Nintedanib-Cyclodextrin Complex to Improve Bio-Activity and Intestinal Permeability

Abstract
Cyclodextrin complex of nintedanib was prepared aiming for increased bio-activity and improved transport across intestinal membrane with reduced p-glycoprotein (p-gp) efflux. Based on preliminary phase solubility studies and molecular modeling, sulfobutyl ether derivative of β- cyclodextrin (SBE-β-CD, Captisol®) was selected to prepare inclusion complex. Complexation was confirmed using FTIR, 1H NMR, DSC, and XRD. Bioactivity of the formed complex was tested using lung fibroblast cells, WI-38 for anti-proliferative activity and effect on collagen deposition and cells migration. In-vitro permeability studies were performed using epiIntestinal tissue model to assess the effect of complexation on transport and p-gp efflux. Results of the study demonstrated that cyclodextrin complexation increased stability of nintedanib in PBS (pH 7.4) and simulated intestinal fluid (SIF). Further, bioactivity of nintedanib also improved. Interestingly, complexation has increased transport of nintedanib across intestinal membrane and reduced efflux ratio, suggesting the role of cyclodextrin complexation in modulating pgp efflux.

1.Introduction:
Nintedanib, a kinase inhibitor, has been approved by the Food and Drug Administration (FDA) for treatment of idiopathic pulmonary fibrosis (IPF) (Mazzei, Richeldi, & Collard, 2015). It inhibits multiple receptor tyrosine kinases and non-receptor tyrosine kinases (Richeldi et al., 2014; Vaidya, Patel, Muth, & Gupta, 2017), including platelet-derived growth factor receptor (PDGFR), fibroblast growth factor receptor (FGFR), vascular endothelial growth factor receptor (VEGFR) and Fms-like tyrosine kinase-3 (FLT3) (Roth et al., 2015). PDGFR, FGFR and VEGFR have been reported to play a critical role in the pathogenesis of IPF (Chaudhary et al., 2007; Coward, Saini, & Jenkins, 2010; Richeldi et al., 2014). Nintedanib (Vargatef®) has also been approved, incombination with docetaxel, by the European Agency Medicines (EMA) for the treatment of non- small cell lung cancer (NSCLC) (Espinosa Bosch, Asensi Diez, Garcia Agudo, & Clopes Estela, 2016). It is also in clinical trials to be used as single agent or in combination with other chemotherapeutic agents as first-line treatment for a variety of cancers, including breast, ovarian, NSCLC, brain are among some of them (www.clinicaltrials.gov). Currently, nintedanib is available by prescription as oral capsules (Ofev® and Vargatef®) and is recommended to use for twice daily dosing. Oral bioavailability of the 100 mg nintedanib oral capsules was reported to be approximately 5% and is believed to be significantly low due to substantial first-pass metabolism and efflux by transporter pumps (Dallinger et al., 2016).

Nintedanib is known to be a substrate for p-glycoprotein (p-gp) and it was reported clinically that p-gp inhibitors increased the bioavailability of nintedanib when administered simultaneously in healthy subjects (Luedtke, Marzin, Jungnik, von Wangenheim, & Dallinger, 2018).Cyclodextrins (CDs) are cyclic polysaccharides consisting of 6-8 D-glucose monomers linked by α-1,4-glucosidic bonds. The structures of CDs are arranged in such a way that they have hydrophobic cavities with hydrophilic outer surfaces. Due to the presence of hydrophobic cavity, CDs have capability to form inclusion complex with less water soluble compounds, which are retained in the cavity of CDs by different molecular interaction (Davis & Brewster, 2004; Yang, Lin, Chen, & Liu, 2009). Among different CDs, β-CDs are most commonly used because of their low price, easy availability and structural orientation favorable for inclusion complex formation; and are well reported to improve in-vivo stability and bioavailability of small molecules (Lima et al., 2016). Increase in the bioavailability of CD-drug inclusion complex has been reported to occur by different mechanisms including solubility enhancement, increasing stability of the compounds in the intestinal environment, increasing the interaction with cell membrane, decreasing the barrierfunction of lipophilic membrane, by modulating p-gp activity, and combinations of these mechanisms (Loftsson, Jarho, Masson, & Jarvinen, 2005; Masson, Loftsson, Masson, & Stefansson, 1999; Nakanishi, Nadai, Masada, & Miyajima, 1992; Rong et al., 2014; Zhang, Cui, Gao, & Jiang, 2013).

