Synthesis and Antiproliferative Activity of Novel A-Ring Cleaved Glycyrrhetinic Acid Derivatives

A series of new glycyrrhetinic acid derivatives was synthesized via the opening of its ring A along with the coupling of an amino acid. The antiproliferative activity of the derivatives was evaluated against a panel of nine human cancer cell lines. Compound 17 was the most active compound, with an IC50 of 6.1 µM on Jurkat cells, which is 17-fold more potent than that of glycyrrhetinic acid, and was up to 10 times more selective toward that cancer cell line. Further biological investigation in Jurkat cells showed that the antiproliferative activity of compound 17 was due to cell cycle arrest at the S phase and induction of apoptosis.


Introduction
Cancer is a leading cause of death worldwide. Its global incidence continues to rise due to the aging and growth of the world population and the increasing adoption of lifestyle choices associated with cancer in developed countries [1]. Plants have been a major source of highly effective conventional drugs for cancer treatment and nowadays they play an important role as a source of leads for the development of potential new agents [2,3]. The plant-derived triterpenoids are proving to be interesting leading compounds as reported in a large number of scientific papers emerging in this field [4][5][6][7][8][9][10][11][12][13][14][15].
Glycyrrhetinic acid (GA) 1 is the hydrolyzed metabolite of glycyrrhizin, a major pentacyclic triterpenoid saponin obtained from the roots of licorice (Glycyrrhiza species) in high yields up to 24% [16,17]. This compound has been shown to inhibit tumor initiation [18][19][20][21][22] and proliferation in several cancer cell lines; its antiproliferative activity is mediated by cell cycle arrest [23,24] and induction of apoptosis [21,[25][26][27][28]. The antitumor effects of GA 1 were also observed in animal models [20,29,30]. Nevertheless, it lacks potency and selectivity as an antitumor agent. Many derivatizations have been performed in order to enhance the potency of GA 1 [31][32][33][34][35][36]. However, the cleavage of its ring A is still poorly explored [37]. On the other hand, is well known that the conjugation of an amino acid moiety to pentacyclic triterpenoids improves their cytotoxicity and their selectivity towards tumor cells [38][39][40]. These findings prompted us to synthesize new GA 1 derivatives via the opening of its ring A along with the coupling with an amino acid. The novel semisynthetic derivatives were tested for their antiproliferative activity against a panel of nine human cancer cell lines. Further biological assays were conducted for the most potent compound 17 in the cancer cell line that yielded the best results (Jurkat cells), to investigate its preliminary mechanism of action. The study of selectivity was performed on human fibroblasts (BJ).

Chemistry
The synthesis of the glycyrrhetinic acid 1 derivatives is outlined in Schemes 1-3. Full structural elucidation of the new glycyrrhetinic acid derivatives was achieved using nuclear magnetic resonance (NMR), mass spectrometry (MS) and elemental analysis. The analytical data obtained for the known compounds 1-5 and 8-10 were in agreement with those reported in the literature [39,[41][42][43].
