European Journal of Pharmaceutical Sciences 97 (2017) 237–246
`
`Contents lists available at ScienceDirect
`
`European Journal of Pharmaceutical Sciences
`
`j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / e j p s
`
`The formation and physical stability of two-phase solid dispersion
`systems of indomethacin in supercooled molten mixtures with different
`matrix formers
`Kristian Semjonov a,⁎, Karin Kogermann a, Ivo Laidmäe a, Osmo Antikainen b, Clare J Strachan b, Henrik Ehlers b,
`Jouko Yliruusi b, Jyrki Heinämäki a
`a Institute of Pharmacy, Faculty of Medicine, University of Tartu, Nooruse 1, 50411 Tartu, Estonia
`b Division of Pharmaceutical Chemistry and Technology, Faculty of Pharmacy, Viikinkaari 5E, 00014, University of Helsinki, Finland
`
`a r t i c l e
`
`i n f o
`
`a b s t r a c t
`
`Article history:
`Received 29 February 2016
`Received in revised form 10 November 2016
`Accepted 16 November 2016
`Available online 24 November 2016
`
`Keywords:
`Solid dispersion
`Two-phase systems
`Supercooled molten mixture
`Indomethacin
`Soluplus®
`Xylitol
`Solid-state stability
`
`Amorphous solid dispersions (SDs) are a promising approach to improve the dissolution rate of and oral bioavail-
`ability of poorly water-soluble drugs. In some cases multi-phase, instead of single-phase, SD systems with
`amorphous drug are obtained. While it is widely assumed that one-phase amorphous systems are desirable,
`two-phase systems may still potentially exhibit enhanced stability and dissolution advantages over undispersed
`systems. The objective of the present study was to understand the solid-state properties of two-phase SDs with
`amorphous drug and their relation to physical stability. Two different types of excipients for SD formation were
`used, one being a polymer and the other a small molecule excipient. The supercooled molten SDs of a poorly
`water-soluble indomethacin (IND) with a graft copolymer, Soluplus® (SOL) and sugar alcohol, xylitol (XYL)
`were prepared. Supercooled molten SDs of IND with SOL were two-phase glassy suspension in which the amor-
`phous drug was dispersed in an amorphous polymer matrix. A short-term aging of the SDs led to the formation
`of glassy suspensions where the crystalline drug was dispersed in an amorphous polymer matrix. These were
`physically stable at room temperature for the time period studied (RT, 23 ± 2 °C), but aging at high-humidity
`conditions (75% RH) recrystallization to metastable α-IND occurred. Interestingly, the SDs with XYL were two-
`phase amorphous precipitation systems in which the drug was in an amorphous form in the crystalline sugar alco-
`hol matrix. The SDs of IND and XYL exhibited fast drug recrystallization. In conclusion, the preparation method of
`two-phase systems via co-melting in association with the rapid quench cooling is a feasible method for the for-
`mulation of poorly water-soluble drugs. The physical stability of these two-phase systems, however, is depen-
`dent on the carrier material and storage conditions.
`
`© 2016 Elsevier B.V. All rights reserved.
`
`1. Introduction
`
`The adoption of new formulation strategies for poorly water-soluble
`drugs are currently of great interest, since the majority of the new drug
`candidates emerging from the drug development pipeline are poorly
`water-soluble compounds (Brough and Williams III, 2013). The com-
`monly used formulation approaches for poorly water-soluble drugs in-
`clude solid dispersions (SDs), and to date the advantages of SDs over
`other strategies are well-documented (Vasconcelos et al., 2007). During
`the past two decades, the fabrication of SDs has been substantially de-
`veloped due to the availability of new types of carriers (vehicles) and
`
`⁎ Corresponding author.
`E-mail addresses: kristian.semjonov@ut.ee (K. Semjonov), karin.kogerman@ut.ee
`(K. Kogermann), ivo.laidmae@ut.ee (I. Laidmäe), osmo.antikainen@helsinki.fi
`(O. Antikainen), clare.strachan@helsinki.fi (C.J. Strachan), henrik.ehlers@helsinki.fi
`(H. Ehlers), jouko.yliruusi@helsinki.fi (J. Yliruusi), jyrki.heinamaki@ut.ee (J. Heinämäki).
`
`new manufacturing technologies (Serajuddin, 1999; Vasconcelos et al.,
`2007; Janssens and Van den Moorter, 2009).
`The SDs for pharmaceutical applications are commonly fabricated by
`melting (fusion), melting solvent, solvent evaporation, extrusion, spray-
`drying or supercritical fluid method. In SDs, the drug is dispersed within
`a carrier (vehicle) matrix (either polymeric or small molecular matrix).