While the mechanisms for p-gp activity modulation by CDs are not clear, results of different studies reveal following point (Zhang et al., 2013). CDs are poor substrates for p-gp because these are hydrophilic, neutral, and high molecular weight compounds with less cell permeability. It has also been suggested that CDs (specially lipophilic dimethyl-β-CD) release p- gp transporters from the apical membrane by depleting cholesterol from the cell membrane and thus reduce the p-gp function of intestinal epithelial cells (Yunomae, Arima, Hirayama, & Uekama, 2003). However, it is to be noted that hydroxypropyl-β-CD and sulfobutyl ether-β-CD show less cholesterol depletion activity compared to methyl- and dimethyl-β-CD and inhibition of P-gp ATPase activity is demonstrated to be a mechanism of p-gp function rather than changing the cell membrane fluidity (Zhang, Meng, Cui, & Song, 2011). Hence, two both HP--CD and SBE--CD are considered to be safe in terms of cytotoxicity/hemotoxicity and nephrotoxicity (Nagase et al., 2003; Rajewski et al., 1995; Wang et al., 2011).It has also been reported earlier that many compounds have shown improved biological activity after encapsulation in β-CDs (Dandawate et al., 2012; Nguyen, Liu, Zhao, Thomas, & Hook, 2013; Pinho, Grootveld, Soares, & Henriques, 2014; Yee et al., 2017). While nintedanib has been approved by the US-FDA for last 4 years, there are no studies reported demonstrating a methodology, either cyclodextrin complexation or any other novel carriers, to overcome known p- gp efflux and in-turn reduced oral bioavailability. In the present study, we hypothesize that by forming inclusion complex with CD, permeability/bioavailability of nintedanib could be improvedby increasing stability in intestinal tract and also by modulating p-gp efflux. Further, bio-activity of nintedanib could also be improved by forming inclusion complexes with cyclodextrins.

2.Materials and Methods
Nintedanib free base (>99%) and nintedanib ethanesulfonate (EHS) salt (>99%) were purchased from LC Laboratories, Woburn, MA. Sulfobutyl ether-β-CD (SBE-β-CD, Captisol®, average mol. wt. 2163, average degree of substitution=6.6) was provided as a gift sample from Cydex Pharmaceutical Inc. KS, USA. Hydroxypropyl-β-CD (HP- β-CD, Cavasol®, W7 HP, average mol. wt. 1410, degree of substitution =4.1-5.1) was purchased from Ashland (produced by Wacker Chemie AG, Burghausen, Germany). HPLC/LCMS grade solvents, and other reagents and chemicals were obtained from Fisher Scientific.Phase solubility studies were done according to previous reported method (Higuchi & Connors, 1965). Briefly, an excess amount of nintedanib base was added to aqueous solution of CDs having different concentrations (0-200 mM). Suspensions were bath sonicated for 30 min and left for 24 hours under continuous stirring for equilibration. After 24 hours, uncomplexed nintedanib was separated by filtering the solution with 0.45 µm polyvinylidene fluoride (PVDF) syringe filter. Filtered solutions were quantified for nintedanib using UPLC analysis. (See Supplementary Materials).Continuous variation method has been used by the researchers for calculation of stoichiometry of the chemical reactions. Here in the present study, we performed Job’s plot analysis to confirm the stoichiometry of nintedanib and CD during complex formation as per earlier reported methods (Upadhye et al., 2010).

