Permeability of glibenclamide through a PAMPA membrane: the effect of co-amorphization

UEF//eRepository

DSpace https://erepo.uef.fi

Rinnakkaistallenteet Terveystieteiden tiedekunta

2018

Permeability of glibenclamide through

a PAMPA membrane: the effect of co-amorphization

Ruponen, Marika

Elsevier BV

Tieteelliset aikakauslehtiartikkelit

© Elsevier B.V.

CC BY-NC-ND https://creativecommons.org/licenses/by-nc-nd/4.0/

http://dx.doi.org/10.1016/j.ejpb.2018.06.007

https://erepo.uef.fi/handle/123456789/6902

Downloaded from University of Eastern Finland's eRepository

1

Permeability of glibenclamide through a PAMPA membrane: the effect of co-amorphization

Marika Ruponen^a, Maiju Visti^a, Rami Ojarinta^a, Riikka Laitinen^a,*

aSchool of Pharmacy, University of Eastern Finland, P.O.Box 1627, 70211, Kuopio, Finland

*Corresponding author:

Riikka Laitinen School of Pharmacy

University of Eastern Finland P.O.Box 1627

70211 Kuopio FINLAND

Phone:+358 50 5695 303 e-mail: riikka.laitinen@uef.fi

2 Abstract

Co-amorphous systems are an attractive alternative for amorphous solid polymer dispersions in the formulation of poorly soluble drugs. Several studies have revealed that co-amorphous formulations can enhance the dissolution properties of poorly-soluble drugs and to stabilize them in the

amorphous form. However, the interplay between the drug dissolution rate, drug supersaturation and different co-formers on membrane permeability of the drug for co-amorphous formulations remains unexplored.

By using side-by-side chambers, separated by a PAMPA (parallel artificial membrane permeability assay) membrane, we were able to simultaneously test dissolution and passive membrane permeability of the co-amorphous combinations (1:1 molar ratio) of a poorly soluble drug

glibenclamide (GBC) in combination with two amino acids, either serine (SER) or arginine (ARG). In addition, a known passive permeability enhancer sodium lauryl sulfate (SLS) was included in the co- amorphous mixtures at two concentration levels. The mixtures were also characterized with respect to their solid-state properties and physical stability.

It was found that GBC mixtures with ARG and SLS had superior dissolution and physical stability properties which was attributable to the strong intermolecular interactions formed between GBC and ARG. These formulations also had optimal permeability properties due to their high

concentration gradient promoting permeation and possible permeation enhancing effect of the co- formers ARG and SLS. Thus, simultaneous testing of dissolution and permeation through a PAMPA membrane may represent a simple and inexpensive tool for screening the most promising

amorphous formulations in further studies.

Keywords: co-amorphous, solubility, dissolution, permeability, PAMPA, stability

3 1. Introduction

Since an estimated 40% of approved drugs and nearly 90% of the developmental pipeline drugs are poorly soluble molecules [1], there is a clear need for formulation strategies that ensure efficient drug dissolution, subsequent absorption and good distribution to the target in the body. Co- amorphous systems are an attractive alternative for amorphous solid polymer dispersions in the formulation of poorly soluble drugs. Co-amorphous systems can be classified into a solid-dispersion subtype of glass solutions along with polymer-based or mesoporous silica-based systems. These are a combination of two or more low molecular weight components that form a homogeneous

amorphous single-phase system [2]. Co-amorphous combinations of either an active molecule and an excipient (e.g. an amino acid) or two active drug compounds have been shown to represent interesting candidates for combination therapy (drug-drug) or more general formulation options (drug-excipient). These approaches been found to enhance the dissolution properties of poorly- soluble drugs compared to their crystalline counterparts or individual amorphous forms, and to stabilize them in an amorphous form [2-5]. Often dissolution studies investigating co-amorphous systems have been conducted under sink conditions, but in order to investigate their

supersaturation-abilities, non-sink dissolution testing is important. Supersaturation maintained over a sufficiently long time may offer a bioavailability advantage over crystalline drugs by increasing the driving force for absorption [6, 7].

There is one example of a co-amorphous system showing a long-lasting supersaturation ability in buffer and biorelevant media, i.e. combinations of glibenclamide (GBC) with two amino acids, serine (SER) and/or threonine (THR) [8]. GBC-SER and GBC-THR combinations at a 1:1 molar ratio and 1:1:1 GBC-SER-THR formed homogenous amorphous blends after cryomilling as shown by a single glass-transition (Tg) value obtained for the mixtures [9]. The formation of co-amorphous GBC- SER provided a physical stability advantage over an amorphous GBC.

The permeability properties of a drug compound through artificial or cell membranes are generally determined by dissolving the drug into the testing medium at a predetermined concentration and

4

then measuring the Papp, i.e. the apparent permeability coefficient [10-12]. However, this experimental setup does not determine the interplay between drug (or formulation) dissolution properties and permeation. This would be particularly important for amorphous formulations, which may display a spring and parachute (supersaturation with varying level and durations) effect to different extents and then undergo subsequent recrystallization during dissolution, which may affect their flux across the membrane [13-15]. Thus, experimental setups, capable of measuring drug dissolution and permeation have been developed; in these, both different artificial membranes [16- 19] and cell layers [20-24] have been used as permeation barriers. However, as far as we are aware, simultaneous drug dissolution and permeation testing has not been examined before as a way of evaluating the interplay between the drug dissolution rate, drug supersaturation and different co- formers on membrane permeability of the drug for co-amorphous formulations.

In this study, we used side-by-side chambers, separated by a PAMPA (parallel artificial membrane permeability assay) membrane to simultaneously test dissolution properties of co-amorphous formulations in supersaturated conditions and the passive membrane permeability of the model drug GBC, which has poor solubility but high intestinal permeability [9]. Amorphous GBC, co- amorphous GBC-SER at a 1:1 molar ratio, GBC- l-arginine (ARG) at a 1:1 molar ratio and all these three combined with sodium lauryl sulphate (SLS) were processed by cryomilling and used as model formulations. ARG was combined with GBC as previously it has been observed to enhance GBC dissolution [25]. SLS was incorporated into the co-amorphous formulations, since it is known to act as an absorption enhancer in vivo by modulating the cellular tight junctions below its critical micelle concentration (CMC, i.e. approx. 8 mmol, [26]) where the drug is in a molecularly dissolved state, i.e.

not associated with micelles [27-29]. Physical characterization and stability studies were conducted with the co-amorphous combinations, except for GBC-SER which has been investigated in a previous study [9].

