Nanoformulation design and therapeutic potential of a novel tubulin inhibitor in pancreatic cancer

Abstract

Successful treatment of pancreatic cancer remains a challenge due to desmoplasia, development of chemoresistance, and systemic toxicity. Herein, we synthesized (6-(3-hydroxy-4-methoxylphenyl)pyridin-2-yl) (3,4,5-trimethoxyphenyl)methanone (CH-3–8), a novel microtubule polymerization inhibitor with little susceptible to transporter-mediated chemoresistance. CH-3–8 binding to the colchicine-binding site in tubulin protein was confirmed by tubulin polymerization assay and molecular modeling. CH-3–8 disrupted microtubule dynamics at the nanomolar concentration in MIA PaCa-2 and PANC-1 pancreatic cancer cell lines. CH-3–8 significantly inhibited the proliferation of these cells, induced G2/M cell cycle arrest, and led to apoptosis. CH-3–8 is hydrophobic with an aqueous solubility of 0.97 ± 0.16 μg/mL at pH 7.4. We further conjugated it with dodecanol through diglycolate linker to increase hydrophobicity and thus loading in lipid-based delivery systems. Hence, we encapsulated CH-3–8 lipid conjugate (LDC) into methoxy poly(ethylene glycol)-block-poly(2-methyl-2-carboxyl-propylene carbonate-graft-dodecanol) (mPEG-b-PCC-g-DC) polymeric nanoparticles (NPs) by solvent evaporation, resulting in a mean particle size of 125.6 ± 2.3 nm and drug loading of 10 ± 1.0% (w/w) while the same polymer could only load 1.6 ± 0.4 (w/w) CH-3–8 using the same method. Systemic administration of 6 doses of CH-3–8 and LDC loaded NPs at the dose of 20 mg/kg into orthotopic pancreatic tumor-bearing NSG mice every alternate day resulted in significant tumor regression. Systemic toxicity was negligible, as evidenced by histological evaluations. In conclusion, CH-3–8 LDC loaded NPs have the potential to improve outcomes of pancreatic cancer by overcoming transporter-mediated chemoresistance and reducing systemic toxicity.

1. Introduction

Pancreatic cancer is highly fatal and accounts for nearly 8% of all cancer-related deaths despite only comprising of 3% of all new cancer cases diagnosed in the United States [1]. Surgery is feasible only in the early stage of the disease and is often not curative. Surgery needs to be supplemented with either chemotherapy or radiotherapy [2]. Gemcitabine (GEM) is the FDA approved single agent for the treatment of pancreatic cancer, but its therapeutic efficacy over time is limited due to its rapid metabolism, systemic toxicity and chemoresistance [3,4]. For the late-stage pancreatic cancer, mono- or combination chemotherapy with radiation therapy is the only option. Apart from the combination with GEM, FDA has approved a FOLFIRINOX, which is a four-drug combination (Leucovorin calcium, Fluorouracil, Irinotecan HCl, and Oxaliplatin), but due to its severe side effects, its use is often limited to relatively healthy patients. With the acquired resistance to existing therapeutic agents, developing more effective drugs that can effectively circumvent multi-drug resistance (MDR) will provide significant benefits to pancreatic cancer patients.

Anti-tubulin agents can be classified into two major categories: microtubule stabilizing agents and microtubule destabilizing agents. These agents arrest the cell division and initiate one of the apoptosis pathways [5]. The use of anti-tubulin agents are limited by chemoresistance due to overexpression of transmembrane efflux pumps, including p-glycoprotein (Pgp) [6], multidrug resistance-associated proteins (MRPs) [7], or over expression of class III β-tubulin (TUBB3) which reduces the effectiveness of anti-tubulin agent [8]. In contrast, drugs targeting the colchicine binding site are reported to be less susceptible to clinically observed resistance [5]. The use of colchicine as an anticancer drug, however, has intrinsic limitations including its toxicity, low therapeutic index, and the lack of tumor specificity [9,10]. Therefore, we utilized computer-aided drug design and developed novel colchicine binding site inhibitors (CBSIs) including SMART-OH, LY293, ABI–III, and QW-296. These compounds showed improved anticancer activities, ability to overcome multidrug resistance and limited toxicity [11–16].

Nanoparticulate systems are commonly used for drug delivery to target sites and preventing off-target toxicity due to their enhanced delivery to the tumor site via the enhanced permeability and retention (EPR) effect. Nanoparticles also improve pharmacokinetic profiles by improving aqueous solubility of hydrophobic drugs, extending circulation time, and improving their bioavailability. We successfully applied micellar delivery for 4-substituted methoxybenzoyl-aryl-thiazole-100 (SMART-100) and bicalutamide to improve their aqueous solubility using methoxy poly(ethylene glycol)-block-poly(L-lactide) (mPEG-PLA) polymer [14]. To increase drug loading we utilized methoxy poly (ethylene glycol)-block-poly(carbonate-co-lactide) [mPEG-b-P(CB-co-LA)] based nanoparticles to deliver (2-(1H-indol-5-yl) thiazol-4-yl) 3, 4, 5-trimethoxyphenyl methanone (abbreviated as LY293) to treat melanoma [13].

In this study, we have synthesized a new tubulin polymerization inhibitor 6-(3-hydroxy-4-methoxylphenyl)pyridin-2-yl) (3,4,5-tri-methoxyphenyl)methanone (code name CH-3–8) which is more potent than our previously reported compound, QW-296 [11,16]. To enhance drug loading, we conjugated CH-3–8 to a lipid chain via a cleavable diglycolate linker.. The lipid conjugate of CH-3–8 was loaded into our previously optimized methoxy poly(ethylene glycol)-block-poly(2-methyl-2-carboxyl-propylene carbonate-graft-dodecanol copolymer (mPEG-b-PCC-g-DC) [17]. We evaluated the efficacy of CH-3–8 and its lipid conjugate both in vitro and in vivo using an orthotopic pancreatic cancer mouse model.

2. Materials and methods

2.1. Materials

PANC-1 and MIA PaCa-2 cells were purchased from American Type Culture Collection (ATCC). Dulbecco’s modified eagle’s medium was purchased from Invitrogen (Carlsbad, CA). Heat inactivated fetal bovine serum (FBS), 1% antibiotic-antimycotic mixture and trypsin EDTA were obtained from ThermoFisher Scientific (Waltham, MA). Primary antibodies were obtained from Cell Signaling (Beverly, MA), Santa Cruz Biotechnology (Dallas, TX), and Abcam (Cambridge, MA), while secondary antibodies were obtained from Licor (Lincoln, NE). The chemical supplies for the synthesis, 2,2-bis(hydroxymethyl) propionic acid, 8- diazabicycloundec-7-ene (DBU), benzyl bromide, L-lactide, and all other reagents were purchased from Sigma-Aldrich (St. Louis, MO). Methoxy poly (ethylene glycol) (mPEG, 5,000 Da) was obtained from Alfa Aesar (Tewksbury, MA). Diglycolic anhydride, sodium sulfate, N-(3-dimethyl aminopropyl) -N′-ethylcarbodiimide hydrochloride (EDC·HCl) were purchased from ACROS Organics (Fair Lawn, NJ). Pyridine, 1-dodecanol, N, N′-dicyclohexylcarbodiimide (DCC), 4-di-methylaminopyridine (DMAP) were purchased from Sigma Aldrich (St. Louis, MO). Methylene chloride (DCM), ethyl acetate, n-hexane were obtained from Fisher Scientific (Waltham, MA). Hydrochloric acid (HCl) was purchased from Alfa Aesar (Ward Hill, MA). Thin layer chromatography (TLC) was performed on Merck precoated plates (silica 60 F254, 0.25 mm), and bands were visualized under ultraviolet (UV) light or stained with a solution of phosphomolybdic acid. ¹H NMR spectra were recorded on a Bruker Avance-III HD 500MHz or 400MHz NMR spectrometer, and data were processed by TopSpin 3.5 (Bruker). Chemical shifts are reported in δ value (ppm) relative to tetra-methylsilane as an internal standard. Matrigel matrix basement membrane was procured from Corning (Chicago, IL). All other reagents were purchased from Sigma-Aldrich and used without further purification.

