Acid–Base Interactions of Polystyrene Sulfonic Acid in Amorphous

Dec 30, 2015 - Haichen Nie , Yongchao Su , Mingtao Zhang , Yang Song , Anthony Leone , Lynne S. Taylor , Patrick J. Marsac , Tonglei Li , and Stephen ...
0 downloads 0 Views 3MB Size
Article pubs.acs.org/molecularpharmaceutics

Acid−Base Interactions of Polystyrene Sulfonic Acid in Amorphous Solid Dispersions Using a Combined UV/FTIR/XPS/ssNMR Study Yang Song,† Dmitry Zemlyanov,‡ Xin Chen,§ Haichen Nie,† Ziyang Su,∥ Ke Fang,† Xinghao Yang,⊥ Daniel Smith,† Stephen Byrn,† and Joseph W. Lubach*,# †

Department of Industrial and Physical Pharmacy, Purdue University, West Lafayette, Indiana 47907, United States Birck Nanotechnology Center, Purdue University, West Lafayette, Indiana 47907, United States § GlaxoSmithKline, Collegeville, Pennsylvania 19426, United States ∥ Center for Drug Evaluation and Research, Food and Drug Administration, Silver Spring, Maryland 20993, United States ⊥ College of Life Sciences, Nanjing Normal University, Nanjing 210046, P. R. China # Small Molecule Pharmaceutical Sciences, Genentech, Inc., South San Francisco, California 94080, United States ‡

S Supporting Information *

ABSTRACT: This study investigates the potential drug− excipient interactions of polystyrene sulfonic acid (PSSA) and two weakly basic anticancer drugs, lapatinib (LB) and gefitinib (GB), in amorphous solid dispersions. Based on the strong acidity of the sulfonic acid functional group, PSSA was hypothesized to exhibit specific intermolecular acid−base interactions with both model basic drugs. Ultraviolet (UV) spectroscopy identified red shifts, which correlated well with the color change observed in lapatinib−PSSA solutions. Fourier transform infrared (FTIR) spectra suggest the protonation of the quinazoline nitrogen atom in both model compounds, which agrees well with data from the crystalline ditosylate salt of lapatinib. X-ray photoelectron spectroscopy (XPS) detected increases in binding energy of the basic nitrogen atoms in both lapatinib and gefitinib, strongly indicating protonation of these nitrogen atoms. 15N solid-state NMR spectroscopy provided direct spectroscopic evidence for protonation of the quinazoline nitrogen atoms in both LB and GB, as well as the secondary amine nitrogen atom in LB and the tertiary amine nitrogen atom in GB. The observed chemical shifts in the LB−PSSA 15N spectrum also agree very well with the lapatinib ditosylate salt where proton transfer is known. Additionally, the dissolution and physical stability behaviors of both amorphous solid dispersions were examined. PSSA was found to significantly improve the dissolution of LB and GB and effectively inhibit the crystallization of LB and GB under accelerated storage conditions due to the beneficial strong intermolecular acid−base interaction between the sulfonic acid groups and basic nitrogen centers. KEYWORDS: lapatinib, gefitinib, amorphous, solid dispersion, physical stability, dissolution rate, XPS, solid-state NMR, ionic interaction, salt



INTRODUCTION

in hydrophilic polymers, have been extensively used to stabilize amorphous products.5 However, there is not a conclusive universal mechanism for stabilizing amorphous drugs by solid dispersions. Many authors have suggested that high glass transition temperature (Tg) polymers could stabilize amorphous drugs because of their antiplasticizing effects.6 In addition, many studies have indicated that strong intermolecular interactions like hydrogen bonding7−9 and acid−base interactions10−16 between polymers and APIs are the driving

With the development of high throughput screening and combinatorial chemistry, more than 40% of newly discovered active pharmaceutical ingredients (API) have poor aqueous solubility, which may lead to low oral bioavailability.1 Amorphization has become one of the most popular and effective approaches to increase drug solubility and dissolution rate.2 Amorphous materials have higher enthalpy, entropy, Gibbs free energy, and molecular mobility compared to their counterpart crystalline forms because of the lack of long-range molecular order.3 Since amorphous forms possess higher energy, they are prone to crystallize to a lower energy state during manufacturing, dissolution, and long-term storage.4 Amorphous solid dispersions, which molecularly disperse APIs © XXXX American Chemical Society

Received: September 15, 2015 Revised: December 11, 2015 Accepted: December 18, 2015

A

DOI: 10.1021/acs.molpharmaceut.5b00708 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Molecular Pharmaceutics Table 1. Characteristics of PSSA and Model Compounds Functional Groups Relevant for Hydrogen Bonding

a

H-bond acceptor strength was determined using the pKBHX scale.33 The strengths used the following scale: weak < 0.75 < medium < 1.5 < strong < 2.25 < very strong.15 bNo data available, strengths are estimated.

modern APIs dispersed in polymers. Fortunately, other analytical techniques are also available to provide more information about drug−excipient interactions. Ultraviolet spectroscopy (UV) has been shown to be very useful for detecting energy level transitions of compounds containing conjugated structures.19 For example, a red shift in the UV spectrum has been observed for the ionization of isonicotinic acid20 and clofazimine.21 X-ray photoelectron spectroscopy (XPS) is a surface sensitive technique that has become widely used for studying physical and chemical phenomena on the surfaces of solids but has seen only limited usage in pharmaceutical research. Generally speaking, intensities of core-level photoelectron peaks are used for quantitative analysis, and the chemically induced binding energy (BE) shifts of core-level photoelectrons are used to identify chemical states (qualitative analysis). XPS characterization has mainly been limited to relatively simple inorganic reactions, and few problems relevant to organic materials have been approached using XPS. The major problems associated with XPS studies of organic materials are, first, possible radiation damage of the sample from the Xrays, and second, the C 1s region, which is most informative for organic chemistry samples, is narrow and the photoemission peaks can overcrowd the region. Thus, the XPS technique has not been widely applied to drug analysis. However, XPS was successfully used to show the susceptibility of hydrochloride pharmaceutical salts to water-induced dissociation at the surface.22,23 XPS was also critical to characterize the surface chemistry of peptide modified sol−gel thin films.24−26 Recently, several studies have indicated the potential of XPS for exploring intermolecular drug−polymer interactions.27 Of particular interest is the high sensitivity of XPS for the detection of the degree of proton transfer in acid−base systems by measuring