Solid nintedanib-CD inclusion complex was prepared by freeze drying method as reported earlier (Zhang et al., 2013). (See Supplementary Materials)Characterization of Solid Nintedanib-CD Inclusion Complex: Inclusion complex was characterized by different techniques to confirm the formation of inclusion complex including Fourier Transform Infrared (FT-IR) Spectroscopy, Proton (1H) Nuclear Magnetic Resonance (NMR) Spectroscopy, Differential Scanning Calorimetry (DSC), and X-ray Diffraction (XRD). (See Supplementary Materials).Molecular modeling studies were carried out on a Dell Precision work station with Intel (R) Xeon(R) CPU E5-1620 v3 @3.50GHz processor. Structure building, docking and analysis were carried out using Accelrys Discovery Studio, GOLD (Genetic Optimization for Ligand Design) suite v5.3 protein ligand docking package and the PyMol molecular graphics software (The PyMOL Molecular Graphics System, Version 1.7.4 Schrödinger, LLC).Structure Preparation: The 3D structure of β-CD co-crystalized with α-amylase (PDB: 1JL8) was downloaded from the Protein Data Bank. β-CD was then extracted out of the α-amylase enzyme using GOLD software (Jones, Willett, & Glen, 1995). Since crystal structures of other CDs were not available, they were built in Accelrys Discovery Studio 4.1 visualization software (Discovery Studio Visualizer 4.1; Accelrys, Inc.: San Diego, CA). Geometry of the added functional groups on the β-CD was optimized using Dreiding like force field. All CDs were saved in mol2 format. 2D structures of all CDs are shown in Supplementary Fig. 1.

There are 21 total hydroxyl groups on CD over seven sugar molecules, including 7 primary hydroxyls on the C-6 and 16 secondary hydroxyls, 7 each on C-2 and C-3 of each sugar molecule.Based on alkyl substitution on the CDs (Bansal et al., 1998) the CDs shown in Supplementary Fig. 1 were used for docking. SBE7-β-CD having seven sulfobutyl groups showed highest drug loading capacity (Vangara et al., 2014). Based on their regional selectivity and reactivity between C-6 and C-2, C-3, 4 isomers were built in the order their probability of alkylation (Isomer 1 through 4). Each sugar molecule is restricted to only one Sulfobutyl group to avoid any steric hindrance. 5 hydroxy propyl groups were placed on CD (HP5CD) as the cavasol® comes with a degree of hydroxypropyl substitution from 4.1 to 5.1.Nintedanib: Nintedanib structure was built in ChemBioDraw Ultra v13.0.2.3021. The energy minimization was done using MM2 force field.Docking: Nintedanib was docked onto various CDs to explore its conformational space within the CD binding pocket using GOLD program. Default GOLD settings were used with 100 GA runs and GOLD score as a scoring function. Binding site was defined using a centroid point for each CD. GOLD score gives information about how good the ligand pose in the binding pocket based on various factors including H-bonding energy, van der waals energy, metal interaction and ligand torsion strain. Best binding pose for each ligand-CD interaction was identified through docking scores and visual inspection considering possible H-bonding interactions.Stability studies of plain drug and drug CD inclusion complex was done in different simulated bodily fluids at different pH conditions including simulated gastric fluid (SGF) (pH 2), simulated intestinal fluid (SIF) (pH 6.5) and phosphate buffered saline (PBS) (pH 7.4) at 37ºC.

Simulated fluids were prepared according to published literature (Marques, Loebenberg, & Almukainzi, 2011). Briefly, plain drug or complex were dissolved in methanol or water, respectively and werediluted with simulated fluids to make final concentration of 100 µg/mL. Drug solution was kept in the incubator at 37°C with continuous stirring and aliquots were withdrawn at different time intervals to measure remaining drug in the solution.EpiIntestinal tissue model (SMI-196, MatTek Corporation, Ashland, MA) was used to study intestinal permeability of nintedanib as per the supplier’s protocol. Briefly, the plate containing epiIntestinal tissue samples was equilibrated overnight with media supplied with package (SMI- 100-MM) in a humidified 37°C, 5% CO2 incubator. Following overnight incubation, media was aspirated from both sides and was replenished with 75 µL of A-side transport buffer (1.98 g/L glucose in 10 mM HEPES, HBSS pH 6.5) to the top (apical) chamber, and 200 µL of B-side transport buffer (1.98 g/L HEPES, HBSS pH 7.4) in bottom (basal) chamber. Plain nintedanib (free base or EHS salt) or nintedanib-CD complex (1 µM equivalent) was added to the top or bottom chamber according to experimental design, i.e., to the top if Apical to basolateral (A to B) or to the bottom if basolateral to apical (B to A). Samples were collected from the respective receiver chambers after 2 hours and amount of nintedanib was measured using LC-MS/MS method (See Supplementary Material for method details). Permeability coefficient was calculated using following equation: