Quassinoid analogs exert potent antitumor activity via reversible protein biosynthesis inhibition in human colorectal cancer

Abstract

Cellular protein synthesis is accelerated in human colorectal cancer (CRC), and high expression of protein synthesis regulators in CRC patients is associated with poor prognosis. Thus, inhibition of protein synthesis may be an effective therapeutic strategy for CRC. We previously demonstrated that the quassinoid bruceantinol (BOL) had antitumor activity against CRC. Herein, potent tumor growth suppression (>80%) and STAT3 inhibition was observed in two different mouse models following BOL administration. Loss of body and spleen weight was observed but was eliminated upon nanoparticle encapsulation while maintaining strong antitumor activity. STAT3 siRNA knockdown exhibited modest suppression of cell proliferation. Surprisingly, STAT3 inhibition using a PROTAC degrader (SD-36) had little effect on cancer cell proliferation suggesting the possibility of additional mechanism(s) of action for quassinoids. BOL-resistant (BR) cell lines, HCT116^BR and HCA7^BR, were equally sensitive to standard CRC therapeutic agents and known STAT3 inhibitors but resistant to homoharringtonine (HHT), a known protein synthesis inhibitor. The ability of quassinoids to inhibit protein synthesis was dependent on the structure of the C15 sidechain. Of note, BOL did not inhibit protein synthesis in normal human colon epithelial cells whereas HHT and napabucasin remained effective in these normal cells. Novel quassinoids were designed, synthesized, and evaluated in pre-clinical CRC models. Treatment with the most potent analog, 5c, resulted in significant inhibition of cell proliferation and protein synthesis at nanomolar concentrations. These quassinoid analogs may represent a novel class of protein synthesis inhibitors for the treatment of human CRC.

1. Introduction

Colorectal cancer (CRC) ranks as the second most common cause of cancer death in the U.S. [1]. It is estimated that in 2023 nearly 150,000 new cases will be diagnosed and > 52,000 will die from the disease in the U.S. Currently, the 5-year overall survival rate for metastatic CRC (mCRC) is less than 15% [2]. Although the FDA has approved 12 new drugs over the past 20 years for the treatment of mCRC, these drugs remain limited in their clinical efficacy [3]. While immune checkpoint inhibitors have clinical activity in mCRC, the positive response to immune checkpoint blockade therapy was only observed in DNA mismatch repair deficient/microsatellite instability-high (MSI-H) tumors, which are limited to approximately 5% of CRC [4]. Thus, there remains an urgent need to develop novel agents that can be used alone or in combination with current chemotherapeutic/targeted agents or immune checkpoint inhibitors to improve antitumor activity and overcome drug resistance in human CRC.

To date, over 40% of all anticancer agents are derived from natural products [5]. Natural compounds and their derivatives from plants/medicinal herbs have unique mechanisms of action in cancer cells and can serve as lead compounds for the development of novel therapeutic agents. Our lab recently identified the plant quassinoid bruceantinol (BOL) as a potent inhibitor of CRC tumor growth [6]. Further study demonstrated that BOL blocked STAT3 activation and suppressed expression of STAT3-mediated downstream targets, including MCL-1, PTTG1, survivin, and c-MYC. Constitutive activation of STAT3 impacts several of the key hallmarks of cancer, including cell growth, proliferation, survival, immune evasion, invasion/metastasis, and angiogenesis [7]. Significant resources have focused on the design and development of STAT3-specific inhibitors. Several inhibitors such as Stattic [8], BP1–102 [9], and S3I-201 [10], BOL [6], SD-36 [11], YHO-1701 [12], ODZ10117 [13], 6Br-6a [14] and LLL12 [15] have been evaluated in preclinical models. Additional STAT3 inhibitors have been evaluated in early-phase clinical trials, such as STAT3 decoy oligonucleotides [16], OPB-31121 [17], OPB-51602 [18], atovaquone [19], and OPB-111077 [20]. The most advanced candidate, napabucasin (NAPA), is an orally bioavailable cancer stemness inhibitor selected in a STAT3-mediated transcription high-throughput screen [21]. A Phase III randomized study showed that NAPA did not increase overall survival (OS) compared with placebo in the overall population of mCRC patients, but significantly improved OS among a subset of patients with pSTAT3-positive mCRC [22].

It has been reported that quassinoids may function through inhibition of ROS signaling pathways, Nrf2-mediated defense mechanisms, and cell signaling pathways such as AKT, MEK, and c-MYC [23–27]. Other studies have shown that quassinoids may inhibit the proliferation of cancer cells through inhibition of protein synthesis [28,29]. Protein synthesis, including ribosome biogenesis, is one of the most essential processes in living cells and strictly regulated [30]. Dysregulation of protein synthesis is a crucial driving force of tumorigenesis and tumor progression. mCRC accumulates multiple genetic changes leading to hyperactivation of oncogenic signaling pathways, including WNT, c-MYC, BRAF, and KRAS signaling pathways. These alterations enhance cell growth and proliferation requiring increased protein synthesis. It has been shown that the rate of protein synthesis in CRC is significantly greater than in normal colon tissue [31]. This is further supported by evidence demonstrating that the global protein synthesis landscape is strongly accelerated in CRC [32]. Moreover, altered expression of critical factors such as eIF-4E and EEF2K necessary for translational control has been shown to be associated with CRC tumor progression and poor prognosis, suggesting CRC is dependent on a high rate of protein synthesis and may, therefore, be susceptible to protein synthesis inhibitors [33,34]. While initially considered to be non-specific as potential therapeutic agents, inhibitors of protein synthesis and the translational machinery have demonstrated significant efficacy against a range of different human cancers [35]. Additional inhibitors of protein synthesis are currently being evaluated in clinical trials against solid cancers (NCT04092673; NCT03522649).

In this study, we confirmed that BOL exerted potent antitumor activity in in vivo CRC tumor models. Quassinoids such as BOL block protein synthesis in CRC, but not in normal colon epithelial cells. New C15-modified quassinoid analogs were designed, synthesized, and evaluated in our preclinical CRC models. Our results confirm that the C15 side chain of the quassinoid is a key determinant for its pharmacological activity. These findings provide strong evidence that targeting protein synthesis may represent a novel strategy for colorectal cancer chemotherapy.

2. Materials and methods

2.1. Chemicals and reagents

A Click-iT assay kit was purchased from ThermoFisher Scientific (Waltham, MA). CellTiter-Glo^® 3D Cell Viability Assay kit and NE-PER nuclear and cytoplasmic extraction kit were purchased from Promega (Madison, WI). e-Myco^™ VALiD² Mycoplasma PCR Kit was obtained from iNtRON Bio (Burlington, MA). The IRDye 700 STAT3 consensus oligonucleotide and EMSA kit were purchased from LI-COR (Lincoln, NE). STAT3 siRNAs (Cat#: D-00354–02 - target sequence: GGAGAAGCAUCGUGAGUGA; D-00354–03 – target sequence: CCACUUUGGUGUUUCAUAA; D-00354–04 - target sequence: UCAGGUUGCUGGUCAAAUU) were purchased from Dharmacon, Inc. Lipofectamine 2000 was purchased from ThermoFisher. STAT3 inhibitors (S3I-201, C-188–9 and napabucasin (NAPA)), homoharringtonine (HHT), cycloheximide (CHX), and torin1 were obtained from Selleckchem (Houston, TX). SD-36 was a gift from Dr. S. Wang (University of Michigan). Bruceantinol was purchased from LabNetwork (Portland, ME). Bruceine A was purchased from MedChemExpress (Monmouth Junction, NJ). All other chemicals were purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO).