force for inhibiting crystallization of amorphous APIs. Since acid−base interactions are much stronger than hydrogen bonding, they have recently attracted significant interest. However, there are only a few acidic polymers suitable for use in oral solid dosage forms, such as hydroxypropylmethylcellulose acetate succinate (HPMCAS), hydroxypropylmethylcellulose phthalate (HPMCP), poly(methacrylic acid-co-ethyl acrylate) or Eudragit L, and poly(acrylic acid) (PAA). Moreover, phthalates are rapidly losing favor in pharmaceutical use due to safety concerns. All of these polymers contain weakly acidic functional groups; hence, they are not very prone to protonate weakly basic groups of many modern APIs. Thus, there exists clear interest for evaluation of the feasibility of a strongly acidic polymer such as polystyrene sulfonic acid (PSSA) in formulating amorphous solid dispersions with weakly basic APIs. PSSA is currently widely used in ion exchanging and proton conducting membranes, and sodium or calcium salts have been administered to humans to treat hyperkalemia via a potassium binding mechanism. However, it is rarely used in pharmaceutical formulations, possibly due to potential gastrointestinal side effects, limited safety data, and lack of historical pharmaceutical usage. Van Eerdenbrugh and Taylor have shown that PSSA can inhibit crystallization of several compounds and have studied its acid−base interactions with model compounds by using Fourier transform infrared spectroscopy (FTIR).15 FTIR spectroscopy is the most commonly used approach for investigating drug− polymer interactions as it is able to measure the vibrational changes of functional groups such as carbonyls, amines, and carboxylates, which are typically involved in hydrogen bonding or acid−base interactions.17,18 However, it has limitations when used in exploring complicated structures, which have overlapping regions in the spectrum and is often the case for B

DOI: 10.1021/acs.molpharmaceut.5b00708 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Molecular Pharmaceutics the BE shift of the involved atoms.28,29 For example, a strong N 1s shift of +2 eV toward higher BEs has been observed for both the protonation of an aromatic nitrogen in theophylline28 as well as the protonation of the aliphatic nitrogen in piperidine groups.30 Solid-state NMR spectroscopy (ssNMR) continues to see increasing use for characterizing pharmaceutical formulations, particularly amorphous materials. For example, 15N ssNMR has shown large upfield shifts of 80−100 ppm upon protonation of nitrogen atoms in heterocyclic aromatic systems, while relatively smaller upfield shifts are seen for hydrogen bonding.31 In contrast, this trend reversed for aliphatic nitrogen atoms: the tertiary nitrogen atoms in piperidine groups show only a 1 ppm downfield shift in sildenafil citrate.32 In general, 15N chemical shift changes for aliphatic amines are much smaller than those of aromatic 15N nuclei. Recently, authors of the present work found a downfield shift on protonation of a secondary amine nitrogen atom in lapatinib,16 showing that one could take advantage of the atomic specificity and selectivity of ssNMR to explore the potential intermolecular interactions between PSSA with model weakly basic compounds in the amorphous state. Based on the strong acidity of sulfonic acid side chain groups in PSSA, we hypothesized the potential for an acid−base interaction between PSSA and nitrogen atoms in both lapatinib and gefitinib. We further hypothesized that this strong intermolecular interaction would effectively inhibit recrystallization during storage and dissolution. The present study investigates the potential drug−excipient interactions in amorphous solid dispersions using a combined UV, FTIR, XPS, and 13C and 15N ssNMR study. Dissolution properties of LB and GB in the solid dispersions were tested in vitro under nonsink conditions using pH-neutral medium. The physical stability of the solid dispersions was also evaluated under accelerated storage conditions with powder X-ray diffraction (PXRD) over six months (Table 1).

Figure 1. Chemical structures of (a) lapatinib, (b) gefitinib, (c) toluenesulfonic acid (counteracid in lapatinib ditosylate), and (d) polystyrene sulfonic acid (PSSA).

Fourier Transform Infrared Spectroscopy (FTIR). FTIR measurements were performed using a Thermo Nicolet Nexus FTIR Spectrometer at ambient room temperature with the following settings: 400−4000 cm−1, 128 scans, resolution of 2 cm−1. OMNIC software was used for analysis of the spectra. Solid dispersions of LB and GB with 10%, 20%, 40%, and 60% drug loading were examined and compared with pure amorphous drug and PSSA. In addition, solid dispersions with 40% drug loading were compared with pure amorphous drug−PSSA and crystalline drug−PSSA physical mixture of the same ratio. X-ray Photoelectron Spectroscopy (XPS). XPS data were obtained using a Kratos Axis Ultra DLD spectrometer with monochromic Al Kα radiation (1486.6 eV) at pass energy of 20 and 160 eV for high-resolution and survey spectra, respectively. A commercial Kratos charge neutralizer was used to avoid nonhomogeneous electric charge of nonconducting powder and to achieve better resolution. The resolution measured as full width at half-maximum of the curve fitted C 1s and N 1s peaks was approximately 1 eV. Binding energy (BE) values refer to the Fermi edge and the energy scale was calibrated using Au 4f7/2 at 84.0 eV and Cu 2p3/2 at 932.67 eV. XPS data were analyzed with CasaXPS software version 2313 Dev64 (www.casaxps.com). Prior to data analysis, the C−C component of the C 1s peak was set to a binding energy of 284.8 eV to correct for charge on each sample (NIST X-ray Photoelectron Spectroscopy Database, Version 4.1, National Institute of Standards and Technology, Gaithersburg, 2012); http://srdata.nist.gov/xps/). Curve fitting was performed following a linear (N 1s) or Shirley (C 1s) background subtraction using asymmetric Lorentzian lines with 80% Gaussian contribution (CasaXPS line shape: LF(1,1.8,10,80)). The atomic concentrations of the elements in the near-surface region were estimated taking into account the corresponding Scofield atomic sensitivity factors and inelastic mean free path (IMFP) of photoelectrons using standard procedures in the CasaXPS software. Solid-State NMR Spectroscopy (ssNMR). Solid-state NMR data were acquired using a Bruker Avance III HD spectrometer operating at 500.13 MHz for 1H, 125.77 MHz for 13 C, and 50.69 MHz for 15N, along with a Bruker 2-channel (HX) solids probe equipped with a 4 mm stator (Bruker BioSpin Corp., Billerica, MA). The pulse sequence for 13C