The synthesis of compounds 2-7 is summarized in Scheme 1. Methyl ester 2 was obtained from the reaction of compound 1, the starting material, with methyl iodide in the presence of potassium carbonate [39]. The 3β-hydroxyl group of compound 2 was then oxidized using the Jones reagent [41] to give the 3-keto derivative 3. The reaction of this derivative with m-chloroperbenzoic acid (m-CPBA) provided lactone 4. The lactone ring of 4 was opened by treatment with p-toluenesulfonic acid (p-TSA) in dichloromethane [42]. Reaction of compound 5 with bis(2-methoxyethyl)aminosulfur trifluoride (Deoxo-Fluor ® ) [44] provided the acyl fluoride intermediate which was reacted either with glycine methyl ester hydrochloride or with L-alanine methyl ester hydrochloride to afford compounds 6 and 7, in yields of 69% and 61%, respectively . We found that the acyl fluoride, in this position of the structure, decomposes on standing. For that reason, the crude compound was employed without further purification, and immediately, in the subsequent reactions. The preparation of compounds 6 and 7 was confirmed by the presence of the proton signals of the amino acid side chains. On the 1 H NMR spectrum of compound 6, the δ signals of the glycine methyl ester side chain were observed around 6.1 ppm (NH), 4.0 ppm (NCH2) and 3.7 ppm (CH3). Compound 7, with an alanine methyl ester side chain, had δ signals around 6.1 ppm (NH), 4.6 ppm (NCH) and 3.7 ppm (CH3). lactone ring of 9 was cleaved by treatment with p-TSA in methanol and dichloromethane [42] to provide compound 10. The derivative 11 and the three pairs of compounds synthesized in the following steps were prepared to explore the influence of the keto group in position C-11 on the antiproliferative activity. The removal of the keto group was performed by a Clemmensen reduction [45] with zinc dust and concentrated HCl in dioxane at room temperature to afford 11 (75%). The reduction was confirmed on the 13 C NMR spectrum, by the absence of the δ signal around 200 ppm, which corresponds to the carbonyl group in ring C. Acyl fluorides 12 and 13 were obtained from the reaction of compounds 10 and 11 with Deoxo-Fluor, in yields of 75% and 61%, respectively. The synthesis of acyl fluorides was detected on the 13 C NMR spectra. The carbon C30 appeared as a doublet with a δ signal around 166 ppm and a coupling constant of 375 Hz, in both compounds 12 and 13. These derivatives were reacted either with glycine methyl ester hydrochloride or with L-alanine methyl ester hydrochloride to afford compounds 14-17, in yields ranging from 43% to 83%. The glycine methyl ester side chain of compounds 14 and 15 was detected on the 1 H NMR spectra. Its proton signals were observed around 6.2 ppm (NH), 4.0 ppm (NCH 2 ) and 3.8 ppm (CH 3 ). Compounds 16 and 17, with an alanine methyl ester side chain, had δ signals around 6.2 ppm (NH), 4.6 ppm (NCH) and 3.8 ppm (CH 3 ).
Deprotection of the carboxyl group of the amino acid chain was performed in compounds 6, 7, 14 and 16, by alkaline hydrolysis (Scheme 3). This reaction also caused deprotection of the other carboxyl group on compounds 14 and 16. Compounds 18-21 were obtained in yields ranging from 94% to 98%. Deprotection of the carboxyl group of the amino acid chains was confirmed by the absence of δ signals around 3.7-3.8 ppm, on the 1 H NMR spectra of compounds 18-21. The loss the other methyl group of compounds 14 and 16 was detected by the absence of the δ signal around 3.6 ppm, on the 1 H NMR spectra of compounds 20 and 21. Compounds 8-17 were synthesized as depicted in Scheme 2. Compound 1 was oxidized using the Jones reagent [41] to afford compound 8, which was reacted with m-CPBA to give the derivative 9. The lactone ring of 9 was cleaved by treatment with p-TSA in methanol and dichloromethane [42] to provide compound 10. The derivative 11 and the three pairs of compounds synthesized in the following steps were prepared to explore the influence of the keto group in position C-11 on the antiproliferative activity. The removal of the keto group was performed by a Clemmensen reduction [45] with zinc dust and concentrated HCl in dioxane at room temperature to afford 11 (75%). The reduction was confirmed on the 13 C NMR spectrum, by the absence of the δ signal around 200 ppm, which corresponds to the carbonyl group in ring C. Acyl fluorides 12 and 13 were obtained from the reaction of compounds 10 and 11 with Deoxo-Fluor, in yields of 75% and 61%, respectively. The synthesis of acyl fluorides was detected on the 13 C NMR spectra. The carbon C30 appeared as a doublet with a δ signal around 166 ppm and a coupling constant of 375 Hz, in both compounds 12 and 13. These derivatives were reacted either with glycine methyl ester hydrochloride or with Lalanine methyl ester hydrochloride to afford compounds 14-17, in yields ranging from 43% to 83%. The glycine methyl ester side chain of compounds 14 and 15 was detected on the 1 H NMR spectra. Its proton signals were observed around 6.2 ppm (NH), 4.0 ppm (NCH2) and 3.8 ppm (CH3). Compounds 16 and 17, with an alanine methyl ester side chain, had δ signals around 6.2 ppm (NH), 4.6 ppm (NCH) and 3.8 ppm (CH3).