2.2. Synthesis of CH-3–8

CH-3–8 was synthesized as outlined in Fig. 1A. Briefly, compound 1 was reacted with 3,4,5-trimethoxybenzaldehyde in the presence of n-BuLi to give intermediate 2. Subsequently, compound 2 was reacted with Dess–Martin periodinane (DMP) in dichloromethane (DCM) to afford compound 3 as a major product. The Suzuki coupling reaction of compound 3 with 2-methoxy-5-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)phenol under tetrakis(triphenylphosphine)palladium [Pd (PPh₃)₄] condition generated the final compound CH-3–8 as a white solid. The chemical structure of the CH-3–8 was confirmed by ¹H NMR and a Waters Xevo G2-S qTOF high-resolution mass spectrometry (HRMS). The purity of CH-3–8 (≥95%) was verified by analytical high-pressure liquid chromatography (HPLC) using a BEH C18 (2.1 mm × 50 mm, 1.7 μm particle size) column using a mixture of solvent acetonitrile (ACN)/water (with 0.1% formic acid) at a flow rate of 0.3 mL/min. To determine the aqueous solubility of CH-3–8, it was placed in water at pH 7.4 or 5.0 in triplicate. The samples were incubated at 37 °C under constant shaking at 125 rpm for 24 h and centrifuged at 5000 rpm for 10 min. The drug concentrations of the supernatant were determined relative to the peak areas of drug standards (0.2–100μg/mL) by reverse-phase high-performance liquid chromatography (RP-HPLC) at λ_max of 304.1 nm using a C₁₈ column (150 mm × 4.6 mm, 5 μm, Phenomenex, Torrance, CA) and a solvent mixture of acetonitrile: water (70:30 v/v) as a mobile phase.

Fig. 1.

Synthesis and characterization of CH-3–8 by ¹H NMR. A) Synthetic scheme and B) ¹H NMR pectra.

2.3. Molecular modeling

Molecular modeling study of CH-3–8 binding at the colchicine site in tubulin was performed using a recently published X-ray crystal structure (PDB ID: 6AGK) containing a very close analog of CH-3–8 [18]. We used Schrödinger Molecular Modeling Suite 2019 (Schrödinger LLC, New York, NY) for this modeling study, following similar procedures described before [18–20]. Briefly, both CH-3–8 and the native ligand in 6AGK (CH-2–77) were built and prepared using LigPrep. The protein was prepared using the Protein Preparation Wizard workflow, including preprocessing, adding missing chains, optimizing the hydrogen-bonding interactions, and minimizing the complex to refine the structure. The receptor grid in the αβ-tubulin dimer was generated by centering the grid box on the native ligand binding site using the Receptor Grid Generation panel. Prepared compounds were then docked into the binding site of 6AGK with the Ligand Docking module. Hydrogen bonds and data analysis were carried out by using the Maestro interface of Schrödinger software. The docking score obtained attempts to estimate the free energy of binding, and a lower docking score (more negative) indicates more favorable interaction.

2.4. Tubulin polymerization assay

The ability of CH-3–8 to inhibit tubulin polymerization was assessed through a fluorescence-based tubulin polymerization assay kit (Cytoskeleton, Inc. Denver, CO). The manufacturer protocol was followed to perform the assay. Briefly, 5 μL of 10× concentrations of DMSO control, paclitaxel (PTX), colchicine, QW-296, and CH-3–8 were aliquoted into prewarmed (37 °C) half area 96-well plate (Corning Costar, Corning, NY). Tubulin mix containing GTP, glycerol and the tubulin protein (50 μL) was added to the plate, which was immediately placed in the pre-set fluorescence microplate reader (SpectraMax M5e, Molecular Devices, San Jose, CA), and reaction kinetics was measured for one and a half hours with readings every 30 s. The fluorescence excitation was measured at 360 nm, and the emission was measured at 450 nm.

2.5. Immunofluorescence confocal microscopy for microtubule subcellular localization

Immunofluorescence was performed to determine the effect of CH-3–8 on the microtubule subcellular localization. PANC-1 cells were seeded at a density of 500 cells/well on eight-chambered cover glass coverslips and incubated overnight. Cells were treated with DMSO, PTX, colchicine, or CH-3–8 for 24 h, washed twice with ice-cold phosphaste-buffered saline (PBS), fixed with 4% formaldehyde solution overnight at 4 °C and washed with PBS to remove formaldehyde. Cells were then incubated in a 1:1 mixture of PBS containing 2.0% BSA and 0.1% Triton X-100 for an hour at room temperature. Cells were washed thrice with PBS and incubated overnight with α/β-tubulin antibody (Cell Signaling Technology, Danvers, MA) at 4 °C. Cells were then washed thrice with PBS and incubated with secondary antibody (goat anti-rabbit IgG-FITC, Santa Cruz Biotechnology, Dallas, TX) for an hour at room temperature. Cells were again washed (3×) with PBS to re move the secondary antibody and stained with a solution of DAPI (in 0.1% Triton X-100) followed by 10 min incubation at room temperature. Cells were then washed (3×) with PBS, and images were acquired with a Zeiss 710 confocal microscope and Zen imaging software (Zeiss, Thornwood, NY).

2.6. Cell viability assay

The effect of CH-3–8 and the NP formulation on the survival of the immortalized pancreatic cancer cell lines was determined by MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay. PANC-1 and MIA PaCa-2 cells were seeded overnight at 2500 cells/well and treated with different concentrations of the drug and the formulation for 48 h. Cells were then treated with MTT reagent to obtain the final concentration of 0.5 mg/mL and incubated at 37 °C for 4 h. After the plates cooled down to room temperature, they were centrifuged at 2000 rpm for 5 min, and the supernatant was removed. 150 μL of DMSO was added to dissolve the purple colour formazan crystals, and absorbance was measured at 570 nm and 630 nm. Similarly, the IC₅₀ value for CH-3–8 lipid conjugate on MIA PaCa-2 and PANC-1 cells were obtained following the same protocol. To determine the resistance index for PTX and CH-3–8, cell viability assay was also performed on taxane resistant (TXR) cells following the aforementioned protocol. The cell viability assays were performed in triplicate and reported as the mean ± standard deviation (S.D.).

2.7. Spheroid formation

We generated tumor spheroids using MIA PaCa-2 cells. Approximately 500 cells were plated on the ultra-low attachment 96-well plate, and the plate was briefly centrifuged to settle cells at the bottom of the wells. Formation of the 3D structure was monitored everyday over a period of 7 days. On the 7th day, spheroids were treated with 6 nM and 9 nM of CH-3–8 and the equivalent DMSO. The spheroids were then monitored everyday till the 7th day on which Live/Dead assay (ThermoFisher Scientific) was performed to stain the cells according to the manufacturer’s protocol. The spheroids were then observed under the fluorescent microscope (ZEISS, Germany) and imaged.

