Synthesis of Paclitaxel Derivatives for Remote Loading into Liposomes and Improved Therapeutic Effect

A series of novel paclitaxel derivatives modified by boronic acid according to the characteristics of the interaction between RB(OH)2 and different strapping agents of intraliposomal aqueous phase were designed and synthesized, which were then used to develop remote poorly water-soluble drugs loading into liposomes. Meanwhile, we screened nineteen paclitaxel boronic acid derivatives for their cytotoxic activities against three cancer cell lines (A549, HCT-116 and 4T1) and one normal cell line (LO2), and performed liposome formulation screening of active compounds. Among all the compounds, the liposome of 4d, with excellent drug-encapsulated efficiency (>95% for drug-to-lipid ratio of 0.1 w/w), was the most stable. Furthermore, the liposomes of compound 4d (8 mg/kg, 4 times) and higher dose of compound 4d (24 mg/kg, 4 times) showed better therapeutic effect than paclitaxel (8 mg/kg, 4 times) in the 4T1 tumor model in vivo, and the rates of tumor inhibition were 74.3%, 81.9% and 58.5%, respectively. This study provided a reasonable design strategy for the insoluble drugs to improve their drug loading into liposomes and anti-tumor effect in vivo.


Introduction
Paclitaxel, a natural product with effectiveness, is one of the most widely used clinical anticancer agents [1,2] and shows commendable antitumor activity against various cancers, especially breast cancer [3,4]. The mechanism of paclitaxel is to inhibit the proliferation of tumor cells by stabilizing the microtubules and inhibiting the G2 or M phases of the cell cycle [5]. It has been reported that Taxol ® is the first marketed preparation of paclitaxel, but histamine may be produced with the degradation of its cosolvent Cremophor EL(CrEL), which results in allergic reactions in patients [6,7]. Similar to Taxol ® , and although other marketed paclitaxel formulations have shown potent antitumor activity against various cancers, there are still problems which limit their clinical applications, such as poor water solubility and low bioavailability, as well as drug resistance [8][9][10][11].
The advantage of a drug carrier is to improve the targeting ability, water solubility and PK properties of a drug [12][13][14][15][16][17]. Paclitaxel, as a non-ionizable drug, is difficult to stably encapsulate in a drug carrier. Fascicle chemical modification or derivatization of drugs is effective in developing remote poorly water-soluble drugs loading into liposomes. In our previous study, we found that weak acidic or basic derivatives of insoluble drugs can be actively encapsulated into liposomes by pH gradients [18][19][20]. In addition to pH gradients, there are many other remote loading gradients, such as copper ion gradients, which can complex with and encapsulate drugs into liposomes [21]. Boronic acid RB(OH) 2 can both spontaneously complex with copper ion and undergo dynamic covalent reactions with

Chemistry
The synthetic routes of the target compounds are illustrated in Scheme 1. Anhydrides 1a-1d were obtained via a dehydration reaction of alkyldicarboxylic acids, then reacted with 4-(hydroxymethyl)benzeneboronic acid pinacol ester to obtain intermediates 2a-2d, and then esterified with paclitaxel or docetaxel to obtain key intermediates 3a-3e. Finally, the target compounds 4a-4e and 6a-6b were obtained by oxidation hydrolysis.
The key intermediates 7a-7b were obtained by Arbuzov reaction of bromoalkyl alcohols, and then reacted with 4-nitrophenyl chloroformate to give 8a-8b, followed by esterification and oxidative hydrolysis to give the target compounds 10a-10c (Scheme 2). The synthetic routes of the target compounds 13a-13b and 17a-17g were similar to those of the compounds 10a-10c. The structures of target compounds, as listed in Table 1, were confirmed by 1 H-NMR, 13 C-NMR and HRMS.

In Vitro Cytotoxicity Studies
MTT assay was applied to verify the rationality of the molecular combination strategy. The cytotoxicity of the compounds was evaluated on lung cancer cell line A549, colon cancer cell line HCT-116, breast cancer cell line 4T1 and normal hepatocyte cell line LO2. Paclitaxel was used as a positive drug. The results showed that, compared with LO2, the tested compounds were more sensitive to HCT-116, A549 and 4T1, which suggested the cell selectivity of these compounds. The therapeutic index (TI) for LO2/HCT-116, LO2/A549 and LO2/4T1 of active compounds 4a, 4d, 13a, 17a, 17c and paclitaxel are shown in Table 2. According to the results of cytotoxicity, we roughly summarized the structure-activity relationship (SAR): (a) Compound 4d with disulfide bonds showed better selectivity for tumor cells than compound 4a without disulfide bonds. (b) The selectivity of aromatic boronic acid compound 13a was much better than that of alkyl boronic acid compounds (10a-10c). (c) Compound 17c (n = 3) showed the best selectivity, but when the carbon chain was further increased, the compounds 17d-17e (n = 4, 5) showed no selectivity.