5 2. Materials and methods

2.1. Preparation of amorphous materials

Physical mixtures (PMs) of glibenclamide (GBC, 494.0 g/mol, Hangzhou DayangChem Co., Ltd , Hangzhou City, China) and L-arginine (ARG, 174.2 g/mol, Sigma-Aldrich ApS, St. Louis, USA) or L- serine (SER, 105.1 g/mol, Hangzhou DayangChem Co., Ltd, Hangzhou City, China) at molar ratios of 1:1 were prepared (Table 1 in Supplementary material). Additionally, low (l) and high (h) amounts of sodium lauryl sulfate (SLS, 288.4 g/mol, Sigma-Aldrich ApS, Brøndby, Denmark), both producing concentrations below the CMC in a side-by-side dissolution chamber, were mixed together with pure GBC as well as with the GBC-ARG and GBC-SER mixtures (Table 1 in Supplementary material). The chemical structures of the starting materials are shown in Fig. 1. In the co-amorphous blends, a total mass of 500 mg of each of the abovementioned mixtures (and also GBC alone) was prepared and transferred to 25 ml milling jars with two 15 mm stainless steel balls. Cryomilling (CM) was

performed in an oscillatory ball mill (Mixer Mill MM400, Retsch GmbH & Co., Haan, Germany) at 30 Hz for 60 min. The milling jars were cooled with liquid nitrogen for 2 min prior to and at 10 min intervals during the milling. After milling, the chambers were first allowed to reach the

environmental temperature in a desiccator before opening, preventing moisture adsorption. The amorphous mixtures were stored at 4°C/0% relative humidity (RH) until analysis (i.e. over P2O5).

2.2. X-Ray powder diffraction

X-ray powder diffraction (XRPD) measurements were performed with a Bruker D8 DISCOVER system (Bruker AXS GmbH, Karlsruhe, Germany) using Cu Kα radiation with λ = 1,5418 Å and a motorized slit. The samples were analyzed at 40 kV and 40 mA from 5 to 35 º 2θ using a scanning speed of 0.120s/step and a step size of 0.013°. The scattered radiation was collected with a 1D LYNXEYE detector fully open.

6 2.3. Differential scanning calorimetry

Thermograms of the formulations were measured with a Mettler Toledo DSC823^e (Mettler Toledo, Schwerzenbach, Switzerland) equipped with an intercooler (Mettler Toledo, METT-FT900 Julabo, Switzerland) and an autosampler (TS080IRO Sample Robot, Mettler Toledo, Schwerzenbach, Switzerland). A nitrogen flow of 50 ml/min was used during the measurements. Temperature and heat flow calibrations were carried out with indium, lead, zinc, and highly purified water standards.

Approximately 10 mg of each powder sample was weighed with a Sartorius SE2 microbalance (Sartorius AG, Göttingen, Germany) and analyzed in 40 µl aluminium pans (Mettler Toledo,

Switzerland) which were sealed with a pierced lid. The measurements were performed in duplicate.

The samples were rapidly cooled down to -50 °C and kept there for 15 min. Then the temperature was raised at 10 °C/min until 150 °C was reached. STARe software (Mettler Toledo, Schwerzenbach, Switzerland) was used for data collection (Tg, midpoint) and melting temperature (Tm, onset).

2.4. Fourier-transform infrared spectroscopy

The Fourier-transform infrared (FTIR) spectra were obtained using a Nicolet Nexus 870 FTIR spectrometer (Thermo Electron Corp, Madison, WI, USA) equipped with an attenuated total reflectance (ATR) accessory (Smart Endurance, single-reflection ATR diamond composite crystal).

Spectra were collected over the range 600-4000 cm^-1 using a resolution of 4 cm^-1 and taking an average of 64 scans per sample.

2.5. Stability studies

The amorphous samples were stored under the following conditions: 4 °C/0% RH, ambient temperature (approximately 22°C)/ 60%RH, and 40°C/0% RH. RH of 60% was obtained with a saturated NaBr solution and RH 0% with solid P2O5. The samples were analyzed regularly with XRPD and FTIR, as described above, until the onset of recrystallization could be observed.

7 2.6. Simultaneous dissolution/permeation testing 2.6.1. Dissolution/permeation studies

PAMPA membranes were prepared by moistening 0.1 µm Durapore PVDF (polyvinylidene fluoride) - membrane filters (Millipore Corporation, Bedford, MA USA) with 10 µl of a 10% (m/V) solution of L- α-phosphatidylcholine in dodecane (both from Sigma-Aldrich, St. Louis, USA) (according to [30]).

Immediately after membrane preparation, they were placed between a pair of jacketed 3ml side-by- side diffusion cells with an effective permeation area of 0.64 cm² (PermeGear Inc., Hellertown, PA, USA).

The diffusion cells were attached to a circulating water bath (M3, Lauda, Köningshofen, Germany) equipped with a thermostat (MS, Lauda, Köningshofen, Germany) with a temperature setting of 37

°C. A powdered sample corresponding to 10 mg of GBC was transferred into the donor (dissolution testing) cell through the sampling port. Then, both the donor and acceptor (permeation testing) cells were filled simultaneously with 3 ml of USP phosphate buffer, pH 7.2. Both cells were stirred with the cylindrical magnetic stir-bars at a fixed speed of 500 rpm (H-3 stirrer, PermeGear Inc.,

Hellertown, PA, USA).

During dissolution testing, aliquots of 300 µl were withdrawn from the donor compartment at pre- determined time points (15, 30, 45, 60, 75, 90, 120, 150, 180, 210, 240, 270, 300, 330 and 360 min) and filtered immediately through a 30 mm polyethersulfone (PES) 0.22 µm diameter membrane filter (Guangzhou JET Bio-Filtration Co., Ltd, Guangzhou, China). Simultaneously, the acceptor cell was emptied completely, but its contents were not analyzed at this stage. The volumes removed from the donor and acceptor compartments were replaced with the buffer.

Analysis of drug permeation was conducted as described above, but without taking samples from the donor side in order not to disturb the dynamic dissolution/permeation process. Both the

8

dissolution and permeability experiments were run in triplicate for the PMs and six parallel measurements were conducted for the amorphous samples.