2.8. Cell cycle analysis

Cell cycle analysis was performed to understand the effect of CH-3–8 on different stages of pancreatic cancer cell growth. PANC-1 and MIA PaCa-2 cells were seeded (3 × 10⁵/well) on a 6-well plate overnight and treated with CH-3–8 for 24 h. After treatment, cells were harvested, fixed in 70% ice-cold ethanol for 1 h, and washed with PBS. Approximately 1 × 10⁶ cells were then suspended in 0.5 mL of FxCycleTM PI/RNase staining solution for 15 min at room temperature away from light. The cell cycle was measured using a flow cytometer (BD FACS Calibur, NJ). Cell cycle analysis was performed in triplicate and reported as mean ± S.D.

2.9. Apoptosis assay

MIA PaCa-2 and PANC-1 cells were seeded (3 × 10⁵/well) on 6-well plates overnight and were treated with CH-3–8 for 24 h. After treatment, cells were harvested, washed with cell staining buffer, and then resuspended in 100 μL Annexin V binding buffer. The APC Annexin V and propidium iodide (PI) solutions (BioLegend, San Diego, CA) were added to the cell suspension, gently vortexed and incubated for 15 min at room temperature in dark. 300 μL of Annexin V binding buffer was added and analyzed by flow cytometer (Ex-488, Em-610). Each experiment was performed in triplicate, and data reported as mean ± S.D.

2.10. Western blot analysis

PANC-1 and MIA PaCa-2 cells were cultured overnight in a 6-well plate at the density of 2 × 10⁵ and 2.5 × 10⁵ cells per well, respectively. After treating with DMSO, CH-3–8 or colchicine for 48 h, cells were washed twice with cold PBS and lysed with RIPA buffer. Protein concentrations were measured with bicinchoninic acid (BCA) protein assay kit (Pierce, Rockford, IL). The lysate was boiled for 5 min, subjected to a 4–15% SDS-PAGE, and transferred to a PVDF membrane using the iBlot™ system (Invitrogen, Carlsbad, CA). Membranes were blocked with blocking solution (Li-COR, Lincoln, NE) at room temperature for an hour and then incubated with primary antibodies at 4 °C overnight, followed by incubation with anti-mouse, anti-goat or anti-rabbit IRDye 800CW secondary antibodies (Li-COR, Lincoln, NE) for an hour at room temperature in dark conditions. β-actin was used as a control, and the signal of target proteins was detected using Li-COR Odyssey® infrared imaging system (Li-COR, Lincoln, NE).

2.11. Cell migration studies

Scratch assay was performed to determine the migration ability of cells in the presence of CH-3–8. Briefly, PANC-1 and MIA PaCa-2 cells were seeded at 3 × 10⁵ cells per well in a 24-well plate and incubated till they reached 100% confluence. A scratch was made on the cell monolayer by dragging a sterile 200 μL pipette tip through the center of the well. The process removed the cell monolayer from the area. The scratched cells were removed by washing twice with the media. The remaining cells in the well were then treated with the media containing 6 nM colchicine, 4 or 6 nM CH-3–8, or the equivalent DMSO. Images of the scratch were obtained immediately after washing and at 6 and 20 h of incubation at 37 °C using the brightfield microscope at 10× magnification (Zeiss, Germany).

Boyden chamber assay was performed to determine the migratory behavior of cells after treatment. 3 × 10⁵ cells/well were seeded in 6-well plates, and plates were incubated overnight. Cells were then treated with 4 or 6 nM of CH-3–8, 6 nM of colchicine, or the equivalent DMSO for 24 h. Cells were harvested, counted, and seeded in equal number in the chamber with 100 μL of serum-free DMEM media. Underneath the chamber, 600 μL of DMEM media with 20% FBS was added. The added volume just touched the membrane on the chamber. The plates with the Boyden chambers were then incubated for 20 h. Next, the cells on the upper chamber were removed by washing and gently scrubbing with cotton tip applicator. Care was taken not to disturb cells on the lower side of the membrane. The chambers were then dipped on to 70% alcohol solution for fixing cells followed by air drying for 20 min. The dry chambers were then placed on a 0.2% solution of crystal violet for staining and then washed multiple times to remove the stain. The membranes on the chamber were imaged under the inverted brightfield microscope.

2.12. Colony formation studies

The effect of CH-3–8 on the tumorigenic potential of PANC-1 and MIA PaCa-2 cells was determined by clonogenic assay. The assay was carried out by treating 1 × 10³ of these cells with 4 or 6 nM CH-3–8, 6 nM colchicine, or the equivalent DMSO for 7 days. The colonies were washed with PBS, fixed with 10% v/v formaldehyde solution, washed again with PBS, and stained with crystal violet (0.2% w/v) in 10% ethanol. The wells were then imaged, 10% acetic acid solution was added into the wells to dissolve the crystal violet, and absorbance was measured at 590 nm using Epoch plate reader (BioTek, Winooski, VT, USA) [21].

2.13. Pgp ATPase activity assay

Pgp-Glo assay system (Promega, Madison, WI) was used to evaluate the effect of CH-3–8 on Pgp ATPase activity according to the manufacturer’s protocol. Briefly, CH-3–8 was incubated at the dose of 10, 100, or 1000 nM with 25 μg of recombinant human Pgp membrane in a white untreated 96-well plate. Pgp-GIo assay buffer was used as the untreated control, 200 μM verapamil was used as the positive control of drug-induced Pgp ATPase activity, and 100 μM sodium orthovanadate was used as the selective inhibitor of Pgp ATPase activity. 5 mM Mg ATP was added to the mixture to initiate ATPase activity and was incubated at 37 °C for 40 min. Further, 50 μL of ATP detection reagent was added to initiate luminescence, followed by plate scanning after 20 min of incubation at room temperature. The luminescence of each well was measured on SpectraMax® i3x microplate reader (Molecular Device, San Jose, CA).

2.14. Synthesis of CH-3–8 lipid conjugate

A lipid conjugate of CH-3–8 (LDC), was synthesized in the following two steps:

2.14.1. Synthesis of diglycolate dodecanol

1-Dodecanol (0.26 g; 1.39 mmol) was dissolved in 3 mL of anhydrous pyridine and added dropwise 3 equivalents of diglycolic anhydride (0.49 g; 4.19 mmol). The reaction was continued for 24 h under mild agitation at room temperature, and the conversion rate was monitored by TLC. The reaction mixture was concentrated in vacuum, and the resulting product was diluted by dichloromethane (50mL) and washed with an aqueous solution of hydrochloric acid (0.05%; 4 × 50 mL), deionized water (50 mL) and brine (50 mL). The organic phase was dried over anhydrous sodium sulfate and filtered. The solvent was then removed under vacuum. The final product was dried in high vacuum and used in the subsequent steps without further purification. The yield was approximately 98.6%. The chemical structure of the compound was confirmed by ¹H NMR (CDCl₃; 500 MHz).