Design and Optimization of Compound 4d Liposome Formulation
Since boronic acid RB(OH) 2 is not only a weak acid, but also a structure that can complex with copper ions, ATP and saccharides, we investigated whether different inner phases could be employed to encapsulate compounds (4a, 4d, 13a, 17a, 17c) with good therapeutic index. Firstly, the encapsulated efficiency of calcium acetate gradients was systematically explored and it was found that it was not feasible with the amount of sediment precipitated after loading (Supplementary Materials: Table S1).
The reasons may be that logD  4d, 13a, 17a, 17c) were not suitable for calcium acetate gradients (−2.5 < logD < 2; pKa < 11) [23]. Vyxeos ® is a liposomal formulation of a fixed molar ratio (1:5) combination of the antineoplastic drugs daunorubicin (DN) and cytarabine (Cyt) [24]. Daunorubicin is subsequently actively encapsulated with high efficiency through complexation with intra-liposomal copper. The copper gluconate buffer forms a complex with daunorubicin at a ratio of 1:1 to 1:2 [25]. Referring to Vyxeos ® , copper ion was selected as a complexed metal ion in the inner phase. In addition to compound 4d, most derivatives (4a, 13a, 17a, 17c) failed in loading into the liposomes, showing obvious sedimentation. Therefore, compound 4d was selected to investigate the effect of the anions on the loading efficiency. Among them, only the neutral gluconate copper was a suitable inner phase to facilitate the encapsulation of compound 4d (Table 3). The reason for the stability of the neutral formulations was potentially due to the fact that PBA was effectively protonated and interacted with copper more tightly in the neutral condition compared with the acidic condition [26]. Thus, the neutral Cu-gluconate was selected as the inner phase. The concentration of copper ions was also investigated, showing that the higher the copper concentration, the higher the EE. Considering the safety of copper ions, 300 mM was chosen. For lipid composition, the content of Chol was 45% (molar ratio), which revealed higher EE. The EE of compound 4d active loading liposome was more than 95% with copper ion gradient. Furthermore, we also evaluated the ATP and saccharides as remote gradients with the optimal formulation of copper ion gradients. The results of encapsulation efficiency are shown in Table 4. The inner phase of ATP was not successful, which may be caused by the weak binding ability between PBA and ATP. The results of different saccharides gradients showed that saccharides could be selected as the inner phase of liposomes to load PBA-modified derivates, but the EE of compound 4d was not high enough. Thus, the copper ion gradient was chosen for further experiments.
The optimal formulation was shown in Figure 2. The drug to lipid ratio was more than 1:10, drug loading efficiency was up to 9.02% and the prepared liposomes remained stable for more than two weeks at 4 • C. The morphology of compound 4d liposome is shown in Figures 3 and 4, with spherical particles, a narrow size distribution at~110 nm and zeta potential at~−15 mV (Table S2).

In Vitro Release of PTX Derivative 4d Liposomes
In vitro release study of compound 4d liposome ( Figure 5) showed that compound 4d was a ROS-sensitive paclitaxel derivative. When the content of H 2 O 2 increased from 0 to 10 mmol/L, the paclitaxel released by compound 4d significantly increased. Paclitaxel release at high H 2 O 2 content (1 mM and 10 mM) for 24 h (75.3% and 67.7%) was much better than that at low H 2 O 2 content (0.1 µM and 1 µM, 19.1% and 29.6%). The results showed that compound 4d liposome was stable in plasma and could reduce the toxicity of paclitaxel in systemic circulation, indicating that compound 4d liposome could effectively release paclitaxel in tumor tissue to achieve targeted anti-tumor effect.  Based on the availability of the tumor model, mice bearing a 4T1 xenograft tumor were used to study the in vivo antitumor efficacy of the compound 4d liposome. As depicted in Figure 6a, there was difference between compound 4d (8 mg/kg) and the paclitaxel control group (8 mg/kg). Furthermore, compound 4d liposome (8 mg/kg) significantly inhibited tumor growth, which may be due to its better stability, better drug delivery advantages and more effective drug release from tumor cells. When administered with a higher dose of compound 4d liposome (24 mg/kg), the tumor volume further decreased, showing obviously anti-tumor effect in vivo. In addition, according to changes in animal weight, all liposome preparations showed no obvious toxic effects ( Figure 6b). (b) Body weight. All data are presented as mean ± SD (n = 5), * p < 0.05 and **** p < 0.0001 by two-tailed Student's t test. "ns" mean "no significance".