2.6.2. High performance liquid chromatography (HPLC)

GBC concentrations were quantified using HPLC. The Gilson HPLC equipment consisted of a Gilson 321 pump, a Gilson UV–vis 151 detector (Both from Gilson Inc., Middleton, WI, USA), a Gilson 234 autoinjector (Gilson, Roissy-en-France, France), and a reversed phase column (Phenomenex Gemini NX 5u C18 110A, 250x4,60 mm, USA) with a precolumn. The mobile phase consisted of acetonitrile (ACN, VWR Chemicals, Fontenay-sous-Bois, France) (70%), H2O (Milli-Q® water purification system (Merck Millipore, Darmstadt, Germany) (30%) and trifluoroacetic acid (TFA, Alfa Aesar GmbH & Co., Germany) (0.1%) and the flow rate was adjusted to 1.2 ml/min. GBC was detected at a wavelength of 225 nm and its retention time was 4.0 min. The results were analyzed using Gilson Unipoint software (version 3.01, Gilson Inc., Middleton, WI, USA).

For analysis, the samples from dissolution studies of amorphous compounds were diluted 7:3 with ACN and if necessary, further by 1:2. Standard series were prepared in ACN/water 70:30 (V/V) over the range of 0.1-100 µg/ml.

2.6.3. Data evaluation and statistical analysis

Area under curve (AUC) values for dissolution and permeation were calculated from the total amount of dissolved and permeated drug against time curves (between 0-360 min) using Origin Pro 2015 (64-bit) Sr2 by linear trapezoidal integration (OriginLab Corporation, Northampton, MA, USA).

Differences were considered significant with p-values <0.05 (95% confidence level).

9 3. Results and discussion

3.1. Preparation and characterization of the co-amorphous formulations 3.1.1. X-ray diffraction and Fourier-transform infrared measurements

GBC alone and GBC-SER in a 1:1 molar ratio are known to transform to an amorphous form upon cryomilling [9]. The X-ray diffractograms of the other cryomilled combinations (GBC-SLSl, GBC-SLSh, GBS-SER-SLSl, GBC-SER-SLSh, GBC-ARG, GBC-ARG-SLSl and GBC-ARG-SLSh) and crystalline starting materials are shown in Fig. 2. As can be observed from Fig. 2b, all the cryomilled combinations were X-ray amorphous after cryomilling.

The FTIR measurements showed typical changes for GBC after amorphization (Fig. 3a.). These included merging and shifting of the N−H stretching bands associated with the amide moiety and urea group (to 3375 cm^-1 in amorphous GBC), shifting of the carbonyl stretch of the benzoyl carbonyl group (to around 1627 cm⁻¹) with the appearance of a shoulder at 1656 cm⁻¹, indicative of the formation of a less stable imidic acid tautomer of GBC [9]. When a co-amorphous mixture was formed with SLS, changes in the spectra were minor when compared to amorphous GBC alone, i.e.

only the shoulder at 1656 cm⁻¹ in amorphous GBC was slightly shifted to 1650 cm^-1 in the GBC-SLS- mixtures. With the formation of GBC-SER (Fig.3b.), the changes were similar to those previously reported, most importantly, there was the disappearance of the imidic acid shoulder indicating that there had been an inhibition of the conversion of the amide group of GBC to the thermodynamically less stable form [9,31]. The same phenomenon was observed when SLS was mixed with GBC and SER.

The formation of GBC-ARG evoked significant changes compared to the spectra of amorphous GBC and crystalline ARG (Fig.3c.). In the wavenumber region of approx. 1500-1700 cm^-1, peaks arising from the carbonyl groups of both GBC and ARG and the amino and guanidyl groups of ARG, had merged into a large absorption area from which three small maxima could be distinguished at

10

wavenumbers of 1634 cm^-1, 1595 cm^-1 and 1532 cm^-1. These maxima, however, corresponded rather well to the wavenumbers of the guanidyl group stretching peak of amorphous ARG (Fig. 3c) at 1632 cm^-1 [32] and the peaks of amorphous GBC at 1596 cm^-1 and at 1531 cm^-1 (urea NH bending [33]), which indicates that these changes originated mostly from the transformation from the crystalline to the amorphous form.

The most evident change in the spectrum of GBC-ARG when compared to the spectra of amorphous GBC and amorphous ARG was the absence of the peak at 1711 cm^-1 (GBC carbonyl stretching at the sulfonylurea end of the molecule [31]. Additionally, the C-N stretching peak of amorphous ARG at 1540 cm^-1 seemed to have shifted to a shoulder at 1565 cm^-1. Furthermore, the imidic acid shoulder of the amorphous GBC appeared to have disappeared (as with GBC-SER), but the interpretation was difficult due to the overlapping of peaks. At lower wavenumbers, the COO^- stretching band of amorphous ARG shifted from 1405 cm^-1 to 1399 cm^-1, the vibrations of sulphonyl groups of GBC (at 1340 cm^-1 and 1157 cm^-1 in amorphous GBC) shifted to 1320 cm^-1 (and possibly combined with the aliphatic chain vibration of ARG [34] and 1162 cm^-1 in the spectrum of co-amorphous GBC-ARG.

Since GBC is a weak acid and ARG is a weak base, and the difference between their pKa values is well above 2 (GBC pKa 5.1 [35] and ARG pKa 12.5 [36]), the absence of carbonyl stretching might be attributed to salt formation between GBC and ARG [37]. In studies examining different crystalline GBC salts [38] and salts and co-crystals of different antidiabetic sulfonylurea drugs with different coformers [39], salt formation between the drugs and different coformers caused major shifts to lower wavenumbers in the urea and amide carbonyl stretching bands of the drugs. Additionally, the proton transfer from the acidic N-H group of the drugs could be observed as peak shifts in the N-H stretching area above 3000 cm^-1. This was also seen with GBC-ARG, as the two N−H stretching bands of GBC (Fig. 3a.), which had combined to one peak at 3375 cm^-1 in the amorphous GBC, shifted to 3353 cm^-1 in the amorphous GBC-ARG (Fig. 3c.). However, this shift may also simply be caused by merging of the N-H peaks of amorphous GBC and amorphous ARG. Additionally, amorphous ARG

11

salts have been prepared with several drug molecules, and similar changes in the IR peaks of ARG between the spectra of amorphous salts and amorphous ARG (i.e. shifts in the C-N stretching, COO^- stretching and aliphatic chain vibrations) have been observed previously [32, 40-45]. Thus, in addition to salt formation, also hydrogen bonding between GBC and ARG [46, 47] as well as an interaction between the guanidinium group of ARG (vibrations at 1674 cm^-1 and 1614 cm^-1 in crystalline ARG) and the aromatic moiety of GBC [41, 46, 48, 49] should be considered as a possible source for the changes in the IR spectrum of the co-amorphous GBC-ARG mixture. Similarly to GBC- SER systems, addition of SLS did not induce changes in the spectra with respect to GBC-ARG (Fig. 3b and c).