2.14.2. Synthesis of CH-3–8 lipid conjugate (LDC)

CH-3–8 (140 mg, 0.355 mmol), diglycolate dodecanol (213 mg, 0.71 mmol) and DMAP (43.4 mg, 0.355 mmol) were dissolved in anhydrous methylene chloride (40 mL). Then EDC·HCl (204 mg, 1.06 mmol) was added, and the reaction was continued for 24 h at room temperature. The reaction progress was monitored by thin layer chromatography (TLC). After completion, the reaction mixture was diluted with methylene chloride (50 mL), washed with deionized water (4 × 50 mL), dried over anhydrous sodium sulfate, filtered, and concentrated in vacuum. The product was purified by column chromatography using n-hexane/ethyl acetate (5/1) as eluent. Collected fractions containing the product were concentrated in vacuum and dried. The final practical yield of the conjugate was 70%. The chemical structure of conjugate was confirmed by ¹H NMR (CDCl₃; 500 MHz).

2.15. Partition coefficient, log P, solubility of LDC and serum stability

The partition coefficients of free and lipid conjugated CH-3–8 were determined by adding the compounds to a 1:1 mixture of water and 1-octanol and shaking for 48h at room temperature. Samples were centrifuged at 30,000 rpm for 30min to an insoluble pellet drug. Samples from 1-octanol and water were diluted with ACN, vortexed, and analyzed for drug content by HPLC. For the serum stability study, LDC was incubated at 37 °C with 50% fetal bovine serum (FBS). Aliquots of 50 μL were taken at various time points and immediately mixed with ACN. The mixture was then vortexed, centrifuged at 10,000 rpm for 10 min and analyzed for LDC cleavage by HPLC.

2.16. Synthesis of mPEG-b-PCC-g-DC

mPEG-b-PCC-g-DC polymer was synthesized as reported in our earlier publications [14,22]. Briefly, benzyl-2,2-bis(methylol)propionate was synthesized by reacting 2,2-bis(hydroxymethyl)propionic acid with benzyl bromide at 100 °C for 15 h. The product was reacted with triphosgene at −70 °C in the presence of pyridine in dry dichloromethane to obtain the monomer 5-methyl-5-benzyloxycarbonyl-1,3-dioxane-2-one (MBC). The monomer MBC was polymerized with mPEG in anhydrous DCM with 8-diazabicycloundec-7-ene (DBU) as a catalyst at room temperature for 3.5 h. The benzyloxy group of mPEG-b-PMBC was hydrolyzed using Pd/C under H₂ environment to give mPEG-b-PCC. Dodecanol lipid was grafted on to the copolymer by EDC conjugation reaction at room temperature for 24 h. The final product poly(ethylene glycol)-block-poly(2-methyl-2-carboxyl-propylese carbonate-graft-dodecanol) mPEG-PCD was characterized by ¹H NMR (Bruker 500, MA) and gel permeation chromatography (GPC)(Shimadzu, Japan) for the molecular weight. GPC system was equipped with a Styragel HR 4E GPC column and a differential refractive index detector with 0.1% LiBr containing dimethylformamide (DMF) as a solvent system running at 0.6 mL/min at 40 °C. PEG standards (3770–71,000 g/mol) from American Polymer Standards Corp. were used for generating a standard curve, and data were processed using Lab Solutions software version 5.84.

2.17. Preparation and characterization of nanoparticles

Nanoparticles (NPs) were prepared by slight modification in nano-precipitation method as described in our previous report [23]. The polymer (4 mg) and the LDC (400 μg) were dissolved in 100 μL of acetone (common solvent). The mixture of organic solvents was added to the PBS (2 mL) in a 5 mL glass vial dropwise under constant stirring (800 rpm). The mixture was stirred overnight with the lid of the vial kept open for the organic solvent to evaporate. Residual organic solvents were removed by using a rotary evaporator. The formulation was centrifuged at 5000 rpm for 10 min to remove unencapsulated LDC. The hydrodynamic particle size and zeta potential of the particles were measured by dynamic light scattering (DLS) using a Zetasizer Nano ZS (Malvern PANalytical, Westborough, MA) without diluting the formulation. Particle size and surface morphology of nanoparticles were further determined by Atomic Force Microscopy (AFM). Briefly, PBS diluted nanoparticle samples were deposited on freshly cleaved mica, rinsed with de-ionized (DI) water, and dried with a gentle flow of argon. Images were collected with the MultiMode Nanoscope IV system (Bruker Instruments, Santa Barbara, CA) in Tapping Mode at ambient conditions. Silicon probes RTESPA-300 (Bruker Nano Inc., CA) with a resonance frequency of ~300 kHz and a spring constant of ~40 N/m were used for imaging at scanning rate for about 2.0 Hz. FemtoScan software package (Advanced Technologies Center, Moscow, Russia) was used to process the images.

The encapsulation efficiency of CH-3–8 lipid conjugate was estimated by reverse-phase high-performance liquid chromatography (RP- HPLC) at λ_max of 286.3 nm using a C₁₈ column (150 mm × 4.6 mm, 5 μm, Phenomenex, Torrance, CA) and a solvent mixture of acetonitrile: water (99:1 v/v) as a mobile phase. For LDC content determination, 500 μL of NPs were mixed with 500 μL of acetonitrile, vortexed for 5 min, and centrifuged at 14,000 rpm at room temperature for 5 min, and the supernatant was injected into the HPLC.

To determine the drug release profile, 1 mL formulation was taken in the Float-A-Lyzer® and dialyzed against 20 mL of PBS at pH 7.4 and 5.0. The dialysis temperature was maintained at 37 °C, and the setup was shaken at 120 rpm for 48 h. 50 μL aliquot was drawn at pre-determined time points from the Float-A-Lyzer®, and the release media was replenished with PBS. The samples were then centrifuged at 5000 rpm for 10 min. 25 μL supernatant was diluted with 100 μL acetonitrile, vortexed, centrifuged again and analyzed by HPLC. The drug content in all these experiments was determined relative to the peak areas of drug standards (0.2–200μg/mL) in acetonitrile: Water (1:1). The study was performed with three different formulations and reported as mean ± S.D.

2.18. In vivo studies

All animal experiments followed the protocol approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Nebraska Medical Center (UNMC), Omaha, NE. The orthotopic pancreatic cancer mouse model was established in NSG mice (10 weeks old). We used MIA PaCa-2 cells expressing GFP and luciferase (1 × 10⁶) in Matrigel (BD Biosciences, San Diego, CA) and injected a total volume of 30 μL into the tail of the pancreas. Based on bioluminescence signal measured by IVIS instrument (1 × 10⁷), mice were randomly divided into the following treatment groups: a) control (CTR), b) free drug (FD), c) lipid drug conjugate (LDC), and d) nanoparticles of lipid drug conjugates (NP). FD, LDC and NP were injected into these tumor-bearing mice intravenously at the dose of 20 mg/kg. The FD co-solvent formulation was prepared by dissolving the drug in 3:5:2 ratio of cremophor EL, propylene glycol, and 70% ethanol and then adding 5% dextrose. The LDC group was treated with drug conjugate in olive oil emulsion, because it was not soluble in the above co-solvent formulation. The emulsion was prepared by adding water for injection into the mixture of emulsifying agents (polysorbate 80 and propylene glycol in 1:1 v/v) and olive oil containing the lipid conjugate. The mixture was vortexed for one hour to obtain the emulsion. A similar system has been used to improve the dissolution rate of hydrophobic molecule for in vivo study [24]. The control group was treated with PBS. Six doses were administered to all the groups every alternate days. Bioluminescence of the tumor was measured every 3 days using the IVIS® Spectrum imaging system (PerkinElmer Inc., MA). At the end of the study, mice were sacrificed, tumors and other major organs such as liver, spleen, kidney, and heart were harvested. Three representative tumor tissues were collected per group and fixed with 10% buffered formalin for 24 h. The fixed samples were embedded in paraffin, thin sections were obtained and immunostained for hematoxylin and eosin (H&E), Ki67 and cleaved Caspase-3.