2.2. Cell lines and cell culture

HCT116, RKO, SKBR3, H522, HepG2, and CCD-841 CoN cells were purchased from American Type Culture Collection (ATCC, Manassas, VA). HCT116^BR and HCA7^BR cells were generated by growing parental HCT116 and HCA7 cells in the presence of 20 nM BOL. Over a two-month period, the BOL concentration was increased to 200 nM. For all experiments, BOL-resistant cells were grown in BOL-free medium for one week. 201 T cells were provided by Dr. Jill Siegfried and were established from a primary non-small cell lung cancer tumor [36]. HCA-7 cells were obtained from Dr. Susan Kirkland (Imperial College, London). HCT116, RKO, SKBR3, H522, HEPG2, 201 T, and CCD-841 CoN cells were maintained in RPMI-1640 (Invitrogen; Carlsbad, CA) with 10% (v/v) fetal bovine serum (FBS) (Gemini Bio Products; Sacramento, CA) at 37 °C in a humidified incubator with 5% CO₂. Mouse colon 38 cancer cells (MC-38) were obtained from the Biological Testing Branch, DCTD (NCI, Bethesda, MD) and cultured in DMEM medium with 10% fetal bovine serum, 2 mM glutamine, 0.1 mM non-essential amino acids, 1 mM sodium pyruvate, and 5 mM HEPES. Cell lines were authenticated by human and mouse STR profiling performed by the University of Pittsburgh Cytogenetics Facility, LabCorp (Burlington, NC), and IDEXX BioResearch (Westbrook, ME), respectively. Cells were tested monthly for mycoplasma by RT-PCR.

2.3. Cell proliferation assay

Cell viability was determined by the WST-1 assay (Takara). HCT116 and HCA-7 cells were plated in 96-well plates. On the following day, cells were treated with various concentrations of quassinoid analogs or STAT3 inhibitors for 72 h. The IC₅₀ value of each compound was calculated by Graphpad Prism 7.0 software.

2.4. Immunoblot analysis

Immunoblot analysis was performed as previously described [6].The following antibodies were used: anti-pTyr705-STAT3 (#9145; Cell Signaling, Danvers MA), anti-STAT3 (#9139; Cell Signaling), anti-pTyr701-STAT1 (#9167; Cell Signaling), anti-STAT1 (#14994S; Cell Signaling), anti-STAT5 (#9363; Cell Signaling), anti-STAT6 (#5397; Cell Signaling), anti-JAK2 (#3230, Cell Signaling), anti-MCL-1 (#5453; Cell Signaling), anti-PTTG1 (#13445; Cell Signaling), anti-c-MYC (#5605; Cell Signaling), anti-TYRO-3 (#5585; Cell Signaling), anti-AXL (#8661; Cell Signaling), anti-cyclin D1 (#2978; Cell Signaling), anti-ABCB1 (#517310; Calbiochem, Burlington, MA), anti-ABCG2 (# 4477; Cell Signaling), and anti-α-tubulin (#2125; Cell Signaling). Signal intensities was quantified by using Image J software. Relative protein expression was normalized by tubulin expression.

2.5. Electrophoretic mobility shift assay

Nuclear protein extract was obtained from HCT116 cells using the NE-PER extraction kit. DNA binding reactions were performed using the LI-COR EMSA Buffer Kit. Nuclear extract (5 μg) was added to the binding reactions containing double-stranded STAT3 consensus IRDye700-labeled oligos and 1 mM MgCl₂. After incubation for 20 min at room temperature, reactions were resolved on 5% acrylamide TBE precast gels (Bio-Rad, Hercules, CA) and visualized using a LI-COR Odyssey near-infrared Imager and Odyssey Infrared Imaging System Software version 3.0.

2.6. Epoxy-activated Sepharose 6B-based pull-down assay

Quassinoids were conjugated to Sepharose beads as previously described [37]. Epoxy-activated Sepharose 6B beads were washed and swelled in deionized water overnight. On the following day, the beads were washed with the coupling buffer (0.1 M NaHCO₃, pH 11, containing 0.5 M NaCl). Quassinoid analogs (1 mg) were dissolved in 1 mL of coupling buffer and conjugated with beads and rotated at 4 °C overnight. After washing, unoccupied binding sites were blocked with 0.1 M Tris-HCl (pH 8) for 2 h at room temperature. Analog-conjugated beads were washed with three cycles of alternating pH wash buffers (buffer 1: 0.1 M acetate and 0.5 M NaCl, pH 4; buffer 2: 0.1 M Tris-HCl and 0.5 M NaCl, pH 8). The control unconjugated beads were prepared as described above in the absence of analogs. The HCT116 nuclear cell lysate (2 mg/mL) was mixed (ratio = 1:1) with analog-conjugated beads or beads alone at 4 °C overnight. The beads were washed with TBST 3 times. The bound proteins were eluted with 2 × SDS loading buffer and were analyzed by Western blot analysis.

2.7. Click-iT assay

Cells were seeded into 6-well plate at 4×10⁵/mL. After 24 h, the medium was replaced with methionine-free RPMI-1640 with 5% dialyzed FBS for 30 min. L-homopropargylglycine (HPG) was added to wells for 2 h in the absence or presence of inhibitors. Cells were harvested, fixed by formalin for 5 min, and rinsed twice in phosphate-buffered saline (PBS) with 0.1% dialyzed FBS. Cells were permeabilized in 1 mL PBS (0.1% saponin; 0.1% dialyzed FBS) for 30 min at room temperature. The permeabilized cells were rinsed and suspended in 200 μL of Click-iT reaction buffer. Alexa-488-conjugated HPG incorporation was determined by flow cytometry.

2.8. Human colorectal cancer organoid culture

Human colon cancer organoids (P#127) were kindly provided by Dr. Lin Zhang [38]. The complete culture medium for organoids contains advanced DMEM/F12 (12634–010, Invitrogen, Waltham, MA) supplemented with 1 UI penicillin/streptomycin (15140–122, Invitrogen), 10 mM HEPES (15630–106, Invitrogen), 2 mM GlutaMAX (Invitrogen), 1 UI B27 (17504–044, Invitrogen), 1 UI N2 Supplement (100X) (17502–048, Invitrogen), 1 mM N-Acetylcysteine (A0737, Sigma), 10 nM [leu-15]-Gastrin (G9145, Sigma), 10 mM nicotinamide (N0636, Sigma), 10 μM SB202190 (S7067, Sigma), 50 ng/ml recombinant murine EGF (315–09, Peprotech, Cranbury, NJ), 0.5 μM A83–01 (2939, Tocris Bioscience, Minneapolis, MN), 10 nM prostaglandin E2 (22–961–0, Tocris Bioscience), and 50% WRN-conditioned medium (vol/vol; obtained from WRN-treated HEK-293 cells). Freshly passed organoids are cultured in the same medium with the addition of 10 μM Y-27632 and 100 μg/mL Primocin to protect the stem cell population.

2.9. Cell Titer-Glo 3D cell viability assay

Organoids were split and replated on 100 μL Matrigel in 24-well plates. Once organoids reached ~ 100 μm, they were incubated with quassinoid analogs for 7 days. Every 2–3 days, the quassinoid-containing medium was replaced. Dimethyl sulfoxide (0.1%) was used as a control. CellTiter-Glo^® 3D Reagent (100 μL) was added to each well, shaken for 5 min, and incubated for 25 min at room temperature. Luminescence was detected by the Tecan Infinite M1000 Pro microplate reader. All results were obtained from at least three independent experiments with duplicate wells in each experiment.

2.10. HCT116 xenograft model

All in vivo animal experiments were approved by the Institutional Animal Care and Use Committee at the University of Pittsburgh. HCT116 cells (5 × 10⁶) cells in 0.1 mL of PBS were implanted subcutaneously on the back of athymic nude female mice. When the tumor volume reached approximately 100 mm³, mice were randomized to 2 groups (6 mice per group): (A) control (vehicle; 0.1% ethanol) and (B) BOL 8 mg/kg given i. p. thrice per week. Tumor volume was measured as V = 1/2ab², in which “a” and “b” represent length and width of tumor, respectively. Tumor volumes were monitored 3 times per week. One mouse was euthanized on day 15 due to body weight loss. Animals were sacrificed on day 22, two h after the last BOL treatment. Tumors were snap-frozen in liquid nitrogen and stored at −80 °C. The weight of spleen, liver and kidney were measured.