EXPERIMENTAL SECTION Materials. Methanol and dichloromethane were obtained from Macron Fine Chemicals (Center Valley, PA). PSSA solution was purchased from Sigma-Aldrich Corporation (St. Louis, MO), lyophilized, and cryomilled to fine particles. Lapatinib and lapatinib ditosylate were purchased from Attix Corporation (Toronto, Canada). Gefitinib was purchased from TOKU-E (Bellingham, WA) (Figure 1). Rotary Evaporation. Drug and polymers were dissolved in a 1:1 (v/v) mixture of dichloromethane (DCM) and methanol. Homogeneous solutions containing drug and PSSA were evaporated on a Buchi rotary evaporator under reduced pressure at 50 °C. The drug content in solution was varied at 10, 20, 40, and 60%. The evaporated samples were dried under vacuum overnight, then cryomilled for 2 min to obtain fine particles. Powder X-ray Diffraction (PXRD). The X-ray diffraction patterns were collected using a Siemens D5000 X-ray diffractometer. Measurements were made using Cu Kα radiation. The data were collected at ambient temperature with a tube power of 40 kV/40 mA and scanning speed of 2°/ min, in the angular range of 4−40° 2θ. Ultraviolet Spectroscopy (UV). UV spectra were obtained using a CCD array UV/vis spectrophotometer (S.I. Photonics, Inc.) in the wavelength range of 250 to 850 nm for methanol solutions of defined concentration. C

DOI: 10.1021/acs.molpharmaceut.5b00708 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Molecular Pharmaceutics acquisition employed ramped cross-polarization (CP)34−36 with a 70−100% ramp on the 1H channel, 5-π total sideband suppression (TOSS),37,38 and high power 1H decoupling with a SPINAL6439 scheme and field strength of 89 kHz. 15N experiments utilized an analogous ramped CP pulse sequence, but without TOSS. Magic-angle spinning (MAS) was performed at 8000 ± 3 Hz for all experiments. The 1H 90° pulse width was 2.8 μs and the TOSS sequence employed 13C 180° pulses of 6.5 μs. Each 13C and 15N experiment utilized a CP contact time of 5 ms and recycle delays of 2.5−5 s, depending on T1 of the sample. A total of 7776 scans were averaged for each 13C spectrum shown. A total of 75000− 100000 scans were averaged for each amorphous 15N spectrum shown, 19440 scans for crystalline LB freebase, and 25000 scans for crystalline LB phthalate. 15N dipolar dephasing experiments utilized a dephasing time of 120 μs and 4 kHz MAS. Line broadening of ∼10−15% of the natural line width was employed for the amorphous 15N spectra shown. All ssNMR data were collected at 298 K. 13C chemical shifts were externally referenced by setting the methyl peak of 3methylglutaric acid to 18.84 ppm relative to tetramethylsilane,40 while 15N chemical shifts were externally referenced to nitromethane by setting the amine peak of glycine to −347.58 ppm. Data were analyzed used Bruker TopSpin 3.2 software (Bruker BioSpin Corp., Billerica, MA), and relaxation data were fitted using KaleidaGraph 4.1 software (Synergy Software, Reading, PA). Dissolution. Dissolution tests were carried out by using a Vankel dissolution apparatus at 37 °C in 500 mL of 0.2% (w/v) aqueous SDS for 2 h. Solid dispersions (15 mg) were put into the basket with a rotation speed of 100 rpm. Dissolution was monitored by UV spectroscopy (S.I. Photonics, Inc.) with spectra taken every 5 min over the course of the experiment. The absorbance at 330 nm was plotted vs time and each experiment was repeated in triplicate. Physical Stability Evaluation. Solid dispersions of PSSA with LB and GB of 40% drug load were stored at the ICH accelerated stability condition of 40 °C/75% RH. After 3 and 6 months, samples were characterized by powder X-ray diffraction (PXRD) to detect crystallization.

Figure 2. PXRD patterns of (a) LB−PSSA and (b) GB−PSSA solid dispersions with drug loadings of 10%, 20%, 40%, and 60%.

change suggests a significant change in the outer electrons in the conjugated system of lapatinib. UV spectroscopy was used to further investigate this color change by determining whether or not a potential red or blue shift was occurring in the drug− polymer solutions. As shown in Figure 3a, there is clear redshift from 360 to 390 nm as the amount of PSSA increases in solution, which corresponds to the shift from a clear to light green solution. Green color can be produced by absorption at ∼400 nm as well as ∼800 nm. The UV spectrum correlates well with the observed color change in methanol solutions containing lapatinib. Similar UV shifts were also detected in methanol solutions containing GB and PSSA. For both compounds, such UV red-shifts indicate the presence of strong intermolecular interactions between the quinazoline of the model compounds and sulfonic acid groups of PSSA. FTIR Spectroscopy. While the fingerprint region of an IR spectrum can provide a wealth of information about a molecule, the inherent complexity of this region can make it difficult to tease out pertinent information. It is common to use simpler model compounds that contain some of the same functional groups to assign unknown peaks. This methodology was used to identify peaks that could indicate if LB and GB were protonated in the solid dispersions. Literature reports indicate that pyridine has two peaks around 1600 cm−1 arising from aromatic C−C vibration and



RESULTS AND DISCUSSION Powder X-ray Diffraction (PXRD). The crystallinity of the drugs was determined by PXRD right after the solid dispersions were prepared. The lack of distinct crystalline peaks in the PXRD is indicative of amorphous drug within the dispersion. As shown in the Figure 2, all LB and GB solid dispersions were amorphous. Solid-state NMR relaxation measurements were also performed on the 40% drug load samples and indicated the LB−PSSA dispersion to be a single phase, with 1H T1 values of 0.87 and 0.88 s for LB and PSSA, respectively, and 1H T1ρ values of 9.21 and 9.61 ms. Relaxation times also indicated that the 40% GB−PSSA dispersion was a single phase, with 1H T1 values of 0.72 and 0.70 s for GB and PSSA, respectively, and 1H T1ρ values of 7.62 and 8.21 ms. Additionally, the 13C and 15N ssNMR spectra showed no crystalline material present in these dispersions. Investigations of Drug−PSSA Interactions. UV−vis Spectrophotometry. Visually, methanolic lapatinib solutions are transparent and primarily colorless; however, as PSSA is added into the solution, it develops a light green color. Since the absorption of UV and visible radiation corresponds to the excitation of outer electrons in conjugated systems, this color D

DOI: 10.1021/acs.molpharmaceut.5b00708 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Molecular Pharmaceutics

Figure 4. FTIR spectra from 1550 to 1650 cm−1 of (a) amorphous lapatinib (LB), crystalline LB−PSSA physical mixture (PM), amorphous LB−PSSA PM, LB−PSSA solid dispersion (SD) at 40% drug load, PSSA; and (b) amorphous gefitinib (GB), crystalline GB− PSSA physical mixture (PM), amorphous GB−PSSA PM, GB−PSSA solid dispersion (SD) at 40% drug load, and PSSA.

Figure 3. UV−vis spectra of (a) LB−PSSA and (b) GB−PSSA in methanol as the drug−polymer ratio is varied.

aromatic C−N vibration.41 As shown in the Figure 4 for LB (4a) and GB (4b), the region between 1580 and 1620 cm−1 clearly shows two distinct peaks that are likely to represent the aromatic C−C and aromatic C−N vibrations. While LB is protonated, in the case of the ditosylate salt, a new peak appears around 1620 cm−1 and a significant decrease in the peak around 1605 cm−1 is seen (Figure 4a). The LB ditosylate salt used in the commercial formulation is a 1:2 salt of LB with toluenesulfonic acid, resulting in the two most basic nitrogen atoms in LB being protonated. Similar changes are observed in the IR spectrum of a 40% LB−PSSA solid dispersion, strongly suggesting that the LB is protonated in the PSSA matrix in a manner analogous to the ditosylate salt. The observed changes can be explained by a reduced electron density on the protonated nitrogen in the quinazoline ring system. Similar changes were observed in the GB IR spectrum when a solid dispersion was made with PSSA (Figure 4b). In this case, the appearance of a new peak at 1637 cm−1 and the disappearance of the peak at 1620 cm−1 indicate the protonation of the quinazoline ring. The changes in the IR spectra of both drugs are not seen in either of the pure amorphous forms or in the drug−PSSA physical mixtures, suggesting that protonation is a solid-state reaction as a result of the manufacturing process creating a new amorphous salt.

X-ray Photoelectron Spectroscopy (XPS). Figure 5 shows the N 1s peak in the XPS spectra of lapatinib freebase, lapatinib−PSSA solid dispersion at 40% drug load, and lapatinib ditosylate salt. The N 1s peak of lapatinib was curve-fitted with two well-resolved components at 399.8 and 398.6 eV for secondary amine and quinazoline nitrogen, respectively. The ratio between these components was close to unity, which is consistent with two secondary amine nitrogen atoms and two pyridine nitrogen atoms in the lapatinib molecule (Figure 1). The shape of the N 1s peak is different for LB−PSSA and lapatinib ditosylate (Figure 5): the spectra demonstrated a more complex multicomponent structure rather than just two simple components in the case of lapatinib freebase. New N 1s components at higher binding energies (BE) should be formed due to nitrogen protonation of the lapatinib molecules. The LB−PSSA and lapatinib ditosylate N 1s spectra were fitted with four components corresponding to secondary amine (NH), quinazoline nitrogen (CN), protonated secondary amine (NH2+), and protonated quinazoline nitrogen (CNH+). The BE of the components and their relative areas are provided in Table 2. The protonated groups E

DOI: 10.1021/acs.molpharmaceut.5b00708 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Molecular Pharmaceutics

redistribution of electron density in the molecule, which is confirmed by the high BEs of the F 1s and Cl 2p peaks by 0.2 eV for lapatinib−PSSA DL40% and lapatinib ditosylate. However, the S 2p peaks shifted by 1.0 eV toward lower BE. The similarity of changes in the XPS spectra of LB−PSSA and lapatinib ditosylate compared to lapatinib freebase points to the same doubly protonated state of lapatinib in the ditosylate salt and the amorphous dispersion in PSSA. The N 1s peaks in the XPS spectra of the gefitinib and gefitinib−PSSA solid dispersion at 40% drug load are shown Figure 6. The summary of the N 1s peak curve fitting is Table