2.19. Statistical analysis

Results were presented as the mean ± S.D. A two-tailed Student’s t- test was used to compare results between groups. p < 0.05 was considered statistically significant.

3. Results

3.1. Synthesis and characterization of CH-3–8

The chemical structure and purity of CH-3–8 were confirmed by NMR and HRMS. Synthetic scheme and proton NMR of CH-3–8 are shown in Fig. 1A and B. ¹H NMR (400 MHz, Chloroform‑d) δ 8.01–7.85 (m, 3H), 7.72 (d, J = 2.2 Hz, 1H), 7.68 (s, 2H), 7.63 (dd, J = 8.4, 2.2 Hz, 1H), 6.94 (d, J = 8.5 Hz, 1H), 5.67 (s, 1H), 3.98 (s, 3H), 3.95 (s, 3H), 3.91 (s, 6H). ¹³C NMR (101 MHz, Chloroform‑d) δ 191.66, 155.15, 154.85, 152.58, 147.83, 145.95, 142.42, 137.84, 131.86, 131.29, 122.59, 121.74, 118.89, 112.95, 110.62, 109.08, 61.00, 56.25, 56.03. HRMS [C₂₂H₂₂NO₆⁺]: calculated 396.1447, found 396.1462. HPLC purity 95.8% (t_R = 3.57 min). We determined that CH-3–8 is hydrophobic with aqueous solubility of 1.14 ± 0.06 μg/mL and 0.97 ± 0.16 μg/mL at pH 5.0 and 7.4, respectively at 37 °C.

3.2. CH-3–8 inhibits tubulin polymerization by binding to the colchicine site in tubulin

CH-3–8 and CH-2–77, a previously published native ligand, were docked into 6AGK to confirm that CH-3–8 maintains its mode of action as a colchicine binding site inhibitor [18]. As anticipated, CH-3–8 (gold ball-and-stick model) overlapped with CH-2–77 (light grey model) very well (Fig. 2A) at the colchicine binding site in tubulin, forming two strong hydrogen bonds (blue dashed lines): one between the -OH moiety in CH-3–8 and Thr179 in α-tubulin monomer, the other between the carbonyl of CH-3–8 and Asp249 in β-tubulin monomer. The trimethoxy moiety fit tightly in the hydrophobic pocket (Fig. 2A) deep in the β-monomer, formed by helix-7 (H7), loop 7 (T7), helix 8 (H8), and sheet 9 (S9). The glide docking score of CH-3–8 (−9.5) was comparable with that of the native ligand CH-2–77 (−10.0), indicating they had similar tubulin-binding affinities.

Fig. 2.

CH-3–8 effectively inhibits microtubule polymerization. A) Proposed binding poses of CH-2–77 and CH-3–8 in the tubulin crystal structure (PDB code: 6AGK). Superposition of CH-2–77 (light grey tube model; glide docking score − 10.0) with CH-3–8 (gold ball-and-stick model; glide docking score − 9.5). B) CH-3–8 inhibited tubulin polymerization in vitro as efficiently as colchicine but much more efficiently than QW-296, while PTX increased tubulin polymerization. C) Confocal microscopic images of PANC-1 cells at 24 h post-incubation with DMSO, 8 nM colchicine, 8 nM PTX and 8 nM CH-3–8 show reduced microtubule network scattered in the cytoplasm of the cells treated with colchicine and CH-3–8, but no disorganized microtubule network observed in the cells treated with DMSO. In contrast, PTX treated cells show enhanced microtubule polymerization and thick microtubule bundle.

The inhibition efficacy of tubulin polymerization by CH-3–8 was comparable to colchicine, but much higher than QW-296. Colchicine and CH-3–8 treated cells demonstrated reduced and fragmented microtubule network scattered in the cytoplasm, indicating the destabilization of microtubules. On the other hand, the microtubule-stabilizing agent PTX showed an enhanced tubulin polymerization effect and formed thick microtubule bundles encircling the nucleus (Fig. 2B). We also determined the effect of CH-3–8 on microtubule networks by confocal microscopy of PANC-1 cells at 24 h post-incubation with colchicine, PTX or CH-3–8. There was a big difference in the organization of microtubules among the cells from different treatment groups (Fig. 2C). The effect of CH-3–8 on tubulin polymerization, even at a lower concentration of 4 nM was similar to that of colchicine. There was no disorganized microtubule network observed in the cells treated with DMSO. These results were consistent with tubulin polymerization assay, and CH-3–8 is more effective than colchicine.

3.3. CH-3–8 inhibits pancreatic cell proliferation and tumor spheroids

Cytotoxicity of CH-3–8 was determined by incubating MIA PaCa-2 and PANC-1 cells at different drug concentrations. CH-3–8 resulted in a dose-dependent cell killing, with IC₅₀ values of 5.7 ± 0.5 and 6.1 ± 0.45 nM for PANC-1 and MIA PaCa-2 cells, respectively (Fig. 3A and B). In comparison, QW-296, another tubulin inhibitor, had IC₅₀ values of 19.9 nM and 149.4 nM for PANC-1 and MIA PaCa-2 cells, respectively. Similarly, PTX had an IC₅₀ value of 116.2 nM and 28.1 nM in PANC-1 and MIA PaCa-2 cells, respectively.

Fig. 3.

CH-3–8 effectively inhibits pancreatic cancer proliferation and tumor spheroids. A) CH-3–8 increases cell killing efficiently than QW-296 and PTX, as determined by MTT assay of MIA PaCa-2 and PANC-1 pancreatic cell lines. B) Determination of resistance index and possible mechanism to overcome chemoresistance by CH-3–8. C) Effect of CH-3–8 on Pgp ATPase activity. Change in luminescence (ΔRLU) compared to 100 μM Na₃VO₄ treated samples was plotted (mean ± S.D., n = 3). CH-3–8 at three concentrations showed a non-significant effect compared to NT control on Pgp ATPase activity. *, p < 0.05. D) Fluorescence and bright-field microscope images of the tumor spheroids generated with MIA PaCa-2 cells after 7 days of treatment with CH-3–8 cells compared to vehicle control (DMSO). Tumor spheroids were grown for 7 days before the treatment.

Taxane derivatives and vinblastine are known to develop chemoresistance in cancer cells [25]. Since CH-3–8 belongs to the same family of compounds as QW-296, we hypothesized that CH-3–8 would retain its ability to circumvent theTXR mechanisms, as we have shown in our previous studies [11,16]. So, we treated CH-3–8 against TXR cell line, PANC-1-TXR cells. CH-3–8 was found to be potent in both PANC-1 and PANC-1-TXR cells with a resistance index of only 5.26, whereas PTX showed significantly low potency (IC₅₀ ≈ 1703 ± 467 nM) in PANC-1-TXR cells with a resistance index of 14.68 (Fig. 3B).

