Glochodpurnoid B from Glochidion puberum Induces Endoplasmic Reticulum Stress-Mediated Apoptosis in Colorectal Cancer Cells

Glochidpurnoids A and B (1 and 2), two new coumaroyl or feruloyl oleananes, along with 17 known triterpenoids (3–19) were obtained from the stems and twigs of Glochidion puberum. Their structures were elucidated by extensive spectroscopic data analyses, chemical methods, and single crystal X-ray diffraction. All compounds were screened for cytotoxicity against the colorectal cancer cell line HCT-116, and 2, 3, 5, 6, 11, and 17 showed remarkable inhibitory activities (IC50: 0.80–2.99 μM), being more active than the positive control 5-fluorouracil (5-FU). The mechanistic study of 2, the most potent compound, showed that it could induce endoplasmic reticulum (ER) stress-mediated apoptosis and improve the sensitivity of HCT-116 cells to 5-FU.


Introduction
Colorectal cancer (CRC) is the third most commonly diagnosed and the second deadliest cancer worldwide [1]. According to global cancer statistics, approximately 1.9 million new CRC cases and 935,000 CRC deaths occurred in 2020, and this number is predicted to be increasing annually [2]. In recent years, although some novel therapies, such as targeted therapy, immunotherapy, and gene therapy, have shown certain benefits in CRC treatment, fluoropyrimidine (5-FU)-based chemotherapy still remains the first-line treatment for CRC [3]. Patients with poor performance or low risk of deterioration are recommended 5-FU single therapy. As for 5-FU-resistance patients, multiple-agent regimens, such as 5-FU + oxaliplatin, 5-FU + irinotecan (IRI), are often adopted. Choosing additive agents appears to be similar in efficacy, with only side effects varying [4]. However, inevitable deficiencies, such as systemic toxicity, low response rate, as well as acquired resistance, limit the clinical application of 5-FU [5]. Thus, developing more effective therapeutic drugs with less side effects for CRC is still urgent.
Natural products have been evolutionarily selected to bind to biological macromolecules and, thus, represent a valuable source of "privileged structures" in drug discovery [6]. Such role is especially prominent in anticancer research, with ca. 70% of FDA-approved anticancer drugs being either natural products or their derivatives [7]. In the treatment of CRC, many natural products are currently undergoing different phases of clinical trials, such as andrographolide (phase II), berberine (phase II/III), curcumin (phase I), epigallocatechin gallate (phase II), resveratrol (phase I), and silymarin (phase IV) [8]. [6]. Such role is especially prominent in anticancer research, with ca. 70% of FDA-approved anticancer drugs being either natural products or their derivatives [7]. In the treatment of CRC, many natural products are currently undergoing different phases of clinical trials, such as andrographolide (phase II), berberine (phase II/III), curcumin (phase I), epigallocatechin gallate (phase II), resveratrol (phase I), and silymarin (phase IV) [8]. These compounds have the potential to exert anti-CRC effects by interfering with the pathways of metastasis, invasion, apoptosis, and angiogenesis. In the past few years, our group has endeavored to discover anticancer agents from Euphorbiaceae plants, and has proven that terpenoids from Euphorbiaceae are a rich source of privileged structures in anticancer drug development [9].
Glochidion puberum (L.) Hutch (Euphorbiaceae) is a shrub widely distributed in southwest China [10]. The leaves, stems, roots, and fruits of G. puberum are used in traditional Chinese herbal drugs to treat dysentery, diarrhea, influenza, fever, cough, and impaludism [11]. According to literature researches, the chemical components isolated from this genus include triterpenoids, alkaloids, flavonoids, steroids, ligans, phenolic and so on [12]. Especially, some triterpenoids from Glochidion possessed antitumor-promoting and cytotoxic activities [13,14]. In our continued efforts toward discovering structurally diverse and bioactive triterpenoids from Euphorbiaceae plants [15][16][17], two new oleanane derivatives, glochidpurnoids A (1) and B (2), along with 17 known analogues were obtained from the stems and twigs of G. puberum. All their cytotoxicity was measured using MTT assay, and six compounds showed pronounced anti-CRC activities against HCT-116 cells. Derivative 2, with the best activity, was selected for in-depth mechanism exploration. Herein, details of the isolation, structural identification, along with anti-CRC potential of these compounds were reported.