`Different types of SDs can be classified into one-phase or two-phase sys-
`tems (Janssens and Van den Mooter, 2009). In one-phase systems, the
`drug is molecularly dispersed within the carrier and both components
`exist in the amorphous phase (or as a co-crystal if crystalline). Such for-
`mulations are widely considered as the most desired form, however
`one-phase systems limit drug loading, may not always be achievable,
`and the dissolution may be even worse if the intermolecular binding be-
`tween the drug and matrix material is too strong. Poor dissolution has,
`for example, been observed with one-phase indomethacin and
`Soluplus® solid dispersions (Surwase et al., 2015). With these factors
`in mind, one can also consider two-phase systems. In two-phase
`
`http://dx.doi.org/10.1016/j.ejps.2016.11.019
`0928-0987/© 2016 Elsevier B.V. All rights reserved.
`
`Purdue 2023
`Collegium v. Purdue, PGR2018-00048
`
`

`

`238
`
`K. Semjonov et al. / European Journal of Pharmaceutical Sciences 97 (2017) 237–246
`
`systems, the components can be present as independent crystalline
`and/or crystalline/amorphous phase. Thommes et al. (2011) prepared
`crystalline suspension by hot melt extrusion (HME), where crystalline
`drug was molecularly dispersed in crystalline sugar. The formulation
`can act as physical mixture, nevertheless with significant enhancing
`effect on dissolution properties. (Thommes et al., 2011; Thommes,
`2012;)
`The fabrication of SDs by the melting (or fusion) involves heating the
`drug and carrier above their melting or glass transition temperatures,
`followed by mixing and cooling (Janssens and Van den Moorter,
`2009). The properties of SDs can vary greatly depending on the method
`of preparation, process conditions (heating rate, melting temperature,
`cooling method and rate, grinding, etc.), and materials selection (phys-
`icochemical properties of the drug and carrier, drug-carrier interac-
`tions) (Serajuddin, 1999; Thommes et al., 2011). For obtaining the
`one-phase molecularly dispersed system, the complete miscibility of
`the drug and the carrier in the molten form is crucial (Leuner and
`Dressman, 2000). Recently, Thommes et al. (2011) prepared by a hot-
`melt extrusion method a two-phase SD (crystalline suspension), in
`which crystalline drug was dispersed in crystalline sugar. The present
`SDs significantly enhanced the dissolution properties of a poorly
`water-soluble drug (Thommes et al., 2011; Thommes, 2012).
`Since SDs are typically thermodynamically unstable systems, there is
`an increased risk for the spontaneous recrystallization of the drug to
`crystalline polymorph, and consequently, the loss of its favourable per-
`formance (Vasconcelos et al., 2007; Williams et al., 2013). Several stabi-
`lization mechanisms of amorphous drug in the molecular dispersion
`have been described including (1) reduced molecular mobility due to
`increased local viscosity (Aso et al., 2004; Van Den Mooter, 2012), (2)
`intermolecular interaction between the drug and carrier polymer
`(Sethia and Squillante, 2004) and (3) prevention of nuclei formation
`and crystal growth (Marsac et al., 2009). Nevertheless, most pharma-
`ceutical SDs are only partially miscible systems and thus are supersatu-
`rated with respect to the drug in the polymer, and phase separation is
`dependent on the physicochemical properties of the drug, the drug
`weight fraction and the carrier used in the SD (Janssens and Van den
`Moorter, 2009). In summary, the SDs can exist as one-phase or two-
`phase systems, and the physical state/stability of SDs are dependent
`on both materials and the method of preparation (process and technol-
`ogy). Molecular level interactions, steric hindrance and reduced inter-
`face molecular mobility (compared to an air-water interface) all play a
`significant role in the physical stabilisation of SDs (Janssens and Van
`den Moorter, 2009).
`Indomethacin (IND) is a widely-used non-steroidal anti-inflam-
`matory drug, and is a commonly used model drug in the develop-
`ment of amorphous systems for poorly water-soluble drugs (Hart
`and Boardman, 1963; Ewing et al., 2014; Dimensional et al., 2015).
`IND has two crystalline polymorphs (α and γ forms), and an amor-
`phous form (Savolainen et al., 2007; Strachan et al., 2007; Atef et
`al., 2012). Recently, Surwase et al. (2013) discovered and character-
`
`ized several new polymorphs (ɛ, ζ, η) for IND, which were prepared
`
`under different crystallization conditions (Surwase et al., 2013).