Change in the luminescence of CH-3–8 treated samples and the controls were plotted to illustrate the stimulation or inhibition of Pgp ATPase activity (Fig. 3C). Verapamil, a well-known Pgp substrate, stimulated Pgp ATPase activity, consumed a significant amount of ATP, resulting in a sharp increase in luminescence. There was no statistically significant difference in luminescence changes between the vehicle control-treated group or CH-3–8 treated groups at 10, 100, and 1000 nM concentrations, suggesting that CH-3–8 has no effect on Pgp ATPase activity. These results strongly indicate that CH-3–8, unlike verapamil, is not a substrate of Pgp and has no effect on Pgp efflux. This might partially explain the mechanism of CH-3–8 to overcoming TXR.

3D spheroids of MIA PaCa-2 cells were generated to emulate the 3D nature of tumors in the body. Spheroids treated with 6 or 9 nM of CH-3–8 for 7 days were smaller and had lower growth rate, compared to DMSO treated spheroids (Fig. 3D and Fig. S1). CH-3–8 demonstrated concentration-dependent cell killing with live-dead stain. As expected, DMSO treated control only showed a few dead cells. Of note, although CH-3–8 (6 nM) treated spheroid look smaller than CH-3–8 (9 nM), because of difference in their initial size, however; the rate of spheroids size growth was comparable within the drug treated groups.

3.4. CH-3–8 inhibits cell migration, invasion, and colony formation of pancreatic cancer cells

CH-3–8 inhibited cell migration, invasion, and colony formation after treatment of MIA PaCa-2 and PANC-1 cells with this drug (Fig. 4). CH-3–8 effectively inhibited cell invasion at a dose of 6 nM but failed to completely inhibit cell invasion at 4 nM after treatment of MIA PaCa-2 and PANC-1 cells in 20 h (Fig. 4A). CH-3–8 significantly inhibited PANC-1 cell migration across the Transwell membrane at 6 nM but not at the dose of 4 nM. In contrast, colchicine (6 nM) and DMSO control were not as effective in inhibiting cell migration (Fig. 4B). ImageJ analysis of the chamber membrane also confirmed the migration inhibitory effect of CH-3–8 in dose-dependent manner (Fig. 4C). These results demonstrate that CH-3–8 strongly controls the aberrant cell proliferation and hinders cell migration efficiently in pancreatic cancer. We also estimated the levels of E-cadherin, an epithelial cell marker, in PANC-1 cells after treatment with 4 nM and 6 nM concentrations of CH-3–8. There was an increase in E-cadherin protein expression level in dose-dependent manner (Fig. 4D), indicating that CH-3–8 resists the formation of mesenchymal cells, which are the usual culprits for the metastasis of cancer.

Fig. 4.

CH-3–8 inhibits cell migration, invasion and colony formation of pancreatic cancer cells. A) MIA PaCa-2 and PANC-1 cells were plated to confluency, scratched, and treated with CH-3–8 (4 nM or 6 nM) and colchicine (6 nM) for 20 h. Images were taken at 0 and 20 h. B) Transwell membrane assay for cell migration. MIA PaCa-2 and PANC-1 cells were cultured in the upper chambers of the Transwell membrane, treated with CH-3–8 (4 nM or 6 nM) or colchicine (6 nM), incubated for 20 h and visualized under a microscope. C) Quantification of the % area covered by migrated cells using ImageJ analysis software. D) Effect of CH-3–8 on cell migration and EMT marker, E-cadherin, a dose-dependent increase in the protein concentration. E) Colony formation assay after incubation of PANC-1 and MIA PaCa-2 cells with CH-3–8 at the doses of 4 and 6 nM for a week. F) quantification of colonies by measuring the optical density of crystal violet at 590 nm.

We determined the effect of CH-3–8 at doses of 4 and 6 nM on the tumorigenic potential of PANC-1 and MIA PaCa-2 cells by colony formation assay and compared the results with that of colchicine at a dose of 6 nM. CH-3–8 significantly reduced colony formation at 6 nM, but less effective at 4 nM (Fig. 4E). CH-3–8 at a dose of 6 nM, compared to DMSO, reduced colony formation of PANC-1 and MIA PaCa-2 cells by 84.9% and 79.7%, respectively, whereas the treatment of these cells with 6 nM colchicine reduced colony formation by 3.3% and 11.7%, respectively (Fig. 4F).

3.5. CH-3–8 arrests cell cycle in G2/M phase and induces apoptosis in pancreatic cancer cells

CH-3–8 arrested cells in the G2/M phase when MIA PaCa-2 and PANC-1 cells were incubated with this drug for 24 h as determined by flow cytometry after propidium iodide (PI) staining (Fig. 5A). In the case of MIA PaCa-2 cells, 30% and 70% were arrested in the G2/M phase, and 35% and 16% were in the S phase, respectively (Fig. 5A). After the treatment of PANC-1 cells with 4 nM and 6 nM CH-3–8, 31% and 58% of cells were arrested in the G2/M phase, whereas 41% and 25% were in S phase, respectively (Fig. 5B). As expected, we observed the concentration dependency of G2/M phase arrest after the treatment in both the cell lines.

Fig. 5.

CH-3–8 arrests cell cycle in the G2/M phase and induces dose-dependent apoptosis in pancreatic cancer cells. A and B) Cell cycle arrest analysis of MIA PaCa-2 and PANC-1 pancreatic cancer cell lines after treatment with CH-3–8. C) Effect of CH-3–8 treatment on the induction of apoptosis. D) A dose-dependent increase in cleaved Caspase-3 and decrease in PARP in PANC-1 cells. Data represented as the mean ± S.D. (n = 3).

CH-3–8 induced apoptotic cell death when PANC-1 and MIA PaCa-2 cells were incubated with this drug at the doses of 4 nM and 6 nM for 24 h as determined by Annexin V assay (Fig. 5C). Total apoptotic cell population after treatment of PANC-1 with 4 nM and 6 nM CH-3–8 went up from 14% to 19% and 21%, respectively. In the case of MIA PaCa-2 cells, the numbers changed from 8% to 13% and 20%, respectively, with 4 nM and 6 nM CH-3–8 treatment groups (Fig. 5C). Similarly, one characteristic event of apoptosis is the proteolytic cleavage of poly (ADP-ribose)polymerase-1, which is a nuclear enzyme responsible for the DNA repair, genomic stability, and transcriptional regulation [26]. Caspase-3 is an enzyme, which cleaves the PARP. So, we have established the expression level of these proteins by Western blot analysis in MIA PaCa-2 and PANC-1 cells. After treatment with CH-3–8, the cleaved Caspase-3 expression was upregulated concentration-dependent without affecting the total Caspase-3, followed by inverse down-regulation of PARP1 protein (Fig. 5D). These observations showcase the mechanism of cell death after the treatment with CH-3–8.

3.6. Lipid conjugation of CH-3–8 enhances its hydrophobicity and its loading, with no loss in cell killing

CH-3–8 lipid conjugate was synthesized as illustrated in Fig. 6A by carbodiimide coupling of CH-3–8 with the lipid anchor. The structure of the lipid anchor was confirmed by ¹H NMR. Protons corresponding to diglycolic acid (−CO−CH₂−O−CH₂−CO−) were observed at – δ 4.30–4.35 and 1-dodecanol: (−CH₃) at δ 0.88 and (−CH₂−) at – δ 1.2, δ 1.4, δ 1.6 and δ 4.1. The structure of the LDC confirmed by ¹H NMR showed additional peaks (−O−CH₃) at δ 3.91, (−O−CH₃) at δ 3.99 and aromatic hydrogens at δ 7–8.1 (Fig. 6B).