General Synthesis Procedure for Intermediates 1a-1d
Compounds 1a-1d were prepared according to the literature [27]. Diacids (2.89 mmol) were added in acetic anhydride (5 mL), followed by nitrogen protection. The reaction was stirred at room temperature for 2 h. Then, dry toluene (10 mL) was added and the solvent was removed by reduced pressure. The product did not need to be purified.

General Synthesis Procedure (A) for Esterification
(4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)phenyl)methanol (2.03 mmol), acid anhydride (6.09 mmol) and DMAP (0.61 mmol) were added to dry DCM (10 mL), followed by nitrogen protection. The reaction was stirred at room temperature for 3 h and monitored by TLC. The mixture was washed with 2 M HCl solution and extracted with DCM three times. The combined organic layer was dried over Na 2 SO 4 and evaporated in a vacuum. The crude product was purified by silica gel column chromatography (petroleum ether: acetone = 10:1) to obtain the product.

General Synthesis Procedure (B) for Esterification
Carboxylic acid (0.35 mmol) and DMAP (0.04 mmol) were added to dry DCM (10 mL). The mixture was stirred at −20 • C for 10 min, followed by adding EDCI (0.35 mmol). Then, the reaction was stirred at −20 • C for 30 min, followed by adding paclitaxel or docetaxel (0.12 mmol). Then, the reaction was stirred at −20 • C for 60 min, and transferred to room temperature for 5-8 h. The mixture was washed with 2 M HCl solution and extracted with DCM three times. The combined organic layer was dried over Na 2 SO 4 and evaporated in a vacuum. The crude product was purified by thin layer silica gel column chromatography (dichloromethane: methanol = 20:1) to obtain the product.

General Synthesis Procedure (C) for the Hydrolysis Reaction
To a solution of boronic acid ester (0.13 mmol) in acetone (5 mL), a solution of sodium periodate (0.52 mmol) and ammonium acetate (0.65 mmol) in water (5 mL) was added and the reaction was stirred at room temperature for 8 h. The solid was precipitated after acetone was evaporated in a vacuum. The crude product was purified by thin layer silica gel column chromatography (dichloromethane: methanol = 20:1) to obtain the product.

General Synthesis Procedure (D) for the Arbuzov Reaction
CuCl (0.39 mmol), bis(pinacolato)diboron (1.97 mmol) and PCy 3 (0.39 mmol) were added to DCM (20 mL), followed by nitrogen protection. Then, t-BuOK (1.97 mmol) was added and the reaction was at 0 • C for 30 min. Finally, bromide (1.31 mmol) was added and the reaction was at 0 • C for 30 min. The mixture was washed with water and extracted with DCM three times. The combined organic layer was dried over anhydrous sodium sulfate and evaporated in a vacuum. The crude product was purified by silica gel column chromatography (petroleum ether: acetone = 100:1) to obtain the product.

General Synthesis Procedure (E) for Esterification
Alcohol (5.13 mmol), pyridine (14.89 mmol) and 4-nitrophenyl chloroformate (7.66 mmol) were added to dry DCM (10 mL), followed by nitrogen protection. The reaction was stirred at −20 • C for 5 h. The mixture was washed with 2 M HCl solution and extracted with DCM three times. The combined organic layer was dried over Na 2 SO 4 and evaporated in a vacuum. The crude product was purified by silica gel column chromatography (petroleum ether: acetone = 100:3) to obtain the product.

General Synthesis Procedure (F) for Esterification
Paclitaxel or docetaxel (0.12 mmol), DMAP (0.36 mmol) and ester (0.36 mmol) were added to dry DCM (10 mL). Then, the reaction was stirred at room temperature for 2 h. The mixture was washed with 2 M HCl solution and extracted with DCM three times. The combined organic layer was dried over Na 2 SO 4 and evaporated in a vacuum. The crude product was purified by thin layer silica gel column chromatography (dichloromethane: methanol = 20:1) to obtain the product.

