Metabolites of Ginger Component [6]-Shogaol Remain Bioactive in Cancer Cells and Have Low Toxicity in Normal Cells: Chemical Synthesis and Biological Evaluation

Our previous study found that [6]-shogaol, a major bioactive component in ginger, is extensively metabolized in cancer cells and in mice. It is unclear whether these metabolites retain bioactivity. The aim of the current study is to synthesize the major metabolites of [6]-shogaol and evaluate their inhibition of growth and induction of apoptosis in human cancer cells. Twelve metabolites of [6]-shogaol (M1, M2, and M4–M13) were successfully synthesized using simple and easily accessible chemical methods. Growth inhibition assays showed that most metabolites of [6]-shogaol had measurable activities against human cancer cells HCT-116 and H-1299. In particular, metabolite M2 greatly retained the biological activities of [6]-shogaol, with an IC50 of 24.43 µM in HCT-116 human colon cancer cells and an IC50 of 25.82 µM in H-1299 human lung cancer cells. Also exhibiting a relatively high potency was thiol-conjugate M13, with IC50 values of 45.47 and 47.77 µM toward HCT-116 and H-1299 cells, respectively. The toxicity evaluation of the synthetic metabolites (M1, M2, and M4–M13) against human normal fibroblast colon cells CCD-18Co and human normal lung cells IMR-90 demonstrated a detoxifying metabolic biotransformation of [6]-shogaol. The most active metabolite M2 had almost no toxicity to CCD-18Co and IMR-90 normal cells with IC50s of 99.18 and 98.30 µM, respectively. TUNEL (Terminal deoxynucleotidyl transferase dUTP nick end labeling) assay indicated that apoptosis was triggered by metabolites M2, M13, and its two diastereomers M13-1 and M13-2. There was no significant difference between the apoptotic effect of [6]-shogaol and the effect of M2 and M13 after 6 hour treatment.


Introduction
Despite enormous efforts made toward the development of cancer therapies over the past several decades, cancer is still a major public health problem worldwide. Increasing evidence has shown that treatments using specific agents or inhibitors that target only one biological event or a single pathway usually fail in cancer therapy [1]. Conventional chemotherapeutic agents have been shown to be associated with unacceptable toxicity. New approaches to the control of cancer are critically needed. Chemoprevention is an innovative area of cancer research that focuses on the prevention of cancer through pharmacologic, biologic, and nutritional interventions [2]. Accumulating studies have shown that dietary phytochemicals present in plants and fruits, which are generally considered as non-toxic agents, can activate or block multiple important pathways that are implicated in cancer cell survival and growth [1,3,4]. Chemoprevention by edible phytochemicals is now considered to be a safe, inexpensive, readily acceptable and accessible approach to cancer prevention, control and management.
A recent study from our group has demonstrated that [6]shogaol is extensively metabolized in mice and in cancer cells, in which thirteen metabolites (M1-M13) were detected and identified [12]. Trace amount of metabolites M6-M12 were purified from fecal samples collected from [6]-shogaol treated mice, and their structures were characterized based on the analysis of 1 H, 13 C, and 2-D NMR data [12]. Structures of the remaining metabolites (M1-M5 and M13) were elucidated based on the analysis of their MS n (n = 123) spectra and comparison to authentic standards. With its a,b-unsaturated ketone structure, we found that [6]shogaol can be reduced to generate metabolites M6-M9 and M11. We also found that [6]-shogaol is the substrate for thiol conjugation through the mercapturic acid pathway to form metabolites M12M5, M10, M12, and M13. Since most of [6]shogaol is metabolized in vivo and in cancer cells, the critical question is whether the metabolites of [6]-shogaol are bioactive. Even if less potent than [6]-shogaol, they may still contribute to the overall biological activity of [6]-shogaol in vivo. The current challenge to study the bioactivity of metabolites is their lack of commercial availability. With trace amounts of metabolites purified from mouse fecal samples, we were only able to investigate the effects of the two major metabolites M9 and M11 on the growth inhibition and induction of apoptosis in human cancer cells. Our results showed that M9 and M11 are bioactive compounds that can inhibit cancer cell growth and induce apoptosis in human lung and colon cancer cells, albeit with less potency than [6]-shogaol. Therefore, it is timely to develop chemical methods to synthesize the metabolites of [6]-shogaol and to further examine their anti-cancer activities. The current study describes a chemical synthesis of the previously identified metabolites of [6]-shogaol and a bioactivity analysis in cancer cells. Also investigated are the toxicities of [6]-shogaol and its metabolites in normal human lung and colon cells, providing a comparison of [6]-shogaol's toxicity in a non-cancer model.