Fig. 6.

Synthesis and characterization of CH-3–8 lipid conjugate (LDC). A) Scheme of dodecanol conjugation to CH-3–8. B) ¹H NMR spectra of CH-3–8 lipid conjugate (LDC). C) Comparative solubility of LDC in 1-octanol and water at room temperature for 48 h. The overlay table shows the apparent log P values of CH-3–8 and LDC. D) Increase in free CH-3–8 concentration and decrease in CH-3–8 lipid conjugate (LDC) concentration with time when CH-3–8 lipid conjugate was incubated at 37 °C in 50% FBS. E and F) Comparable cell killing effect of CH-3–8 and its lipid conjugate after incubation with MIA PaCa-2 and PANC-1 pancreatic cell lines. Results are expressed as the mean ± S.D. (n = 3).

Fig. S2 shows the synthetic scheme for the mPEG-b-PCC-g-DC polymer. After the polymerization reaction, ¹H NMR showed a characteristic peak of phenyl ring at δ 7.3. After hydrogenation, the peak at δ 7.3 disappeared, and a peak appeared instead at δ 13 corresponding to the exposed carboxyl group, indicating complete removal of the benzyl group. The ungrafted polymer, mPEG-PCC, also showed copolymer backbone peaks corresponding to PEG(−CH₂ − CH₂ − O) at δ 3.63, and PCC (−CH₂−) at δ 4.2. Based on the peak integrals of mPEG and PCC protons, a number-average molecular weight (Mn) of the mPEG–PCC copolymer was calculated to be 11,400 g/mol with 40 PCC units. EDC/HOBt coupling reaction was used to conjugate dodecanol (DC) chains to the copolymer. ¹H NMR showed peaks for DC (−CH₃) at δ 0.88 and (−CH₂−) at – δ 1.2 (Fig. S3). Approximately 17 units of DC were present in the final copolymer with the disappearance of the COOH peak, and Mn calculated was approximately 11,000 g/mol. This was also confirmed by GPC with the polydispersity index (PDI), the ratio of weight average molecular weight (Mw) and number average molecular weight (Mn), value of 1.07 indicating a narrow molecular weight distribution of the polymer (Fig. S4).

CH-3–8 lipid conjugate (LDC) was found to be more hydrophobic compared to CH-3–8, as determined by its solubility in octanol and water to calculate the log P value, which increased from 2.54 for CH-3–8 to 3.21 for LDC (Fig. 6C). Conjugation of CH-3–8 to dodecanol provided sustained drug release from the lipid conjugate when incubated in 50% fetal bovine serum (FBS), as evidenced by an increase in % of CH-3–8 in 50% FBS with time due to its release from the lipid conjugate (Fig. 6D). LDC was as effective as CH-3–8 in cell killing when MIA PaCa-2 and PANC-1 cells at different concentrations (Fig. 6E and F).

We then encapsulated LDC into mPEG-b-PCC-g-DC polymeric nanoparticles. The mean hydrodynamic particle size of lipid drug conjugate (LDC) loaded NPs was 125.6 ± 2.3 nm with PDI of 0.1 as determined by DLS (Fig. S5A) while the blank particles were 102.2 ± 2.7 nm (Fig. S5B). These NPs were of spherical shape and had a mean particle size of 60.0 nm, as determined by the atomic force microscopy (AFM) (Fig. S5C). The particles were neutral with ζ potential of −1.08 ± 0.19 (Fig. S5D). The drug loading, as determined by HPLC, was 10.0 ± 1.0% (w/w) with encapsulation efficiency of 99.02 ± 2.59%. There was a sustained release of CH-3–8 and LDC from these NPs (Fig. S6).

3.7. CH-3–8 lipid conjugate loaded nanoparticles exert superior efficacy against orthotopic pancreatic tumor growth

The preclinical effectiveness of CH-3–8 with or without lipid conjugation was evaluated in the orthotopic pancreatic cancer model generated by injecting MIA PaCa-2 cells stably expressing luciferase into NSG mice [27]. Tumor growth was monitored by IVIS imaging. The mice with the bioluminescent radiance of 1.5 × 10⁷ p/s/cm²/sr were injected intravenously every alternate day, at the equivalent dose of 20 mg/kg. Treatment with LDC loaded NPs resulted in significantly smaller tumor size (p < 0.05) than the CH-3–8 and LDC treated groups (p < 0.001) (Fig. 7A–C). The average weight of the tumors also followed the same trend (Fig. 7D). In addition, there was no significant weight change in the treated mice over the course of the treatment (Fig. 7E).

Fig. 7.

CH-3–8 lipid conjugate loaded nanoparticles exert superior efficacy against orthotopic pancreatic tumor growth. A) In vivo representative bioluminescent images at day 1 and day 13 of treatment. Bioluminescent images of mice from control, FD group, LDC group, and LDC loaded in NP groups were taken five times during the treatment (n = 6), B) Radiance intensity plot of treatment groups measured from day 1 of treatment to the day mice were euthanatized. C) The image of isolated tumors from the 4 different groups at the end of the study. D) Tumor weight in the groups at the end of the study (n = 6, *p < 0.05, **p < 0.01) E) % body weight change of mice during treatment for all the groups. Data represented as the mean ± S.D. (n = 6). F) H&E, proliferation marker Ki67, and apoptosis marker cleaved Caspase-3.

H&E staining of tumor tissues confirmed the chaos of the architecture and higher density of nuclei, while the inhibition of proliferation of tumor cells in the treated groups (Fig. 7F). Compared to the control and free CH-3–8 and its lipid conjugate group, there was significant inhibition of proliferation in the NP treated groups represented by Ki67 staining (middle panel). Further, the presence of cleaved Caspase-3 (marked with black arrows) indicated the induction of apoptosis, and we observed significantly increased apoptotic cells in NP treated group compared to the other groups (lower panel).

4. Discussion

Microtubule targeting agents are among the most commonly used drugs for the treatment of a variety of cancers, including pancreatic cancer. Taxanes (PTX, docetaxel) target the taxane site on β-tubulin, facing the lumen of a microtubule and stabilizing the M-loop of tubulin, which prevents disassembly. Introduction of nab-paclitaxel formulation (Abraxane), which is albumin-bound PTX, provided additional support to the treatment with GEM for an additional 2.2 months of survival in combination treatment in advanced pancreatic cancer patients. Abraxane has the favorable pharmacologic characteristics, which enabled the delivery of a higher dose, increased cell killing and also synergized the effects of GEM by enhancing intratumoral accumulation through several mechanisms, including depletion of the stromal matrix, increasing tumor microvasculature, and inhibiting the catabolism of GEM metabolites [28]. However, patients almost invariably succumb to the disease, and little is known about the mechanisms underlying PTX resistance [29]. Apart from PTX and docetaxel, there are other tubulins stabilizing agents, which have extensively been tested for treating pancreatic cancer with variable degrees of success in combination. However, no randomized phase III trials have been performed to demonstrate survival benefit over GEM alone [30].

