatment. As shown in Figure  Compound 4d effectively inhibited cell proliferation and induced cell cycle arrest. The influence on the cell cycle regulatory molecules was further investigated. These include cdc2, cyclin-dependent kinases (CDKs), which work with cyclins to control the progression of the cell cycle, and p21, the CDK inhibitor known to suppress DNA synthesis and cellcycle progression by inhibiting CDKs. Furthermore, proliferating cell nuclear antigen (PCNA), a cofactor of DNA polymerase-delta and essential for cell proliferation [33], was analyzed. The protein expression of these regulators was analyzed to elucidate the molecular mechanism. As shown in Figure 5, treatment of compound 4d induced a remarkable increase of p21 and decrease of PCNA, cyclin B1, cyclin A2, and cdc2 expression in both HSC3 ( Figure 5A) and OECM-1 ( Figure 5B) cells. Negligible or no change in the expression of these proteins as observed in both lines after pterostilbene treatment. Together, these results indicated that compound 4d effectively inhibited cell proliferation and induced cell-cycle arrest at the G2/M phase through modulation of cell cycle regulatory proteins in OECM-1 and HSC3 cells. For cell proliferation analysis, HSC-3 and OECM-1 cells were treated by compound 4d or pterostilbene with 0 (DMSO control), 1, 10, 25, and 50 µM for 24 h followed by BrdU incorporation assay. Incorporated BrdU in cells was determined by measurement of 450 nm absorbance using a spectrophotometer. The data is presented as the mean ± standard deviation of three independent experiments. Statistically significant differences of OD450 are indicated as follows: * p < 0.05, ** p < 0.01, *** p < 0.001 between each group and vehicle control; ### p < 0.001 between compound 4d treatment and the same concentration of pterostilbene treatment.

Compound 4d Induced Apoptosis through Regulation of PARP and Caspase Activation
Compound 4d has been indicated to induce proliferation inhibition and cell cycle arrest in OECM-1 and HSC-3 cells. Whether cell apoptosis, another crucial therapeutic and anticancer index, could be induced by compound 4d was subsequently investigated. Cytotoxicity of compound 4d to both lines was measured by MTT assay. The results showed that, with the treatment of 25 and 50 µM-compound 4d, the cell viability dropped to 40 and 28% in HSC-3 and 51 and 29% in OECM-1 cells, as compared to 100% cell viability in control group ( Figure 6). The IC 50 of 4d at 24, 48, and 72 h were 21, 18, and 18 µM on HSC3 cells, and were 25, 20, and16 µM on OECM-1 cells, respectively (Table 2). No influence on the viability of both cells was observed after treatment of pterostilbene, even with 50 µM. A similar result was observed by the evaluation of Annexin V staining and flow cytometry. As shown in Figure 7A,C, when treated with 1, 10, and 25 µM compound 4d, apoptotic (Annexin V + ) population of HSC-3 cells increased with treatment dose, from 7.2 ± 0.6, 9.5 ± 0.9, to 18.3 ± 0.8%, respectively, as compared to 5.3 ± 0.2% in the control (DMSO) group. For OECM-1 cells, apoptotic populations were 5.4 ± 0.6, 12.4 ± 0.9, and 19.7 ± 2.0% after treatment with 1, 10, and 25 µM compound 4d, compared to 4.2 ± 1.9% in control group ( Figure 7B,C). Note that the cell apoptosis in both lines was induced by pterostilbene treatment with a dose up to 25 µM (9.0 ± 0.9% and 8.7 ± 2.5% apoptotic population in HSC-3 and OECM-1 cells, respectively). To further elucidate the underneath mechanism, the expression of molecules involved in the apoptosis was evaluated. As shown in Figure 5B, the expression of antiapoptotic Bcl-2 decreased after cells were treated with 25 µM compound 4d, while the pro-apoptotic Bax did not change in its expression. When a cell undergoes the final step of apoptosis, cleavage of Caspase 3 and PARP (Poly (ADP-ribose) Polymerase) are prerequisites. Our results showed that treatment of 10 and 25 µM compound 4d remarkably increased the cleavage of PARP and caspase 3 in both lines ( Figure 8A,B), as compared to those after pterostilbene treatment. In addition, the expression of survivin, which inhibits apoptosis protein family, was found decreased after treatment of 25 µM compound 4d. Taken together, these results indicated compound 4d was able to induce apoptosis through regulation of apoptotic related proteins.

General Information
Commercial reagents were used without additional purification. Melting points were tested with an Electrothermal IA9100 (Electrothermal, Staffordshire, UK) micro-melting point apparatus and were uncorrected. NMR spectra were performed with a Varian Unity 400 MHz spectrometer using tetramethylsilane (TMS, Merck, Darmstadt, Germany) as an internal standard and DMSO-d 6 as solvent. Chemical shifts were described as (ppm). Splitting patterns are presented as follows: s = singlet; d = doublet; t = triplet; q = quartet; dd = double doublet; m = multiplet. Analytical thin-layer chromatography (TLC) was carried out by an Art. 5554 Kieselgel 60 GF254 (Merck, Darmstadt, Germany) made by E. Merck. The spots of compounds were examined with a UV light indicator irradiated at 254 and 366 nm. Art. 7734 Kieselgel 60 GF254 (70-400 mesh, Merck, Darmstadt, Germany) produced by E. Merck was used for preparing column chromatography. Waters ZQ-4000 (Waters, Milford, MA, USA) liquid chromatography electrospray ionization mass spectrometry was used to record mass spectra.

