Cholecalciferol complexation with hydroxypropyl-β-cyclodextrin (HPBCD) and its molecular dynamics simulation

Abstract

The focus of the current study is to investigate cholecalciferol (vitamin D3) solubilization by hydroxypropyl-β-cyclodextrin (HPBCD) complexation through experimental and computational studies. Phase solubility diagram of vitamin D3 (completely insoluble in water) has an A_P profile revealing a deviation from a linear regression with HPBCD concentration increase. Differential scanning calorimetry (DSC) is the best tool to confirm complex formation by disappearance of cholecalciferol exothermic peak in cholecalciferol–HPBCD complex thermogram, due to its amorphous state by entering HPBCD inner hydrophobic cavity, similarly validated by Fourier-transform infrared spectroscopy (FTIR) and X-ray diffraction (XRD). A_P solubility diagram profile can be associated with cholecalciferol–HPBCD complex instability in liquid phase requiring spray drying to bring it to a solid dispersion state (always more stable) illustrated by scanning electron microscopy (SEM). Computational studies led to a deeper understanding and clarification, at molecular level, of the interactions within cholecalciferol–HPBCD complex. Thermodynamics and geometry of the complex were investigated by molecular dynamics (MD) simulation.

Introduction

Cholecalciferol (vitamin D3) is an essential vitamins, and plays important roles in maintaining human health; for example, it is one of the primary biological regulators of calcium homeostasis (Norman 2008). Cholecalciferol is inactive in our body. It is converted to its active form 25-hydroxycholecalciferol (calcifediol, 25-OH vitamin D3) in the liver, then converted to 1,25-dihydroxycholecalciferol (calcitriol, 1,25-(OH)₂vitamin D3) in the kidney. Calcitriol is a steroid hormone and functions by interacting with its cognate vitamin D receptor (VDR) (Feldman et al. 2004; Norman 2008). We know that calcitriol and VDR significantly contribute to good bone health (Vitamin D Fact Sheet for Health Professionals 2021). Thus, vitamin D3 is widely used in the prevention and treatment of diseases such as rickets (Stuart et al. 2009; Joint Formulary Committee 2015), osteomalacia, osteoporosis and hypoparathyroidism (HSDB: Cholecalciferol 2022), and Fanconi syndrome (Stuart et al. 2009; Hamilton 2015). In addition to metabolic disorders, recent evidence suggests that vitamin D3 may have a role in colon cancer, prostate cancer, and breast cancer prevention (HSDB: Cholecalciferol 2022). Vitamin D3 is a fat-soluble vitamin which is practically insoluble in water(1.3 × 10⁻⁵ mg/L, 25 °C) (Estimation Program Interface (EPI) Suite 2004), and orally delivered. There is no intramuscular (IM) or intravenous (IV) injection available (Stuart et al. 2009).

Delivery of cholecalciferol is difficult due to its hydrophobicity, current commercial formulations rely on incorporating cholecalciferol oils, emulsions, or micellar formulations, which can limit its therapeutic effectiveness and bioavailability; thus, new formulation strategies to enhance its solubility are needed. There are many techniques available to increase water solubility including surfactants addition (Basalious et al. 2011; Ribeiro et al. 2012), amorphous solid dispersion formulation (Leuner and Dressman 2000; Forster et al. 2001; Nair et al. 2002; Moshe et al. 2019), co-solvents (Seedher and Kanojia 2009), pH control (Tatavarti and Stephen 2006), and cyclodextrin complexation (Holvoet et al. 2005; Loftsson et al. 2005; Brewster and Loftsson 2007; Linda et al. 2014).

β-cyclodextrins (B-CDs) are water-soluble cyclic oligosaccharides consisting of seven glucose subunits. Its macrocyclic ring forms a bucket-like structure having a hydrophobic internal cavity and a hydrophilic exterior. This special structure allows formation of inclusion complexes, where lipophilic or hydrophobic compounds are easily bounded within the cavity leading to aqueous solubility increase (Trapani et al. 2005; Mallick et al. 2008; Yang et al. 2008; Fu et al. 2012; Ozdemir and Erkin 2012). Many applications of B-CDs have also been investigated especially in drug delivery systems (Felton et al. 2002; Godwin et al. 2006; Kear et al. 2008; Wang et al. 2020). In order to improve the aqueous solubility (Loftsson and Brewster 2012) and complexing capability (Tongiani et al. 2009) of native B-CD, side chain substitutions can be used (Malanga et al. 2016). Numerous studies have demonstrated that for a variety of drugs their solubility, permeability, and bioavailability can be enhanced by complexation with β-CD and its derivatives (hydroxypropylated, methylated and sulfobutyl ether) (Brewster and Loftsson 2007; Kratz et al. 2012; Malanga et al. 2016). In some papers, simulation models were used to predict the complexation efficiency of CDs with various drugs (Trapani et al. 2005; Chari et al. 2009; Li et al. 2011).