Figure 5. XPS N 1s peaks obtained from (a) lapatinib freebase, (b) lapatinib−PSSA solid dispersion at 40% drug load, and (c) lapatinib ditosylate salt.

demonstrated higher BE than the corresponding nonprotonated counterparts. The BE shift due to protonation is in the range of 1.7−2.5 eV for quinazoline nitrogen and in the range of 1.3−1.5 eV for the secondary amine. It should be noted that protonation of the quinazoline nitrogen (CNH+) and secondary amine (NH2+) in the lapatinib molecule also affected the BE of nonprotonated groups: the N 1s components of the nonprotonated quinazoline nitrogen (CN) and nonprotonated secondary amine (NH) shifted by 0.6−0.9 eV toward higher BE (Figure 5). This could be a reflection of the

Figure 6. XPS N 1s peaks obtained from (a) gefitinib and (b) gefitinib−PSSA solid dispersion at 40% drug load.

2. The protonated quinazoline and protonated morpholine groups in gefitinib demonstrated shifts to higher BE of +1.65 and +2.1 eV, respectively. According to XPS analysis, the morpholine group was fully protonated. The appearance of the protonated components and the shift of the entire N 1s spectrum toward higher BE for the gefitinib−PSSA solid

Table 2. XPS N 1s Assignments for Each Nitrogen Atom in Lapatinib and Gefitinib sample LB

LB−PSSA DL40% ASD LB-Ditosylate

GB

GB−PSSA DL40% ASD

functional group/binding energy (eV)/area (%) NH 399.8 50% NH 400.55 25% NH 400.5 25% NH 399.87 25% NH 400.3 25%

functional group/binding energy (eV)/area (%) quinazoline 398.57 50% quinazoline 399.5 25% quinazoline 399.2 25% quinazoline 398.4 50% quinazoline 399.3 25%

functional group/binding energy (eV)/area (%)

functional group/binding energy (eV)/area (%)

NH2+ protonated 402.1 25% NH2+ protonated 401.86 25% morpholine N 399.08 25% morpholine NH+ protonated 402.14 25%

quinazoline NH+ protonated 401.2 25% quinazoline NH+ protonated 401.2 25%

N

N

N

N

N

F

quinazoline NH+ protonated 400.95 25% DOI: 10.1021/acs.molpharmaceut.5b00708 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Molecular Pharmaceutics

Figure 7. 13C CPMAS NMR spectra of lapatinib (LB) freebase, LB−PSSA solid dispersion with 40% drug load, gefitinib (GB) freebase, GB−PSSA solid dispersion with 40% drug load, and PSSA.

Figure 8. 15N ssNMR of lapatinib (LB) freebase, LB ditosylate salt, and lapatinib−PSSA amorphous solid dispersion at 40% drug load.

lapatinib and gefitinib are protonated in their respective PSSA dispersions. While it is feasible to deduce which positions are actually protonated through chemical reasoning and computational methods for pKa calculation, neither UV nor FTIR provide direct evidence for which nitrogen atoms are involved in the protonation. Like XPS, ssNMR is sensitive to subtle changes in the electronic environment surrounding nuclei of interest, making it an excellent technique to investigate the protonation sites in these systems. The 13C ssNMR spectra of crystalline LB, GB, PSSA, and their dispersions are depicted in Figure 7. The results for crystalline LB and GB show sharp peaks, while PSSA and the solid dispersions show broad,

dispersion points to a doubly protonated state of the gefitinib molecule in a similar manner as lapatinib−PSSA. In both compounds, the nonbasic nitrogen atoms (other quinazoline N and aniline N) shifted toward high BE by 0.6−0.9 and ca. 0.9 eV for lapatinib and gefitinib, respectively. The electron levels of the halogen atoms, Cl 2p and F 1s, show a similar tendency as well. As speculated above, this high BE shift is likely due to redistribution of the positive charge throughout the entire molecules, and the conjugated quinazoline rings were affected to the greatest extent. Solid-State NMR Spectroscopy. Combined, the results from UV and FTIR strongly support the hypothesis that both G

DOI: 10.1021/acs.molpharmaceut.5b00708 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Molecular Pharmaceutics

Figure 9. 15N ssNMR spectra of gefitinib (GB) freebase, GB−PSSA amorphous solid dispersion at 40% drug load, and dipolar dephased spectrum of the GB−PSSA solid dispersion. The dipolar dephased spectrum is scaled such that intensity of the N3 peak matches that of the standard CPMAS spectrum.

Gaussian resonances indicative of amorphous materials. This correlates well with the PXRD data. While informative in their own respects, the carbon spectra do not provide specific information about the exact sites of drug−excipient interactions between these drugs and PSSA. Previously, we investigated the acid−base interaction between the secondary amine of LB and the phthalic acid groups in HPMCP by using 15N ssNMR.16 This technique was therefore also of interest to examine the potential acid−base interactions in PSSA solid dispersions. 15N solid-state NMR spectra of LB, GB, and amorphous solid dispersions in PSSA are shown in Figures 8 and 9. Assignment of the nitrogen peaks in LB and GB compared well to those previously reported in the literature for a structurally similar compound, E-2-methoxyN-(34-[3-methyl-4-(6-methyl-pyridin-3-yloxyl)-phenylamino]quinazolin-6-yl-allyl)-acetamide.31 For this compound, it was shown that protonation of quinazoline nitrogens can induce an upfield shift of 80−100 ppm. Indeed, the more basic quinazoline N nucleus of LB (N3) and GB (N2) shows large upfield shifts of 86.4 and 79.6 ppm, respectively, indicating the protonation is occurring on this specific nitrogen nucleus in each molecule. Additionally, the aliphatic secondary amine N of LB (N4) shows a downfield shift of 16.6 ppm in the LB−PSSA solid dispersion, relative to the crystalline freebase. This is an excellent indication of protonation for the aliphatic secondary amine nitrogen of LB (N4), as the 15N chemical shifts of aliphatic amines do not change on the magnitude as those of aromatic nitrogen atoms. It is worth noting that the aniline NH (N1 in LB and GB), which is nonionizable and can only participate in hydrogen bonding, also shows a downfield shift of 16.8 ppm relative to the crystalline freebase. As this nitrogen is sandwiched between two electron-rich aromatic ring systems, its chemical shift is highly sensitive to changes in the surrounding electronic environment, such as perturbation of the neighboring aromatic system by protonation of the quinazoline group. While aliphatic amine chemical shifts are also sensitive to conformational changes or H-bonding, the