Isolation and Structure Elucidation
The air-dried powder of the stems and twigs of G. puberum was extracted with 95 EtOH to give a crude extract, which was partitioned between water and EtOAc. The EtOAc extract was separated using multiple column chromatography to obtain triterpenoids 1−19 ( Figure 1).     3 and 132.5), ten sp 3 methylenes, four sp 3 methines (one oxygenated), seven methyls, and six sp 3 quaternary carbons. The abovementioned information revealed an oleanane core, morolic acid [19], with a transcoumaroyl substituent, which was further validated by elucidation of 2D NMR analyses of 1 ( Figure 2). In particular, the HMBC cross-peak from H-3 [δ H 4.90 (dd, J = 11.6, 4.7 Hz)] of the oleanane core to the ester carbonyl (C-1 , δ C 167.7) of the coumaroyl moiety located this group at OH-3. The relative configuration of 1, regarding the oleanane part, was established via interpretation of NOESY data ( Figure 3) and by comparing its NMR date with those of morolic acid. As shown in Figure 3, the strong NOE interactions of H-3/H-5 and Me-24/Me-25 disclosed that these methyls or protons were axially oriented on the chair conformational A-ring. Me-24 and Me-25 were arbitrarily assigned β-orientations, and thus H-3 and H-5 were α-oriented. The NOE interactions of H-9/H-5 and Me-27 implied the α-orientations of these protons and methyl group. Correspondingly, the NOE interactions of Me-26/Me-25 and H-13 suggested that H-13 and Me-26 were β-orientation. Finally, the absolute configuration of 1 was determined by comparing the alkaline hydrolysis products with those of 3, an acetylated derivative of morolic acid also isolated in the current research. Both compounds generated the same morolic acid product. As the absolute configuration of 3 was assigned as 3S,5R,8R,9R,10R,13S,14R,17S by using single crystal X-ray diffraction analysis [Flack parameter = −0.01 (6)] ( Figure 4 in the current study), the structure of 1 with the same absolute configuration was established as depicted and named glochidpurnoid A.       Compound 2 exhibited a molecular formula of C40H56O6 based on HRESIMS ion peak at m/z 655.3972 [M + Na] + (calcd. For C40H56O6Na + , 655.3969). The 1D NMR data of 2 showed high similarity to those of 1, with the only difference being the presence of an additional aromatic methoxyl signal (δH 3.93, δC 55.9) in 2, suggesting that 2 was a methoxylated derivative of 1. The location of methoxyl was authenticated by the HMBC correlation from 6-OCH3 (δH 3.93) to the C-6 (δC 146.7). Derivative 2 had the same relative configuration as that of 1 based on their similar NOE correlations. Similarly, the alkaline hydrolysis of 2 produced the same morolic acid product as those of 1 and 3, suggesting that they possessed the same absolute configuration. Hence, the structure of 2 was determined as shown and named glochidpurnoid B.

2 Inhibited Cell Proliferation in Colorectal Cancer (CRC) Cell Line HCT-116
1-19 were tested for their cytotoxicity against HCT-116 cells (5-FU as the positive drug). To our surprise, compounds 2, 3, 5, 6, 11, and 17 exhibited remarkable inhibitory activities (IC50: 0.80-2.99 μM), being more potent than that of the 5-FU (Table 2). Among them, 2 showed the most potent activity with IC50 = 0.80  0.05 μM, and its activity was further confirmed by colony formation assay and timedose-dependent assays. As illustrated in Figure 5A,B, the cell viability and morphology of HCT-116 cells were dose-dependently and time-dependently deteriorated with the treatment of 2 at 48 or 96 h. Meanwhile, the colony numbers of HCT-116 cells with the administration of 2 in 12-day were dose-dependently decreased, suggesting that 2 could repress the proliferation of HCT-116 cells ( Figure 5C).  Compound 2 exhibited a molecular formula of C 40 H 56 O 6 based on HRESIMS ion peak at m/z 655.3972 [M + Na] + (calcd. For C 40 H 56 O 6 Na + , 655.3969). The 1D NMR data of 2 showed high similarity to those of 1, with the only difference being the presence of an additional aromatic methoxyl signal (δ H 3.93, δ C 55.9) in 2, suggesting that 2 was a methoxylated derivative of 1. The location of methoxyl was authenticated by the HMBC correlation from 6 -OCH 3 (δ H 3.93) to the C-6 (δ C 146.7). Derivative 2 had the same relative configuration as that of 1 based on their similar NOE correlations. Similarly, the alkaline hydrolysis of 2 produced the same morolic acid product as those of 1 and 3, suggesting that they possessed the same absolute configuration. Hence, the structure of 2 was determined as shown and named glochidpurnoid B.

2 Inhibited Cell Proliferation in Colorectal Cancer (CRC) Cell Line HCT-116
1-19 were tested for their cytotoxicity against HCT-116 cells (5-FU as the positive drug). To our surprise, compounds 2, 3, 5, 6, 11, and 17 exhibited remarkable inhibitory activities (IC 50 : 0.80-2.99 µM), being more potent than that of the 5-FU (Table 2). Among them, 2 showed the most potent activity with IC 50 = 0.80 ± 0.05 µM, and its activity was further confirmed by colony formation assay and time/dose-dependent assays. As illustrated in Figure 5A,B, the cell viability and morphology of HCT-116 cells were dose-dependently and time-dependently deteriorated with the treatment of 2 at 48 or 96 h. Meanwhile, the colony numbers of HCT-116 cells with the administration of 2 in 12-day were dose-dependently decreased, suggesting that 2 could repress the proliferation of HCT-116 cells ( Figure 5C).

2 Induced Apoptosis in HCT-116 Cells
Apoptosis is the cell's natural mechanism for programed cell death. While the apoptotic pathway in cancer cells is typically inhibited, allowing cancer cells to survive longer for the accumulation of mutations which can increase invasiveness during tumor progression, stimulate angiogenesis, deregulate cell proliferation and interfere with differentiation [35]. Thus, apoptosis is considered as a promising target for anticancer therapy. Additionally, previous studies reported that lots of plant-derived compounds exhibit anticancer activity through activating the apoptotic pathway. To explore whether 2 could induce apoptosis, we adapted the western blotting to measure the expressions of apoptosis-related proteins cleaved PARP (c-PARP) and cleaved caspase-3 (c-cas-3), pro-apoptotic protein Bax, and anti-apoptotic protein Bcl-2. As displayed in Figure 5D, the expressions of c-PARP, c-cas-3, and Bax were dose-dependently increased, while Bcl-2 expression was downregulated with the treatment of 2 in HCT-116 cells, indicating that 2 could trigger apoptosis. To verify this, we further conducted apoptosis-related characteristic experiments. As shown in Figure 5E, Z-VAD-FMK, an apoptosis inhibitor, could dose-dependently attenuate 2-induced cell death and downregulate the expressions of apoptosis-related proteins (c-PARP and c-cas-3). These results suggested that 2 caused HCT-116 cells death via apoptosis.

2 Stimulated Endoplasmic Reticulum (ER) Stress-Mediated Apoptosis in HCT-116 Cells
A previous study found that pentacyclic triterpenoids could induce endoplasmic reticulum stress, which participates in cell apoptosis in most tumors [36]. We then investigated whether 2 induced apoptosis via ER stress signaling. As shown in Figure 5F, the expressions of ATF4 and CHOP, two marker genes of ER stress, were improved with the addition of 2, suggesting that 2 could induce ER stress in HCT-116 cells. Furthermore, knockdown of CHOP by siRNAs downregulated the expressions of c-PARP and c-cas-3 ( Figure 5G). Meanwhile, 2-induced increased expressions of CHOP and apoptosis-related proteins in HCT-116 cells could be decreased with the treatment of the ER stress inhibitor 4-PBA (2 mM). ( Figure 5H). Taken together, 2-induced apoptosis in HCT-116 cells was related to ER stress.

2 Potentiated the Antitumor Activity of 5-FU in HCT-116 Cells
As 5-FU, the first-line drug for CRC treatment, performs its efficacy via inhibiting DNA biosynthesis [9], the mechanism of which is different from that of 2, we further explored whether the combination of 5-FU and 2 could improve the therapeutic activity. As shown in Figure 5I, 2 dose-dependently increased the cytotoxic activity of 5-FU in HCT-116 cells, with 7 times improvement under the administration of 4 µM of 2. The above result indicated that 2 might possess a synergistic effect with 5-FU in HCT-116 cells, making 2 a promising agent in future combination therap.

Plant Material
The stems and twigs of Glochidion puberum (L.) Hutch were collected in June 2020 from Hengyang City, Hunan Province, China, and measured by Prof. Guihua Tang, one of the authors (Tang, G.H.) in our current research. A voucher specimen (accession number: GB2020-1) has been deposited at the School of Pharmaceutical Sciences, Sun Yat-sen University, Guangzhou, China.

General Experimental Procedures
Details of chemical procedures and instruments were displayed in Supplementary Materials.