General Procedure for the Synthesis of Compound 4a-i
The selected acetophenone (2 mmol) was firstly dissolved in 10 mL methanol and 1 mL 3M NaOH (aq.) at room temperature for 20 min. Then compound 3 (0.28 g, 1 mmol) was slowly added and refluxed 8 h. After the reaction was completed (monitored by TLC), the mixture was cooled to room temperature and concentrated in a vacuum. The crude product was then purified with silica gel column chromatography, using ethyl acetate:hexane = 1:2 as the mobile phase. The target product was concentrated in a vacuum and recrystallized in ethyl acetate and hexane. The pure compound was obtained in 17-45% yield.
well-plate and incubated with test compounds for fixed time intervals. Then, the medium was discarded and rinsed by PBS. Afterward, cells were added with 100 µL fresh serum-free containing 10% MTT solution (Cyrusbioscience, Taipei, Taiwan). After incubation at 37 • C for 1 h, MTT-containing solution was removed and the formazan crystals were solubilized with 120 µL of DMSO to detect activities of succinate dehydrogenase in living cells. The absorption levels of each well was measured at 590 nm by a microplate spectrophotometer (Multiskan GO Microplate Spectrophotometer, Thermo Scientific, Vantaa, Finland).

BrdU Incorporation Assay
HSC-3 and OECM-1 cells (5 × 10 3 cells per well) were seeded in 96 well-plate and incubated with test compounds for 24 h. The effect of test agents on cell proliferation was analyzed using a cell proliferation ELISA kit (Millipore Corp., Bedford, MA, USA) with the incorporation of 5-bromo-20-deoxyuridine (BrdU) for 2 h. The absorbance of BrdU uptake was quantified using an ELISA reader at 450 nm (Multiskan GO Microplate Spectrophotometer, Thermo Scientific, Vantaa, Finland).

Apoptosis Detection by Annexin V Staining and Flow Cytometry
HSC-3 and OECM-1 cell death induced by test compounds was accessed by using an Annexin V/PI staining kit (BioLegend, San Diego, CA) following the manual protocol. Briefly, cells were seeded at a density of 5 × 10 5 cells per 60 mm Petri dish. After treatment with different doses of test drugs for 24 h, cells were detached by trypsin and washed by Cell Staining Buffer. Cells were then stained with the Annexin V-FITC and PI (5 µg/mL) in the dark at room temperature for 15 min. The number of apoptosis cells was analyzed by flow cytometry using a BD LSR II Cell Analyzer (BD Biosciences, San Jose, CA, USA). Data were collected at rates of 70-300 events per second at the lowest or medium flow rate, with 10,000 events being recorded per sample. The proportion of apoptotic cells was evaluated by FlowJo (FlowJo, Ashland, OR, USA).

Cell Cycle Analysis by Flow Cytometry
HSC-3 and OECM-1 cells (5 × 10 5 ) were cultured and incubated with test compounds at different concentrations for 24 h. Cells were harvested using accutase (BioLegend, San Diego, CA, USA) and washed twice with PBS. Next, cells were fixed by 70% ethanol at 4 • C overnight. The fixed cells were washed by PBS and suspended with 500 µL of mixture containing PI (40 µg/mL), RNase A (50 µg/mL). Follow by a 30-min incubation in the dark at room temperature, cell pellets were analyzed by flow cytometry BD LSR II (BD Biosciences, San Jose, CA, USA). Cell cycle profiles were measured at rates of 70-300 events per second at the lowest or medium flow rate from 10,000 events per sample. The percentages of cells in different cell cycle phases were evaluated by FlowJo (FlowJo, Ashland, OR, USA).

Statistical Analysis
All experimental data were repeated at least three times and were presented as the mean ± standard deviation (SD). Comparison of data among different concentrations of pterostilbene, compound 4d, and/or vehicle was analyzed by ANOVA followed by Bonferroni post hoc tests.

Conclusions
In the present work, two series of pterostilbene derivatives were synthesized and evaluated in vitro for their anticancer activities (Figure 9). Among them, compound 4d showed the most potent activity against the cell growth of HSC-3 and OECM-1 with IC 50 values of 16.83 and 18.06 µM, respectively, and therefore, was selected as a new lead for further comprehensive evaluation and mechanism investigation. Results indicated that compound 4d effectively inhibited cell proliferation, induced G2/M cell cycle arrest via modulation of the p21, cyclin B1, and cyclin A2. Compound 4d ultimately induced cell apoptosis by the decrease of antiapoptotic protein Bcl-2 and surviving. In addition, cleavage of PARP and caspase-3 were induced and increased following the treatment of compound 4d with increased dose. It is noteworthy that the anticancer effects above were observed in cells after the treatment of compound 4d with a dose range 10~50 µM, while pterostilbene did not show any effect with the same dose, suggesting a superior anticancer potential of compound 4d over its parent compound, pterostilbene. Data Availability Statement: Not applicable. All data which was showed in this study was originally made by the authors.