2.11. MC-38 mice model

MC-38 cells (2 × 10⁶ cells) were inoculated subcutaneously into the right flank of female C57BL/6 mice. Mice were randomized into 3 groups: (A) control (n = 9; vehicle; 0.1% ethanol); (B) BOL, 4 mg/kg (n = 7) and (C) BOL, 8 mg/kg body weight (n = 8), i.p. 3 times/week. Mice were treated for 17 days. One mouse was euthanized on day 10 due to body weight loss (4 mg/kg group). In the 8 mg/kg group, three mice were euthanized due to distress. On day 12, BOL dose was reduced to 6 mg/kg. On day 17, the remaining animals were euthanized and protein expression in tumor tissues were determined by immunoblot analysis.

2.12. Encapsulation of BOL

POEG-POM was synthesized as described previously [39]. A solution of BOL in dichloromethane (10 mg/mL) was mixed with POEG-POM polymer solution (50 mg/mL in dichloromethane) at a carrier/drug ratio of 10:1 (m/m). The solvent was evaporated by nitrogen flow, and the residual solvent was further removed by vacuum for 2 h. The thin film obtained was hydrated in PBS and gently vortexed to form BOL-loaded POEG-POM nanoparticles (NPs). BOL-free POEG-POM NPs were similarly prepared as described above. BOL NPs were administered i.v. to mice twice a week at a dose of 8 mg BOL/kg in 200 μL.

2.13. 4-Step semi-synthetic route of BOL analogs

The synthetic route to generate quassinoid derivatives with a variety of C15 substituents follows a modified protocol based on the work of Lee K-H, et al [40]. As shown in Fig. 1A, the highly reactive C3 hydroxyl moiety of 1 (i.e., commercially available bruceine A) is protected with a tert-butyldimethylsilyl (TBS) ether to provide intermediate 2. The ester located on C15 is subsequently saponified with potassium methoxide in anhydrous methanol to give alcohol 3. Various C15 side chains are coupled to this intermediate via a Steglich esterification (using 1-ethyl-3-(3-dimethylaminopropryl) carbodiimide (EDCI) and 4-dimethylaminopyridine (DMAP)), resulting in 4. This esterification method is significantly cleaner and more amenable to chemical diversification compared to the published protocol, which employs acid chlorides. Finally, TBS-deprotection provides the desired final product 5. All analogs were chromographically purified and determined by LCMS or NMR to be > 90% pure. It is notable that the C11 and C12 alcohols of the natural product core are unreactive and undergo neither TBS protection nor esterification during the synthesis. Based on the X-ray structure of BOL (CSD entry code KUTHIR), this is not surprising given that these two hydroxyl groups are: (1) sterically less accessible than the C3 and C15 hydroxyls, and (2) engage in intramolecular hydrogen bonds with the exo-ring ether oxygen of the chemotype core and the carbonyl oxygen of the C13 methyl ester.

Fig. 1.

Effect of BOL on the growth of human and mouse colorectal tumors. Mice bearing HCT116 xenografts were administered via the i.p. route vehicle (0.1% ethanol) or BOL (8 mg/kg) thrice per week. (A) Tumor volume and (B) body weight were recorded thrice per week (n = 6). (C-F) After 22 days, tissue weights were measured. Plasma ALT was determined by ELISA. Mice bearing MC38 tumors were i.p. administered vehicle (0.1% ethanol), BOL (4 and 8 mg/kg) thrice per week (n = 9, 7, 8). (G) Tumor volume and (H) body weight were recorded thrice per week. (I) After 17 days, spleen weight was measured. (J, K) Tumors were harvested, and protein expression was determined by immunoblot analysis. (L, M) Mice bearing MC38 tumors were administered via the i.v. route vehicle (PBS) and BOL NPs (8 mg/kg) twice a week (n = 6). Values represent the mean ± SD. *p < 0.05; **p < 0.01 vs. control.

Methyl (1R,2S,3S,3aS,3a ¹R,4R,6aR,7aR,11aS,11bR)-9-((tert-butyldimethylsilyl)oxy)-4-(((2E,4E)-hexa-2,4-dienoyl)oxy)-1,2-dihydroxy-8,11a-dimethyl-5,10-dioxo-1,4,5,6a,7,7a,10,11,11a,11b-decahydro-2H-3,3a¹-(epoxymethano)dibenzo[de,g]chromene-3(3aH)-carboxylate (4a) [41]. A solution of 3 [42] (10 mg, 0.018 mmol), sorbic acid (2.6 mg, 0.024 mmol), and DMAP (2.2 mg, 0.018 mmol) in CH₂Cl₂ (0.20 mL) was cooled to 0 °C for 10 min, treated with N-(3-dimethylaminopropyl)-N’-ethylcarbodiimide hydrochloride (EDCI, 6.8 mg, 0.036 mmol) and stirred at room temperature for 17 h. The reaction mixture was diluted with H₂O (10 mL) and extracted with CH₂Cl₂ (3 × 10 mL). The combined organic layers were washed with brine, dried (Na₂SO₄), and concentrated. The crude residue was purified by chromatography on SiO₂ (CHCl₃:EtOAc, 7:3) to give 4a as a white flaky powder.

Methyl (1R,2S,3S,3aS,3a¹R,4R,6aR,7aR,11aS,11bR)-4-(((2E,4E)-hexa-2,4-dienoyl)oxy)-1,2,9-trihydroxy-8,11a-dimethyl-5,10-dioxo-1,4,5,6a,7,7a,10,11,11a,11b-decahydro-2H-3,3a¹-(epoxymethano) dibenzo[de,g]chromene-3(3aH)-carboxylate (5a) [42]. A solution of ester 4a (9.0 mg, 0.014 mmol) in dry THF was treated with a 1.0 M solution of TBAF in dry THF (28 μL, 0.028 mmol) and stirred under nitrogen at 0 °C for 15 min. The reaction mixture was quenched with cold H₂O (0 °C, 5 mL) and extracted with CHCl₃ (3 × 2 mL). The combined organic layers were washed with 0.0 5 M HCl (5 mL), a 5% solution of NaHCO₃ (5 mL), H₂O, and brine, dried (Na₂SO₄) and concentrated under reduced pressure. A concentrated solution of the oily residue in CHCl₃ was purified by chromatography on SiO₂ (CHCl₃:EtOAc, 2:1) to give 5a (6.1 mg, 82%) as a colorless, amorphous powder: HRMS (ESI⁺) m/z for C₂₇H₃₃O₁₁ [M + H] Calcd 533.2017, Found 533.2015.

Methyl (1R,2S,3S,3aS,3a¹R,4R,6aR,7aR,11aS,11bR)-9-((tert-butyldimethylsilyl)oxy)-4-(((E)-3-cyclohexyl-2-methylacryloyl)oxy)-1,2-dihy-droxy-8,11a-dimethyl-5,10-dioxo-1,4,5,6a,7,7a,10,11,11a,11b-decahydro-2H-3,3a¹-(epoxymethano)dibenzo[de,g]chromene-3(3aH)-carboxylate (4b) [41]. A solution of 3 [42] (15 mg, 0.027 mmol), 3-cyclohexyl-2-methylacrylic acid (5.5 mg, 0.032 mmol), and DMAP (3.3 mg, 0.027 mmol) in CH₂Cl₂ (0.2 mL) was cooled to 0 °C for 10 min, treated with N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDCI, 10 mg, 0.054 mmol) and stirred at room temperature for 14.5 h. The mixture was then diluted with H₂O (10 mL) and extracted with CH₂Cl₂ (3 × 10 mL). The combined organic layers were washed with water, brine, dried (Na2SO4), and concentrated. The crude residue was then purified by chromatography on SiO₂ (CHCl₃:EtOAc (3:1)) to give 4b (14 mg, 70% yield) as a white flaky powder.

Methyl (1R,2S,3S,3aS,3a¹R,4R,6aR,7aR,11aS,11bR)-4-(((E)-3-cyclohexyl-2-methylacryloyl)oxy)-1,2,9-trihydroxy-8,11a-dimethyl-5,10-dioxo-1,4,5,6a,7,7a,10,11,11a,11b-decahydro-2H-3,3a¹-(epoxymethano) dibenzo[de,g]chromene-3(3aH)-carboxylate (5b) [42]. A solution of ester 4b (13 mg, 0.019 mmol) in dry THF (0.5 mL) was treated with a 1.0 M solution of TBAF in dry THF (37 μL, 0.037 mmol) and stirred under nitrogen at 0 °C for 15 min. The reaction mixture was quenched with cold H₂O (0 °C, 5 mL) and extracted with CHCl₃ (3 × 2 mL). The combined organic layers were combined, and washed with diluted HCl (0.05 M HCl, 5 mL), H₂O (5 mL)), a 5% solution of NaHCO₃ (5 mL), H₂O, and brine, dried (Na₂SO₄), and concentrated under reduced pressure. A concentrated solution of the oily residue in CHCl₃ was purified by chromatography on SiO₂ (CHCl₃:EtOAc, 2:1) to give 5b as a colorless, amorphous powder: HRMS (ESI⁻) m/z for C₃₁H₃₉O₁₁ [M−H] Calcd 587.2487, Found 587.2498.

Methyl (1R,2S,3S,3aS,3a¹R,4R,6aR,7aR,11aS,11bR)-9-((tert-butyldimethylsilyl)oxy)-4-(((E)-3-(furan-2-yl)acryloyl)oxy)-1,2-dihydroxy-8,11a-dimethyl-5,10-dioxo-1,4,5,6a,7,7a,10,11,11a,11b-decahydro-2H-3,3a¹-(epoxymethano)dibenzo[de,g]chromene-3(3aH)-carboxylate (4c) [41]. A solution of 3 [42] (10 mg, 0.018 mmol), sorbic acid (3.0 mg, 0.022 mmol) and DMAP (2.2 mg, 0.018 mmol) in DCM (0.2 mL) was cooled to 0 °C for 10 min, treated with N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDCI, 6.9 mg, 0.036 mmol), and stirred at room temperature for 15 h. The reaction mixture was diluted with H₂O (10 mL) and extracted with CH₂Cl₂ (3 × 10 mL). The combined organic layers were washed with brine, dried (Na₂SO₄), and concentrated. The crude residue was then purified by chromatography on SiO₂ (CH₂Cl₂:EtOAc (2.2:1)) to give 4c (63 mg, 52%) as a white, flaky powder.

Methyl (1R,2S,3S,3aS,3a1R,4R,6aR,7aR,11aS,11bR)-4-(((E)-3-(furan-2-yl)acryloyl)oxy)-1,2,9-trihydroxy-8,11a-dimethyl-5,10-dioxo-1,4,5,6a,7,7a,10,11,11a,11b-decahydro-2H-3,3a¹-(epoxymethano) dibenzo[de,g]chromene-3(3aH)-carboxylate (5c) [42]. A solution of ester 4c (14.0 mg, 0.021 mmol) in dry THF (0.2 mL) was treated with a 1.0 M solution of TBAF in dry THF (31 μL, 0.031 mmol) and stirred under nitrogen at 0 °C for 15 min. The reaction mixture was quenched with cold H₂O (0 °C, 5 mL) and extracted with CHCl₃ (3 × 2 mL). The combined organic layers were washed with 0.05 M HCl (5 mL), a 5% solution of NaHCO₃ (5 mL), H₂O, and brine, dried (Na₂SO₄) and concentrated under reduced pressure. A concentrated solution of the oily residue in CHCl₃ was purified by chromatography on SiO₂ (CHCl₃/EtOAc (2:1)) to give 5c (6.1 mg, 82% yield) as a colorless, amorphous powder: ¹H NMR (600 MHz, CDCl₃) δ 7.51 (s, 1H), 7.47 (d, J = 16.2 Hz, 1H), 6.67 (d, J = 3.0 Hz, 1H), 6.49 (s, 1H), 6.40 (bs, 1H), 6.25 (d, J = 15.6 Hz, 1H), 6.08 (s, 1H), 4.80 (s, 1H), 4.74 (d, J = 7.8 Hz, 1H), 4.28–4.26 (m, 1H), 4.21 (s, 1H), 3.82–3.81 (m, 1H), 3.79 (s, 3H), 3.28 (d, J = 2.4 Hz, 1H), 3.17 (broad d, 1H), 3.01–2.97 (m, 2H), 2.43–2.39 (m, 2H), 2.25 (d, J = 8.4 Hz, 1H), 2.14 (s, 1H), 1.85 (s, 1H), 1.80 (s, 3H), 1.41 (s, 3H); HRMS (ESI⁻) m/z for C₂₈H₂₉O₁₂ [M−H] Calcd 557.1654, Found 557.1656.

Methyl (1R,2S,3S,3aS,3a¹R,4R,6aR,7aR,11aS,11bR)-9-((tert-butyldimethylsilyl)oxy)-1,2-dihydroxy-8,11a-dimethyl-4-(((E)-3-(5-methylfuran-2-yl)acryloyl)oxy)-5,10-dioxo-1,4,5,6a,7,7a,10,11,11a,11b-decahydro-2H-3,3a¹-(epoxymethano)dibenzo[de,g]chromene-3(3aH)-carboxylate (4d) [41]. A solution of 3 [42] (40 mg, 0.072 mmol), sorbic acid (0.013 g, 0.087 mmol) and DMAP (9.0 mg, 0.072 mmol) in CH₂Cl₂ (0.75 mL) was cooled to 0 °C for 10 min, treated with N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDCI, 28 mg, 0.15 mmol) and stirred at room temperature for 15 h. The reaction mixture was diluted with H₂O (10 mL) and extracted with CH₂Cl₂ (3 × 20 mL). The combined organic layers were washed with brine, dried (Na₂SO₄), and concentrated. The crude residue was then purified by chromatography on SiO₂ (CHCl₃:EtOAc (2.2:1)) to give 4d as a white, flaky powder.

Methyl (1R,2S,3S,3aS,3a¹R,4R,6aR,7aR,11aS,11bR)-1,2,9-trihydroxy-8,11a-dimethyl-4-(((E)-3-(5-methylfuran-2-yl)acryloyl)oxy)-5,10-dioxo-1,4,5,6a,7,7a,10,11,11a,11b-decahydro-2H-3,3a¹-(epoxymethano) dibenzo[de,g]chromene-3(3aH)-carboxylate (5d) [42]. A solution of ester 4d (24 mg, 0.035 mmol) in dry THF (0.35 mL) was treated with a 1.0 M solution of TBAF in dry THF (52 μL, 0.052 mmol) and stirred under nitrogen at 0 °C for 15 min. The reaction mixture was quenched with cold H₂O (0 °C, 5 mL) and extracted with CHCl₃ (3 × 2 mL). The combined organic layers were washed with 0.05 M HCl (5 mL), a 5% solution of NaHCO₃ (5 mL), H₂O, and brine, dried (Na₂SO₄) and concentrated under reduced pressure. A concentrated solution of the oily residue in CHCl₃ was purified by chromatography on SiO₂ (CHCl₃:EtOAc, 2:1) to give 5d as a colorless, amorphous powder: ¹H NMR (400 MHz, CDCl₃) δ 7.38 (d, J = 15.6 Hz, 1H), 6.56 (d, J = 3.2 Hz, 1H), 6.15 (d, J = 15.6 Hz, 1H), 6.1 (m, 1H), 6.08 (s, 1H), 4.80 (bs, 1H), 4.73 (d, J = 8.0 Hz, 1H), 4.26 (bs, 1H) 4.21 (s, 1H), 3.84–3.80 (m, 1H), 3.78 (s, 1H,), 3.33 (s, 1H), 3.17 (bs, 1H), 3.00–2.95 (m, 2H), 2.43–2.31 (m, 6H), 2.13 (bs, 1H), 1.84 (d, 3H), 1.77 (td, J = 2.8, 13.2 Hz, 2H), 1.40 (s, 3H); HRMS (ESI⁺) m/z for C₂₉H₃₃O₁₂ [M + H] Calcd 573.1967, Found 573.1991.

Methyl (1R,2S,3S,3aS,3a¹R,4R,6aR,7aR,11aS,11bR)-1,2,4,9-tetrahydroxy-8,11a-dimethyl-5,10-dioxo-1,4,5,6a,7,7a,10,11,11a,11b-decahydro-2H-3,3a¹-(epoxymethano)dibenzo[de,g]chromene-3(3aH)-carboxylate (6) [42]. A solution of 1 (52 mg, 0.10 mmol) in CH₃OH (3.5 mL) was added dropwise a solution of cooled potassium methoxide (0.376 g, 5.36 mmol) in CH₃OH (4 mL) The mixture was then stirred at 0 °C, and the reaction was monitored hourly. After 3 h, the reaction mixture was neutralized by the addition of HCl ether solution (2 M), and then concentrated. To the resulting residue was added CH₂Cl₂ (5 mL) and the solution was filtered through a plug of Celite, concentrated, and purified on C18 RP-SiO₂ (0% to 100% MeCN in water) to give 6 (14 mg, 32%) as a white powder: ¹H NMR (400 MHz, CDCl₃) δ 6.09 (s, 1H), 5.28 (d, J = 12.4 Hz, 1H), 4.73 (m, 1H), 4.70 (s, 1H), 4.23 (m, 2H), 3.90 (s, 3H), 3.78 (dd, J = 1.2, 8.0 Hz, 1H), 2.99 (d, J = 16.0 Hz, 1H), 2.89 (dm, 1H), 2.79 (dd, J = 1.6, 12.4 Hz, 1H), 2.40–2.39 (m, 1H), 2.36–2.34 (m, 1H), 2.07 (d, J = 3.6 Hz 1H), 1.84 (d, J = 2.0 Hz, 3H), 1.79 (dd, J = 2.8, 13.2 Hz, 1H), 1.75 (dd, J = 2.8, 13.2 Hz, 1H), 1.39 (s, 3H); HRMS (ESI⁻) m/z for C₂₁H₂₅O₁₀ [M−H] Calcd 437.1442, Found 437.1448.

2.14. Statistical analysis

All experimental data are shown as mean ± SD. Statistical analysis of the data was performed using SPSS software. The Student’s t-test (two-tailed) was used for two group comparisons. One-way ANOVA was used to comparisons between groups of more than two unpaired values. Two-way ANOVA analysis was used to comparison of anti-tumor response in vivo. p < 0.05 was considered statistically significant.

3. Results

3.1. In vivo activity of BOL in human and mouse CRC models

We have previously shown that BOL at a dose of 4 mg/kg inhibited in vivo CRC tumor growth by 58% [6]. This dose did not significantly inhibit growth of STAT3^−/− CRC tumors suggesting STAT3 was a potential target of BOL. At that time, higher doses were not evaluated due to insufficient amount of BOL source material. Attempts to prepare semisynthetic BOL were unsuccessful due to difficulties with the synthesis of the C15 side chain of BOL. Subsequently, we identified a commercial source of BOL, and a higher BOL dose (8 mg/kg) was administered i.p. to nude mice bearing HCT116 xenografts. The higher dose considerably suppressed tumor growth > 90% (Fig. 1A; p < 0.001). At this higher dose, one of the treated mice lost > 20% body weight (Fig. 1B). Liver and kidney weight and plasma ALT levels were unaffected by BOL administration (Fig. 1C–E). However, spleen weight was reduced by 33% (Fig. 1F). We also evaluated the antitumor activity of BOL in an immunocompetent model. MC38 tumor-bearing C57BL6 mice were administered vehicle, 4 mg/kg, or 8 mg/kg of BOL i.p. 3 times per week. BOL exhibited potent in vivo antitumor activity in this model(Fig. 1G). After 17 days, tumor volume was decreased 59% and 77% in the 4 mg/kg and 8 mg/kg BOL-treated groups, respectively (p < 0.01; p < 0.01, versus control group respectively). The 8 mg/kg dose resulted in body weight loss in 3 mice. Both BOL doses resulted in significant spleen weight loss (Fig. 1I), similar to the athymic mouse model, suggesting hematologic toxicities. To demonstrate the in vivo mechanism of action of BOL, tumor protein expression was determined by immunoblot analysis. The proteins, MCL-1 and survivin, were markedly suppressed by BOL administration in a dose-dependent manner (Fig. 1J,K). As observed previously, expression of p-STAT3 protein was significantly reduced in BOL-treated groups.

Encapsulation of potent anticancer drugs has improved the efficacy of multiple agents such as doxorubicin, cisplatin, and paclitaxel by reducing intolerable side effects and toxicities [43–45]. In an attempt to reduce BOL host toxicities, we encapsulated BOL into poly(oligoethylene glycol methacrylate)–co-poly(oleyl methacrylate)(POEG-POM) nanoparticles (NPs) [39]. These particles were injected i.v. via tail vein twice a week at 8 mg/kg into MC38-bearing mice. Significant inhibition of tumor growth was observed (84%) (Fig. 1L; p < 0.001 versus control group). Most importantly, no body weight loss was observed with NP-encapsulated BOL administration (Fig. 1M). However, the group of MC38-bearing mice that received i.v. BOL without encapsulation were clearly distressed after a single injection and were euthanized (data not shown). This observation was unexpected given that i.p. administration at the same dose was tolerable. Thus, encapsulation significantly reduced host toxicities while maintaining potent antitumor activity. Previous studies have suggested that quassinoid inhibition of protein synthesis was irreversible [29]. As shown in Fig. 1L, cessation of BOL-NP treatment resulted in tumor growth. This delay of growth inhibition suggests that suppression of protein inhibition by BOL may be reversible. Additional in vitro studies demonstrated that protein synthesis resumed following BOL removal (data not shown).

3.2. Role of STAT3 in CRC cell proliferation

To further investigate the interaction of BOL and STAT3 protein, we conjugated BOL to epoxy-activated Sepharose 6B beads. Protein pull-down assays were performed with BOL-conjugated beads and HCT116 nuclear extracts. As shown in Fig. 2A, STAT3 protein bound specifically to BOL-conjugated beads in contrast to other STAT family members (STAT1, STAT5, STAT6) and JAK2 where binding was not observed. This finding supports our previous findings suggesting a possible mechanism of action of BOL is STAT3 inhibition. We next sought to confirm the role of STAT3 protein as a modulator of CRC cell growth and proliferation. Since small molecules, such as BOL, may influence cell growth through additional biological mechanisms, we suppressed STAT3 protein expression using STAT3-specific siRNAs and a PROTAC STAT3 inhibitor SD-36 [11]. Levels of STAT3 protein were significantly reduced by siRNAs and SD-36 while those of STAT1 were unchanged (Fig. 2B,E). Cellular proteins presumably dependent upon STAT3 expression (cyclin D1, MCL-1, survivin) were mostly unchanged following siRNA transfection. None of these downstream STAT3 signaling proteins were affect by SD36 treatment (Fig. 2E). While cell proliferation was modestly affected by siRNA treatment, SD-36 treatment had almost no effect on cell growth (Fig. 2D,G). Since detectable STAT3 protein expression remained following siRNA but not for SD36 treatment, the proliferation differences between SD-36 and STAT3 siRNAs might be related to siRNA off-target effects as siSTAT3–4 affected MCL-1 expression which may have contributed to growth suppression. We also investigated the effect of SD-36 in other cancer cell types. SD-36 effectively suppressed STAT3 protein expression in breast cancer SKBR3 cells, lung cancer H522 and 201 T cells, and liver cancer HepG2 cells (Fig. 2H). No effect was observed on STAT1 or on any of the STAT3 target proteins. Similarly, SD-36 had minimal effect on growth of these cells (Fig. 2J). Taken together, these findings suggest that, while quassinoids can directly bind STAT3 and inhibit its phosphorylation, inhibition of STAT3 directly had limited effect on cancer cell growth and proliferation.