extent of shift changes is much less for these functionalities. We believe that this observation, especially when coupled with the XPS data, serves to confirm the doubly protonated state of LB in the solid dispersion with PSSA. Confirmation of a doubly protonated state of gefitinib in the PSSA solid dispersion based on 15N solid-state NMR was somewhat more challenging, as isotropic chemical shifts alone were not sufficient. It has been shown that tertiary aliphatic amines often do not have significantly different chemical shifts when protonated,32 and this is indeed the case with gefitinib. As shown in Figure 9, the tertiary amine in the GB morpholine group (N4) has the same chemical shift (−330.8 ppm) in both the crystalline freebase and GB−PSSA solid dispersion. As with LB, the nonionizable aniline NH (N1) sandwiched between two electron-rich aromatic ring systems does show a change of 18.4 ppm relative to the crystalline freebase, likely due to perturbation of the quinazoline ring system via protonation. To confirm protonation of the aliphatic tertiary amine nitrogen (N4), a dipolar dephasing (interrupted decoupling) experiment was performed on the GB−PSSA dispersion. In this experiment, the 1H decoupler is turned off for a short period of time before signal acquisition, 120 μs in this case. During this time, nitrogen nuclei strongly coupled (i.e., covalently bonded) to protons will dephase, or lose signal intensity, relative to nonprotonated nitrogens. If GB were doubly protonated, only the least basic nitrogen, N3, would remain unprotonated, and the three other N peaks would be reduced or disappear relative to the N3 peak. Indeed this is what was observed in the dipolar dephased 15N spectrum of GB−PSSA shown in Figure 9. The peak for N4 shows significantly reduced intensity relative to N3, confirming protonation of N4. Additionally, the quinazoline N2 shows significant dephasing, confirming that its large chemical shift change is due to protonation. N1 also shows complete dephasing as expected since it is already protonated in the neutral state of the molecule and is nonionizable under normal conditions. It should be noted that optimization of the dephasing time should result in complete disappearance of H

DOI: 10.1021/acs.molpharmaceut.5b00708 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Molecular Pharmaceutics peaks for N2 and N4, but this would have required excessive spectrometer time. The chemical shift change of 79.6 ppm in the quinazoline N2, coupled with the dipolar dephasing data for both N2 and morpholine N4, serve to confirm the doubly protonated state of gefitinib in GB−PSSA amorphous solid dispersions and support the conclusions drawn from the other techniques presented here. Dissolution and Physical Stability. Figure 10 shows that PSSA solid dispersions with both LB and GB give significant

Figure 11. Physical stability testing via PXRD of lapatinib (LB)−PSSA solid dispersion at 40% drug load at t = 0, 3, and 6 months; and gefitinib (GB)−PSSA solid dispersion at 40% drug load at t = 0, 3, and 6 months.

tion of the surrounding electron-rich aromatic systems, which has a large effect on the chemical shielding. XPS is excellent for determining local protonation, but cannot necessarily distinguish between two chemically different NH groups, such as the aliphatic secondary amine NH and aniline NH in LB freebase. Thus, XPS coupled with 15N ssNMR is a powerful combination for protonation state determination in amorphous solids, and both add atomic-level insight in support of the more traditionally used methods of UV and FTIR spectroscopy. The special stability afforded by ionic interactions in amorphous solid dispersions is especially advantageous at moderate to high drug loads and can help reduce the pill burden issue often presented by using amorphous formulations.

Figure 10. Dissolution testing at pH 7 of lapatinib (LB) freebase compared with the LB−PSSA solid dispersion at 40% drug load; and gefitinib (GB) freebase compared with the GB−PSSA solid dispersion at 40% drug load.

improvements in the dissolution performance compared to their crystalline forms. This shows that the amorphous salt formed between each API and PSSA results in greatly improved dissolution and that the drug does indeed release from the polymer in aqueous solution. This solid-state ionic interaction further leads to excellent physical stability of these two amorphous systems, as no crystallization was observed at 40% drug load in each system when stored at 40 °C/75% RH for up to 6 months (Figure 11).



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.molpharmaceut.5b00708. The complete XPS spectra for the systems studied in this manuscript (PDF)



CONCLUSIONS Multiple spectroscopic techniques, including UV−vis, FTIR, XPS, and ssNMR all provide evidence of a strong, specific ionic interaction between two model weakly basic drugs, lapatinib and gefitinib, and the acidic polymer polystyrene sulfonic acid (PSSA). XPS and solid-state NMR both provided the most specific and definitive evidence of protonation of each drug when dispersed in PSSA. XPS showed differences in the N 1s binding energy populations caused by protonation of the two most basic nitrogen atoms in lapatinib and gefitinib. 15N solidstate NMR chemical shifts and dipolar dephasing data helped to confirm the doubly protonated state of each compound, as well as exactly which nitrogen atoms were involved in the ionic interaction. While 15N chemical shifts of aliphatic amines are not always unambiguous in determination of protonation, additional NMR experiments such as dipolar dephasing, together with XPS data can unambiguously determine protonation state of these types of amines in noncrystalline materials. Chemical shifts of aromatic amines show much more drastic changes upon protonation, due to significant perturba-