Liposome Preparation
Blank liposomes were prepared for membrane hydration. HSPC, chol and DSPE-PEG2000 were dissolved using chloroform at a mass ratio of 3:1:0.02 and then spinevaporated at 37 • C. The obtained film was then hydrated with 300 mM Cu-Glu (pH 7.4) at 65 • C for 30 min. The multi-layer vesicles were subsequently put into an extruder and passed through polycarbonate membranes with pore sizes of 400 nm, 200 nm and 100 nm 12 times at 65 • C containing 20 mM HEPES, 15 mM EDTA and 300 mM sucrose for removing the un-encapsulated Cu 2+ , and then EDTA was removed with 20 mM HEPES and 300 mM sucrose buffer (pH 7.4, adjusted with triethylamine). Compound 4d was dissolved in DMSO and slowly added into the liposome at 65 • C for 30 min and then quenched in an ice bath for 5 min (the drug to lipid ratio was 1:10). The encapsulation efficiency and drug loading efficiency of compound 4d were determined using HPLC. The morphologies of liposomes were characterized using a transition electron microscope (TEM, Hitachi HT7700, Tokyo, Japan); then, their size, PDI and Zeta potential were measured using a Zetasizer Nano ZS90 (Malvern, UK).

In Vitro Drug Release Study
The in vitro drug release of compound 4d liposome was carried out by dialysis method at 37 • C under a shaking bed (100 rpm). An amount of 1 mL of compound 4d liposome was added into a dialysis bag and suspended in 25 mL PBS (pH 7.4) containing 15% ethanol (v/v) and 10 mmol/L EDTA. The release behaviors for redox response were also accomplished as described above, except for the addition of 0, 0.02 µmol/L, 0.1 µmol/L, 1 µmol/L, 10 µmol/L, 100 µmol/L, 1 mmol/L and 10 mmol/L H 2 O 2 . At the predetermined time points, 200 µL samples were taken for analysis and an equal volume of medium was replenished. The concentrations of paclitaxel were determined by HPLC.

In Vivo Antitumor Efficacy and Toxicity
The antitumor effect was evaluated by the nude mice armpit tumor model bearing breast cancer 4T1 cells. When the average tumor volume reached 100 mm 3 , the treatment was initiated at a 3-day interval for 12 days. The mice were randomly divided into five groups (n = 5 per group) receiving different palictaxel at 8 mg/kg as follows: saline (the control group), taxol, compound 4d, compound 4d liposome, high dose 4d liposome (3 times the dose, immediately 24 mg/kg). Tumor volume (V = length × width 2 /2) and body weight were recorded. At the end of the experiment, mice were sacrificed. The obtained tumors were accurately weighed to calculate the tumor burden (W tumor /W body ) and tumor inhibition rate [TIR= (W control tumor − W sample tumor )/W control tumor ].

Statistical Analysis
Statistics data were presented as mean ± SD and were representative of at least three independent trials. Multiple-group comparisons were performed using a one-way ANOVA followed by Dunnett's t-test. Two group comparisons were performed using Student's t-test. The treatment effect was considered significant at p < 0.05.

Conclusions
By using a molecular combination strategy to link paclitaxel with boronic acid compounds, a series of new paclitaxel derivatives were designed and synthesized. The antitumor activities of these compounds against A549, HCT-116, 4T1 and LO2 cells in vitro were also determined. As a result, compound 4d, one of the most effective compounds, showed considerable selectivity to tumor cells. The optimal formulation of compound 4d liposome, with a drug-lipid ratio of 1:10, was obtained by screening liposomes prescriptions, and the encapsulated efficiency was more than 95%. Meanwhile, compound 4d liposome (8 mg/kg, four times) and a higher dose of compound 4d liposome (24 mg/kg, four times) showed better therapeutic effect than paclitaxel (8 mg/kg, four times) in 4T1 tumor models in vivo, and tumor inhibition rates were 74.3%, 81.9% and 58.5%, respectively. In sum-mary, compound 4d, which connected paclitaxel with boronic acid by using a combination strategy, showed the potential for further research and laid a foundation for exploring the universality of active drug loading of PBA-modified insoluble drugs.