Fig. 2.

Interaction of BOL and STAT3 protein. (A) BOL-Sepharose 6B beads were incubated with HCT116 nuclear lysate. Bound proteins were determined by immunoblot analysis. (B,C) HCT116 cells were transfected with STAT3 siRNAs (10 nM) for 72 h and processed for immunoblot analysis. (D) HCT116 cells were transfected with 100 nM siRNAs. After 72 h, cell proliferation was quantified using the WST-1 assay. (E,H) Cells were incubated with SD-36 (10 μM) for 24 h and processed for immunoblot analysis. (G,J) Cells were treated with SD-36 (30 μM) for 72 h, and cell proliferation was quantified using the WST-1 assay. Values in the graph represent the mean ± SD from three independent experiments. *p < 0.05; **p < 0.01 vs. control.

3.3. Resistance to quassinoids in CRC cells

Cancer drug resistance remains a major limitation to clinical efficacy of chemotherapy. Thus, we established BOL-resistant HCT116 and HCA7 CRC cell lines and used these resistant cell lines to screen anticancer agents to identify potential common pathways of resistance. Cells were continually exposed to increasing concentrations of BOL (20 nM to 200 nM) for two months. As shown in Table 1, HCT116^BR cells were 3-fold resistant to BOL while HCA7^BR cells were 11-fold resistant. Given our previous findings that BOL effectively inhibited STAT3 phosphorylation [6], we tested the sensitivity of the BOL-resistant cells to other STAT3 inhibitors (NAPA; C-188–9). HCT116^BR cells demonstrated a 4-fold resistance to NAPA, but HCA7^BR cells were equally sensitive to NAPA as the parental cells. No resistance was observed with C-188–9. Drug resistance was also not observed with standard CRC chemotherapeutic agents oxaliplatin, 5-FU, or SN-38 in either BOL-resistant cell line. Comparison of the level of protein expression between parental and resistant cells showed that ABCB1, the multi-drug resistance protein efflux pump, was increased in HCA7^BR cells but not in HCT116^BR cells (Fig. 3A,B), which appears to play a role in their resistance to doxorubicin, a classic ABCB1 substrate (Table 1). Previous studies have suggested that the mechanism of action of quassinoids may be mediated through inhibition of protein synthesis [29,46]. BOL-resistant CRC cells were not cross-resistant to cycloheximide (CHX; Table 1), a well-established protein synthesis inhibitor. The quassinoid bruceantin and the natural product homoharringtonine (HHT) have been reported to bind to the same location within the large ribosomal subunit of H. marismortui [47]. When HHT was tested in BOL-resistant cells, we observed a similar resistance to HHT, suggesting a possible common mechanism of resistance against BOL and HHT (Table 1). To further investigate the signaling pathways affected by HHT and BOL, we performed a Reverse Phase Protein Array (RPPA) analysis following BOL and HHT treatment in HCT116 cells. Both molecules altered expression of the same proteins (Fig. 3C), many known to be directly or indirectly involved in protein synthesis (i.e., p-mTOR, eEF2K, p-p70-S6K, p-S6). However, protein expression differences were also observed, suggesting that these molecules may inhibit the synthesis of different cellular proteins (Fig. 3D).

Table 1.

Drug IC₅₀ values in parental and BOL-resistant CRC cells.

Drugs	Cells
	HCT116	HCT116BR	FR*	HCA-7	HCA7BR	FR*
BOL (nM)	55 ± 3	142 ± 16	2.6	64 ± 6	702 ± 118	11.0
NAPA (μM)	0.4 ± 0.1	1.4 ± 0.4	3.9	0.9 ± 0.2	0.8 ± 0.2	0.9
C-188–9 (μM)	13 ± 2	15 ± 2	1.1	9 ± 4	15 ± 4	1.6
Oxal (nM)	773 ± 39	530 ± 92	0.7	631 ± 79	481 ± 53	0.8
5-FU (μM)	3.1 ± 0.2	2.8 ± 0.6	0.9	16 ± 5	10 ± 2	0.7
SN-38 (nM)	7 ± 1	9 ± 1	1.3	35 ± 13	45 ± 9	1.3
DOX (nM)	50 ± 5	21 ± 1	0.4	129 ± 6	546 ± 41	4.2
CHX (ng/ml)	86 ± 4	90 ± 11	1.0	121 ± 16	120 ± 10	1.0
HHT (nM)	12 ± 1	46 ± 4	3.7	17 ± 1	95 ± 12	5.6
5c (nM)	25 ± 4	94 ± 10	3.8	40 ± 7	344 ± 65	8.6

Values represent mean ± S.D. from at least 3 independent experiments performed in duplicate.

^*Fold resistance = IC₅₀ resistant cells/IC₅₀ parental cells.

Fig. 3.

Potential mechanisms of resistance in BOL-resistant cells. (A,B) Protein expression in parental and BOL-resistant cells. (C,D) HCT116 cells were treated for 24 h with 50 nM BOL and HHT. Cells were harvested and processed for RPPA analysis. Percentage values represent the mean from duplicate wells. Cutoff values for alterations were set at 30% for either treatment.

3.4. The effect of quassinoid analogs on inhibition of protein synthesis

As HHT is a well-known protein synthesis inhibitor, we evaluated the effect of BOL on global protein synthesis with the Click-iT protein synthesis assay. HCT116 cells were incubated with L-homopropargylglycine (HPG) for 2 h in the absence or presence of inhibitors and then fixed, permeabilized, and incubated in the Click-iT reaction to bind Alexa 488 azide to the HPG-alkyne group. Protein-incorporated HPG-Alexa 488 was determined by flow cytometry. BOL (30 nM) inhibited protein synthesis by 75% (Fig. 4A). To determine the extent that other naturally occurring quassinoids inhibited protein synthesis, HCT116 cells were incubated with 30 nM bruceine A, B, C, and bruceantin (Btin). Both bruceine A and C had limited activity while bruceine B was as equally potent as BOL (Fig. 4A). Bruceantin suppressed protein synthesis to the greatest extent (93%). These quassinoids differ only in the length and composition of the C15 side chain, which plays a critical role in determining their inhibitory activity.

Fig. 4.

Effect of quassinoid analogs on protein synthesis. (A) HCT116 cells were pre-incubated for 30 min in methionine-free RPMI1640. HPG was added in the absence or presence of natural quassinoids (30 nM) for 2 h. Cells were harvested, fixed, and processed for the Click-iT assay. (B) Cells were incubated with BOL (30 nM), HHT (30 nM) and NAPA (3 μM) for 2 h and processed for Click-iT assay analysis. Mean signal intensities in the graphs represent the mean ± SD from three independent experiments performed in duplicate. **p < 0.01 vs. + HPG alone. HCA7 (C) and CCD841 (D) cells were incubated with BOL (30 nM), HHT (30 nM) and NAPA (3 μM) for 2 h and processed for Click-iT assay. Values in the graph represent the mean ± SD from three independent experiments. *p < 0.05; **p < 0.01 vs. control.