3.1.2. Differential scanning calorimetry measurements

DSC measurements revealed a single Tg for the amorphous GBC-ARG and GBC-ARG-SLS combinations (Table 1), indicating that the mixtures were homogenous, single phase systems [9, 50, 51]. Instead, for the GBC-SLS and GBC-SER-SLS combinations, two Tg-values were observed. The higher Tg-values of GBC-SLS and GBC-SER-SLS corresponded approximately to the previously reported Tg-values of GBC and GBC-SER, respectively [9]. The addition of SLS seemed to lower the higher Tg-values of the mixtures, while the lower values remained approximately unchanged upon addition of SLS.

Furthermore, both GBC-SLS and GBC-SER-SLS formulations showed clear recrystallization (Trc) events and increasing the SLS content appeared to lower the Trc. After Trc, two melting events were

observed. Similarly as with the Tgs, the lower Tm-values were lowered only slightly, when the amount of SLS was increased, and the higher Tm-values were lowered significantly with increasing SLS

content. This might then indicate that the higher-melting phase (with the higher Tg) was a SLS rich phase. SLS is a known plasticizer [52]; here it displayed a small plasticizing effect on the higher Tg- values of the GBC-SLS and GBC-SER-SLS mixtures. Instead, Trc was not observed for GBC-ARG. The single Tg value observed was found to be lower than for pure GBC and higher than that previously

12

estimated for ARG (18.4°C [41], 34.0°C [32] and 55°C [53]). When SLS was added to the mixture, the value of Tg was lowered somewhat further. In addition, in GBC-ARG-SLS mixtures an endothermic event was observed (no clear preceding recrystallization) which may be melting of SLS and arginine and/or degradation. However, as XRPD indicated, these mixtures were amorphous.

Theoretical Tg-values for amorphous mixtures can be estimated by the Gordon-Taylor equation [54].

However, this equation cannot be properly applied in the mixtures investigated, as amorphous SLS and SER, and thus their Tgs, could not be obtained and their Tg values were not found in the literature. The Tg value for a freeze dried solution containing 5% gelatin and 10% SER in water has been found to be −18.75 °C, however due to the plasticizing water and the presence of gelation, this value is not a good estimate for pure SER. However, using an average of the literature values for ARG mentioned above, a theoretical estimate for the Tg of GBC-ARG-mixture is 52 ^°C, which is

approximately 10 ^°C lower than the observed value (Table 1). This may be an indication of the strong molecular interactions between GBC and ARG, as suggested by FTIR [42, 55].

3.2. Physical stability of the amorphous formulations

The physical stability of the formulations containing ARG and/or SLS was tested in this study under different temperature and humidity conditions. The physical stability of amorphous GBC and GBC- SER has been studied previously [9]. It was found that at 4 °C/0% and at 40 °C/0%, amorphous GBC had started to recrystallize at the five months’ time point, whereas GBC-SER remained stable for six months. At conditions of elevated humidity (ambient/60% RH), GBC CM and GBC−SER CM showed signs of recrystallization at 5 months.

When SLS was used as a co-former with GBC (i.e. GBC-SLS mixtures), the mixtures remained

amorphous at 4 °C/0% and at 40 °C/0% for at least 11 months according to the XRPD measurements (Fig. 1, Supplementary material). At the ambient/60% RH –condition, the mixtures remained

13

amorphous for six months (data not shown) and recrystallization was observed in XRPD at the 11 months’ time point (more pronounced with GBC-SLSl, Fig. 1, Supplementary material). However, the FTIR measurements did not show any changes indicative of recrystallization at the time points mentioned above (Fig. 2, Supplementary material). Thus, SLS was found to be a stabilizing co-former for GBC as compared to amorphous GBC alone. It has been previously observed that SLS is a non- hygroscopic material at RH values less than 75% [56] and it may thus protect the amorphous material from moisture [57].

When SLS was incorporated into mixtures with GBC and SER, then the product, i.e. GBC-SER-SLSl, remained amorphous at 4 °C/0% and at 40 °C/0% for at least 11 months (Fig 2, Supplementary material). At the ambient/60% RH condition, this mixture was found to have started to recrystallize by two months. Instead, the GBC-SER-SLSh mixture was less stable, i.e. had started to recrystallize at six months at 4 °C/0% and at 40 °C/0% and at two months at ambient/60% RH. Thus, here SLS did not confer any moisture-protecting effect as the stability of these mixtures was more or less similar to amorphous GBC-SER, except it was poorer under elevated humidity conditions. This may be due to the lack of intimate mixing of SLS with both of the components, which was manifested by the two Tgs observed in DSC measurements (Table 1). No major changes were observed in FTIR at these timepoints, except for GBC-SER-SLSl at two months under the ambient/60% RH condition that showed sharpening of the peak at approx. 1715 cm^-1 as compared to the freshly prepared mixture (Fig, 2 Supplementary material).

GBC-ARG was found to be stable for at least 18 months and GBC-ARG-SLS mixtures for at least 11 months under all storage conditions. No recrystallization was observed in the XRPD measurements (Fig.4a) and the FTIR spectra remained unchanged (Fig. 4b-d.). These observations can be considered to be directly related to the strength and amount of stabilizing intermolecular interactions between the formulation components [4]. In GBC-ARG-mixtures, the peak shifts in observed FTIR were indicative of strong interactions (Fig. 3, supported by DSC measurements (Table 1)), while the other,

14

less stable formulations showed only minor changes in FTIR, indicative of absence of strong interactions, when compared to amorphous GBC alone.

3.3. Dissolution and permeation properties

The equilibrium solubility of GBC at pH 7.2 has been previously found to be 9.6±0.4 µg/ml, i.e. 28.8 µg dissolved in the donor chamber at equilibrium [8]. This can be considered as the equilibrium solubility value also in the presence of the different formulation components, as SER has been previously found not to solubilize GBC when physically mixed with the drug [8]. It has been reported that ARG can solubilize GBC at a 1:0.5 molar ratio [25]. However, in the current study, the maximum concentrations of ARG in the donor cell were always less than 0.1% (m/V), and we have found that at concentrations less than 1% (m/V), ARG was not able to increase the equilibrium solubility of GBC (results not shown). In addition, the SLS concentration in the donor cell was always less than 0.05%

(m/V) which is not likely to be able to increase the equilibrium solubility of GBC.