For these studies, the goal was to improve the aqueous solubility of cholecalciferol. HPBCD is added as a solubilizing excipient. The complexation mechanism between cholecalciferol and HPBCD was evaluated experimental and by computer simulation revealing a rare type A_P complexation behavior (Takeru and Kenneth 1965; Phennapha et al. 2018) through both methods. Cholecalciferol is wildly used in many clinical research studies, and IV administration can lead to a faster collection of biological data compared to other dosage like oral dose forms. Worth mentioning that at present time there is no related literature data on liquid based parenteral cholecalciferol formulation. We developed a new parenteral cholecalciferol formulation based on its complexation with HPBCD (with a molar substitution (MS) of 0.58–0.68). In preliminary studies, the complex liquid formulations data of 6-months stability demonstrated that high concentration HPBCD could preserve cholecalciferol, but could not keep it stable in the liquid formulation for long time storage. To increase cholecalciferol–HPBCD complex stability, spray-dried dispersions (SDDs) were made as dry powder. It would be used as reconstitutable powder for injection using sterile water. Complex formation and its amorphous character in SDDs was validated by DSC and concurrently by FTIR, XRD and illustrated by SEM. Meanwhile, molecular dynamics (MD) simulation evaluation is based on 1:1 complexation interaction between cholecalciferol and HPBCD.

Materials and methods

Materials

Cholecalciferol (vitamin D3, MW: 384.64) (HPLC, assay ≥98%, Sigma, St. Louis, MO) was purchased from Sigma-Aldrich (St. Louis, MO), hydroxypropyl-β-cyclodextrin (HPBCD, M.W.: 1400) (Kleptose^® HPB parenteral grade, Roquette Pharmaceutical Ltd., Lestrem, France) was provided as a gift sample. Sterile Empty Vial Amber 5 mL (Health Care Logistics Inc., Circleville, OH) for stability studies.

Phase solubility studies

The phase solubility approach described by Higuchi and Connors (Takeru and Kenneth 1965) is a well-established method to evaluate complexation. The saturation solubility for cholecalciferol is detected at HPBCD at 10, 20, 40, 50, 70, 100, 150, 170, and 200 mM concentrations in water. Excess amounts of cholecalciferol were added to each aqueous solution of the different HPBCD concentrations. Deionized water without HPBCD was used as the control. The suspensions were sealed in 20 mL amber vials. All the vials were secured on the Titer Table Shaker (LAB-LINE Instruments, Inc., Dubuque, IA) by a holder, covered with aluminum foil to protect from light, and shaken at speed 500 rpm for three days at room temperature. The aqueous solutions were filtered (0.45 μm, nylon syringe filter) and immediately tested via UPLC for API concentration. For studies done at a later date, the samples were stored in amber vials at 4–8 °C; in these vials no signs of precipitation was observed.

The filtered supersaturated solutions were transferred into amber vials for cholecalciferol analysis and stability testing by UPLC (Waters^® AcQuity, Milford, MA). A BEH C18 Column (130 Å, 3.5 μm, 4.6 mm × 100 mm, XBridge (Waters^® Milford, MA)) was used, and mobile phase was 0.08 M phosphoric acid (pH = 5.5):methanol at a ratio of 5:95. After spray drying of these filtered supersaturated solutions, an amorphous white powder was obtained, which was used for FTIR, XRD, DSC, and SEM analysis.

Phase solubility diagram resulted by plotting the total concentration of cholecalciferol versus the concentration of HPBCD. Based upon theory (Takeru and Kenneth 1965), a type A_P diagram (positive deviation from linearity) is assumed to come from two complexes (1:1 complex and 1:2 complex) of cholecalciferol:HPBCD_i. The equation for the combination of these two species in solution is for A_P profile (Takeru and Kenneth 1965).

$$y=S_0+K_{1:1}{S}_{0}x+K_{1:1}{K}_{1:2}{S}_0{x}^2+K_{1:1}{K}_{1:2}{K}_{1:3}{S}_0{x}^3+\cdots$$ (1)

where S₀ is the solubility of cholecalciferol in water, K_1:1 is the stability constant of the cholecalciferol:HPBCD (1:1) complex, K_1:2 is the stability constant of the cholecalciferol:HPBCD₂ (1:2) complex, K_1:3 is the stability constant of the cholecalciferol:HPBCD₃ (1:3) complex, x is the concentration of the relatively soluble ligand (HPBCD), y is the total concentration of cholecalciferol in solution. The stability constants were estimated by regression of total cholecalciferol versus HPBCD (Takeru and Kenneth 1965).

The free energy of complexation is calculated from K_1:1 and K_1:2 using:

$$\Delta G_{1:1}=-\mathit{\text{RTln}}{K}_{1:1}$$ (2)

$$\Delta G_{1:2}=-\mathit{\text{RTln}}{K}_{1:2}$$ (3)

Using the notation of Higuchi and Connors (Takeru and Kenneth 1965), A_P represents a positive deviation type phase-solubility diagram, K is the stability constant, R = 8.314 J mol⁻¹ K⁻¹, and T is the temperature on the Kelvin scale.

Cholecalciferol–HPBCD complex stability

Cholecalciferol–HPBCD complex stability in aqueous solution was evaluated under ICH conditions 25 °C/60% relative humidity (RH) and 40 °C/75% RH for 6 months. The filtered supersaturated solutions were sealed in amble vials by rubber stopper and aluminum crimped cap, and then kept in two different stability chambers. To assess the extent of degradation, after 6 months, all the samples were measured for the concentration of cholecalciferol.

Differential scanning calorimeter (DSC)

The powder samples were analyzed using differential scanning calorimeter TA DSC2500 (TA Instruments, New Castle, DE). Approximately, 5.0 mg of sample was placed in a sealed aluminum pan and heated at a rate 10 °C/min within the temperature range of 30–200 °C, and the nitrogen flow was at 9 PSI pressure that base purge was about 270 mL/min.

Fourier-transform infrared spectrophotometric (FTIR) analysis

The IR spectrum of the solid dispersion samples were measured by Fourier-transform infrared spectrometer (JASCO, FT/IR-6100, Easton, MD). Sample preparation was by KBr disc method and included about 10 mg sample for each test, with the wave number range of 4000–400 cm⁻¹.

Powder X-ray diffraction (XRD)

All powder samples (cholecalciferol, HPBCD, cholecalciferol:HPBCD complexes spray dry powder and cholecalciferol with HPBCD physical mixture) were studied using a Bruker D8 Advance Bragg-Brentano Diffractometer (Bruker AXS, Inc., Madison, WI) with Kbeta filter, using Cu radiation and wavelength WL = 1.5406 Å, with a voltage 40 kV and a current 30 mA, using LynxEye position sensitive detector. There were about 200 mg of each sample in X-ray holder, scanned from 5 to 70° (2θ) with a step of 0.021°, and total exposure of 180 s per step. The results were processed by EVA software (Bruker AXS, Inc., Madison, WI).

Scanning electron microscopy (SEM)

Tablets were mounted on to SEM specimen holders with conductive carbon adhesive tabs (Ted Pella, Inc., Redding, CA) and sputter coated with 10–20 nm of platinum/palladium in a sputter coater EMS 150T ES (Electron Microscopy Sciences, Hatfield, PA). SEM images were taken in a scanning electron microscope Quanta 200 (FEI Co., Hillsboro, OR) under the conditions specified in the images (operating voltage 200 kV).

Spray drying

A BUCHI 290 Mini Spray Dryer was used to spray dry the cholecalciferol:HPBCD complexes. The filtered supersaturated solution (pass the suspension through a syringe tip filter of 0.45-μm pore size) was dispersed to control spray drop size the by spray nozzle, rapidly drying by hot gas, and the inlet temperature was controlled at 105 °C, aspirator was 75%, spray airflow was 473 L/h and flow rate 12%. In collection vessel, a white fine SDDs powder was produced.

Computational methods

Computational models were built to initiate MD simulations of the complexation of HPBCD and cholecalciferol. According to the information from the vendor (Roquette Pharmaceutical Ltd., Lestrem, France), the molar hydroxyl groups′ substitution ratio ranges from 0.58 to 0.68 for HPBCD sample. Thus, on average each molecule of BCD contains four or five substituted hydroxypropyl groups. Considering the symmetry of the molecule, four combinations for 4-substituted and three for 5-substituted BCD are possible. This includes combinations of hydroxyl groups on different sugar units as (1,2,3,4), (1,2,3,5), (1,2,4,5), and (1,2,4,6) for 4-substituted and (1,2,3,4,5), (1,2,3,4,6), and (1,2,3,5,6) for 5-subsititued case as shown in Figure 1. All these HPBCD molecular models were constructed for further consideration.

Figure 1.

2D structure of a BCD molecule with hydroxyl groups labeled, which is substituted by hydroxypropyl groups in HPBCD case. Considering the symmetry, for example, substitutions on (1,2,3,4) will be similar to that on (2,3,4,5), (3,4,5,6), (4,5,6,7), (5,6,7,1), (6,7,1,2), and (7,1,2,3), thus only (1,2,3,4) will be modeled for such substitution case.

To put the cholecalciferol into the HPBCD molecular cavity, Molecular Operating Environment (MOE) software (Molecular Operating Environment (MOE) 2022) was employed to manually dock one cholecalciferol molecule into all seven possible HPBCD models. As previous studies have shown (Lauren et al. 2013), guest molecule can bind to host cyclodextrin molecules in two orientations as up-state and down-state defined by the orientation of polar function groups in the guest molecule. From these docked molecules, the most favorite docking poses of each orientation for each HPBCD model were selected. These models were used to initiate calculations to evaluate the binding orientations and relative affinities using MD simulations.

Each complex model was solvated in a water box with the system extrema separated from the box edge by 12 Å on all sides. Each system was minimized for 50 000 steps with the steepest descent (SD) algorithm (Levitt and Lifson 1969) in the presence of periodic boundary conditions (PBCs) and was followed by a 100 ps MD equilibration under NVT ensemble and another 100 ps equilibration under NPT ensemble. During these two rounds of equilibration, the complex system was restrained using harmonic restraints with a force constant of 1 kcal/mol/Å (Feldman et al. 2004), and all nonhydrogen atoms of cholecalciferol and HPBCD with water were optimized. Additional NPT equilibration was conducted with no restrains to further equilibrate the whole system. The final production MD run was maintained for 20 ns. During MD, the Nosé–Hoover method (Nosé 1984; Hoover 1985) was used to maintain the temperature at 298 K and pressure was maintained at 1 bar using the Parrinello–Rahman barostat (Parrinello and Rahman 1981). CHARMM36 carbohydrate force field (Olgun et al. 2008), CHARMM General Force Field (CGenFF) (Vanommeslaeghe et al. 2010; Wenbo et al. 2012) and modified TIP3P water model (Stewart et al. 1994) were used to describe HPBCD, cholecalciferol, and water during the simulation, respectively. All MD simulations were conducted using the GROMACS program (Berk et al. 2008) and Verlet integrator was used with a 2 fs time step to propagate each system. The SHAKE algorithm (Ryckaert et al. 1977) was applied to constrain bonds to hydrogen atoms to their equilibrium lengths, and long-range electrostatic interactions were treated with the particle mesh Ewald (PME) method (Darden et al. 1993) with a real space cutoff of 12 Å, and a switching function (Steinbach and Brooks 1994) was applied to Lennard-Jones interactions in the range of 10–12 Å.

To study the interaction between cholecalciferol and HPBCD, the GROMACS Molecular Mechanics Poisson–Boltzmann Solvent Accessibility (g_mmpbsa) tool (Kumari et al. 2014) was used to analyze each MD trajectory to estimate relative free energies of binding based on the MM-PBSA approach (Samuel and Uif 2015).

Results and discussion

Phase solubility studies

The total concentration of cholecalciferol was measured in HPBCD solutions of various concentrations. The phase solubility diagram shows an A_P type (1:1 and 1:2 complexation) profile of cholecalciferol–HPBCD complex (see Figure 2). Based on the solubilization profile, we fit Equation (1) and the best-fit equation was found to be:

$$y=0.0016+0.0024x+0.0003x^2$$ (4)

where S₀=0.0016 mM, the stability constants K_1:1=7.19 × 10⁴ mM⁻¹ (7.19 × 10⁷ M⁻¹) and K_1.2=0.125 mM⁻¹ (1.25 × 10² M⁻¹).

Figure 2.

Total vitamin D3 solution concentration (mM) at different HPBCD concentrations (mM) in deionized water (under vitamin D3 saturation conditions and room temperature) shows a type A_P diagram.

The accepted calculated value for cholecalciferol solubility in water is S₀=1.3 × 10⁻⁵ mg/L (3.38 10⁻⁸ mM). While there are literature reports that try and measure cholecalciferol solubility in water, this value is extremely difficult to measure accurately due to the very low solubility being lower than the detection limit of most analytical methods; thus, the typically used accepted value is calculated using EPI Suite^™ (Syracuse Research Corp., Cicero, NY); a screening-level tool that is used if acceptable measured values are not available, and is accepted by the US EPA and PubChem (Estimation Program Interface (EPI) Suite 2004). From Figure 2, our estimated value for S₀ is 0.0016 mM, which is much higher than the accepted literature value. The difference is the result of the actual solubility being so difficult to measure, so we use the theoretic value instead of our calculated intercept value for the stability constant calculation. The corresponding free energies of complexation are ΔG_1:1=−10.68 kcal/mol and ΔG_1:2=−2.85 kcal/mol.

The solubility increase at different HPBCD molarities is presented in Table 1, which shows the solubility increase ratio is enhanced by five or more orders of magnitude. Clearly, the solubility of cholecalciferol in water is greatly enhanced by HPBCD.

Table 1.

Cholecalciferol solubility at different HPBCD concentrations.

HPBCD concentration (mM)	Saturated concentration of vitamin D3 in HPBCD solution (mM) (S)	Solubility increase ratio (S/S0)
0 (control, D.I. H20)	3.38×10−8 (S0) (Hamilton 2015)	1
10	0.02	5.92×105
40	0.48	1.42×107
50	0.78	2.29×107
70	1.54	4.56×107
100	3.14	9.30×107
120	3.75	1.11×108
150	6.41	1.90×108
170	7.81	2.31×108
200	11.12	3.29×108

Cholecalciferol–HPBCD complex stability

The cholecalciferol–HPBCD complex in aqueous solution was analyzed after 6 months of storage under room temperature and accelerated ICH storage conditions. Figure 3 shows the percent degradation ratio (6-month concentration divided by t = 0 concentration) of cholecalciferol versus to HPBCD molarity. From Figure 3, it is evident that as the concentration of HPBCD increased, the stability of cholecalciferol generally increases and percent maintained at 25 °C was higher than at 40 °C. Moreover, the degradation rate at 25 °C was lower than at 40 °C with the same HPBCD molarity. The 6-month stability study shows that higher concentrations of HPBCD reduced the amount of vitamin D3 degradation at 25 °C and 40 °C in a liquid formulation. While these rates of degradation are still too fast for a commercially viable formulation, the results do show the HPBCD can improve vitamin D3 stability.

Figure 3.

The cholecalciferol–HPBCD complex solution 6-month stability evaluation at 25 °C/60% RH and 40 °C/75% RH. The Y-axis represents the concentration ratio of the percent cholecalciferol remaining compared to the original concentration at day 0.

Characterization of complexation

During the phase solubility study, the A_P type of profile was observed. The thermodynamic parameters of this complexation experiment – two stability constants and the free energies of complexation were calculated, and an aqueous formulation with cholecalciferol was successfully produced. However, preliminary stability studies indicated that the cholecalciferol–HPBCD complex degraded during long-term storage. To increase cholecalciferol stability, the spray-dry technical was used to produce SDDs powder of cholecalciferol–HPBCD.

To characterize the solid-state properties of this complex, DSC, FTIR, XRD, and SEM applied in the analysis. These analysis evidences supported the interaction between cholecalciferol and HPBCD after the cholecalciferol–HPBCD complexes formation. In addition, MD simulations on cholecalciferol and HPBCD complex also conducted to evaluate the interactions at the molecular level.

DSC analysis

Thermograms of cholecalciferol, HPBCD, and cholecalciferol–HPBCD complex are shown in Figure 4. DSC is a thermal analysis technique that used to test the heat capacity (Cp) of materials with the temperature change. A crystal material will exhibit a melting or crystallization event associated with an enthalpy change. As the melting temperature (T_m) is reached, there will be an endothermic peak associated with the melting transition and this event shows up in the DSC curve. Amorphous materials do not have an organized solid state. When heating amorphous materials there is a second-order phase change in which the heat capacity abruptly changes, and this transition from the glassy to rubbery state is called the glass transition and is characteristic of amorphous materials. This transition is broad shift in the baseline of the DSC curve, and T_g is the glass transition temperature.