`Soluplus® (SOL) is a new amorphous block copolymer consisting of
`hydrophilic repeating units of polyethylene glycol 6000 and hydro-
`phobic segments of polyvinyl acetate and polyvinyl caprolactam.
`Originally, SOL was designed for preparing the solid solutions of
`poorly water-soluble drugs by hot-melt extrusion technology (Basf,
`The Chemical Company, Soluplus, Technical Information, 2010). SOL
`has been studied as a stabilizer of amorphous systems, providing hy-
`drogen bond donors (\\OH) and acceptors (C_O) (Zhang et al.,
`2012; Kogermann et al., 2013; Paaver et al., 2014; Lust et al., 2015).
`Xylitol (XYL) is sugar alcohol, which has gained interest in the uses
`as a small-molecular matrix of SDs for improving the solubility of
`poorly water-soluble drugs (Mummaneni and Vasavada, 1990;
`Sjökvist and Nyström, 1991; Singh et al., 2011). The chemical struc-
`ture of IND, SOL and XYL is shown in Fig. 1.
`
`Fig. 1. Molecular structures of pure materials: a) indomethacin (IND); b) Soluplus® (SOL);
`Xylitol (XYL).
`
`The aim of the present study was to investigate the formation and
`physical state/solid-state stability of SDs of a poorly water-soluble
`drug (IND) in two-phase supercooled molten mixtures with a graft co-
`polymer, Soluplus® (SOL) and a sugar alcohol (XYL). Special attention
`was paid on (1) the physical state of the SDs (i.e., the supercooled mol-
`ten mixtures with either a polymer or small molecular matrix former);
`(2) the solid-state stabilisation mechanisms of the SDs where the drug
`is dispersed within a polymeric or small molecular matrix; and (3) the
`significance of matrix former in the stability of such systems (the hy-
`pothesis is that the present matrix formers can attribute as solid-state
`stabilizers of IND in the present SDs).
`
`2. Materials and methods
`
`2.1. Materials
`
`Indomethacin, IND (1-(p-chlorobenzoyl)-5-methoxy-2-methylin
`dole-3-acetic acid) was purchased from Hawkins (USA) as the γ poly-
`morphic form (γ-IND). The α-IND was obtained by dissolving the
`γ-IND in ethanol at 80 °C and precipitating with Milli-Q-water at
`room temperature (RT, 23 ± 2 °C). The precipitated mass was vac-
`uum filtered and dried overnight (Savolainen et al., 2007). The
`amorphous form of IND was prepared by melting the γ-IND on a
`stainless steel dish at approximately 175 °C and quench cooled
`(QC) by pouring liquid nitrogen on to it. Soluplus® (SOL) graft co-
`polymer and xylitol (XYL) were obtained from BASF (Germany)
`and Yliopiston Apteekki (Helsinki), respectively. All molecular
`structures of pure materials are presented on Fig. 1.
`
`2.2. Methods
`
`2.2.1. Preparation of solid dispersions (SD) and corresponding physical
`mixtures (PM)
`Supercooled co-melted solid dispersion (SD) mixtures were pre-
`pared by melting the γ-IND in a stainless steel dish at approximately
`175 °C and then adding SOL/XYL. The molten mass was solidified by
`pouring liquid nitrogen onto the stainless steel dish. The molten mass
`was then gently ground in a mortar. Corresponding physical mixtures
`(PMs) of drug and carrier with a batch size of 5 g were prepared by
`mixing the components with a pestle in a mortar and using geometric
`dilution. PMs were prepared with SOL/XYL and both IND polymorphs
`(γ-IND and α-IND). All samples (supercooled co-melted SD mixtures
`and PMs) were sieved using a mesh of 450 μm and the fraction passing
`through was used. The drug:carrier weight ratios used in the
`supercooled co-molten SDs and corresponding PMs were 1:3, 1:6 and
`1:9 (w/w). The SDs and PMs with a drug-carrier weight ratio 3:1 and
`1:1 (w/w) were also prepared, but since these SDs showed a poor phys-
`ical stability (early-stage recrystallization within 48 days), these formu-
`lations were not further tested.
