Preparation and Characterization of Inclusion Complexes of N-Substituted-benzenesulfonyl Heterocycles with Cyclodextrins

Abstract

As part of our ongoing drug discovery program we have prepared a library of N-substituted-benzenesulfonyl hetero cycles. Both the heterocyclic and the benzenesulfonyl moiety are considered privileged structures. The insolubility of these compounds has hampered the study of their biological properties. We have prepared and studied the inclusion complexes of representative N-benzenesulfonyl hetero cycles with α-, β-, and γ-cyclodextrins. The complexes have been characterized by spectroscopic (1H-NMR, IR, UV and Raman) and chromatographic methods (HPLC/MS), as well as phase solubility techniques. Molecular modeling techniques have been correlated with the experimental results in order to have a predictive tool to include in the development of novel compounds. The results show that the benzenesulfonyl moiety interacts with the hydrophobic pocket of the cyclodextrin molecule. The phase-solubility diagrams show that complexation with cyclodextrins, in particular β-cyclodextrin can significantly increase the apparent aqueous solubility of the N-benzenesulfonyl heterocycles.

Keywords: Cyclodextrins; Inclusion complexes; Spectroscopy; Phase Solubility Diagrams; Benzenesulfonyl heterocycles

Introduction

As part of our ongoing drug discovery program we have prepared a library of N-benzenesulfonyl derivatives of indoline and 2-methyl indoline (BSH) [1]. From this library we have identified two compounds with moderate antimicrobial activity. The low water solubility of these compounds has hampered the study of their biological properties and the development of SAR/QSAR models that would help in the design of more active antimicrobial compounds. Therefore, it was necessary to enhance the water solubility of these antimicrobial compounds. The goal of the present study was to prepare and characterize cyclodextrin complexes of N-benzenesulfonyl derivatives of indolines (BSHs).

Among the several possible approaches to improve the solubility of drugs, complexation with cyclodextrins is one of the most successful [2,3]. Cyclodextrins (CD) are cyclic oligosaccharides of glucose with a cone-like structure. They consist of (α-1,4)-linked α-D-glucopyranose unit with a lipophilic central cavity and a hydrophilic exterior surface. The naturally occurring cyclodextrins areα, β, and γ types containing 6, 7, or 8 glucopyranose units, respectively. In aqueous solutions, CDs are able to form inclusion complexes with many ligands by taking up the lipophilic portion of the molecule into the central cavity [4]. These oligosaccharides have been used not only to increase aqueous solubility but they have proven to be useful in increasing stability and bioavailability of drugs [5].

In this work we have studied the formation of complexes of CDs with N-benzenesulfonyl derivatives of in do line and 2-methylindoline by phase-solubility diagrams, NMR (1D and 2D), DSC, X-ray diffraction, Raman and IR spectroscopy and molecular modeling techniques.

Materials and Methods

The series of substituted-benzenesulfonyl derivatives of indoline and 2-methylindoline (BSH) were previously prepared and characterized [1]. α-, β-, and γ-CDs (Cavamax W6, W7, and W8, respectively) were generously donated by ISP Technologies, 99.99% deuterium oxide was purchased from Across Organics. All water used in the present studies was distilled and deionized using a Milli-Q Synthesis A10 system (Millipore, Billerica, MA, USA). Methanol was HPLC grade and purchased from Fisher Scientific.

Preparation of inclusion complexes

Inclusion complexes were prepared by dispersing 3mL of β-CD (0.1mol) in water into 3ml of BSH (0.1mmol) in acetone. The mixtures were sonicated for 1 hour at room temperature and stirred overnight. The samples were filtered and the resulting aqueous solution was frozen and lyophilized.

Phase solubility studies

The phase solubility diagrams (PSD) were constructed according to the method proposed by Higuchi and Connors [6]. A series of α-, β-, and γ-CDs solutions were prepared with increasing concentrations. Three concentrations of α-CD were prepared (5, 10 and 15mM), and five concentrations of β-CD (1, 2, 3, 4, and 5mM), and γ-CD (2, 4, 6, 8, 10mM) were prepared. A constant mass of each compound was added to 4mL of each CD solution and was analyzed in triplicates. Samples were sonicated for one hour at 30-50°C followed by overnightmechanical stirring. After this period all suspensions were allowed to stand for at least one hour and then filtered through 0.2μM nylon membranes (Fisher Scientific). This method resulted in reproducible equilibrium concentrations. The supernatants were analyzed by HPLC and quantitated from the ligand peak area (absorbance at 254 nm). The intrinsic water solubility of each compound was determined by HPLC after saturating 4mL of de-ionized water and stirring overnight. The intrinsic solubility was also calculated from the PSD (Table 2). HPLC spectra were recorded on a Shimadzu LC20AT equipped with a SPD M20A diode array detector, and a SIL-20A auto sampler. The column used for the HPLC analysis was a Waters X Bridge column (RP18, 3.5μ, 4.6 x 50mm) and it was eluted at 1 ml/ min with methanol/water (60:40).

Solubility test

A visual observation of the solubility of the inclusion complexes of BSHs and CD was performed. Samples of complexes, BSH, CD, and a physical mixture were prepared in a 1-wt% in distilled water in transparent test tubes.

