Development of novel 9-O-substituted-13-octylberberine derivatives as potential anti-hepatocellular carcinoma agents

Abstract A series of novel 9-O-substituted-13-octylberberine derivatives were designed, synthesised and evaluated for their anti-hepatocellular carcinoma (HCC) activities. Compound 6k showed the strongest activity against three human hepatoma cells including HepG2, Sk-Hep-1 and Huh-7 cells with IC50 values from 0.62 to 1.69 μM, which were much superior to berberine (IC50 >50 μM). More importantly, 6k exhibited lower cytotoxicity against normal hepatocytes L-02 with good lipid-water partition properties. The mechanism studies revealed that 6k caused G2/M phase arrest of the cell cycle, stabilised G-quadruplex DNA, and induced apoptosis via a mitochondrial apoptotic pathway. Finally, the in vivo anti-HCC activity of 6k was validated in the H22 liver cancer xenograft mouse model. Collectively, the current study would provide a new insight into the discovery of novel, safe and effective anti-HCC agents.


Introduction
Liver cancer is the seventh most common cancer and the third leading cause of cancer-related death around the world, with an estimation of 905677 new cases and 830180 new deaths in 2020. 1 Liver cancer patients are increasingly diagnosed per year, with an estimated incidence of >1 million cases by 2025. 2 Hepatocellular carcinoma (HCC) constitutes over 80% of all liver cancers, possessing both high metastatic spread and extremely poor prognosis. 3 Despite the substantial advances in drug discovery and development, the current treatment options for HCC are still limited. Surgical resection, local ablation and liver transplantation are applied with curative intention at the early stage of HCC, 4,5 while most patients do not present with symptoms until their tumours develop rapidly into an advanced, often incurable, stage. The molecular targeted drugs, such as sorafenib and regorafenib, are used to treat patients with advanced HCC, however, they are associated with deleterious clinical side effects and high drug resistances. [6][7][8] Hence, the development of novel, safe and effective agents in HCC therapy is urgently required.
Historically, natural products played a key role in drug discovery and development. There are over 80% of the approved small molecule antitumor agents deriving from natural products in the past 39 years from 1981 to 2019. 9 Berberine (1, Figure 1), a natural isoquinoline alkaloid isolated from Coptidis Rhizoma, has shown a variety of pharmacological activities including anti-bacterial, anti-cancer, anti-inflammatory, anti-Alzheimer, anti-diabetic, anti-obesity, antiviral, anti-hypertensive, anti-hyperlipidemic, etc. 10,11 The antitumor effects of berberine have been demonstrated in various types of cancers, such as colorectal carcinoma, melanoma, gastric and breast cancers as well as HCC. 12,13 Numerous studies revealed berberine exerted the anti-HCC activity via multiple mechanisms of action, including suppression of cell proliferation, migration and invasion, regulation of cell cycle, induction of cell apoptosis and autophagy, etc. 14 Besides, several studies have shown that berberine could selectively inhibit hepatoma cell invasion and induce hepatoma cell apoptosis while eliciting less cytotoxicity in normal hepatocytes, 15,16 and on the contrary, berberine could be combined with radiotherapy or chemotherapy drugs to neutralise their toxicity and enhance their therapeutic activities. [17][18][19] Based on these attractive anticancer properties, berberine seems to be a promising candidate for the treatment of HCC, however, the poor absorption and moderate anticancer efficacy seriously hamper its further development and application. 20 Berberine possesses poor lipid solubility, which leads to decreased effective concentration and minimal absorption in the gastrointestinal tract, thus exhibiting low antitumor effects. 21 As a result, many efforts including derivatization have been devoted to enhancing berberine's anticancer potential. 22 Accumulating evidence has demonstrated that the introduction of an alkyl or arylalkyl group at the C9 or C13 position of berberine could improve the lipophilicity, making berberine easier to penetrate the cytomembrane, so as to enhance the antitumor efficacy. 22 And our previous studies also confirmed that the introduction of alkyl chains of various lengths into the C9 position could effectively strengthen the anti-HCC activity of NO-donating berberine derivatives. 23 To our knowledge, however, few studies were carried out to improve the antitumor potency on structural modification at both C9 and C13 positions of berberine.
In order to find potential anti-HCC agents, in our ongoing efforts on structural modification of berberine, firstly, several 9,13dialkyl-substituted berberine derivatives were synthesised with high lipid-water partition (LWP) coefficients. While administrated against HepG2 cells, the derivatives were easy to precipitate in MTT assays due to low water solubility, leading to the test errors (data not provided). To overcome the above drawback, a series of novel 9-O-13-disubstituted berberine derivatives were designed to step by step ( Figure 1). We first introduced various alkyl or arylalkyl groups into the C13 position (3) of berberine to obtain the most potent 13-alkyl or -arylalkyl berberine derivative by in vitro activity screening, then treating that as a lead compound, again introduced diverse aminoalkyl groups into its C9 position (6) to remain the LWP coefficient and further enhance the anti-HCC potency. Herein, we report the synthesis, in vitro and in vivo anti-HCC activity and mechanism studies for these novel derivatives.

Chemistry
The synthesis of compounds 3a-e and 6a-n is illustrated in Scheme 1. The reaction of commercially available berberine (1) with acetone under the base produced compound 2a, which was treated with benzyl bromide in the presence of NaI to give compound 3a. On the other hand, 1 was reduced using NaBH 4 under the base to obtain compound 2b, followed by the reaction of diverse aldehydes in a mixed solvent of EtOH and AcOH to give compounds 3b-e. Next, selective demethylation of 3c through pyrolysis produced compound 4c, which was directly reacted with different dibromoalkanes to form compounds 5a-d, followed by treatment with various amines to afford compounds 6a-n.

In vitro anti-HCC activity against HepG2 cells
Compounds 3a-e were first synthesised and screened for their antiproliferative activity against HepG2 cells. As seen in Table 1, compound 3c exhibited the strongest activity among the obtained C13-substituted berberine derivatives. Then taking 3c as a precursor, compounds 6a-h were prepared and evaluated. The results indicated compounds bearing terminal six-membered heterocycles such as morpholinyl (6f), piperidinyl (6g) and N-methylpiperazinyl (6h) groups displayed more potent activities than those with other substituted amino groups such as diethylamino (6d) and pyrrolidinyl (6e) groups (Table 1). Moreover, it was found that the activity increased by extending the alkyl chain length from C 3 to C 6 (e.g. 6d vs 6a, 6e vs 6b, 6f vs 6c, Table 1). Therefore, 9-O-substituted-13-octylberberine derivatives 6i-n with longer  alkyl chains and various six-membered heterocycles were further synthesised. The bioassay results indicated the anticancer activity gradually strengthened with the extension of the alkyl chain, however, the chain length had a little detrimental effect on the activity from C 8 to C 10 (e.g. 6i vs 6l, 6j vs 6m, 6k vs 6n, Table 1). While compounds bearing morpholinyl (e.g. 6i), piperidinyl (e.g. 6j) and N-methylpiperazinyl (e.g. 6k) groups displayed comparably potent activity. Among all the derivatives, compound 6k showed the strongest anti-HCC activity with an IC 50 value of 0.76 lM, which was much superior to that of positive control cisplatin (IC 50 6.87 lM).

2.3.
In vitro anti-HCC activity of 6k against a wide range of hepatoma cell lines Compound 6k was selected to examine the antiproliferative activity against other hepatoma cell lines including Sk-Hep-1, Huh-7 and Hep3B cells. As presented in Table 2, 6k showed potent anti-HCC activities against the three cell lines with IC 50 values of 1.69, 0.62 and 1.60 lM, respectively, much higher than berberine with IC 50 values of >50 lM, even superior to positive control cisplatin with IC 50 values of 6.52, 3.03 and 11.19 lM, respectively. More importantly, 6k showed lower cytotoxicity towards normal hepatocytes L-02 compared to other hepatoma cell lines ( Table 2). As a result, 6k would be a potential anti-HCC agent and deserve further antitumor mechanism and in vivo studies.

Effect of 6k on cell cycle progression
To determine whether 6k inhibited hepatoma cell proliferation by cell cycle regulation, the cell cycle distribution was analysed by flow cytometry after staining cellular DNA with propidium iodide (PI). As shown in Figure 2(A,B), 6k caused cell cycle arrest at G2/M phase in a dose-dependent manner. Treatment of HepG2 cells with 6k at concentrations from 0 to 1.0 lM increased the G2phase cell percentages from 18.00% to 32.16%, while the total cell percentages of G1 and S phases decreased from 82.00% to 67.83% concomitantly. Then, the mechanism of 6k-induced cell cycle arrest in HepG2 cells were further explored. The suppression of cell cycle progression from G2 to M phase in eukaryotic cells has been suggested to be associated with expressions of several regulatory proteins including cdc25c, cdc2 and cyclin B1 24 . As seen in Figure 2(C,D), 6k significantly promoted cdc25c phosphorylation, inhibited cdc2 kinase activation and reduced cyclin B1 binding in concentration-dependent manners. These results indicated that 6k blocked G2/M phase of the cell cycle possibly by regulating the expressions of p-cdc25c, cdc2 and cyclin B1.

Effect of 6k on G-quadruplex DNA
It is well known that G2 phase is the late stage of DNA synthesis and the preparation period for mitosis. During the progression of DNA replication, there are a number of G-quadruplex (G4) conformation present in oncogene (e.g. c-MYC) region, which are formed by diverse guanine-rich sequences, to maintain the stable existence of the replication action. 25 G4 stabilisation by a smallmolecule ligand could repress transcription of certain oncogenes and/or induce DNA damage at telomeres and oncogenes leading to replication defects and cancer cell death. 26 Berberine has been reported to exert the anticancer activity related to its binding to G4 DNA. 27 To investigate whether the anti-HCC activity of 6k is involved with G4 DNA-binding effects, FRET-melting experiments were carried out with berberine as a reference. The fluorescent labelled oligomers FPu18T and F10T were employed as c-MYC G4 DNA and duplex DNA model, respectively. As shown in Table 3, 6k could effectively stabilise G4 structure with a DTm(A) value of 13.8 C, which was much higher than berberine (DTm: 1.2 C). Whereas the melting temperature of F10T was not obviously elevated by 6k with a DTm(B) value of 0.8 C (Table 3), suggesting  6k possessed lower binding affinity towards duplex DNA than G4 DNA. Besides, the G4 selectivity of 6k was further examined by FRET-based competition assays using a nonfluorescent duplex DNA (ds26) as a competitor. The presence of excess ds26 didn't obviously influence FPu18T stabilisation by 6k with a DTm(C) value of 12.4 C (Table 3), indicating the duplex DNA had negligible effect on the binding of 6k to G4 DNA. Altogether, 6k possessed high G4 DNA stabilising efficacy with better selectivity over duplex DNA.

Effect of 6k on cell apoptosis and mitochondrial membrane potential
To verify whether cancer cell viability decreased by 6k is associated with apoptosis, an Annexin V-FITC/PI binding assay was performed and the results were analysed by flow cytometry. As shown in Figure 3(A,B), 6k markedly induced apoptosis in a dosedependent manner. HepG2 cells were treated with 6k at concentrations of 0.25, 0.50 and 1.0 lM for 72 h, which resulted in 16.42%, 34.29% and 59.17% total apoptotic cells (Q2 þ Q3), as compared with 6.28% in the control group, indicating that 6k exerted its anti-HCC effect possibly by inducing apoptosis.
It is well known that mitochondria play an important role in the process of apoptosis. Mitochondrial dysfunction triggers an alteration in mitochondrial structure, which can result in disruption of mitochondrial membrane potential (MMP) to initiate the apoptosis pathway 28 . In order to explore whether 6k could induce mitochondrial dysfunction, the lipophilic mitochondrial probe JC-1 was employed to measure MMP. As shown in Figure 3(C,D), treatment with 6k at concentrations from 0.25 to 1.0 lM for 72 h, the percentages of HepG2 cells with green fluorescence intensity correspondingly increased from 10.33% to 37.52%, as compared with 1.90% in the control group, suggesting that 6k caused MMP collapse and mitochondrial dysfunction in the process of HepG2 apoptosis.