`
`Purdue 2023
`Collegium v. Purdue, PGR2018-00048
`
`

`

`K. Semjonov et al. / European Journal of Pharmaceutical Sciences 97 (2017) 237–246
`
`239
`
`2.2.2. Storage conditions and physical stability measurements
`For assessing the physical stability of prepared multi-phase SD sys-
`tems (possible phase transitions and/or polymorphic changes), the
`supercooled molten SDs were stored in open glass petri dishes at 0%
`RH, 50% RH and 75% RH at RT (23 ± 2 °C) for up to 2 months. The sam-
`ples of the aged SDs stored under 50% and 75% RH were taken periodi-
`cally every 24 h during the first week, and subsequently after 10 days,
`14 days, 21 days and 27 days. The samples stored at 0% RH were moni-
`tored for up to 2 months. In addition, the SD of IND with XYL (a drug-
`carrier ratio 1:3) was taken for the short accelerated stability testing
`and stored only at 50% RH/RT up to 4 days.
`
`2.2.3. Scanning electron microscopy (SEM)
`The physical appearance of pure materials, SDs and PMs were inves-
`tigated with a high resolution SEM (FEI Quanta™ 250, Netherlands). The
`samples were sputter coated with a thin platinum layer in nitrogen at-
`mosphere on a carbon tape and examined with a large field low vacuum
`secondary electron detector (LFD) at different magnifications. The SEM
`images were analyzed using ImageJ software (version 1.50b). Martin's
`diameter was used for calculating the average particle size and present-
`ed as the mean ± standard deviation (SD) (n = 60).
`
`2.2.4. Solid state characterisation
`
`variable-temperature XRPD (VT-XRPD) experiments were conducted
`with pure materials (γ-IND, amorphous IND, XYL, QC XYL) and the
`supercooled molten SDs in 1:3 weight ratio (IND:SOL and IND:XYL).
`XRPD (VT-XRPD) was used for tracking the heat induced solid-state
`transitions and understanding the miscibility of the systems and
`revealing the presence of multi-phase system. Data were collected
`from 8 to 22.5° 2θ, with a step size of 0.0195° 2θ. The heating process
`started from RT at a rate of 0.5 °C/min and the first scan was taken at
`30 °C and other scans were measured at intervals of 10 °C up to
`150 °C for IND:SOL systems and after every 5 °C until 160 °C for all sys-
`tems. The heated samples were then cooled down to 30 °C and
`remeasured. The total scan time at each investigated temperature was
`ca 16 min. Amorphous IND and XYL were prepared prior inserting the
`sample to the VT-XRPD.
`
`2.2.4.2. Fourier transform infrared spectroscopy (FT-IR). A FT-IR spectrom-
`eter (Vertex 70, DTGS detector, Bruker, Germany) with a MIRacle™
`(PIKE Technologies, Inc., US) single reflection attenuated total reflection
`(ATR) crystal was used for collecting the spectra. The spectral region
`from 650 to 4000 cm−1 was collected as an accumulation of 64 scans.
`A total of 3 spectra were taken from each sample and presented as an
`average. The spectra were analyzed and processed in OriginPro software
`(version 9.1).
`
`2.2.4.1. X-ray powder diffraction (XRPD). The samples were studied with
`XRPD using the Bruker D8 Advanced diffractometer (Bruker AXS GmbH,
`Germany) with Cu K-alpha1 radiation λ = 1.5406 Å, operated at 40 kV
`and 40 mA. A Vario1 focusing primary monochromator, two 2.5° Soller
`slits and LynxEye line detector were used. Data were collected in θ–2θ
`geometry in the range of 5–30° 2θ, with the step size of 0.0195° 2θ
`from 5 to 30° 2θ and a total counting time of 49.8 s per step. The
`
`2.2.5. Thermal analysis
`Differential scanning calorimetry (DSC) thermograms were obtain-
`ed with a DSC823e (Mettler Toledo AG, Switzerland) and analyzed
`using STARe software (version 9.0). Samples with the weight of 6–
`7 mg were encapsulated into aluminium pans with two holes in the
`lid and heated with the isothermal step at 30 °C for 5 min up to
`200 °C. The heating rate of 20 °C/min in the inert nitrogen atmosphere
`
`Fig. 2. Scanning electron microscopy (SEM) micrographs of the solid dispersions (SDs) and respective physical mixtures (PMs) of γ-indomethacin (IND) with carrier materials. Key: A. The
`1:3 SD of IND with Soluplus® (SOL) (drug:polymer); B. The 1:3 PM of γ-IND with SOL (drug:polymer); C. The 1:3 SD of IND with xylitol (XYL) (drug:sugar alcohol); D. The 1:3 PM of γ-IND
`with XYL (drug:sugar alcohol). Magnification of SEMs ×100. Numbers in the up right corner indicate the average particle size ± SD.