Fig. 5 shows the total cumulative amounts (µg) of GBC dissolved from crystalline and amorphous GBC, the different physical mixtures and their corresponding co-amorphous mixtures. From Fig. 5a, it can be observed that all of the PMs, with the exception of GBC-SER, exceeded the equilibrium solubility value of GBC during the test. This is due to the dynamic nature of the experimental setup, i.e. part of the dissolved drug is constantly removed from the donor cell and transferred through the PAMPA membrane into the acceptor cell. Instead, when examining the concentrations present in the donor cell in the case of PMs (Fig. 3a in the supplement), it can be seen that the concentrations remained below the equilibrium solubility value of GBC. Only the GBC-ARG-SLS formulations reached the equilibrium solubility of GBC, which indicates that the formulation components had not

solubilized GBC when physically mixed with the drug. The two GBC-ARG-SLS formulations exhibited the highest total amounts of dissolved GBC (Fig. 5a). The other PMs showed dissolution properties similar or even worse than crystalline GBC, which is in accordance with previous observations with

15

amino acids [8]. The drug solubilizing capacity of surfactants, such as SLS, is well-known [29], however in the case of GBC, SLS was not able to markedly increase the amount of the dissolved drug, probably due to the small SLS concentration. Only in combination with ARG, was the dissolution enhanced, i.e. the amounts of dissolved GBC were approximately double that was observed with GBC alone.

In the case of amorphous formulations, the amorphous GBC-ARG and GBC-ARG-SLS formulations showed significantly, i.e. over 10-fold higher, amounts of dissolved GBC than the other formulations (Fig. 5b). In addition, a clear supersaturation behavior was observed, i.e. approx. 34-, 37- and 29-fold for GBC-ARG, GBC-ARG-SLSl and GBC-ARG-SLSh, respectively (Fig. 3b in the supplement). The

dissolution and supersaturation properties were found to be even poorer than with amorphous GBC in all of the formulations containing SER and with GBC-SLSh. Previously, co-amorphous GBC-SER has been found to produce a similar supersaturation but to dissolve slightly faster than amorphous GBC [8]. As shown in Fig. 5b, GBC-SLSl was able to produce similar total amounts of dissolved GBC as the amorphous GBC alone. The significant increase of dissolved GBC by ARG when compared to the other formulations may be due to the stronger interactions between GBC-ARG formed upon cryomilling, as observed by the FTIR measurements (Fig. 3), as compared to interactions in the rest of the formulations (Fig. 3., [9]). The formation of an amorphous ARG salt is known to enhance the dissolution of drugs [32, 34] and furthermore, interactions of the guanidinium group of ARG with aromatic rings have been observed to increase the solubility of aromatic molecules and this has been referred to as the “arginine-assisted solubilization system” [48]. However, when the higher amount of SLS was added to the mixture, the amounts of dissolved GBC remained slightly lower.

Similarly to GBC, SLS also contains a sulphonyl group (Fig. 1) and thus it may interfere with the sulfonyl-mediated interactions of GBC with ARG [58-60], leading to lower amounts of dissolved GBC.

Fig. 6 shows the total amounts of GBC permeated (µg) through the PAMPA membrane in the simultaneous dissolution/permeation test. In the case of crystalline GBC and PMs, no clear

16

differences were observed in the permeability between the formulations (Fig. 6a). Both GBC-ARG- SLS formulations and GBC-SER-SLSh showed the highest amounts of permeated GBC which were approx. half of the amounts of GBC dissolved from the formulations (Fig. 5a). In the case of

amorphous formulations (Fig. 6b), it was observed that the same formulations that showed superior dissolution and supersaturation properties, i.e. GBC-ARG and both GBC-ARG-SLS formulations, displayed the highest amounts of permeated GBC. The permeated amounts of GBC were virtually identical in the three formulations until 150 min, i.e. over three times higher than with GBC alone.

After 150 min, the amount of permeated GBC remained lower with GBC-ARG-SLSl, while the other two formulations showed permeated GBC amounts which were approx. five times higher than obtained with amorphous GBC alone. This observation is in accordance with the fact that a high drug concentration gradient can sustain high flux values through membranes, provided that the drug is in the free form instead of being solubilized, e.g. not incorporated in surfactant micelles [13, 15, 27, 61]. Thus, the co-formers’ (ARG and SLS) ability to stabilize supersaturation represents a high driving force for permeation of GBC. Furthermore, SLS may be able to extract lipids from membranes which can make the PAMPA barrier more permeable [62] and would most likely increase the in vivo intestinal permeability of drugs by weakening the integrity of the membrane [28, 29]. However, SLS did not clearly show this effect in this study. In addition, ARG has been found to be able to act as a permeation enhancer for insulin across a Caco2 cell membrane [63]. This effect has been considered to be attributable to the ability of arginine to bind to cell membrane phospholipids via stacking interactions between guanidium moieties [64]. Thus, it may be possible that ARG also makes the PAMPA membrane more permeable through similar interactions with lecithin.

Area under curve (AUC) values were calculated for the total amounts of dissolved and permeated GBC (Table 2). The table reveals that the physical mixtures of GBC-SLSl, GBC-SER-SLSl, GBC-ARG-SLSl and GBC-ARG-SLSh had slightly higher (statistically significant) AUC values for dissolution than the crystalline GBC. However, of these, only GBC-ARG-SLSh had also a higher AUC value for permeation when compared to the crystalline drug. Thus, it can be concluded that when the excipients are

17

simply physically mixed with a crystalline drug, this will not simultaneously increase the dissolution and permeation of GBC, with the exception of the combination of ARG and SLS.

Instead, when co-amorphous mixtures were formed, AUC values for dissolution were statistically significantly higher than their crystalline counterparts in the case of all formulations and for the co- amorphous GBC-ARG, GBC-ARG-SLSl and GBC-ARG-SLSh mixtures, the increase was 74, 75 and 56- fold, respectively. When compared to amorphous GBC alone, all ARG containing co-amorphous GBC- mixtures displayed approx. 10-fold and statistically significantly higher AUC values for dissolution whereas the dissolution AUC values of SER containing mixtures and GBC-SLSh was 2-fold lower than for amorphous GBC alone. Similarly to dissolution, the AUC values for permeation of all of the amorphous formulations were statistically significantly higher than for the crystalline counterparts.

Furthermore, GBC-SLSl, GBC-ARG, GBC-ARG-SLSl and GBC-ARG-SLSh showed statistically significantly higher AUC-values than amorphous GBC alone, i.e. 2.3-, 7.2-, 5.7- and 7.0-fold, respectively. This was particularly interesting for GBC-SLSl, as its AUC value for dissolution was similar to that of

amorphous GBC alone.