Figure 4.

DSC thermograms of cholecalciferol (vitamin D3), HPBCD and spray dried cholecalciferol–HPBCD complex sample.

In Figure 4, the thermal curve of (red line) exhibited a sharp exothermic peak at 83 °C, corresponding to the melting point of crystalline cholecalciferol. The blue line represented the thermogram of HPBCD with a broad baseline shift peak at about 130 °C T_g, indicating its amorphous nature. The cholecalciferol–HPBCD complex (green line) showed a similar behavior as HPBCD, and a broad peak was observed at 110 °C T_g, which proved that it had amorphous character as well (cholecalciferol being hold inside HPBCD hydrophobic inner cavity). Meanwhile, different T_g of these two compounds demonstrated their different physical structures.

Fourier-transform infrared spectrophotometric analysis

The interaction between cholecalciferol and HPBCD was examined by FTIR spectroscopy. Pure cholecalciferol, pure HPBCD, a physical mixer and a complex are characterized in Figure 5. In the spectrum of cholecalciferol (Figure 5(A)), the characteristic peak at 3306 cm⁻¹ is attributed to the hydroxyl groups corresponding to the O–H stretch. Another absorption band 2935 cm⁻¹ corresponds to C–H aliphatic stretch, band 1648 cm⁻¹ corresponds to conjugated diene C=C stretching vibration, 1457 cm⁻¹ indicates the methylene C–H stretching vibrations, 1165 cm⁻¹ and 1052 cm⁻¹ correspond to the C–H and C–O stretching vibrations, respectively. The spectrum HPBCD (Figure 5(B)) exhibited the characteristic absorption band at 3429 cm⁻¹ for O–H stretching vibration, 2935 cm⁻¹ due to the C–H stretching vibration, 1653 cm⁻¹ to absorbed water contribution, 1156 and 1035 cm⁻¹ corresponding to the C–H and C–O stretching vibrations, respectively (Jing et al. 2011; Brain 2017). The FTIR spectrum of the physical mixture (i.e. not complexed, Figure 5(C)) did not show significant differences in the common spectral peaks present in pure cholecalciferol and HPBCD. In contrast, the spectrum of the cholecalciferol–HPBCD complex (Figure 5(D)) showed different features compared to the pure cholecalciferol, but similar to HPBCD. The bands 1648 cm⁻¹, 1164 cm⁻¹, and 1052 cm⁻¹ are shifted to 1654 cm⁻¹, 1157 cm⁻¹, and 1035 cm⁻¹, the characteristic peak at 3306 cm⁻¹ of cholecalciferol corresponding to the O–H stretch moved to 3387 cm⁻¹ between the positions of cholecalciferol (3306 cm⁻¹) and pure HPBCD (3429 cm⁻¹). This means new bond interaction between the O–H of cholecalciferol and HPBCD during the complex formation. The total hydroxyl groups in the complex are shifted from cholecalciferol and HPBCD. The significant change in peak intensity, position, and shape of the bands indicated interactions between molecules and indicates the presence of the inclusion compound (Xianhong et al. 2004). Therefore, the hydroxyl group’s interaction between cholecalciferol and HPBCD is very important in the complex formation, that it correspondingly increases the solubility of cholecalciferol in water.

Figure 5.

FTIR spectrum of (A) cholecalciferol; (B) HPBCD; (C) cholecalciferol and HPBCD physical mixture (1.6:10 mole ratio); (D) spray dried cholecalciferol–HPBCD complex (1.1:20 number of moles ratio).

The analysis of the IR spectrum is based upon relative peak shifts; the preparation of all the samples in KBr pellets was the same so any effect of the KBr pellet would be constant and thus not affect the relative comparisons. In addition, the use of KBr disc as matrix for all samples did not result in bands appearing or disappearing as evident from the spectra of pure components and their physical mixture. To further confirm this, a spectrum generated by mathematical sum of pure spectra of cholecalciferol and HPBCD coincided with the spectrum of physical mixture of the two components indicating using a KBr disc did not significantly affect the IR spectra or our analysis.