Effect of 6k on mitochondrial apoptotic pathway
The above results demonstrated that apoptosis induced by 6k was likely involved in a mitochondrial pathway. It has been found that when mitochondrial dysfunction occurs, cytochrome C releases from mitochondria into the cytoplasm across the membrane, which promotes the formation of a caspase activating complex that triggers the activation of downstream caspases 29 . Caspase-3 acts as a final executor in apoptosis and can be irreversibly activated through cleavage of procaspase-3 30 . Bcl-2, an anti-apoptotic protein, preserves mitochondrial structure and function. 31 Bax, a dominant-negative inhibitor of Bcl-2, increases mitochondrial membrane permeability and promotes apoptosis. 31 Thus, the expressions of Bax, Bcl-2, cytochrome C and caspase-3 were tested by western blot analysis. As shown in Figure 4(A,B), treatment with 6k in HepG2 cells promoted cytochrome C release,    activated caspase-3, downregulated Bcl-2 expression and upregulated Bax expression in concentration-dependent manners. The results indicated that 6k-induced apoptosis could be mediated via a mitochondrial-dependent apoptotic pathway.

LWP property and aqueous solubility of 6k
In order to verify whether the physicochemical properties of berberine were improved, compounds 3c and 6k were selected to determine the LWP coefficient and aqueous solubility by the shake flask 32 and HPLC 33 methods, respectively. As expected in Table 4, compound 6k could not only enhance the anti-HCC activity but also improve LWP property (Log D 3.87) compared to berberine (Log D À0.98). The water solubility of 6k was 1.08 mg/mL, which was comparable to that of berberine (1.36 mg/mL).

In vivo anti-HCC activity of 6k
Encouraged by the above results, 6k was selected to further test it is in vivo anti-HCC efficacy in mice liver cancer xenograft model established by subcutaneous inoculation of H22 cells into the right flank of mice. The tumour sizes and body weights of the mice were monitored and recorded every 2 days. All the mice were sacrificed after 3 weeks, the tumours were excised and weighed ( Figure 5A). As illustrated in Figure 5(B,C), 6k markedly inhibited the tumour growth (both volume and weight) in a dosedependent manner. At a dose of 30 mg/kg, 6k possessed a much higher tumour inhibitory rate (TIR 61.1%) than berberine (TIR 47.3%). Besides, 6k didn't cause significant loss in mice body weights during treatment ( Figure 5D).

Conclusion
In summary, a series of novel 9-O-substituted-13-octylberberine derivatives were designed to step by step, prepared and their in vitro anti-HCC activities were initially evaluated. The results indicated that all the derivatives exhibited markedly improved antiproliferative activity compared to the parent compound berberine. Among them, compound 6k showed the most potent activity (IC 50 0.62 $ 1.69 lM) against a wide range of human hepatoma cells while displaying lower cytotoxicity towards normal hepatocytes. The mechanism studies indicated that 6k blocked G2/M phase of the cell cycle possibly by influencing the expressions of p-cdc25c, cdc2 and cyclin B1, stabilised G4 DNA with higher selectivity over duplex DNA, caused MMP loss and induced apoptosis through a mitochondrial apoptotic pathway. Besides, 6k possessed good lipid-water partition property (log D 3.87) and aqueous solubility (1.08 mg/mL). Furthermore, 6k effectively suppressed the tumour growth by 61.1% (w/w) in the H22 liver cancer xenograft mouse model, which was significantly superior to berberine (47.3%, w/w) at the same dose. Altogether, compound 6k deserves to be further investigated as a potential anti-HCC agent for cancer therapy.

General
Most chemicals and solvents were purchased from commercial sources. Further purification and drying by standard methods were employed when necessary. Melting points (m.p.) were taken on an XT-4 micro melting point apparatus and uncorrected. 1 H NMR and 13 C NMR spectra were recorded on Bruker-300 spectrometers in the indicated solvents (CDCl 3 or DMSO-d 6 , TMS as internal standard): the values of the chemical shifts are expressed in d values (ppm) and the coupling constants (J) in Hz. High-Resolution Mass measurement was performed on Agilent QTOF 6520 mass spectrometer with electron spray ionisation (ESI) as the ion source. The purity of all tested compounds was !95%, as estimated by HPLC analysis. Reactions were monitored by thin-layer chromatography (TLC) on 0.25 mm silica gel plates (GF254) and visualised under UV light. Flash column chromatography was carried out using commercially available silica gel (200-300 mesh) under pressure.