Alkaline Hydrolysis of 1-3
Potassium carbonate drying granules (5 mg) and a drop of water were added into a stirred solution of 3 (5 mg) in CH 3 OH (2 mL). After stirring at rt for 6 h, the organic phase was extracted and evaporated in vacuo to obtain the hydrolysis product 3a (2.0 mg). Correspondingly, sodium ethoxide drying granules (5 mg) and a drop of water were added into a solution of 2 (5 mg) in CH 3 OH (2 mL). The reaction was stirred at 120 • C for 2 h and then directly subjected to HPLC (CH 3 CN/H 2 O, 90%, 3 mL/min) to obtain hydrolysis product 2a (1.6 mg). Treating 1 in the same way as 2 afford the hydrolysis product 1a (1.3 mg). These hydrolysis products 1a-3a were finally identified as the same compound, morolic acid [19], by comparison of their 1 H NMR data.

Cell Culture
The HCT-116 (human colorectal carcinoma cell line) cells were purchased from the Laboratory Animal Service Center at Sun Yat-sen University (Guangzhou, China). Cells were maintained in 1640 medium containing 10% FBS (Gibco BRL, Waltham, MA, USA) and incubated at 37 • C in a cell incubator with 5% CO 2 .

Cell Viability Assay
MTT assay was applied to examine cytotoxicity of the candidates in HCT-116 cells. Cells seeded into 96-well plates at a density of 1 × 10 3 cells/well were incubated for 24 h, then cells were treated with or without indicated concentrations of the candidates for 48 or 96 h. After incubation, 10 µL per well of MTT solution (5 mg/mL) was supplied and incubated for 4 h. After discarding the suspension, each well was added into 100 µL of DMSO, and the absorbance was determined by a multifunction microplate reader (Biotek., USA) at 570 nm. The assays were carried out in triplicate.

Colony Formation Assay
HCT-116 cells were plated in 6-well plates at a density of 1 × 10 3 cells/well for 24 h and later added into various concentrations of 2. Cells incubated for 12 days were fixed with 4% paraformaldehyde (Biosharp, Hefei, China) for 30 min and stained with Ciemsa stain (Beyotime, Haimen, China).

Western Blot Analysis
The cells were seeded evenly at a density of 2 × 10 5 cells/well in 6-well plates and were allowed to attach for 24 h. After treated with 2 for 12 h, the cells were rinsed three times with ice-cold PBS and lysed in RIPA buffer (Beyotime, China) containing protease inhibitor cocktails (Roche Life Science, Branford, CT, USA). The lysates were centrifugated at 12,000× g for 30 min at 4 • C, and the total protein concentration was determined using a BCA protein assay kit (Beyotime, China). The cell lysates were mixed with sample dye (Cwbio, Beijing, China) and boiled at 95 • C for 10 min. The mixture was resolved by SDS-PAGE electrophoresis and transferred to PVDF membrane. The blots were blocked with TBST (Tris-Buffered Saline Tween-20) containing 5% skim milk for 2 h. The antibodies were diluted using a Primary Antibody Dilution Buffer (Beyotime, China) at a ratio of 1:1000, and the blots were incubated with specific antibodies at 4 • C for 12 h. Subsequently, the blots were detected using enhanced chemiluminescence detection kit (Thermo, Waltham, MA, USA). The density of the immunoblot was analyzed using ImageJ software (National Institute of Health, Bethesda, MD, USA).

Conclusions
Generally, diterpenoids are recognized as the characteristic metabolites of the Euphorbiaceae plants [37]. Interestingly, previous and current chemical investigation suggested that triterpenoids were the main metabolites of G. puberum. In the current study, two new oleanolanes, designated as glochidpurnoid A (1) and B (2), along with seventeen known lupane triterpenoids were isolated from the whole herbs of G. purberum, of which six compounds showed high cytotoxicity against HCT-116 cells. Among them, 2, a new oleanane possessing a feruloyl moiety at OH-3, showed the most pronounced activity with IC 50 = 0.80 ± 0.05 µM, being more potent than the positive control 5-FU. Compared with 1 and 3, it seems that the trans-feruloyl substituent group in 2 is a key pharmacophore for the activity. Mechanistic study implied that 2 could induce apoptosis in HCT-116 cells via ER stress signaling and improve the sensitivity of CRC to 5-FU. Notably, although limited amount of 2 was found from the nature, the easily obtained trans-ferulic acid and morolic acid may provide an opportunity for us to mass-produce 2 by semi-synthesis and verify its in vivo anti-CRC effect in the future. This research not only enlarges the structural diversity of triterpenoids from the Euphorbiaceae, but also provides a potent agent for further developing anti-CRC drugs.