The natural product HHT suppressed protein synthesis to a greater extent than BOL at equimolar concentrations (Fig. 4B). Interestingly, the STAT3 inhibitor NAPA (3 μM) demonstrated significant protein synthesis inhibition similar to HHT. This inhibitory effect was also observed in HCA7 cells (Fig. 4C). Previously, we demonstrated that normal colon epithelial CCD841 cells were resistant to the growth inhibitory effects of BOL [6]. BOL was unable to inhibit protein synthesis in CCD841 cells while HHT and NAPA maintained their strong inhibitory effects (Fig. 4D). These results suggest that normal cells may have reduced toxicities to these quassinoids.

Our results demonstrate that BOL can both inhibit STAT3 phosphorylation and suppress global protein synthesis. To examine whether these two observations are biologically connected, we explored a possible link between protein synthesis and phosphorylation of STAT3. We treated HCT116 cells with two protein synthesis inhibitors CHX and HHT for 24 h. Surprisingly, we observed significant inhibition of p-STAT3 by these two inhibitors (Fig. 5A). In support of this, Cao et. al. demonstrated p-STAT3 inhibition by HHT [48]. HHT and CHX treatment also had considerable inhibitory effects on the expression of the downstream proteins MCL1, PTTG1, survivin, and c-MYC. As the RPPA analysis demonstrated that both HHT and BOL activate the mTOR signaling pathway, we next treated cells with Torin1, a potent mTOR inhibitor [49]. As shown in Fig. 5B, Torin1 treatment inhibited constitutive expression of p-STAT3 but to a smaller extent than the protein synthesis inhibitors. Thus, these results suggest suppression of protein synthesis may lead to the downstream consequence of inhibition of STAT3 phosphorylation.

Fig. 5.

Effect of protein synthesis inhibitors on protein expression. (A) HCT116 cells were treated with HHT and CHX for 24 h and processed for immunoblot analysis. (B) Cells were treated with Torin1 for 24 h and processed for immunoblot analysis. Values in the graph represent the mean ± SD from three independent experiments. *p < 0.05; **p < 0.01 vs. control.

3.5. Synthesis and activity of quassinoid analogs

We demonstrated that natural quassinoids with different C15 side chains had altered biological activity (Fig. 4A). To further investigate the structural diversity at the C15 position, a semi-synthetic route (Fig. 6A) was developed to design novel quassinoid analogs based on a modified protocol by Lee K-H et al [40]. We designed and synthesized five analogs and analyzed their cytotoxicity against human CRC cells. The analog incorporating a sorbic acid (2E,4E)-hexadiene side chain (5a) had similar antiproliferative activity as BOL (Table 2). While addition of a cyclohexyl group on C15 reduced antiproliferative activity (5b), attaching a furan ring (5c), which was chosen as a bioisostere of the terminal ester on BOL, resulted in a more active analog with 2-fold greater cytotoxicity (Table 2). The BOL-resistant cell lines remained resistant to this potent analog (Table 1). To further explore the furan-based derivative, a methylated analog (5d) was synthesized, with the hypothesis being that the methyl moiety might mimic the methyl group of the terminal acetate of the BOL C15 side chain. However, this analog was found to be somewhat less active than 5c. To demonstrate the critical importance of the side chain for biological activity, we synthesized an analog with only a hydroxyl group on C15 (6; also known as bruceolide). Notably, 6 was ~ 250-fold less potent than 5c. The effect of these analogs on clonogenic growth was also evaluated. As with the cell proliferation assays, 5c was the most potent analog with 6 have no effect (Fig. 6B).

Fig. 6.

Synthetic route and SAR of quassinoid analogs. (A) Analogs were synthesized using a 4-step semi-synthetic route with bruceine A as the starting material. (B) Effect of quassinoid analogs on colony formation. HCT116 and HCA7 cells were incubated with quassinoids for 24 hr. After 10–14 days, colonies were fixed.

Table 2.

IC₅₀ values of quassinoid analogs.

Analog (nM)	HCT116	HCA-7
5a	47.7 ± 6.6	103.7 ± 32.5
5b	84.8 ± 2.2	165.3 ± 42.3
5c	24.5 ± 3.9	39.8 ± 6.6
5d	36.9 ± 3.8	69.8 ± 12.6
6	10170 ± 1185	10546 ± 2721
BOL	55.3 ± 3.2	63.6 ± 6.2

Values represent the mean ± S.D. from at least 3 independent experiments performed in duplicate.

3.6. Effect of quassinoid analogs on alteration of cancer signaling pathways

To determine how C15 modifications may affect cellular protein expression, HCT116 cells were treated with these analogs for 24 h, and changes in protein expression were subsequently analyzed by immunoblot analysis. The ability of these analogs to suppress proteins such as MCL-1, PTTG1, and c-MYC correlated with their cytotoxic effects (Fig. 7A). Analog 5c inhibited protein expression to the greatest extent with 6 having no effect (5c > 5a = 5d > 5b > 6). Similar to BOL, the analogs also inhibited STAT3 phosphorylation. The same inhibitory pattern on protein expression was observed in HCA7 cells (data not shown). Furthermore, treatment of normal colon epithelial CCD841 cells with 5c had little effect on protein expression (Fig. 7B) which is likely due to ABCB1 overexpression in these normal cells (7C). To investigate the alterations in protein expression in a CRC three-dimensional organoid model, P#127 organoids were treated with BOL, 5c, and 6 for 48 h, followed by IL-6 stimulation. While both BOL and 5c reduced expression of the proteins MCL1, PTTG1, and p-STAT3, BOL was more effective than 5c in this organoid model (Fig. 7D). Analog 6 had no effect on protein expression. BOL was more potent than 5c at inhibition of organoid proliferation (IC₅₀: 84 ± 40 nM vs. 219 ± 130 nM; Fig. 7E). As expected, 6 had no activity in CRC organoids.

Fig. 7.

Effect of quassinoid analogs on protein expression. (A) HCT116 cells were incubated with quassinoid analogs for 24 h. Cells were harvested and processed for immunoblot analysis. (B) CCD841 cells were treated with BOL and 5c for 24 h and processed for immunoblot analysis. (C) ABCB1 expression in HCA7, HCA7^BR and CCD841 cells. (D) CRC organoids were treated with drugs for 48 h followed by IL-6 incubation. Organoids were harvested and processed for immunoblot analysis. (E) CRC organoids were treated with quassinoids for 7 days. Cell proliferation was determined using CellTiter Glo 3D assay. Values in the graph represent the mean ± SD from three independent experiments. *p < 0.05; **p < 0.01 vs. control.

For comparison with the quassinoid analogs, we also evaluated the effect of three STAT3 inhibitors on protein expression: napabucasin (NAPA), S3I-201, and C188–9 [10,22,50]. While these compounds inhibited the expression of multiple cellular proteins, the concentrations required were 100-fold (NAPA) or 1000-fold (S3I201; C188–9) >5c (Fig. 8). The concentration of NAPA necessary for inhibition of p-STAT3 expression was 3 μM while cell proliferation was suppressed at considerably lower concentrations (IC₅₀, 0.37 μM; Table 1) suggesting that STAT3 inhibition plays only a minor role in the activity of this compound in CRC cells.

Fig. 8.

Effect of STAT3 inhibitors on protein expression. HCT116 cells were treated with NAPA, S3I-201, and C-188–9 for 24 h. Cells were harvested and processed for immunoblot analysis. Values in the graph represent the mean ± SD from three independent experiments. *p < 0.05; **p < 0.01 vs. control.

To determine whether the analogs bound to STAT3 protein, protein pull-down assays were performed with analog-conjugated Sepharose beads [37]. While 5c readily bound to STAT3 protein, 6 demonstrated a limited ability to interact with the protein (Fig. 9A). To confirm this analog-protein interaction, we performed electrophoretic mobility shift assays (EMSA). Analogs 5c and 5d inhibited STAT protein binding to the STAT3 DNA binding oligo similarly while 6 had no binding activity (Fig. 9B). We next determined the effect of the C15 modifications on protein synthesis using the Click-iT assay. 5a and 5b had weak suppressive activity while 6 completely lacked activity (Fig. 9C). The most cytotoxic analog, 5c, inhibited protein synthesis to the greatest extent (89%). Similar observations were seen in HCA7 cells. These findings provide evidence that the C15 side chain on the quassinoid determines both the interaction with STAT3 protein and the interaction with the protein synthesis machinery.

Fig. 9.

Effect of quassinoid analogs on STAT3 binding and protein synthesis. (A) Analog-conjugated 6B beads were incubated with HCT116 nuclear lysate. (B) Effect of quassinoid analogs on STAT3 oligo DNA binding assay. EMSA assay values represent the mean ± S.D from three independent experiments. (C) Cells were incubated with BOL and analogs (30 nM) for 2 h and processed for Click-iT assay analysis. (D) Effect of quassinoids on protein synthesis. Luciferase mRNA and quassinoids were added to RBC reticulocyte lysate translation reaction. After 90 min at 30 °C, reactions were analyzed for luciferase activity by luminescence. Values represent the mean ± S.D. from 3 individual experiments.

4. Discussion

We recently showed that the quassinoid BOL is a potent suppressor of STAT3 phosphorylation with promising antitumor activity against in vivo CRC models [6]. Given their preclinical antitumor activity, this class of compounds has remained of therapeutic interest [26,51,52]. While bruceantin exhibited clinical activity in early-phase clinical trials [53,54], subsequent phase II trials revealed only modest activity but substantial toxicities [55]. Clinical interest was also hampered by unclear mechanisms of action suggesting additional preclinical studies are necessary. Upon obtaining BOL from a commercial source, higher BOL doses were evaluated in vivo. In nude mice, we demonstrated that BOL had potent antitumor activity (>90%) but some body weight loss and a 33% reduction in spleen weight. In the MC38 syngeneic tumor model, significant suppression of tumor growth was observed along with host toxicities (body and spleen weight loss), especially at the higher BOL dose. Host toxicities were completely eliminated following encapsulation of BOL into POEG-POM nanoparticles. These BOL-containing nanoparticles maintained effective tumor growth suppression.

To date, the precise mechanism of action of quassinoids has not been well established. This class of compounds exerts its cytotoxic effects through several different mechanisms, including inhibition of ROS signaling pathways, Nrf2-mediated defense mechanisms, and inhibition of key cell signaling pathways such as AKT, MEK, and c-MYC [23–27]. Our previous work suggested the possibility of STAT3 inhibition in the antitumor activity of BOL. Other studies have shown that quassinoids inhibit the proliferation of cancer cells through inhibition of protein synthesis [28,29]. We addressed whether STAT3 phosphorylation played a role in the antiproliferative activity of these molecules in CRC cells. STAT3 siRNAs had modest effects on cell proliferation while a STAT3 degrader (SD-36) had little to no effect on CRC cell proliferation. In fact, STAT3′s role in cell growth appeared to be limited in other cancer cell types such as breast, lung, and liver. This finding may explain the lack of clinical efficacy of investigational STAT3 inhibitors against solid tumors [17,18,20,22]. However, Lu et. al. demonstrated that growth of STAT3^−/−ovarian cancer cells was suppressed compared to parental cells suggesting STAT3 may play different roles depending upon the cellular context [56]. Interestingly, we showed that several STAT3 inhibitors can effectively inhibit protein synthesis. Zuo et. al. also demonstrated that NAPA blocked protein synthesis [57]. Furthermore, we demonstrated that protein synthesis inhibitors (HHT; CHX) and an mTOR inhibitor (Torin1) significantly inhibit STAT3 phosphorylation, which when, taken together, suggests that inhibition of protein synthesis may lead to downstream consequences that influence and/or regulate STAT3 phosphorylation.

The BOL-resistant cell lines demonstrated cross-resistance to the natural plant alkaloid homoharringtonine (HHT; omacetaxine). HHT is a protein synthesis inhibitor that was approved by the US FDA in 2012 for the treatment of chronic myeloid leukemia [58]. We confirmed that natural quassinoids effectively inhibit protein synthesis in CRC cells but not in normal colon epithelial cells. To improve the pharmacologic properties of natural quassinoids and further evaluate the structure–activity relationship (SAR), we designed and synthesized novel quassinoid analogs and evaluated them against CRC cells. The analog 5c with a C15 furan ring side chain exhibited more potent cytotoxicity and greater protein synthesis inhibition than the natural compound BOL. Scale-up synthesis of this active analog is currently being explored to permit future in vivo studies. It was previously suggested that the C15 side chain may play a role in cellular uptake [29]. Compound 6 (bruceolide), which lacks a C15 side chain, was significantly less active than 5c. While intracellular analog concentrations were not measured, cell-free translation experiments demonstrated that 6 could inhibit mRNA translation suggesting that the C15 side chain plays a role in cellular uptake of these molecules (Fig. 9D). Thus, it is clear that the mechanism of action of quassinoids and the synthesized analogs is due to inhibition of protein synthesis in agreement with investigations reported by others [28,29]. While quassinoids bind and suppress phosphorylation of nuclear STAT3, this inhibition does not appear to influence cell proliferation in the cell lines evaluated.

Protein synthesis is a highly regulated process and generally increased in cancer cells to cope with a rise in metabolism to sustain unrestricted growth. It was been shown that colorectal cancer has a higher rate of protein synthesis than normal colon tissue [31]. Several investigators have demonstrated that the development of CRC depends upon the translational capacity of intestinal cells [32,59]. The dependence of CRC on translation initiation factors has been shown by others [60]. The novel quassinoid analogs exhibited potent inhibition of protein synthesis, with 5c being the most promising analog. Our results demonstrate that the C15 side chain of the quassinoid is the key determinant factor for its pharmacological activity. Further, nanoparticle-encapsulation of the quassinoid protected host tissues from toxicity while maintaining potent antitumor activity. Thus, targeting CRC with protein synthesis inhibitors, such as quassinoid analogs, may represent a novel therapeutic strategy [30]. However, the broad spectrum of biological effects highlights the need for more accurate and systematic screening of these molecules. A rational drug design-based study could identify leads for the synthesis or semi-synthesis of new analogs with increased activity, decreased toxicity, and/or improved pharmacological profiles for the development of future cancer therapeutics.

Abbreviations:

ABCB1

ATP binding cassette subfamily B member 1

ATCC

American Type Culture Collection

Btin

bruceantin

BOL

bruceantinol

BR

BOL-resistant

CHX

cycloheximide

CRC

colorectal cancer

mCRC

metastatic CRC

DMAP

4-dimethylaminopyridine

EDCI

1-ethyl-3-(3-dimethylaminopropryl) carbodiimide

EMSA

electrophoretic mobility shift assays

FBS

fetal bovine serum

HHT

homoharringtonine

HPG

L-homopropargylglycine

i.p.

intraperitoneal

i.v.

intravenous

MSI-H

microsatellite instability-high

NAPA

napabucasin

NP

POEG-POM nanoparticles

OS

overall survival

PBS

Phosphate-buffered saline

POEG-POM

poly(oligoethylene glycol methacrylate)-co-poly(oleyl methacrylate

PROTAC

Proteolysis-targeting chimeras

ROS

reactive oxygen species

RPPA

Reverse Phase Protein Array

SAR

structure-activity relationship

siRNA

Small interfering RNA

STAT3

Signal Transducer and Activator of Transcription 3

TBS

tert-butyldimethylsilyl

Data availability

Data will be made available on request.