`
`Purdue 2023
`Collegium v. Purdue, PGR2018-00048
`
`

`

`240
`
`K. Semjonov et al. / European Journal of Pharmaceutical Sciences 97 (2017) 237–246
`
`with a flow rate of 50 ml/min was used. Indium was used for the valida-
`tion of the DSC system. Modulated temperature DSC (MT-DSC) was
`used to determine the glass transition temperature (Tg) of SDs from 0
`to 180 °C/min at the heating rate of 2 °C/min and pulse height of
`0.5 °C and 30 s pulse width. Conventional DSC was not sensitive enough
`to separate overlapping events at Tg region.
`Hot-stage microscopy (HSM) was performed with a Mettler Toledo
`F82 (Switzerland) and visualized with a polarized light microscope
`(Leica DM/LM, Germany) under 5× magnification (HSM-PLM). The
`video and images were captured using full HD-video and 13.0 mega-
`pixel equipped camera with full auto focus function. The temperature
`range and heating rate were chosen as in the DSC experiments from
`30–200 °C and 20 °C/min. All temperatures are given as onset tempera-
`tures unless mentioned otherwise. The sample was run twice for pure
`materials and physical mixtures, once for SDs.
`
`2.2.6. Data analysis
`ImageJ (version 1.50b) software was used to measure the particle
`size (Martin's diameter) and particle size variation. ChemBioDraw
`Ultra (version 13.0) drawing program was used for structure generation.
`
`3. Results and discussion
`
`3.1. Particle size, shape and surface morphology
`
`Comparison of the SEM results of pure materials with those obtained
`with the supercooled molten SDs can help to indicate the influence of
`the method of preparation and carrier on the particle size, shape and
`
`surface morphology. All mixtures (1:3, 1:6, 1:9) were initially prepared,
`but only 1:3 was tested further. It was confirmed that the γ-IND exhib-
`ited flat, rectangular and sharp edged tiny crystals with the average SEM
`particle size of 51 ± 20 μm. Both stabilizing carrier materials, SOL graft
`copolymer and XYL, presented oblong shape, smooth surfaced particles
`with much larger average particle size (285 ± 68 μm for SOL and 235 ±
`89 μm for XYL) (Appendix 1). The average particle size of IND
`supercooled molten mixtures with SOL and XYL as well as the corre-
`sponding PMs was approximately 70 μm (Fig. 2). The SEM micrographs
`of the 1:3 SDs suggest that the γ-IND particles partially melted inside or
`were attached on the surface of SOL graft copolymer (Fig. 2A). The cor-
`responding micrographs of SDs containing IND and XYL at the weight
`ratio of 1:3 showed similar results, thus indicating that the drug parti-
`cles were dispersed in a XYL matrix (Fig. 2C).
`
`3.2. Physical solid-state properties
`
`The XRPD patterns of pure materials (γ-IND, α-IND, SOL and XYL),
`SDs and PMs are shown in Fig. 3. Characteristic XRPD reflections of all
`materials are summarized in Table 1. These results are in good accor-
`dance with the literature (Crowley and Zografi, 2003; Masuda et al.,
`2006; Bogdanova et al., 2007; Ueda et al., 2014). As seen in Fig. 3, pure
`SOL was XRPD amorphous with a broad halo across the entire pattern,
`XYL exhibited strong crystalline reflections.
`The XRPD patterns of the freshly prepared supercooled molten
`amorphous SDs and their corresponding PMs with γ-IND and SOL are
`shown in Fig. 3A–B. The absence of characteristic crystalline reflections
`of γ-IND in the supercooled molten SDs of IND with SOL at all
`
`Fig. 3. X-ray powder diffraction (XRPD) patterns of γ and α-indomethacin (IND), and the solid dispersions (SDs) and physical mixtures (PMs) of γ-IND with carrier materials. A. SDs with
`Soluplus® (SOL) IND:SOL in different weight ratios (1:3, 1:6, 1:9 (drug:polymer)), and B. its respective PMs; C. SDs with Xylitol (XYL) IND:XYL in different weight ratios (1:3, 1:6, 1:9
`(drug:polymer), and D. its respective PMs. In addition, pure material XRPD patterns (γ-IND, α-IND, SOL, XYL) are shown as references. All patterns are normalized. The (*) and (#)
`marks denote the reflections unique to γ-IND and α-IND, respectively.
`
`Purdue 2023
`Collegium v. Purdue, PGR2018-00048
`
`

`

`K. Semjonov et al. / European Journal of Pharmaceutical Sciences 97 (2017) 237–246
`
`241
`
`Table 1
`Summary of the physical storage stability of the supercooled molten SDs stored at 0% RH / RT; 50% RH / RT and 75% RH / RT for 1–8 weeks. The results are based on the XRPD, ATR-FTIR and
`DSC analyses (more detailed data are given in Appendix 3).
`
`Crystalline IND/storage conditions
`
`XRPD/
`
`DSC/°C
`
`ATR- FT-IR/cm−1
`
`Pure materials
`γ-IND
`α-IND
`
`11.7°, 19.6°, 21.8°, 26.6°
`8.4°, 14.4°, 18.5°, 22.0°
`
`Pure amorphous IND
`
`No reflections
`
`159.8 ± 0.4 (Tm)
`153.5 ± 0.5 (Tm)
`
`44.8 ± 0.8 (Tg)
`131.7 ± 0.5 (Tc)
`157.5 ± 0.1 (Tm)
`
`1714 cm−1, 1689 cm−1
`1734 cm−1, 1689 cm−1,
`1679 cm−1, 1649 cm−1
`1707 cm−1, 1680 cm−1
`
`Stored samples
`0%RH/RT
`IND:SOL 1:3
`
`50%RH/RT
`IND:SOL 1:3
`
`IND:XYL 1:3
`
`75%RH/RT
`IND:SOL 1:3
`
`NA- non-applicable
`
`First IND reflections detected after 1 week
`21.9°
`(γ-IND/α-IND)
`
`Tm detected after 2 months
`
`Changes in spectra detected after 2 months
`1680 cm−1 N 1686 cm−1
`
`152.5 ± 0.9
`(α-IND)
`
`NA
`
`First IND reflections detected after 1 week
`
`Tm detected after 1 week
`
`21.9°
`(γ-IND/α-IND)
`XYL reflections
`14°, 14.6°,17,6°,19,8°
`First IND reflections
`Detected after 4 days
`11.7°
`(γ-IND)
`
`155.9 ± 0.4
`(α-IND)
`Tm of XYL
`89.0 ± 1.4
`Tm of IND
`152.2 ± 0.1
`
`(α-IND)
`
`Changes in spectra detected after 2 weeks
`1682 cm−1N1688 cm−1
`NA
`
`NA
`
`First IND reflections detected after 1 week
`
`21.9°
`
`(γ-IND/α-IND)
`
`Tm detected after 1 week
`155.9 ± 0.4
`
`(α-IND)
`
`Changes in spectra detected after 2 and 10 days
`
`1632 cm−1N1612N1618
`1683 cm−1N1685N1691
`1732 cm−1N1722N1718
`NA
`
`drug:polymer weight ratios (1:3, 1:6, 1:9) indicated the formation of
`amorphous IND (Fig. 3A). The XRPD patterns of PMs of γ-IND and SOL
`showed characteristic crystalline reflections of γ-IND (Fig. 3B). As ex-
`pected, due to a dilution effect, the intensity of the reflections decreased
`as the amount of SOL in the PMs was increased. However, in spite of the
`reduced intensity the characteristic reflection of γ-IND at 11.7° 2θ was
`still observed at all weight ratios.
`The XRPD patterns for the supercooled molten SDs and their corre-
`sponding PMs with γ-IND and XYL are shown in Fig. 3C–D. Interestingly,
`the XRPD patterns at all tested drug:polymer weight ratios (1:3, 1:6,
`1:9) showed only crystalline XYL reflections with some amorphous
`halo. It is evident that the SDs containing IND and XYL formed two
`separate phases: an amorphous phase composed largely of IND and a
`crystalline phase composed of XYL (Fig. 3C). While the molecular size
`and arrangement of these molecules in XYL would likely preclude an
`(significant) incorporation of IND molecules into the crystalline phase
`(i.e. eutectics), one cannot exclude the presence of a certain amount
`XYL molecularly dispersed within the amorphous phase of the matrix,
`which is presumably IND rich. In the corresponding PMs, the character-
`istic reflections for both γ-IND and XYL were detected in all weight
`ratios.
`
`3.3. Thermal behaviour
`
`Understanding of the thermal phase behaviour and polymer-drug
`interactions of the two-phase binary system is of key importance for
`the selection of the most suitable carrier. Thermal stability of the fresh
`supercooled molten SDs at the weight ratio of 1:3 (drug:polymer) as
`well as drug-polymer miscibility was further studied using DSC, MT-
`DSC, HSM and VT-XRPD. As a reference, the thermal behaviour of all
`pure materials and corresponding PMs was studied. All thermal param-
`eters for different samples are depicted in Table 1, and the thermal
`
`behaviour of pure materials was in agreement with the literature
`(BASF, 2012; Bogdanova et al., 2007; Singh et al., 2011).