Deprotection of the carboxyl group of the amino acid chain was performed in compounds 6, 7, 14 and 16, by alkaline hydrolysis (Scheme 3). This reaction also caused deprotection of the other carboxyl group on compounds 14 and 16. Compounds 18-21 were obtained in yields ranging from 94% to 98%. Deprotection of the carboxyl group of the amino acid chains was confirmed by the absence of δ signals around 3.7-3.8 ppm, on the 1 H NMR spectra of compounds 18-21. The loss the other methyl group of compounds 14 and 16 was detected by the absence of the δ signal around 3.6 ppm, on the 1 H NMR spectra of compounds 20 and 21.
Compounds 1, 5-7 and 10-21 were screened for their antiproliferative activity on A549 (lung adenocarcinoma) and HT-29 (colon adenocarcinoma) cell lines (Table 1). Compounds 2-4, 8 and 9 were not evaluated because they have already been tested with no improvements in potency and/or selectivity [37,39,46,47]. Intermediates 5 and 10, afforded by the cleavage of the ring A, were more potent than the parental compound GA 1. Removal of the keto group from ring C, that provided compound 11, resulted in an increment of cytotoxicity. Acyl fluorides 12 and 13 were less potent compared to their substrates. Analysis of the IC50 values of the derivatives 6, 7 and 14-17 showed that the conjugation of an amino acid methyl ester provided more potent compounds. Deprotection of the carboxyl groups resulted in a loss of cytotoxicity. Compounds 18-21, afforded by the alkaline hydrolysis, were further tested in Jurkat (acute T-cell leukemia) and MOLT-4 (acute lymphoblastic leukemia) cell lines ( Table 2). The results of these assays confirmed that the deprotection provided less active compounds. Derivatives 6, 7 and 14-17 and the parental compound GA 1 were also screened for their antiproliferative activity in seven additional human cancer cell lines: Jurkat, MOLT-4, MIAPaca 2 (pancreas adenocarcinoma), MCF7 (breast adenocarcinoma), HeLa (cervix adenocarcinoma), A375 (melanoma) and HepG2 (hepatocellular carcinoma). Comparing IC50 values of compounds 6 and 7 with those obtained for compounds 14 and 16, no significant differences were found regarding the position in which the amino acid methyl ester was introduced. Derivatives 15 and 17 were respectively more potent than compounds 14 and 16 in all tested cell lines, which confirmed that the removal of the keto group from ring C enhanced the cytotoxicity. Compounds 7, 16 and 17, which have an alanine methyl ester chain, were more active than compounds 6, 14 and 15, with a glycine methyl ester chain, respectively. These results suggest that the type of amino acid moiety introduced influences the antiproliferative activity. Within the newly synthesized derivatives, compound 17, with a reduced ring C and with an alanine methyl ester chain, was the most potent derivative. This compound was 5 to 17-fold more active than GA 1, depending on the cancer cell line.