X-ray diffraction analysis

X-ray powder diffraction is a rapid and primary technique for detecting the presence of crystalline structure in a fine powder. Veiga et al. found that for CD complexes, the superposition of each component’s diffraction pattern is different from the X-ray pattern of the complex (Veiga et al. 1996). The XRD patterns are shown in Figure 6; the sharp peaks of cholecalciferol (Figure 6(A)) illustrate the crystalline nature of the compound, with peaks at 2θ of 5.1°, 6.7°, 13.7°, 15.7°, 18.1°, and 21.8°, which coincide with the literature PXRD data of cholecalciferol (Gopal et al. 2019) and identify our material with good crystallinity. The diffraction pattern of HPBCD (Figure 6(B)) displays a broad amorphous halo in the 15–25° (2θ) range, verifying its amorphous character. The XRD pattern of cholecalciferol and HPBCD physical mixture with a molar ratio of 1.6:10 (Figure 6(C)) is the superposition of cholecalciferol and HPBCD X-ray patterns. The sharp peaks at 15.7°, 18.1°, and 21.8° revealed the cholecalciferol crystalline structure in the mixture. For the complex of cholecalciferol with HPBCD, the XRD pattern shows a large and broad peak similar to the amorphous HPBCD (Figure 6(D)), indicting the amorphous character of the spray dried material.

Figure 6.

X-ray powder diffraction pattern of (A) cholecalciferol; (B) HPBCD; (C) cholecalciferol and HPBCD physical mixture (1.6:10 mole ratio); (D) spray dried cholecalciferol–HPBCD complex (1.1:20 mole ratio).

SEM analysis

Images of SEM revealed the external morphology/microstructures of cholecalciferol, HPBCD, physical mixture of vitamin D3 and HPBCD, and a spray dried cholecalciferol–HPBCD complex. As Figure 7 shows, the crystal structure of cholecalciferol had a needle shape (Figure 7(A)). While, from the photo of neat HPBCD (Figure 7(B)), it was typical of materials with an amorphous character (also confirmed by DSC), and no regular structure pattern was observed. In the physical mixture photo (Figure 7(C)), the crystalline cholecalciferol could be found intermixed with HPBCD particles. However, the inclusion complex of cholecalciferol–HPBCD displayed the typical morphology of spray drying amorphous materials and the particle size was less than 10 μm (Figure 7(D)). The comparison of these four pictures indicated that the complex had different microstructure compared to cholecalciferol, HPBCD and their physical mixture, which further supports the presence of the inclusion complex and its amorphous nature.

Figure 7.

SEM micrographs of (A) cholecalciferol (vitamin D3), (B) HPBCD, (C) physical mixture of cholecalciferol and HPBCD, and (D) spray dried cholecalciferol–HPBCD complex particles.

Molecular dynamic simulations of cholecalciferol–HPBCD complex

A previous study (Molecular Operating Environment (MOE) 2022) showed that the hydrophobic guest molecule could adopt two binding orientations when bounding to a cyclodextrin host system. The up-state binding mode referred to the binding orientation when polar functional group in the guest molecule aligned with the secondary hydroxyls of BCD, or the down-state mode where the guest polar functional group aligned with the primary hydroxyls of BCD. Both binding modes were considered in our MD simulations. The lowest energy complex structures extracted from trajectories are shown in Figure 8. From inspection of these structures, it could be concluded that the interaction of hydroxyl head in cholecalciferol with secondary hydroxyls of HPBCD for up-state or with hydroxypropyls of HPBCD for down-state drives the binding of the complex.

Figure 8.

The lowest energy complex structure for both up-state (A) and down-state (B) mode selected from seven MD simulations. Cholecalciferol shown with carbon as a ball and stick representation. HPBCD shown with carbon is a light shade colored stick representation. Both top view (left) and side view (right) are present.

MM-PBSA calculations were undertaken to understand the relative free energies of binding of the up versus down states and reveal the energetic contributions to the binding free energy. MM-PBSA calculations were performed on snapshots every 10 ps from the 20 ns MD simulations with results listed in Table 2. For both up- and down-state binding modes, a large stabilizing hydrophobic contribution was seen with a small favorable electrostatic component in cholecalciferol consisting of a large hydrophobic part and a small polar hydroxyl head. The total dispersion and electrostatic binding energy favored the up-state over the down-state for nearly 4 kcal/mol. However, the trade-off between such favorable components with the loss in aqueous salvation yielded about the same net binding free energy with only about 0.1 kcal/mol predicted difference. This indicated that there would be an equilibrium between the two binding modes and cholecalciferol–HPBCD complex could exist in both orientations.