We have fully characterized the structures of M5-M12 using their 1-D and 2-D NMR and mass spectral data in our previous study [12]. Therefore, the structures of these synthetic compounds were confirmed by comparison of their 1 H and 13 C NMR spectra with those of authentic standards obtained from mouse fecal samples. Structures of the remaining synthetic metabolites (M1, M2, M4, and M13), previously deduced by multi-stage mass spectrometry techniques [12], were further confirmed by their 1-D and 2-D NMR spectra data for the first time.
M2 showed the molecular formula C 20 H 31 NO 5 S on the basis of positive APCI-MS at m/z 398 [M+H] + and its 1 H and 13 C NMR data. The molecular weight of M2 was 42 mass units less than that of N-acetylcysteine conjugated [6]-shogaol (M5) [12] indicating M2 was the cysteine conjugated [6]-shogaol. This was in agreement with the fact that M2 was made by [6]-shogaol and L-cysteine. This was also supported by the observation of the absence of an acetyl group in the 1 H and 13 C NMR spectra of M2. The linkage of an L-cysteinyl moiety to the [6]-shogaol residue at C-5 was established by HMBC cross-peaks between H Cys -b (d H 3.18 and 2.84) and C-5 (d C 42.3) (Figure 1). Therefore, M2 was confirmed to be 5-cysteinyl-[6]-shogaol.
M1 had the molecular formula of C 20 H 33 NO 5 S on the basis of positive APCI-MS at m/z 400 [M+H] + and its 1 H and 13 C NMR data. The molecular weight of M1 was 2 mass units higher than that of M2, matching with the fact that M1 was a ketone-reduced product of M2, and also supported by the appearance of oxygenated methines ( (Figure 1). The structure of the cysteinyl residue (Cys) was established by 1 H-1 H COSY cross-peaks at H Cys -a/H Cys -b in combination with HMBC correlation between H Cys -b (two sets of protons at d H 3.05-2.95 and 2.84-2.80) to Cys a-CON (d C 175.2) ( Figure 1). Subsequently, the connection of the glutamyl residue with the cysteinyl moiety was established between Glu c-COOH and Cys a-NH 2 through an amide bond, by HMBC correlations found at H Cys -a (d H 4.50) to Glu c-CON (d C 175.2). The attachment of a glycinyl moiety to the cysteinyl residue was found between Cys a-COOH and Gly a-NH 2 , by HMBC correlations observed at H Gly -a (d H 3.80) to Cys a-CON (d C 175.2). Thus, the GSH residue was undoubtedly identified as c-glutamyl-cysteinylglycine. Consequently, linkage of the GSH moiety to the [6]shogaol residue was established at C-5 through a thioether bond by HMBC correlations found at H Cys -b (two sets of protons at d H 3.05-2.95 and 2.84-2.80) to C-5 (d C 42.2). Therefore, M13 was confirmed to be 5-glutathiol-[6]-shogaol.