Compounds 1, 5-7 and 10-21 were screened for their antiproliferative activity on A549 (lung adenocarcinoma) and HT-29 (colon adenocarcinoma) cell lines (Table 1). Compounds 2-4, 8 and 9 were not evaluated because they have already been tested with no improvements in potency and/or selectivity [37,39,46,47]. Intermediates 5 and 10, afforded by the cleavage of the ring A, were more potent than the parental compound GA 1. Removal of the keto group from ring C, that provided compound 11, resulted in an increment of cytotoxicity. Acyl fluorides 12 and 13 were less potent compared to their substrates. Analysis of the IC 50 values of the derivatives 6, 7 and 14-17 showed that the conjugation of an amino acid methyl ester provided more potent compounds. Deprotection of the carboxyl groups resulted in a loss of cytotoxicity. Compounds 18-21, afforded by the alkaline hydrolysis, were further tested in Jurkat (acute T-cell leukemia) and MOLT-4 (acute lymphoblastic leukemia) cell lines ( Table 2). The results of these assays confirmed that the deprotection provided less active compounds. Derivatives 6, 7 and 14-17 and the parental compound GA 1 were also screened for their antiproliferative activity in seven additional human cancer cell lines: Jurkat, MOLT-4, MIAPaca 2 (pancreas adenocarcinoma), MCF7 (breast adenocarcinoma), HeLa (cervix adenocarcinoma), A375 (melanoma) and HepG2 (hepatocellular carcinoma). Comparing IC 50 values of compounds 6 and 7 with those obtained for compounds 14 and 16, no significant differences were found regarding the position in which the amino acid methyl ester was introduced. Derivatives 15 and 17 were respectively more potent than compounds 14 and 16 in all tested cell lines, which confirmed that the removal of the keto group from ring C enhanced the cytotoxicity. Compounds 7, 16 and 17, which have an alanine methyl ester chain, were more active than compounds 6, 14 and 15, with a glycine methyl ester chain, respectively. These results suggest that the type of amino acid moiety introduced influences the antiproliferative activity. Within the newly synthesized derivatives, compound 17, with a reduced ring C and with an alanine methyl ester chain, was the most potent derivative. This compound was 5 to 17-fold more active than GA 1, depending on the cancer cell line.  The selectivity towards cancer cells was studied for GA 1 and compound 17 by incubating them with a human nontumoral cell line (BJ) ( Table 2). GA 1 and compound 17 showed IC 50 values that were 1.6 and more than 16.4 times lower on Jurkat cells than on the nontumoral BJ cells, respectively. Therefore, the novel derivative 17 was up to 10 times more selective towards malignant cells than its parental compound 1. This compound also showed a significant improvement in selectivity compared to the chemotherapy agent cisplatin. Considering also the Jurkat cell line, cisplatin presented an IC 50 value that was 5.3 times lower than on BJ cells (Table 2); therefore, compound 17 was up to 3 times more selective than cisplatin towards Jurkat cells.

Analysis of Cell Cycle Distribution and Apoptosis
The Jurkat cell line which was the most susceptible to these derivatives was selected to investigate the mechanism of action of compound 17. To evaluate the effects on the cell cycle distribution, Jurkat cells were treated with compound 17, at a concentration corresponding to its IC 50 value at 72 h of treatment, for 24, 48 and 72 h and then analyzed by flow cytometry. The calculation of the fraction of cells in G0/G1, S and G2/M phases was performed using the fraction of live cells. Treatment for 24 h induced significant increase in the population at S phase with respect to untreated cells ( Figure 1); after 48 h this effect has decreased and after 72 h it was no longer observed. DNA fragmentation was detected after 72 h of incubation based on the appearance of a sub-G0 peak. This sequence of events suggests that the cell cycle arrest at S phase may have led cells to undergo apoptosis. The Jurkat cell line which was the most susceptible to these derivatives was selected to investigate the mechanism of action of compound 17. To evaluate the effects on the cell cycle distribution, Jurkat cells were treated with compound 17, at a concentration corresponding to its IC50 value at 72 h of treatment, for 24, 48 and 72 h and then analyzed by flow cytometry. The calculation of the fraction of cells in G0/G1, S and G2/M phases was performed using the fraction of live cells. Treatment for 24 h induced significant increase in the population at S phase with respect to untreated cells ( Figure 1); after 48 h this effect has decreased and after 72 h it was no longer observed. DNA fragmentation was detected after 72 h of incubation based on the appearance of a sub-G0 peak. This sequence of events suggests that the cell cycle arrest at S phase may have led cells to undergo apoptosis. Apoptosis assays were then performed to better elucidate the mechanism of cell death involved in the cytotoxic effect of compound 17. The Annexin V-FITC/PI flow cytometric assay employs the property of fluorescein isothiocyanate (FITC) conjugated to Annexin V (Annexin V-FITC) to bind to phosphatidylserine (PS) and the property of propidium iodide (PI) to enter cells with damaged cell membranes and to bind to DNA. Early apoptosis is characterized by the loss of membrane asymmetry, with translocation of PS from the inner to the outer membrane, prior to the loss of membrane integrity. Therefore, this assay allows the discrimination of live cells (Annexin-V -/ PI -) from early apoptotic (Annexin-V + /PI -), late apoptotic (Annexin-V + /PI + ) or necrotic cells (Annexin-V + /PI + ). The experiments were conducted on Jurkat cells treated with compound 17 at a concentration corresponding to its IC50 value at 72 h of treatment (6.1 µM) for 24 and 48 h, and at concentrations of 6.1 µM and 12.2 µM for 72 h. Exposure to this compound for 24 and 48 h did not change significantly the apoptotic (Figure 2A) and necrotic (data not shown) populations. Treatment for 72 h with compound 17 at concentrations of 6.1 µM and 12.2 µM increased the early apoptotic population by 19% and 30%, respectively. No significant changes were observed in the late apoptotic population. These results were in good agreement with those obtained in the cell cycle experiments.
The induction of apoptosis was further confirmed by the observation of its characteristic morphological changes. Hoechst 33342 staining showed volume reduction, chromatin condensation and apoptotic bodies in Jurkat cells treated in the same conditions for 72 h ( Figure 2B). In contrast, untreated cells presented a normal morphological profile. Taken together, the results described above suggest that compound 17 inhibits cell growth through cell cycle arrest at the S phase and induction of apoptosis. Its mechanism of action needs to be studied further. The induction of apoptosis was further confirmed by the observation of its characteristic morphological changes. Hoechst 33342 staining showed volume reduction, chromatin condensation and apoptotic bodies in Jurkat cells treated in the same conditions for 72 h ( Figure 2B). In contrast, untreated cells presented a normal morphological profile.
Taken together, the results described above suggest that compound 17 inhibits cell growth through cell cycle arrest at the S phase and induction of apoptosis. Its mechanism of action needs to be studied further.

Biology
All cell cultures were performed at 37 • C in an atmosphere of 5% CO 2 .

Antiproliferative Activity Assays
The antiproliferative activity of the synthesized compounds on A549, HT-29, MIA Paca 2, MCF7, HeLa, A375, HepG2 and BJ adherent cells was determined by the MTT assay. Exponentially growing cells were plated in 96-well plates at a density of 1-8 × 10 3 cells/ well. After 24 h, cells were attached to the plate, and the growth medium was replaced with fresh medium containing either the compounds dissolved in DMSO at different concentrations or only DMSO, in triplicate, and the cells were continued to culture for 72 h. After incubation with the compounds, the medium was removed and 100 µL of MTT solution (0.5 mg/mL) were added to each well and the plates were incubated for 1 h. MTT was removed and 100 µL of DMSO was added to dissolve the formazan crystals. The absorbance was immediately read at 550 nm on an ELISA read plater (Tecan Sunrise MR20-301, TECAN, Austria). For Jurkat and MOLT-4 non-adherent cells, the antiproliferative activity was determined by XTT assay. These cell lines were plated with 5.5 × 10 3 and 1 × 10 4 cells/well, respectively, in 96-well plates in 100 µL medium. The seeding was executed simultaneously with the addition of the different concentrations of compounds or vehicle, in triplicate, and cells were allowed to incubate for 72 h. After that incubation period, 100 µL of the XTT labelling mixture were added to each well and the plates were incubated again for 4 h. Then, the absorbance was read at 450 nm on the ELISA plate reader.
Concentrations that inhibit cell proliferation by 50% (IC 50 ) represent an average of a minimum of three independent experiments and were expressed as means ± standard deviation (SD).

Cell Cycle Analysis
Cell cycle was assessed by flow cytometry using a fluorescence-activated cell sorter (FACS). Jurkat cells were plated in six-well plates at a density of 1.6 × 10 5 cells/well, simultaneously with the addition of compound 17, at a concentration corresponding to its IC 50 value at 72 h of treatment, or with only the vehicle, in a total volume of 2 mL of medium. The cells were allowed to incubate for 24, 48 and 72 h. After incubation, cells were collected and centrifuged. The supernatant was removed and the pellet was resuspended in 1 mL of TBS containing 1 mg/mL PI, 10 mg/mL RNase free of DNase and 0.1% Igepal CA-630, for 1 h, at 4 • C. FACS analysis was performed at 488 nm in an Epics XL flow cytometer (Coulter Corporation, Hialeah, FL, USA). Data were collected and analyzed using the Multicycle software (Phoenix Flow Systems, San Diego, CA, USA). Three independent experiments were performed, with two replicates per experiment.

Annexin V-FITC/PI Flow Cytometry Assay
Apoptosis was assessed by flow cytometry using a FACS. Jurkat cells were plated in six-well plates at a density of 1.6 × 10 5 cells/well, simultaneously with the addition of compound 17, at specified concentrations, or with only the vehicle, in a total volume of 2 mL of medium. The cells were allowed to incubate for 24, 48 and 72 h. After incubation, cells were collected and centrifuged. The supernatant was removed and the pellet was resuspended in 95 µL of binding buffer (10 mM HEPES/NaOH, pH 7.4, 140 mM NaCl, 2.5 mM CaCl 2 ). Annexin V-FITC conjugate (3 µL) was added and cells were incubated for 30 min, at room temperature, in darkness. After incubation, 0.8 mL of binding buffer were added. Just before the FACS analysis, cells were stained with 20 µL of 1 mg/mL PI solution. Three independent experiments were performed, with two replicates per experiment.

Hoechst 33342 Staining
The morphological changes were observed by fluorescence microscopy using Hoechst staining. Jurkat cells were plated in six-well plates at a density of 1.6 × 10 5 cells/well, simultaneously with the addition of compound 17, at specified concentrations, or with only the vehicle, in a total volume of 2 mL of medium. The cells were incubated for 72 h. After incubation, cells were collected by centrifugation, washed twice with PBS and stained with 500 µL of Hoechst 33342 solution (2 µg/ml in PBS), for 15 min, at room temperature, in darkness. Finally, cells were washed and resuspended in 10 µL PBS. The samples were mounted on a slide and observed with a fluorescence microscope (DMRB, Leica Microssystems, Wetzlar, Germany) with a 4 ,6-diamidine-2 -phenylindole dihydrochloride (DAPI) filter. Three independent experiments were conducted.

Conclusions
In summary, we synthesized a series of new GA derivatives via the opening of its ring A along with the coupling of an amino acid. Antiproliferative activity assays in a panel of nine human cancer cell lines showed that the most potent compound 17 was 5 to 17-fold more active than GA 1. The study of selectivity revealed that this new derivative was up to 10 times more selective towards malignant cells than its parental compound. Preliminary mechanism investigation indicated that compound 17 may act through arresting cell cycle progression at the S phase and inducing apoptosis. The enhanced potency and the high selectivity of this new GA derivative warrant further biological evaluation. Funding: Jorge A. R. Salvador thanks PT2020 (Programa Operacional do Centro 2020), project nº 3269, drugs2CAD, and the financial support by FEDER (European Regional Development Fund) through the COMPETE 2020 Programme (Operational Programme for Competitiveness and Internationalisation). Jorge A. R. Salvador also wishes to thank Universidade de Coimbra for financial support. Daniela P. S. Alho thanks FEDER (Programa Operacional Factores de Competitividade-COMPETE 2020) and Fundação para a Ciência e Tecnologia (FCT) through Projecto Estratégico: UID/NEU/04539/2013 and the financial support for the PhD grant SFRH/BD/66020/2009. Marta Cascante and Silvia Marin thank MINECO-European Commission FEDER-Una manera de hacer Europa (SAF2017-89673-R) and Agència de Gestió d'Ajuts Universitaris i de Recerca (AGAUR)-Generalitat de Catalunya (2017SGR1033). Marta Cascante also acknowledges the support received through the prize "ICREA Academia" for excellence in research, funded by ICREA Foundation-Generalitat de Catalunya. The authors acknowledge UC-NMR facility, which is supported by FEDER and FCT funds through the grants REEQ/481/QUI/2006, RECI/QEQ-FI/0168/2012, and CENTRO-07-CT62-FEDER-002012, and Rede Nacional de Ressonância Magnética Nuclear (RNRMN), for NMR data.