`
`3.3.1. Indomethacin-soluplus® (IND-SOL) solid dispersion (SD)
`In line with the XRPD results, the absence of a melting endotherm on
`the DSC thermograms of freshly prepared supercooled SDs with
`
`Fig. 4. Modulated-temperature differential scanning calorimetry (MT-DSC) thermograms
`of the fresh solid dispersions (SDs) of indomethacin (IND). Key: (A) SDs with Xylitol (XYL)
`IND:XYL in weight ratio of 1:3 (drug:polymer), (B) SDs with Soluplus® (SOL) IND:SOL in
`weight ratio of 1:3 (drug:polymer).
`
`Purdue 2023
`Collegium v. Purdue, PGR2018-00048
`
`

`

`242
`
`K. Semjonov et al. / European Journal of Pharmaceutical Sciences 97 (2017) 237–246
`
`IND:SOL confirmed the presence of XRPD amorphous systems (Table 1).
`Upon heating no heat-induced recrystallization was observed in the
`DSC thermograms. Due to the broad endothermic artifact from 30–
`100 °C in the conventional DSC, the Tg-s of amorphous IND and SOL
`were indistinguishable and it was impossible to understand the pres-
`ence or absence of multiple phases of this drug-carrier system. In
`order to investigate this further and unambiguously identify the Tg-s
`as well as classify the SDs according to their structure (one or multi-
`phase systems), MT-DSC experiments were conducted. Hence the MT-
`DSC with supercooled SDs of IND:SOL (1:3) system showed a two-
`phase SDs system, where on a reversing thermogram two Tg-s were
`clearly observed, one at 40.3 °C for amorphous IND and a second at
`approximately 88 °C for SOL (Fig. 4). Clear shift in Tg-s of IND-SOL SDs
`compared to pure materials confirmed some level of drug-polymer
`miscibility important for the stability of amorphous systems (Table 1).
`HSM-PLM was also used to visualize the DSC/MT-DSC results and get
`more insight into the phenomena occurring at the interfaces within SD
`and PM mixtures (Fig. 5). The supercooled molten IND:SOL mixtures
`were heterogeneous and consisted of darker particles and more yellow-
`ish particles, hence no birefringence was observed confirming the pres-
`ence of two-phase amorphous SDs system (Fig. 6A). Upon heating no
`crystals were observed in these systems, only the drug dissolution into
`a polymer melt was detected. Similar phenomenon has been described
`by Fini et al. (2008) with ibuprofen and diclofenac, when they
`
`investigated the interactions between these two drugs with different
`types of polyvinylpyrrolidones (PVP) in PMs and SDs (Fini et al., 2008).
`Interestingly, according to the VT-XRPD results, amorphous SDs of
`IND and SOL started to recrystallize as α-IND at 70 °C, which was also
`noticed for amorphous IND alone (Fig. 5A vs). The most pronounced re-
`flections of α-IND were detected at 100 °C (denoted with # on Fig. 5B).
`This confirmed that SOL is not able to entirely prevent the heat-induced
`phase separation and recrystallization of α-IND. Differences between
`VT-XRPD and other thermal techniques (DSC, MT-DSC, HSM-PLM) can
`be explained with much longer measurement time in the VT-XRPD ex-
`periments, which may induce more pronounced changes in the sample
`upon heating. (Mirza et al., 2006)
`
`3.3.2. Indomethacin-xylitol (IND-XYL) solid dispersion (SD)
`The DSC and MT-DSC thermograms of supercooled molten IND:XYL
`SDs showed two endothermic events (one for XYL at 89 °C and one for
`α-IND at 152 °C) and no change in Tg compared to pure materials, thus
`suggesting the presence of a poorly miscible two-phase drug-carrier
`system (Table 1). Since the recrystallization of IND was also detected
`at 124 °C, it is evident that XYL was not able to prevent the heat-induced
`solid-state changes of IND in these SDs. Furthermore, also VT-XRPD
`results using IND:XYL SDs (Figs. 5C and D) verified the DSC/MT-DSC
`results showing a two-phase system (Fig. 4).
`
`Fig. 5. Variable temperature X-ray powder diffraction patterns (VT-XRPD) (all patterns are normalized) of indomethacin (IND) fresh solid dispersions (SDs): A–B. SDs with Soluplus®
`(SOL) IND:SOL in weight ratio of 1:3 (drug:polymer) and C-D. SDs with Xylitol (XYL) IND:XYL in weight ratio of 1:3 (drug:polymer) in comparison to XRPD patterns of untreated
`crystalline IND forms, and pure stabilizer (SOL and/or XYL). Heating from 30 °C up to 160 °C. The (*) and (#) marks denote the reflections unique to γ-IND and α-IND, respectively.