One of the main reasons for relapse in the curative pancreatic cancer treatment setting is the development of drug resistance, as the taxane-binding site on microtubules is only present in assembled tubulin. Further, taxanes become ineffective due to the distinct expression profiles of β-tubulin isotypes and high expression of Pgp (ABCB1). Increased β2-, β3- and β4–4 tubulin expression levels contribute to the taxane and vinca resistance but with no effect on the resistance to CBSIs. The CBSIs bind to the colchicine domain of the soluble tubulin dimers and induce their conformational change, thus hinder their incorporate into microtubules. Although their mechanism of action may be different from the taxanes, CBSIs also enhance the vulnerability of cells to spindle assembly checkpoint (SAC) and induce cell death.

We have previously shown that the CBSIs can overcome TXR in prostate cancer. Several molecules such as QW-296, LY293, SMART-100 developed in our laboratory showed efficacy even in PTX resistant cells [11,13,14]. In the present study, we developed a new CBSI, CH-3–8, and examined whether it could overcome taxane-resistant pancreatic cancer cells. CH-3–8 exerts biological effect by inhibiting tubulin assembly and suppressing microtubule depolymerization by binding to tubulin at the colchicine binding site [13–15]. Further, QW-296 is another ABI that was found effective against different cancers [11,16]. In our recent studies, the X-ray crystallographic structure of ABI-231 (PDB ID: 6O61) showed that the binding pocket near the central imidazole ring could accommodate a larger moiety. Therefore, we replaced the central 5-membered imidazole ring with a 6-membered pyridine ring and generated a series of 6-aryl-2-benzoyl-pyridines (ABPs), which were about 5-fold more potent than ABIs, suggesting that the 6-membered pyridine ring can enhance inhibition of tumor progression. Among all the ABP analogs synthesized, analog CH-2–77 showed the most potent antiproliferative activity [18]. In this study, we further modified the structure of CH-2–77 and synthesized a new potent tubulin inhibitor CH-3–8 with 3-hydroxy-4-methoxyl on the phenyl ring. According to molecular modeling, CH-2–77 and CH-3–8 have similar docking scores and should have equivalent binding interaction with the tubulin protein. The mechanism of action study showed that the presence of CH-3–8, QW-296, colchicine, and PTX affect the polymerization of tubulin proteins, wherein CH-3–8 along while QW-296 showed decreased inhibition of tubulin polymerization. These results are consistent with our previous publications on QW-296 [11,16]. In addition, PANC-1 cells treated with PTX presented highly fluorescent polymerized microtubules while those treated with colchicine and CH-3–8 showed depolymerized microtubules scattered throughout the cytoplasm. Tubulin polymerization and immunofluorescence results prove that CH-3–8 binds to the colchicine binding site of tubulin protein resulting in depolymerization of microtubules (Fig. 2B and C).

As can be seen in Fig. 3A and B, CH-3–8 was 3.3 and 24.5 times more potent than QW-296 against PANC-1 and MIA PaCa-2 pancreatic cancer cell lines, respectively. In our earlier publication, we have reported that the cytotoxicity of QW-296, with IC₅₀ of 30 nM, was 10 folds more effective than SMART-OH in melanoma cells [16]. CH-3–8 inhibited the migration of PANC-1 cells, cell mobility, and cellular remodeling by decreasing the rate of wound healing in a concentration-dependent manner (Fig. 4A–B). The compounds which bind to the tubulin protein are known to disturb the ability of the cells to changes the shape according to the changing environment [31]. Here, we evaluated the effect of CH-3–8 treatment on an important biomarker associated with the epithelial form, E-cadherin. Our results show that the treatment of PANC-1 cells with CH-3–8 increased E-cadherin expression compared to the control group (Fig. 4D), suggesting that CH-3–8 has the potential to inhibit the migration and invasion of pancreatic cells.

Tubulin polymerization inhibitors cause mitosis catastrophe by arresting the cells in the G2/M phase. Apoptosis proceeds through different pathways, one such pathway involves cleavage of PARP by protease enzyme called Caspase-3. These protease enzymes undergo autolytic cleavage when stimulated and turn into the active form, cleaved Caspase. So, the cleaved Caspase is a reliable marker for the cells that are dead or dying due to apoptosis [32]. Concentration-dependent upregulation of cleaved Caspase-3 and downregulation of PARP in our experiment (Fig. 5D) points to the fact that CH-3–8 causes cell death through this mechanism.

A lipid conjugate approach has the goal of facilitating the delivery of hydrophobic as well as hydrophilic drugs by countering their solubility, permeability, bioavailability, instability, toxicity, and targeting issues, ultimately increasing the concentration of the active moiety at the target site [33,34]. The NP system employed in this work required the synthesis of CH-3–8 lipid conjugate to increase its hydrophobicity. While CH-3–8 itself is nearly insoluble in water, it possesses sufficient aqueous solubility that it can rapidly partition out of the hydrophobic core of NPs, resulting in low drug loading and rapid release. We selected the lipid conjugate approach with lipid dodecanol to obtain a reduced partition rate from the same NPs composition. The lipid conjugation increased the log P value of the CH-3–8 from 2.5 to 3.2 (Fig. 6C). In case of poor drug loading (<5% w/w of the carrier material), the quantity of the administered drug may be insufficient to reach the required concentration in the tumor or require a too high amount of the polymer, which could lead to toxicity. Conjugation of CH-3–8 to dodecanol changed the encapsulation capacity of the polymer from 1.6% to 10% (w/w) of this drug. Further, in NPs, drug physico-chemical properties largely dictated by the drug release from the carrier rather than degradation of the NP itself. Partitioning rate dependent release of a LDC can be rationalized in terms of the accumulation of NPs in the tumors [35]. NPs showing faster drug release will have rapid clearance from the circulation, thus deliver less drug to the tumor site and may exhibit reduced efficacy. We observed a slow release of LDC from NPs with 75% drug release in 24 h at pH 5.5 and 7.4 (Fig. S5).

The effective cleavage of a lipid drug conjugate is one of the rates limiting factors for the drug activity. The cleavage can be ascribed to different factors, such as the hydrophilic environment between the drug and the lipid and the stability of the conjugating linker [36]. We have used diglycolate linker, which is suitable for rapid cleavage in the low pH environment of the tumor. Therefore, it should induce substantial cytotoxicity in cancer cells. Specifically, the diglycolate linker with 3-oxa moiety increases the susceptibility of 2′-acyl group to hydrolysis relative to the ester-linked lipid. In addition, mPEG-b-PCC-g-DC, offers distinct advantages with PEG corona imparting stealth property and the small size of these micelles can take advantage of the EPR effect to maximize drug delivery to the pancreatic tumor. CH-3–8 lipid drug conjugate (LDC) loaded NPs significantly inhibited tumor growth after systemic administration in orthotopic pancreatic tumor-bearing mice (Fig. 7A and B), with minimum weight loss (Fig. 7E). More importantly, LDC loaded NPs demonstrated a significantly superior antitumor activity compared to free drug and free LDC. One possibility is that the prolonged circulation time enabled NPs to take full advantage of the EPR effect. Thus, more drug amount might have accumulated at the tumor site. The second possibility is that NPs are reported to be able to overcome Pgp mediated MDR by increasing the cellular uptake and retention, inhibition of Pgp, and transient depletion of ATP [37]. At the molecular level, we observed a significant decrease in Ki67 expression but an increase in cleaved Caspase-3 expressions in the combination treatment group (Fig. 7F).

In conclusion, lipid conjugated CH-3–8 loaded NPs have the potential to treat pancreatic cancer after systemic administration into orthotopic pancreatic tumor-bearing mice. Since TXR remains a significant unmet problem, the ability of LDC loaded NPs to address this problem may be promising.