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



REFERENCES

(1) Babu, N. J.; Nangia, A. Solubility Advantage of Amorphous Drugs and Pharmaceutical Cocrystals. Cryst. Growth Des. 2011, 11 (7), 2662−2679. (2) Newman, A.; Knipp, G.; Zografi, G. Assessing the performance of amorphous solid dispersions. J. Pharm. Sci. 2012, 101 (4), 1355−77. (3) Yu, L. Amorphous pharmaceutical solids: preparation, characterization and stabilization. Adv. Drug Delivery Rev. 2001, 48 (1), 27−42. (4) Hilden, L. R.; Morris, K. R. Physics of amorphous solids. J. Pharm. Sci. 2004, 93 (1), 3−12. (5) Janssens, S.; Van den Mooter, G. Review: physical chemistry of solid dispersions. J. Pharm. Pharmacol. 2009, 61 (12), 1571−86.

I

DOI: 10.1021/acs.molpharmaceut.5b00708 Mol. Pharmaceutics XXXX, XXX, XXX−XXX

Article

Molecular Pharmaceutics (6) Ueda, H.; Aikawa, S.; Kashima, Y.; Kikuchi, J.; Ida, Y.; Tanino, T.; Kadota, K.; Tozuka, Y. Anti-plasticizing effect of amorphous indomethacin induced by specific intermolecular interactions with PVA copolymer. J. Pharm. Sci. 2014, 103 (9), 2829−38. (7) Konno, H.; Taylor, L. S. Influence of different polymers on the crystallization tendency of molecularly dispersed amorphous felodipine. J. Pharm. Sci. 2006, 95 (12), 2692−705. (8) Aso, Y.; Yoshioka, S. Molecular mobility of nifedipine-PVP and phenobarbital-PVP solid dispersions as measured by 13C-NMR spinlattice relaxation time. J. Pharm. Sci. 2006, 95 (2), 318−25. (9) Weuts, I.; Kempen, D.; Decorte, A.; Verreck, G.; Peeters, J.; Brewster, M.; Van den Mooter, G. Physical stability of the amorphous state of loperamide and two fragment molecules in solid dispersions with the polymers PVP-K30 and PVP-VA64. Eur. J. Pharm. Sci. 2005, 25 (2−3), 313−20. (10) Weuts, I.; Kempen, D.; Verreck, G.; Peeters, J.; Brewster, M.; Blaton, N.; Van den Mooter, G. Salt formation in solid dispersions consisting of polyacrylic acid as a carrier and three basic model compounds resulting in very high glass transition temperatures and constant dissolution properties upon storage. Eur. J. Pharm. Sci. 2005, 25 (4−5), 387−93. (11) Chokshi, R. J.; Shah, N. H.; Sandhu, H. K.; Malick, A. W.; Zia, H. Stabilization of low glass transition temperature indomethacin formulations: impact of polymer-type and its concentration. J. Pharm. Sci. 2008, 97 (6), 2286−98. (12) Liu, H.; Zhang, X.; Suwardie, H.; Wang, P.; Gogos, C. G. Miscibility studies of indomethacin and Eudragit E PO by thermal, rheological, and spectroscopic analysis. J. Pharm. Sci. 2012, 101 (6), 2204−2212. (13) Kojima, T.; Higashi, K.; Suzuki, T.; Tomono, K.; Moribe, K.; Yamamoto, K. Stabilization of a Supersaturated Solution of Mefenamic Acid from a Solid Dispersion with EUDRAGIT EPO. Pharm. Res. 2012, 29 (10), 2777−2791. (14) Wegiel, L. A.; Mauer, L. J.; Edgar, K. J.; Taylor, L. S. Crystallization of amorphous solid dispersions of resveratrol during preparation and storage-Impact of different polymers. J. Pharm. Sci. 2013, 102 (1), 171−184. (15) Van Eerdenbrugh, B.; Taylor, L. S. An ab initio polymer selection methodology to prevent crystallization in amorphous solid dispersions by application of crystal engineering principles. CrystEngComm 2011, 13 (20), 6171−6178. (16) Song, Y.; Yang, X.; Chen, X.; Nie, H.; Byrn, S. R.; Lubach, J. W. Investigation of drug-excipient interactions in lapatinib amorphous solid dispersions using solid-state NMR spectroscopy. Mol. Pharmaceutics 2015, 12 (3), 857−66. (17) Sathigari, S. K.; Radhakrishnan, V. K.; Davis, V. A.; Parsons, D. L.; Babu, R. J. Amorphous-state characterization of efavirenz–polymer hot-melt extrusion systems for dissolution enhancement. J. Pharm. Sci. 2012, 101 (9), 3456−64. (18) Sarode, A. L.; Sandhu, H.; Shah, N.; Malick, W.; Zia, H. Hot melt extrusion for amorphous solid dispersions: temperature and moisture activated drug-polymer interactions for enhanced stability. Mol. Pharmaceutics 2013, 10 (10), 3665−75. (19) Wang, K.; Hou, Y.; Zhang, W.; Ksebati, M. B.; Xian, M.; Cheng, J. P.; Wang, P. G. 15N NMR and electronic properties of Snitrosothiols. Bioorg. Med. Chem. Lett. 1999, 9 (19), 2897−902. (20) Asuero, A. G.; Navas, M. J.; Herrador, M. A.; Recamales, A. F. Spectrophotometric evaluation of acidity constants of isonicotinic acid. Int. J. Pharm. 1986, 34 (1−2), 81−92. (21) Nie, H.; Mo, H.; Zhang, M. T.; Song, Y.; Fang, K.; Taylor, L. S.; Li, T.; Byrn, S. R. Investigating the Interaction Pattern and Structural Elements of a Drug-Polymer Complex at the Molecular Level. Mol. Pharmaceutics 2015, 12 (7), 2459−68. (22) Guerrieri, P.; Zemlyanov, D.; Taylor, L. S. Study of water adsorption on organics crystal surfaces using a modified X-ray photoelectron spectroscopy instrument. Anal. Chem. 2011, 83 (3), 1144−7.