NMR analysis

1D and 2D-NMR experiments were carried out in order to characterize the complexes. ¹H and ¹³C NMR were first recorded. The complexation induced chemical shift (CIS) was measured for each proton of the CD moiety in each complex. 2D experiments such as COSY, ROESY and HSQC were performed to further characterize the complexes. Acetone was added as a reference for ¹³C-NMR. NMR experiments were performed on a JEOL ECS 400MHz spectrometer, at 400.16(H1) and 100.62(C13). Chemical shift values are reported in ppm (δ) and were taken with D2O as a solvent (referred to residual D₂O at 4.75 ppm for ¹H).

UV spectroscopy

The stability constant of the complex of β-CD with 6b was determined from UV spectra using the method of Rose and Drago [7,8]. Aqueous solutions of 2.50 × 10^-5 M of 6b in the presence of β-CD (from 0.00 to 1.80 × 10^-4 M) were prepared. Spectra were recorded using 1-cm quartz cuvettes with a Cary 50 Bio (Varian). The wavelength range from 255 to 278 nm, which represents 24 unique wavelengths, was used for the determination.

Raman Spectra

Raman spectra were recorded with an excitation wavelength of 785 nm using a Senterra Raman microscope (Bruker). The microscope was used in backscattering mode with a 20× objective lens and a resolution of 3-5 cm^-1. Additional Raman spectra were recorded with an excitation wavelength of 1064 nm using a Multi Ram spectrometer (Bruker). The spectra were recorded in backscattering mode with a defocused excitation beam and a resolution of 1 cm^-1. All samples were analyzed as compressed powders.

Differential scanning calorimetry (DSC)

Thermal analysis was performed on a TA Instruments (New Castle DE) Q100 differential scanning calorimeter calibrated using the recommended indium melting procedure. Samples were accurately weighed on a microbalance before analysis. All analyses were conducted in T-zero® (TA Instruments) aluminum sample pans under dry nitrogen flow at 50 ml/min. The heating rate for all experiments was 10°C/min. The cooling rate during heat-cool-heat experiments was 5°C/min. Data were analyzed using TA Universal Analysis software.

Molecular modeling

SYBYL-X 2.0 was employed in all modeling calculations [9]. The objective was to determine if computational techniques could provide insight regarding the geometry and strength of interaction between b-CD and the BSH library. Specifically, we were interested in exploring the enthalpic contribution responsible for the interactions between b-CD and the BSH library. The entropy was estimated by calculating the integer number of water molecules that would be displaced from both the solvated b-CD molecule and the particular solvated BSH in the optimized interaction geometry.

All BSH molecules were verified for atom and bond assignments, followed by Tripos force field optimization employing Gasteiger- Hückel charges. The dielectric function was set to distance, and the terminal gradient was set to 0.01kcal/mol-Å. The Protein Data Bank was accessed and the b-CD molecule was extracted from the complex with beta-amylase (accession code 1BFN). The molecule was imported into SYBYL and the structure was optimized using the same routines as described above for BSH. The strength of interaction between water and b-CD or BSH was calculated using two routines for the energy calculations: i) the pre-computed box routine where the water was restricted to a monolayer, and ii) the XFIT routine in which small numbers of water molecules were added to specific sites on either b-CD or BSH. The sites were chosen based on the specific areas of interaction that would be involved when b-CD complexed with BSH. The water molecule displacement energies for b-CD:H₂0 and BSH:H₂O were assumed to be equivalent to the opposite sign of the energies of interaction between the same number of water molecules at the specific interaction site [10].

The calculation of the interaction energies between b-CD and BSH was carried out using the Surflex-Dock routine in SYBYL. The calculated energy in the absence of water is summarized by:

$\Delta E_{\mathrm{int}eraction}=E_{\beta -CD:BSH}-(E_{\beta -CD}+E_{BSH}) \left(\text{Equation 1}\right)$

The calculated energy in the presence of water can be summarized by:

$\Delta E_{\mathrm{int}eraction}=E_{hydrated\beta -CD:BSH}-(E_{hydrated\beta -CD}+E_{hydratedBSH})-(E_{\beta -CD:x{H}_{2}O}+E_{BSH:y{H}_{2}O}) \left(\text{Equation 2}\right)$

where x and y are the integer number of water molecules displaced from b-CD and BSH, respectively.

Results and Discussion

Phase solubility diagrams

The corresponding affinity constants, intrinsic solubility and maximum solubility values were calculated from the PSD (Table 2). Compounds 11a,b and 12a,b did not form inclusion complexes. This might be due to the size of the BS moiety, the portion of the molecule that might be included in the hydrophobic cavity. All of the complexes showed AL type solubility diagrams, which indicates the formation of a soluble inclusion complex [6]. These results suggest a 1:1 stoichiometry for the complex. β-CD formed stable complexes with most of the compounds of this series. α-CD complexes exhibited the lowest stability constant as well as the lowest complexation efficiency. g-CD formed complexes with those compounds that have larger substituent’s in the BS moiety, but, as expected, results were erratic. For example g-CD formed a complex with compound 5b but not 5a (Table 2). Similar results were observed with compounds 7a,b, 13a,b and 15a,b.

Stability Constant

The apparent stability constant was calculated by fitting the data to the equation for a 1:1 complex:

$K_{1:1}=\frac{Slope}{S_0(1-Slope)} \left(\text{Equation 3}\right)$

where slope is the value of found in the linear regression and S₀ is the intrinsic solubility of the free drug. Table 1 summarizes the K_1:1 values for each complex and the experimentally determined S₀ was used to calculate K_1:1. As previously stated the β-CD complexes are the ones with the highest K_1:1, while the -CD complexes are the weakest ones.