4.1.2.
General procedure for synthesis of compound 3 4.1.2.1. Synthesis of compound 2. To a stirred 2 N NaOH (8 ml) solution of berberine 1 (1.86 g, 5 mmol) was dropwise added acetone (4 ml), the mixture was stirred at room temperature for 1.5 h. The mixture was then filtered and washed with 80% ethanol (15 ml) to provide the crude compound 2a without further purification for next step directly.
To a stirred solution of berberine 1 (3.72 g, 10 mmol) and K 2 CO 3 (4.14 g, 30 mmol) in MeOH (100 ml) was dropwise added a 5% NaOH (10 ml) solution containing NaBH 4 (570 mg, 15 mmol), the mixture was stirred at room temperature for 2 h. The mixture was then filtered, and washed with water (30 ml) and 80% ethanol (30 ml) to provide the crude compound 2b without further purification for next step directly.

Synthesis of compound 3.
To a solution of compound 2a (1.18 g, 3 mmol) in acetonitrile (10 ml), sodium iodide (450 mg, 3 mmol) and benzyl bromide (1.1 ml, 9 mmol) were added. The reaction mixture was stirred at 85 C for 8 h. The solvent was removed under reduced pressure, the residue was purified by silica gel column chromatography using dichloromethane/methanol as an eluent to give compound 3a.
13-Benzyl-9,10-dimethoxy-5,6-dihydro- [1,3] To a mixed solution of compound 2b (2.02 g, 6 mmol) in 80% ethanol (16 ml) and AcOH (4 ml), different aldehydes (18 mmol) were added. The reaction mixture was stirred at 90 C for 5-8 h. The solvent was removed under reduced pressure, the residue was then acidified with 1 N HCl (20 ml), and stirred at room temperature for 2 h. The solid was collected by filtration and then purified by silica gel column chromatography using dichloromethane/methanol as an eluent to give compounds 3b-e.

In vitro antiproliferative assay
HepG2, Sk-Hep-1, Huh-7, Hep3B and L-02 cells were purchased from Nanjing Key Gen Biotech Co. Ltd. (Nanjing, China). The cytotoxicity of test compounds was determined using an MTT assay. Briefly, the cell lines were incubated at 37 C in a humidified 5% CO 2 incubator for 24 h in 96-microwell plates. After medium removal, 100 lL of culture medium with 0.1% DMSO containing test compounds at different concentrations was added to each well and incubated at 37 C for another 72 h. The MTT (5 mg/mL in PBS) was added and incubated for another 4 h, the optical density was detected with a microplate reader at 490 nm. The IC 50 values were calculated according to the dose-dependent curves. All the experiments were repeated in at least three independent experiments.

Cell cycle analysis
HepG2 cells were seeded into 6-well plates and incubated at 37 C in a humidified 5% CO 2 incubator for 24 h, and then treated with or without 6k at indicated concentrations for another 72 h. The collected cells were fixed by adding 70% ethanol at 4 C for 12 h. Subsequently, the cells were resuspended in PBS containing 100 lL RNase A and 400 lL of PI for 30 min. The DNA content of the cells was measured using a FACS Calibur flow cytometer (BectoneDickinson, San Jose, CA, USA). The melting process of G4 DNA was monitored in the presence or absence of 2 lM compounds by measuring the fluorescence intensity of fluorescein. Measurements were performed with excitation at 470 nm and detection at 530 nm. Fluorescence readings were taken at an interval of 2 C over the range of 30-98 C, with a constant temperature being maintained for 1 min prior to each reading to ensure a stable value. The data was analysed by Origin 8.5 (OriginLab Corp.). To test the selectivity of compounds for G4 structure, 10 lM ds26 DNA (5 0 -GTTAGCCTAGCTTAAGCTAGGCTAAC-3 0 ) was added, followed by the same method as above.

Cell apoptosis analysis
Briefly, after treatment with or without 6k at indicated concentrations for 72 h, the cells were washed twice in PBS, centrifuged and resuspended in 500 lL AnnexinV binding buffer. The cells were then harvested, washed and stained with 5 lL Annexin V-FITC and 5 lL PI in the darkness for 15 min. Apoptosis was analysed using a FACS Calibur flow cytometer (BectoneDickinson, San Jose, CA, USA).

MMP measurement
HepG2 cells were incubated with 6k as described above. After incubation for 72 h, the cells were washed in PBS and resuspended in 500 lL JC-1 incubation buffer at 37 C for 15 min. Then, 6k was immediately assessed for a red fluorescence using a microplate reader (ELx80, Bio-Tek, USA). The fluorescent signal of monomers was measured with an excitation wavelength of 488 nm. The percentage of cells with healthy or collapsed mitochondrial membrane potentials was monitored by flow cytometry analysis (BectoneDickinson, San Jose, CA, USA).

Western blotting analysis
HepG2 cells were incubated with different doses of 6k for 72 h. Cells were collected, centrifuged, and washed two times with icecold PBS. The pellet was then re-suspended in lysis buffer. After the cells were lysed on ice for 20 min, lysates were centrifuged at 13000 g at 4 C for 15 min. The protein concentration in the supernatant was determined using the BCA protein assay reagents. Equal amounts of protein (20 lg) were resolved using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (10-12% acrylamide gels) and transferred to PVDF membrane. Membranes were blocked for 1 h at room temperature. Membranes were then incubated with primary antibodies against p-cdc25c, cdc2, cyclin B1, Bax, Bcl-2, cytochrome C, caspase-3, VDAC1 and b-actin, the membrane being gently rotated overnight at 4 C. The bound antibodies were detected using horseradish peroxidase (HRP)-conjugated second antibodies and visualised by the enhanced chemiluminescent reagent.

4.2.7.
Determination of the LWP coefficient and aqueous solubility 4.2.7.1. LWP coefficient (log D) measurement. Berberine, 3c and 6k of 1 mg were dissolved in three centrifugal tubes (5 ml) containing phosphate buffer (pH 7.4)-saturated 1-octanol solution (2 mL), followed by the addition of 1-octanol saturated PBS (2 mL). Each tube was vigorously mixed for 1 min and then shaken for 6 h at a speed of 1000 rpm at 37 C. The mixtures were then centrifuged. 1 mL of 1-octanol phase and 1 mL of buffer were separately collected. Each phase was diluted to a suitable concentration and analysed by HPLC (Shimadzu DGU-20A3R) on a Shimadzu-GL WondaSil C18-WR column (k ¼ 254 nm) with an eluent of methanol/water (65/35, v/v, 1.0 ml/min). The log D value for each compound was calculated from the ratio of the peak areas using the following equation: Log D ¼ log (Area oct Â x/Area buf Â y). x, y is the dilution factor. 4.2.7.2. Aqueous solubility measurement. Berberine, 3c and 6k (5 mg) were dissolved in 1 ml PBS (50 mM, pH 7.4) at 25 C. The suspension was sonicated for 30 min and then centrifuged at 15000 g for 5 min. The supernatant was filtered through a PTFE filter (0.2 lm) slowly. Then 10 lL filtered solution was injected into the HPLC system (Shimadzu DGU-20A3R) on a Shimadzu-GL WondaSil C18-WR column and quantified using the standard curves made by the corresponding known concentrations of berberine, 3c and 6k. The eluent was methanol/water (65/35, v/v) at a flow rate of 1.0 ml/ min with the detection wavelength at 254 nm.

In vivo anti-HCC assay
Five-week-old male Institute of Cancer Research (ICR) mice were purchased from Shanghai SLAC Laboratory Animals Co. Ltd. Briefly, a total of 1 Â 10 6 H22 cells were subcutaneously inoculated into the right flank of ICR mice according to protocols of tumour transplant research, to initiate tumour growth. After incubation for one day, mice were weighed and randomly divided into 5 groups of 8 animals. The groups were administered berberine (30 mg/kg) and 6k (15 or 30 mg/kg) in a vehicle of 10% DMF/2% Tween 80/ 88% saline, respectively. Vehicle and cyclophosphamide (20 mg/ kg) were employed as negative and positive controls, respectively. Treatments were performed by intravenous injection once per day for a total of 21 consecutive days. Body weights and tumour volumes were measured every 2 days. The mice were sacrificed after the treatments, and the tumours were excised and weighed. The tumour volume and inhibition rate (w/w) were calculated as follows: tumour volume ¼ L Â W 2 /2, where L is the length and W is the width; tumour inhibitory ratio (%) ¼ (1-average tumour weight of treated group/average tumour weight of control group) Â 100%. All procedures were performed following institutional approval in accordance with the NIH Guide for the Care and Use of Laboratory Animals.

Disclosure statement
No potential conflict of interest was reported by the authors.