Separation of M13 isomers on preparative HPLC resulted in two diastereoisomers, M13-1 and M13-2, which had very similar 1 H and 13 C NMR spectra. The major differences were the 1    In order to investigate the influence of stereochemistry on activity, metabolite M13, as a mixture of diastereomers, was separated by reverse phase prep-HPLC into two individual isomers, M13-1 and M13-2. Cancer cells HCT-116 and H-1299 were treated with M13 or its constituent stereoisomers (M13-1 and M13-2) individually, with concentrations ranging from 0 to 80 mM. We found both isomers had similar but slightly less activity than M13, and M13-2 to be slightly more effective than

M2 and M13 Induce Apoptosis in Human Cancer Cells
In many cases, cell death is the result of the activation of the major regulatory pathway called apoptosis, or programmed cell death. We aimed to determine whether apoptosis was triggered following exposure to [6]-shogaol and its metabolites by using a TUNEL assay, which detects the DNA breaks characteristic of cells undergoing apoptosis. In this study, we tested the two most active metabolites against cancer cell growth, M2 and M13, as well as one of the most abundant metabolites, M6, together with [6]shogaol. We incubated H-1299 and HCT-116 cells with them at various concentrations to determine the active range of these compounds versus DMSO only. The results are summarized in Figure 8. After 24 hours of incubation, metabolite M6 did not display any apoptotic effect in HCT-116 and H-1299 cell lines. Significant apoptosis was observed for M2 and M13 in both cell lines, except for M2 in H-1299 cells for the 20 mM concentration. The induction of apoptosis by [6]-shogaol was significantly superior to that of both M2 and M13 at the same concentration  Figure 8A and 8B).
In order to determine if the apoptotic effect observed in Figure 8A and 8B was constant over time, we incubated [6]shogaol, M2 and M13 in HCT-116 cells for only 6 hours ( Figure 8C). After 6 hours of incubation with [6]-shogaol or metabolites M2 or M13, we were able to detect significantly higher levels of apoptosis for all 3 compounds compared to DMSO in HCT-116 cells. Interestingly there was no significant difference between the apoptotic effect of [6]-shogaol and the effects of M2 at 20 and 40 mM and M13 at 40 mM. M13 was significantly more potent than [6]-shogaol at 20 mM. Exposure of HCT-116 cells to M13 isomers M13-1 and M13-2 also showed a higher level of apoptosis, but the isomers' apoptotic effect was significantly inferior compared to [6]-shogaol for both concentrations used. These results show that apoptosis is triggered by [6]-shogaol metabolites M2, M13, M13-1 and M13-2, and is the mechanism responsible, at least partially, for the cell death observed previously.

Discussion
It is now well accepted that natural compounds provide the opportunity to interfere in early stages of cancer or prevent its development altogether [16]. However, these compounds do have their limitations, such as fast in vivo turnover, limited quantities in foodstuffs, and their eventual toxicity at higher doses [17]. It is remarkable that anti-cancer effects for these compounds can be observed in cohort studies despite a short half-life and fast metabolism once ingested [18]. Often, presence of the compound can only be assessed by quantification of metabolites, suggesting that these metabolites circulate in the body for a certain amount of time and potentially interact with biological processes [19]. Consequently, one hypothesis explaining the bioactivity of natural compounds is that the metabolites themselves retain and carry part or all of the original compound's bioactivity. The present study explored this possibility using [6]-shogaol, the main component of dried ginger, and its metabolites.
Our previous study has indicated that [6]-shogaol is extensively metabolized in mice and in cancer cells and more than 90% of [6]shogaol is converted to its metabolites [12]. We also noticed that   human normal cell IMR-90 to some extent metabolized [6]shogaol in a similar way to cancer cells (data not shown). As a result, fast metabolism of [6]-shogaol might offer the chance for its stable metabolites to get involved in biological processes if they possessed bioactivities. In order to verify this hypothesis, it is critical to obtain the metabolites in a stable form. Isolation from their native environment would not be convenient and most likely would not yield the quantities needed for experimental biology. A chemical synthesis is much more desirable, reproducible, and efficient. We have purified M6-M12 in very limited quantities (0.5-17 mg) from fecal samples collected from [6]-shogaol treated mice [12]. In order to have enough quantity to further study the bioactivity of [6]-shogaol metabolites, we describe the chemistry steps allowing the synthesis of the major metabolites of [6]-shogaol in this study. Twelve metabolites of [6]-shogaol (M1, M2, and M4-M13) were synthesized successfully by practicable approaches. The structures of synthetic metabolites were verified using 1-D and/or 2-D NMR data as well as mass spectra.