Fig. 7 depicts the AUC-values for permeation as a function of the AUC-values for dissolution; this summarizes well the dissolution and permeability properties of the studied formulations. In the graph, most of the PMs cluster together with the crystalline drug. Only the PMs with ARG and SLS in combination are able to increase both dissolution and permeation when physically mixed with GBC.

Instead, the formation of co-amorphous mixtures significantly changes the situation. While the amorphous state provides a clear dissolution advantage, the improvement in permeation is modest with the majority of the amorphous formulations. The co-amorphous GBC-SER-SLSl seems to display relatively large AUC values for permeation, with a relatively modest improvement in dissolution, when compared to the other amorphous formulations. However, the co-amorphous mixtures of GBC, ARG and SLS clearly have the highest values of permeability and dissolution and are thus located in the upper right corner of the graph. In general, permeability may be decreased and

18

solubility increased by excipients, but the two effects are most often not equal in magnitude [13].

Avdeef et al. [65]) studied the absorption potential of different sparingly soluble drugs (including GBC) with the PAMPA membrane and it was observed that the presence of an excipient

predominantly lowered permeability, but most often not by the same amount as solubility had been elevated, leading to an increased absorption potential. However, in the case of GBC, the excipients did not considerably increase absorption potential. Previously, it has been found that GBC has a tendency to form monoanionic dimers in solution [66] which may not be able to permeate as efficiently through membranes as the drug monomers, due to their larger size. This tendency was found to be even increased by excipients, especially by the surfactant sodium taurocholate. This may explain why the excipients in our study were not found to increase the permeability of GBC to any significant extent. ARG was the only excipient that may have been able to break down the GBC dimers by forming interactions with the drug and thus leading to the improved permeability of GBC.

Thus, this study showed that detailed investigation of the supersaturation and its effects on drug permeability is needed for co-amorphous formulations, since the components of a co-amorphous mixture may also have unpredictable and unwanted effects on dissolution and absorption potential in vivo and thus potential for enhancing drug bioavailability [67].

3.4. Conclusions

In this study, co-amorphous combinations of GBC-SER, SGB-SER-SLS, GBC-ARG and GBC-ARG-SLS were prepared by cryomilling. A solid-state characterization of the formulations revealed that GBC- ARG and GBC-ARG-SLS combinations were homogenous, single-phase systems, showing single Tg

values and evidence of strong intermolecular interactions between GBC and ARG. These interactions were probably also the reason for the observed good physical stability of the formulations, i.e. GBC- ARG was found to be stable for at least 18 months and GBC-ARG-SLS mixtures for at least 11 months under all storage conditions. Instead, GBC-SLS and GBC-SER-SLS combinations exhibited two Tg-

19

values, only minor changes in the FTIR spectra upon amorphization and thus poorer physical stability.

By carrying out the dissolution/permeation testing simultaneously in side-by-side diffusion cells, separated by a PAMPA membrane, we aimed to investigate the interplay between the GBC dissolution rate, GBC supersaturation and different co-formers on membrane permeability of the drug. The simultaneous test showed that the co-amorphous GBC-ARG, GBC-ARG-SLSl and GBC-ARG- SLSh had superior dissolution and supersaturation properties, reflected as 11, 11 and 8.5 times higher AUC values when compared to amorphous drug alone, while the other amorphous formulations showed AUC values that were even smaller than amorphous GBC. Similarly,

permeation was superior with GBC-ARG, GBC-ARG-SLSl and GBC-ARG-SLSh, i.e. AUC was improved by factors of 7.2, 5.7 and 7.0, respectively, when compared to amorphous GBC alone. In addition, when simply physically mixed with a crystalline drug, the combination of ARG and SLS was found to increase simultaneously the dissolution and permeation of GBC. Thus, based on the PAMPA model, it can be concluded that ARG and SLS may act as dissolution and passive permeation enhancers for GBC. Furthermore, GBC may have a significantly improved absorption potential when formulated as co-amorphous combination with ARG and/or SLS.

It can also be concluded that simultaneous testing of dissolution and permeation through a PAMPA membrane may serve as a simple and inexpensive tool for screening the most promising

formulations in future studies. In the method used here, the material consumption was low.

However, it should also be noted that formation of a supersaturated state in the donor side

demanded a high dose/volume ratio, which was approximately 120-240 times higher than the ratio in clinical use of GBC.

20 Acknowledgements

The authors thank Mrs. Lea Pirskanen and the Master students Louise Lerminiaux, Leon Balters and Caroline Buß for their skillful technical assistance. RO acknowledges the support from the Doctoral Programme in Drug Research in the Doctoral School of University of Eastern Finland.

Declarations of interest: none

21 References

[1] Loftsson, T., Brewster, M.E., 2010. Pharmaceutical applications of cyclodextrins: basic science and product development. J. Pharm. Pharmacol. 62, 1607–21. http://dx.doi.org/10.1111/j.2042-

7158.2010.01030.x

[2] Dengale, S.J., Grohganz, H., Rades, T., Löbmann, K., 2016. Recent advances in co-amorphous drug formulations. Adv. Drug Deliv. Rev., 100, 116-125. http://dx.doi.org/10.1016/j.addr.2015.12.009

[3] Allesø, M., Chieng, N., Rehder, S., Rantanen, J., Rades, T., Aaltonen, J., 2009. Enhanced dissolution rate and synchronized release of drugs in binary systems through formulation:

amorphous naproxen-cimetidine mixtures prepared by mechanical activation. J. Control. Rel., 136, 45-53. http://dx.doi.org/10.1016/j.jconrel.2009.01.027

[4] Korhonen, O., Pajula, K., Laitinen, R., 2017. Rational excipient selection for co-amorphous formulations. Expert Opin. Drug Deliv. 14, 551-569.

http://dx.doi.org/10.1080/17425247.2016.1198770

[5] Löbmann, K., Laitinen, R., Grohganz, H., Gordon, K.C., Strachan, C., Rades, T., 2011. Coamorphous drug systems: enhanced physical stability and dissolution rate of indomethacin and naproxen. Mol.

Pharm. 8, 1919-1928. http://dx.doi.org/10.1021/mp2002973

[6] Newman, A., Knipp, G., Zografi, G., 2012. Assessing the performance of amorphous solid dispersions. J. Pharm. Sci. 101, 1355–1377. http://dx.doi.org/10.1002/jps.23031

[7] Sun, D.D., Lee, P.I., 2013. Evolution of supersaturation of amorphous pharmaceuticals: the effect of rate of supersaturation generation. Mol. Pharm. 10, 4330−4346.