Table 2.

Computed complex interaction energy^a contributed (in kcal/mol) from cholecalciferol with HPBCD.

Energy contribution	Up-state	Down-state
Dispersion/hydrophobic	−32.68 (0.92)	−29.33 (1.41)
Electrostatic/hydrogen bond	−0.98 (0.35)	−0.71 (0.12)
Solvation	12.09 (0.66)	8.62 (0.52)
Totalb	−21.56 (0.56)	−21.42 (1.16)

^aEnergies are averaged over all seven simulations for each state with standard deviations in parentheses.

^bDiffer from binding free energy by entropy contributions.

Further analysis of the MD trajectories was performed to investigate the conformations sampled during complexation. Figure 9 plots the simulation time profile for the distance between centers of mass (COM) of cholecalciferol and HPBCD during one MD simulation for each binding mode. During the 20 ns MD simulation, the cholecalciferol remained in the cavity of HPBCD but with fluctuations to some extent. The COM distance around 0.3 Å represented a complex structure, in which the cholecalciferol molecule was well embedded in HPBCD, and the hydroxyl head of cholecalciferol formed hydrogen bonds with HPBCD hydroxyls or hydroxypropyls. The COM distances around 0.6 Å generally indicated a complex structure that the hydroxyl group of cholecalciferol was most probably exposed in solvent and some hydrophobic region of cholecalciferol was out of the HPBCD cavity. The up-state binding mode samples had much less exposed states (most around COM 0.3 Å, black solid line) than the down-state mode (most around COM 0.6 Å, gray dash line). This indicated that the hydroxyl head group of cholecalciferol had more solvent exposed in the down-state versus the up-state mode, which resulted in the less favorable solvation energy for the down-state as listed in Table 2. Based on all seven MD simulations, during approximately 21% of the simulation time cholecalciferol formed hydrogen bonds with HPBCD, and during 93% of the simulation time formed hydrogen bonds with water in the up-state binding mode. These values are 6% and 95% for the down-state binding mode. Such information is consistent with data shown in Table 2 and dynamics profile in Figure 9. The sums of the HPBCD and water hydrogen bond percentages being greater than 100% are due to the cholecalciferol simultaneously forming hydrogen bonds with both water and HPBCD in some conformations.

Figure 9.

Distances between centers of mass of cholecalciferol and HPBCD during one MD simulation for each state. Representative binding modes are shown for both up (indicated by solid arrows) and down-state (indicated by dotted arrows).

Our simulation was based on 1:1 complex formation model, while our experimental data showed a mixture of 1:1 and 1:2 complex formations (see Figure 1). The free energy of 1:1 complex formation was −10.68 kcal/mol, which is significantly more favorable than the additional energy (−2.85 kcal/mol) to form the 1:2 complex formation. This indicates that the 1:1 complex was the dominant species in this complex system. Also from Figure 9, only around COM 0.6 Å or higher COM structures could possibly lead to 1:2 complex formation because of the more solvent exposure of the cholecalciferol hydroxyl group. Therefore, we could use the simple 1:1 complex model to simulate the complexes. However, use of the 1:1 complex alone may contribute to the calculated free energy of binding from the simulation being a little different from the experimental value.

Conclusions

Cholecalciferol is practically insoluble in water and our study is demonstrating an increase in its aqueous solubility by eight orders of magnitude in the presence of 200 mM HPBCD (molar ratio of 11.1:200). Cholecalciferol complexes with HPBCD are facilitated by long distance interaction forces like: Van der Waals interactions, hydrophobic bonds, hydrogen bonds, etc. never covalent bonds leading to week interactions between them within the complex and can easily be broken in the presence of organic solvent during UPLC evaluation. DSC is conforming the complex formation also supported by FTIR, XRD, and SEM revealing a different physical behavior in comparison with the individual components or their physical mixture. The computational study is based on a 1:1 complex model at the molecular level and the free energy of binding is in a good agreement with the experimental data. Through the COM distances calculation on MD trajectories, two complex structures are observed to coexist with the preferred one being the 1:1 complex. In addition, long COM distances (0.6 Å) observed in the simulations that are twice as long as the COM distance (0.3 Å) for the 1:1 complex support the possibility of 1:2 complex formation. In summary, our study shows that HPBCD can increase the solubility of cholecalciferol in aqueous solutions via complex formation.