We compared the growth inhibitory effects of the synthetic metabolites with [6]-shogaol in two human cancer cells and two human normal cells. Our results showed that two metabolites, M2 and M13, possessed the most comparable growth inhibitory effects to [6]-shogaol towards cancer cells. Most importantly, M2 exhibited a discriminatory effect, as it did not seem to be toxic towards normal cells. This effect was not detected with [6]shogaol. M13 also showed less toxic effects towards normal cells compared to [6]-shogaol. In addition, M5, M6 and M8-M12 also had certain potency against the growth of cancer cells, but showed no toxicity towards normal cells with IC 50 values greater than 100 mM (Figures 5 and 6). Our results clearly indicate that metabolites of [6]-shogaol remain bioactive against cancer cells but are much less toxic than [6]-shogaol to normal cells. This confirms our hypothesis that the metabolites themselves retain and carry part or all of the original compound's bioactivity.
Apoptosis is a mechanism often responsible for the induction of cell death in response to internal or external stress. In order to gain insight into the mechanism of action of the metabolites, we performed a TUNEL assay in HCT-116 cells using M2, M6, M13 and its two isomers, M13-1 and M13-2 and compared their activities to that of [6]-shogaol. It showed that both M2 and M13, but not M6, are capable to induce cancer cell apoptosis in both HCT-116 human colon cancer cells and H-1299 human lung cancer cells ( Figure 8A and 8B). For M13, apoptosis induction could not firmly be attributed to one isomer or the other, suggesting that stereo configuration is not important to the bioactivity of this compound. We observed that both M2 and M13 could trigger apoposis in HCT-116 cells at a level similar to that of [6]-shogaol at the 6 hour time point ( Figure 8C). However, after 24 hours of exposure to the metabolites, the percentage of TUNEL-positive cells was mostly unchanged at 20 mM for both M2 and M13 while the effect of [6]-shogaol was remarkably increased. The induction effect of M2 on apoptosis at a concentration of 40 mM dramatically increased at the 24 hour time point compared to that of the 6 hour time point, which was significantly higher than that of M13 at a 40 mM concentration. We also observed a concentration-dependant effect of [6]-shogaol and its metabolites on cancer cell apoptosis, where an increase in concentration of a compound resulted in a corresponding increased percentage of apoptotic cells. Altogether these results suggests that it is likely that the activation of others mechanisms are involved in triggering cancer cell death, however, apoptosis appears to be one of the major activated pathways for [6]-shogaol metabolites to induce cancer cell death.
It is always a challenge to separate stereo isomers. Among all the synthesized metabolites, M13 was the only diastereomers mixture that we were able to separate using non-chiral preparative HPLC column. Our results indicated that M13 as the mixture of M13-1 and M13-2 had slightly better growth inhibitory effects on cancer cells than the two diastereomers alone and M13-1 and M13-2 had similar activity, though M13-2 was slightly more potent than M13-1. More importantly, we did observe M13 as the metabolite of [6]shogaol in the form of a mixture of M13-1 and M13-2 in HCT-116 human colon cancer cells (data not shown). It would be worthwhile to separate the two diastereomers of M2, the most active metabolite of [6]-shogaol, using a chiral column and further determine the effect of stereo configuration on the activity of this compound in future study. We recently identified M2 as a metabolite of [6]-shogoal in humans (unpublished data). Thus, it is important to determine if M2 exists as a mixture of two diastereomers in mice and in humans.
In conclusion, the present study allowed us to show that metabolites of [6]-shogoal can account for the bioactivity of the parent compound, and specifically triggers molecular pathways responsible for cancer cell death in a similar fashion. It suggests that anti-cancer activity attributed to compounds such as [6]shogoal might have been partly misrepresented in the past. What was originally thought to be an effect of natural compound might have been the direct result of rapid metabolism of the compound, rendering it less capable of triggering molecular mechanisms. However, the apparition of its metabolites in target tissues would then supplement, sustain or replace altogether the bioactive effect of the original compound. Supplementary in vivo studies will be needed to obtain a definitive answer to that question, but as of now the study presented here opens novel research possibilities for the study of bioactive metabolites of [6]-shogaol.