(23) Guerrieri, P.; Zemlyanov, D.; Taylor, L. S. Dissociation of water on the surface of organic salts studied by X-ray photoelectron spectroscopy. Langmuir 2010, 26 (14), 11998−2002. (24) Jedlicka, S. S.; Rickus, J. L.; Zemlyanov, D. Controllable Surface Expression of Bioactive Peptides Incorporated into a Silica Thin Film Matrix. J. Phys. Chem. C 2010, 114 (1), 342−344. (25) Jedlicka, S. S.; Little, K. M.; Nivens, D. E.; Zemlyanov, D.; Rickus, J. L. Peptide ormosils as cellular substrates. J. Mater. Chem. 2007, 17 (48), 5058−5067. (26) Jedlicka, S. S.; Rickus, J. L.; Zemlyanov, D. Y. Surface analysis by x-ray photoelectron spectroscopy of sol-gel silica modified with covalently bound peptides. J. Phys. Chem. B 2007, 111 (40), 11850− 11857. (27) Maniruzzaman, M.; Morgan, D. J.; Mendham, A. P.; Pang, J.; Snowden, M. J.; Douroumis, D. Drug-polymer intermolecular interactions in hot-melt extruded solid dispersions. Int. J. Pharm. 2013, 443 (1−2), 199−208. (28) Stevens, J. S.; Byard, S. J.; Muryn, C. A.; Schroeder, S. L. Identification of Protonation State by XPS, Solid-State NMR, and DFT: Characterization of the Nature of a New Theophylline Complex by Experimental and Computational Methods. J. Phys. Chem. B 2010, 114 (44), 13961−9. (29) Stevens, J. S.; Byard, S. J.; Seaton, C. C.; Sadiq, G.; Davey, R. J.; Schroeder, S. L. Proton transfer and hydrogen bonding in the organic solid state: a combined XRD/XPS/ssNMR study of 17 organic acidbase complexes. Phys. Chem. Chem. Phys. 2014, 16 (3), 1150−60. (30) Luo, X.; Goh, S. H.; Lee, S. Y.; Tan, K. L. Interpolymer Complexation between Poly(N-methyl-4-piperidinyl Methacrylate) and Acidic Polymers. Macromolecules 1998, 31 (10), 3251−3254. (31) Li, Z.; Abramov, Y.; Bordner, J.; Leonard, J.; Medek, A.; Trask, A. Solid-State Acid-Base Interactions in Complexes of Heterocyclic Bases with Dicarboxylic Acids Crystallography Hydrogen Bond Analysis and 15N NMR Spectroscopy. J. Am. Chem. Soc. 2006, 128 (25), 8199−8210. (32) Wawer, I.; Pisklak, M.; Chilmonczyk, Z. 1H, 13C, 15N NMR analysis of sildenafil base and citrate (Viagra) in solution, solid state and pharmaceutical dosage forms. J. Pharm. Biomed. Anal. 2005, 38 (5), 865−70. (33) Laurence, C.; Brameld, K. A.; Graton, J.; Le Questel, J. Y.; Renault, E. The pKBHX Database:Toward a Better Understanding of Hydrogen-Bond Basicity for Medicinal Chemists. J. Med. Chem. 2009, 52, 4073−4086. (34) Metz, G.; Wu, X.; Smith, S. O. Ramped-amplitude cross polarization in magic-angle-spinning NMR. J. Magn. Reson., Ser. A 1994, 110 (2), 219−27. (35) Pines, A.; Gibby, M. G.; Waugh, J. S. Proton-enhanced nuclear induction spectroscopy. Method for high-resolution NMR of dilute spins in solids. J. Chem. Phys. 1972, 56 (4), 1776−7. (36) Stejskal, E. O.; Schaefer, J.; Waugh, J. S. Magic-angle spinning and polarization transfer in proton-enhanced NMR. J. Magn. Reson. (1969-1992) 1977, 28 (1), 105−12. (37) Dixon, W. T.; Schaefer, J.; Sefcik, M. D.; Stejskal, E. O.; McKay, R. A. Total suppression of sidebands in CPMAS carbon-13 NMR. J. Magn. Reson. (1969-1992) 1982, 49 (2), 341−5. (38) Song, Z.; Antzutkin, O. N.; Feng, X.; Levitt, M. H. Sideband suppression in magic-angle-spinning NMR by a sequence of 5 pi pulses. Solid State Nucl. Magn. Reson. 1993, 2 (3), 143−6. (39) Fung, B. M.; Khitrin, A. K.; Ermolaev, K. An improved broadband decoupling sequence for liquid crystals and solids. J. Magn. Reson. 2000, 142 (1), 97−101. (40) Barich, D. H.; Gorman, E. M.; Zell, M. T.; Munson, E. J. 3Methylglutaric acid as a 13C solid-state NMR standard. Solid State Nucl. Magn. Reson. 2006, 30 (3−4), 125−129. (41) Sakurai, K.; Douglas, E. P.; MacKnight, W. J. Spectroscopic study of an ionic blend made from the acid form of sulfonated polystyrene and poly[ethyl acrylate-co-(4-vinylpyridine)]. Macromolecules 1992, 25 (18), 4506−4510.

J

DOI: 10.1021/acs.molpharmaceut.5b00708 Mol. Pharmaceutics XXXX, XXX, XXX−XXX