[8] Heikkinen, A.T., DeClerck, L., Löbmann, K., Grohganz, H., Rades, T., Laitinen, R., 2015. Dissolution properties of co-amorphous drug-amino acid formulations in buffer and biorelevant media.

Pharmazie, 70: 452–457. http://dx.doi.org/10.1691/ph.2015.4210

22

[9] Laitinen, R., Löbmann, K., Grohganz, H., Strachan, C., Rades, T., 2014. Amino acids as co- amorphous excipients for simvastatin and glibenclamide: physical properties and stability. Mol.

Pharm. 11, 2381−2389. http://dx.doi.org/10.1021/mp500107s

[10] Markopoulos, C., Imanidis, G., Vertzoni, M., Symillides, M., Parrott, N., Reppas, C., 2013. In vitro and ex vivo investigation of the impact of luminal lipid phases on passive permeability of lipophilic small molecules using PAMPA. Pharm. Res. 30, 3145–3153. http://dx.doi.org/10.1007/s11095-013- 1141-1

[11] di Cagno, M., Bibi, H.A., Bauer-Brandl, A. , 2015. New biomimetic barrier Permeapad™ for efficient investigation of passive permeability of drugs. Eur. J. Pharm. Sci., 73, 29-34.

https://doi.org/10.1016/j.ejps.2015.03.019

[12] Mesallati, H., Conroy, D., Hudson, S., Tajber, L., 2017. Preparation and characterization of amorphous ciprofloxacin-amino acid salts. Eur. J. Pharm. Biopharm. 121, 73-89.

http://dx.doi.org/10.1016/j.ejpb.2017.09.009

[13] Borbás, E., Sinkó, B., Tsinman, O., Tsinman, K., Kiserdei, É., Démuth, B., Balogh, A., Bodák, B., Domokos, A., Dargó, G.,3, Balogh, G.T., Nagy, Z.K., 2016. Investigation and mathematical description of the real driving force of passive transport of drug molecules from supersaturated solutions. Mol.

Pharm., 13, 3816-3826. http://dx.doi.org/10.1021/acs.molpharmaceut.6b00613

[14] Dahan, A., Miller, J.M., 2012. The solubility–permeability interplay and its implications in formulation design and development for poorly soluble drugs. AAPS J., 14, 244-251.

http://dx.doi.org/10.1208/s12248-012-9337-6

[15] Indulkar, A.S., Gao, Y., Raina, S.A., Zhang, G.G.Z., Taylor, L.S., 2016. Exploiting the phenomenon of liquid−liquid phase separation for enhanced and sustained membrane transport of a poorly water-soluble drug. Mol. Pharm. 13, 2059-2069.

http://dx.doi.org/10.1021/acs.molpharmaceut.6b00202

23

[16] Gantzsch, S.P., Kann, B., Ofer-Glaessgen, M., Loos, P., Berchtold, H., Balbach, S., Eichinger, T., Lehr, C., Schaefer, U.F., Windbergs, M., 2014. Characterization and evaluation of a modified PVPA barrier in comparison to Caco-2 cell monolayers for combined dissolution and permeation testing. J.

Control. Rel. 175, 79-86. http://dx.doi.org/10.1016/j.jconrel.2013.12.009

[17] Kataoka, M., Tsuneishi, S., Maeda, Y., Masaoka, Y., Sakuma, S., Yamashita, S., 2014. A new in vitro system for evaluation of passive intestinal drug absorption: establishment of a double artificial membrane permeation assay. Eur. J. Pharm. Biopharm. 88, 840-846.

http://dx.doi.org/10.1016/j.ejpb.2014.09.009

[18] Sironi, D., Rosenberg, J., Bauer-Brandl, A., Brandl, M., 2017. Dynamic dissolution-/permeation- testing of nano- and microparticle formulations of fenofibrate. Eur. J. Pharm. Sci., 96, 20-27.

http://dx.doi.org/10.1016/j.ejps.2016.09.001

[19] Sironi, D., Christensen, M., Rosenberg, J., Bauer-Brandl, A., Brandl, M., 2017. Evaluation of a dynamic dissolution/permeation model: Mutual influence of dissolution and barrier-flux under non- steady state conditions. Int. J. Pharm. 522, 50-57. http://dx.doi.org/10.1016/j.ijpharm.2017.03.002

[20] Ginski, M.J., Taneja, R., Polli, J.E., 1999. Prediction of dissolution-absorption relationships from a continuous dissolution/Caco-2 system. AAPS Phar. Sci. 1, 27-38. http://dx.doi.org/10.1208/ps010203

[21] Ginski, M.J., Polli, J.E., 1998. Prediction of dissolution-absorption relationships from a dissolution/Caco-2 system. Int. J. Pharm. 177, 117-125.

http://dx.doi.org/10.1016/S03785173(98)00330-5

[22] Kataoka, M., Masaoka, Y., Yamazaki, Y., Sakane, T., Sezaki, H., Yamashita, S., 2003. In vitro system to evaluate oral absorption of poorly water-soluble drugs: simultaneous analysis on dissolution and permeation of drugs. Pharm. Res. 20, 1674-1680.

http://dx.doi.org/10.1023/A:1026107906191

24

[23] Kobayashi, M., Sada, N., Sugawara, M., Iseki, K., Miyazaki, K., 2001. Development of a new system for prediction of drug absorption that takes into account drug dissolution and pH change in the gastro-intestinal tract. Int. J. Pharm. 221, 87-94. https://doi.org/10.1016/S0378-5173(01)00663- 9

[24] Motz, S.A., Schaefer, U.F., Balbach, S., Eichinger, T., Lehr, C.M., 2007. Permeability assessment for solid oral drug formulations based on Caco-2 monolayer in combination with a flow through dissolution cell. Eur. J. Pharm. Biopharm. 66, 286-295. https://doi.org/10.1016/j.ejpb.2006.10.015

[25] Singh, S.K., Srinivasan, K.K., Singare, D.S., Gowthamarajan, K., Prakash, D., 2012. Formulation of ternary complexes of glyburide with hydroxypropyl-β-cyclodextrin and other solubilizing agents and their effect on release behavior of glyburide in aqueous and buffered media at different agitation speeds. Drug Dev. Ind. Pharm. 38, 1328–1336. http://dx.doi.org/10.3109/03639045.2011.650645

http://dx.doi.org/10.1021/mp400439q

[26] Mukerjee, P., Mysels, K.J., 1970. Critical micelle concentrations of aqueous surfactant systems.