General Procedure A for Michael Addition Reaction
A catalyst amount of NaHCO 3 (0.05 eq) was added to a mixture of [6]-shogaol (1.0 eq) and amino acid (3.0 eq) in methanol/water (1:1, v/v). The mixture was stirred at room temperature (rt) for 3-48 h, adjusted pH until 6 with a diluted HOAc solution (0.1 M), and extracted with n-butanol (BuOH) (5 mL 6 3). Combined organic layers were concentrated under reduced pressure at 20uC. The residue was subjected to column chromatography (CC) on Sephadex LH-20, and eluted with 90% ethanol in water, producing the desired thiol conjugates M2, M5, or M13.
General Procedure B for the Synthesis of Ketone Reduced Metabolites using NaBH 4 NaBH 4 (2.5-4.0 eq) was added to a solution of M2, M5 or [6]shogaol (1.0 eq) in methanol at 0uC. After stirring at 0uC for 2 h, the reaction media was neutralized with a diluted HOAc solution (0.1 M) and extracted with n-BuOH (5 mL63). Combined organic layers were concentrated under reduced pressure. The residue was purified by CC on Sephadex LH-20 or preparative TLC to produce the required compounds M1, M4, or M9.

Separation of the M13 Isomers Using Preparative HPLC
Waters preparative HPLC systems with 2545 binary gradient module, Waters 2767 sample manager, Waters 2487 autopurification flow cell, Waters fraction collector III, dual injector module, and 2489 UV/Visible detector, were used to separate M13 isomers. A Phenomenex Gemini-NX C 18 column (250 mm 6 30.0 mm i.d., 5 mm) was used with a flow rate of 20.0 mL/min. The wavelength of UV detector was set at 280 nm. The injection volume was 1.0 mL for each run. The mobile phase consisted of solvent A (H 2 O +0.1% formic acid) and solvent B (MeOH +0.1% formic acid).
M13 (5 mg/mL) was injected to the preparative column and eluted with a gradient solvent system (0% B from 0 to 5 min; 0 to 50% B from 5 to 15 min; 50 to 60% B from 15 to 25 min; 60 to culture medium to the desired final concentrations (final DMSO concentrations for control and treatments were 0.1%). After the cells were cultured for 24 h, the medium was aspirated and cells were treated with 200 mL fresh medium containing 2.41 mmol/L MTT. After incubation for 3 h at 37uC, the medium containing MTT was aspirated, 100 mL of DMSO was added to solubilize the formazan precipitate, and plates were shaken gently for an hour at room temperature. Absorbance values were derived from the plate reading at 550 nm on a Biotek microtiter plate reader (Winooski, VT). The reading reflected the number of viable cells and was expressed as a percentage of viable cells in the control. Both HCT-116 and H-1299 cells were cultured in McCoy's 5A medium. CCD-18Co and IMR-90 cells were cultured in Eagle's modified essential medium (EMEM). All of the above media were supplemented with 10% fetal bovine serum, 1% penicillin/ streptomycin, and 1% glutamine, and the cells were kept in a 37uC incubator with 95% humidity and 5% CO 2 .

TUNEL (Terminal Deoxynucleotidyl Transferase dUTP Nick end Labeling) Assay
HCT-116 and H1299 cells were seeded in 6-well plates at 1.10 5 cells/well and incubated at 37uC in 5% CO 2 incubator. After 24 hours, fresh media supplemented with DMSO (control), [6]shogaol, M2, M6, or M13 metabolites (20 mM or 40 mM) were added to the wells. After 6 or 24 hours incubation at 37uC in 5% CO 2 incubator, cells were washed and pre-treated for 15 min at room temperature with a solution of 20 mg/ml proteinase K. Cells were then washed twice with phosphate buffer saline pH 7.4 (PBS) and fixed for 10 min at room temperature using 10% neutral formaldehyde solution. After 2 washes in ddH 2 O, cells were resuspended in 100 mL ddH 2 O and applied on silanized microscope slides. Slides were incubated overnight at 37uC, and washed twice with PBS. TUNEL assay was then carried out according to the manufacturer's protocol. Cells were observed under 400X power using a Zeiss microscope A1 (Thornwood, NY). 10 fields per slide were evaluated, and TUNEL+ cells (with brown coloration in the nucleus) were expressed as a percentage of the total number of cells contained in a field.

Statistical Analysis
For simple comparisons between two groups, two-tailed Student's t-test was used. A p-value of less than 0.05 was considered statistically significant in all the tests.