`
`Purdue 2023
`Collegium v. Purdue, PGR2018-00048
`
`

`

`K. Semjonov et al. / European Journal of Pharmaceutical Sciences 97 (2017) 237–246
`
`243
`
`3.3.3. Physical mixtures (PMs) of IND:SOL and IND:XYL
`As expected the PMs of both systems (IND-SOL and IND-XYL)
`showed the characteristic thermal behaviour of two-phase systems
`(Table 1). The latter was also verified by HSM-PLM and VT-XRPD.
`HSM-PLM revealed that upon heating these two PMs systems showed
`clear distinguishable boundary of a drug containing phase (yellow)
`and transparent carrier phase (Fig. 6B and C). The corresponding
`drug-carrier incompatibility has been reported also with the SDs of ibu-
`profen and XYL (Greenhalgh et al., 1999).
`
`3.4. Drug-carrier interactions
`
`Since thermal analyses revealed that IND-SOL SD systems have some
`mixing between the components, the ATR-FTIR spectroscopy was used
`to examine further the possible molecular-level interactions between
`the drug and carrier in SD systems. The ATR FTIR spectra for pure mate-
`rials, the supercooled co-melted SDs systems of IND with SOL/XYL (in
`different drug-polymer ratios) and their corresponding PMs with γ-
`IND are shown in Fig. 7. The pure material spectra matched with the re-
`ported spectra in the literature (Bahl and Bogner, 2006; Strachan et al.,
`2007; Lan et al., 2010; Shamma and Basha, 2013).
`The ATR-FTIR spectra of IND-SOL amorphous SDs at all drug:polymer
`ratios showed strong absorption peaks at 1630 cm−1 and 1732 cm−1
`(Fig. 7A) which belong to SOL molecular vibrations (carbonyl groups
`of block sequences –polyvinylacetate O\\C_O\\CH and caprolactam
`C_O\\N vibrations, respectively). Since the pure amorphous IND also
`showed shoulder band at 1734 cm−1 and strong vibrations at
`1689 cm−1 and 1679 cm−1 (data not shown), the spectral overlap
`with strong SOL vibrations covered several amorphous IND peaks. How-
`ever, as evidenced by the MT-DSC results (shift in Tg of IND-SOL SDs
`compared to pure materials), some hydrogen-bond formation at the
`molecular level may have occurred between SOL and IND molecules
`during SD preparation (peak observed at 1680 cm−1). Zhang et al.
`(2013) described similar interaction with the SDs composed of
`itraconazole and SOL. The specific spectral regions of interest for the
`IND and XYL related interactions in SDs and PMs are at 1714 cm−1
`and 1689 cm−1 (both C_O of IND) and 3421 cm−1, 3359 cm−1, and
`3293 cm−1 (OH groups of XYL). As expected based on the thermal anal-
`ysis results, the FT-IR spectra showed limited compatibility between
`those two components since instead of the peak shift only intensity
`changes were detected in carbonyl and hydroxyl group vibrations. Sev-
`eral studies have shown limited ability of XYL to form chemical interac-
`tion with the active ingredients (Mummaneni and Vasavada, 1990;
`Sjökvist and Nyström, 1991; Suzuki and Sunada, 1997; Madgulkar et
`al., 2015).
`
`3.5. Physical stability during storage under different conditions
`
`According to the literature, molecularly dispersed single-phase SDs
`are the most stabilized systems against crystallization (Williams et al.,
`2013). Hence, the degree of miscibility between the drug and polymer
`is important for the formation of a physically stable amorphous system.
`Aging under different humidity conditions can change the physical
`solid-state properties and performance of two-phase SDs systems. In
`our study, the physical stability of the two-phase systems was investi-
`gated to understand their potential stabilizing mechanisms. Table 1
`summarizes the physical storage stability results (XRPD, DSC and ATR-
`FTIR) obtained with the aged supercooled molten amorphous SDs of
`IND:SOL and IND:XYL. Interestingly different analytical methods re-
`vealed somewhat different level of physical stability in these two-
`phase SD systems, thus giving further proof that complementary analy-
`sis methods are needed in order to obtain complete insight into the
`complex recrystallization behaviour.
`According to the XRPD, the supercooled molten SDs (with a IND:SOL
`weight ratio of 1:3