Nat. Stand. Ref. Data Ser. Nat. Bur. Stand.

http://nvlpubs.nist.gov/nistpubs/Legacy/NSRDS/nbsnsrds36.pdf

[27] Fischer, S.M., Brandl, M., Fricker, G., 2011. Effect of the non-ionic surfactant Poloxamer 188 on passive permeability of poorly soluble drugs across Caco-2 cell monolayers. Eur. J. Pharm. Biopharm.

79, 416-422. http://dx.doi.org/10.1016/j.ejpb.2011.04.010

[28] Hamman, J., Steenekamp, J., 2012. Excipients with specialized functions for effective drug delivery. Expert. Opin. Drug Deliv. 9, 219-230. http://dx.doi.org/10.1517/17425247.2012.647907

[29] Miller, J.M., Beig, A., Krieg, B.J., Carr, R.A., Borchardt, T.B., Amidon, G.E., Amidon, G.L., Dahan, A., 2011. The solubility-permeability interplay: mechanistic modeling and predictive application of the impact of micellar solubilization on intestinal permeation. Mol. Pharm. 8, 1848-1856.

http://dx.doi.org/10.1021/mp200181v

25

[30] Kansy, M., Senner, F., Gubernator, K., 1998. Physicochemical high throughput screening: parallel artificial membrane permeation assay in the description of passive absorption processes. J. Med.

Chem. 41, 1007-1010. http://dx.doi.org/10.1021/jm970530e

[31] Wojnarowska, Z., Grzybowska, K., Adrjanowicz, K., Kaminski, K., Paluch, M., Hawelek, L., Wrzalik, R., Dulski, M., 2010. Study of the amorphous glibenclamide drug: analysis of the molecular dynamics of quenched and cryomilled material. Mol. Pharm. 2010, 7, 1692−1707.

http://dx.doi.org/10.1021/mp100077c

[32] Ojarinta, R., Lerminiaux, L., Laitinen, R., 2017. Spray drying of poorly soluble drugs from aqueous arginine solution. Int. J. Pharm. 532, 289-298. https://doi.org/10.1016/j.ijpharm.2017.09.015

[33] Panagopoulou-Kaplani, A., Malamataris, S., 2000. Preparation and characterisation of a new insoluble polymorphic form of glibenclamide. Int. J. Pharm. 195, 239-246.

https://doi.org/10.1016/S0378-5173(99)00401-9

[34] Löbmann, K., Grohganz, H., Laitinen, R., Strachan, C., Rades, T., 2013. Amino acids as co- amorphous stabilizers for poorly water soluble drugs--Part 1: preparation, stability and dissolution enhancement. Eur. J. Pharm. Biopharm., 85, 873-881. http://dx.doi.org/10.1016/j.ejpb.2013.03.014

[35] Avdeef, A., 2007. Solubility of sparingly-soluble ionizable drugs. Adv. Drug Deliv. Rev. 59, 568–

590. http://dx.doi.org/10.1016/j.addr.2007.05.008

[36] Wu, W., Löbmann, K., Rades, T., Grohganz, H., 2018. On the role of salt formations and

structural similarity of co-formers in co-amorphous drug delivery systems. Int. J. Pharm. 535, 86-94.

http://dx.doi.org/10.1016/j.ijpharm.2017.10.057

[37] Elder, P.D., Holm, R., Lopez de Diego, H., 2013. Use of pharmaceutical salts and cocrystals to address the issue of poor solubility. Int. J. Pharm. 453, 88-100.

http://dx.doi.org/10.1016/j.ijpharm.2012.11.028

26

[38] Suresh, K., Rao Khandavilli, U.B., Gunnam, A., Nangia, A., 2017. Polymorphism, isostructurality and physicochemical properties of glibenclamide salts. CrystEngComm 19, 918-929.

http://dx.doi.org/10.1039/C6CE02295B

[39] Samie, A., Desiraju, G.R., Banik, M., 2017. Salts and cocrystals of the antidiabetic drugs

gliclazide, tolbutamide, and glipizide: Solubility enhancements through drug-coformer interactions.

Cryst. Gowth Des. 17, 2406-2417. http://dx.doi.org/10.1021/acs.cgd.6b01804

[40] Jadhav, P., Petkar, B., Pore, Y., Kulkarni, A., Burade, K., 2013. Physicochemical and molecular modeling studies ofcefixime–l-arginine–cyclodextrin ternary inclusion compounds. Carbohydrate Polymers 98, 1317– 1325. http://dx.doi.org/10.1016/j.carbpol.2013.07.070

[41] Jensen, K.T., Löbmann, K., Rades, T., Grohganz, H., 2014. Improving co-amorphous drug formulations by the addition of the highly water soluble amino acid, proline. Pharmaceutics 6, 416- 435. http://dx.doi.org/10.3390/pharmaceutics6030416

[42] Löbmann, K., Laitinen, R., Strachan, C., Rades, T., Grohganz, H., 2013b. Amino acids as co- amorphous stabilizers for poorly water-soluble drugs – Part 2: Molecular interactions. Eur. J. Pharm.

Biopharm. 85, 882-888. http://dx.doi.org/10.1016/j.ejpb.2013.03.026

[43] Ochi, M., Inoue, R., Yamauchi, Y., Yamada, S., Onoue, S., 2013. Development of meloxicam salts with improved dissolution and pharmacokinetic behaviors in rats with impaired gastric motility.

Pharm. Res. 30, 377–386. http://dx.doi.org/10.1007/s11095-012-0878-2

[44] Patel, J.R., Carlton, R.A., Yuniatine, F., Needham, T.E., Wu, L., Vogt, F.G., 2012. Preparation and structural characterization of amorphous spray-dried dispersions of tenoxicam with enhanced dissolution. J. Pharm. Sci. 101, 641-663. http://dx.doi.org/10.1002/jps.22800

[45] Ueda, H., Wu, W., Löbmann, K., Grohganz, H., Müllertz, A., Rades, T., 2018. Application of a salt co-former in a co-amorphous drug system dramatically enhances the glass transition temperature: a

27

case study of the ternary system carbamazepine, citric acid and L-arginine. Mol. Pharm., in press.

http://dx.doi.org/10.1021/acs.molpharmaceut.8b00174

[46] Hirano, A., Kameda, T., Shinozaki, D., Arakawa, T., Shiraki, K., 2013. Molecular dynamics simulation of the arginine-assisted solubilization of caffeic acid: intervention in the interaction. J.

Phys. Chem. B 117, 7518−7527. http://dx.doi.org/10.1021/jp401609p

[47] Shukla, D., Trout, B. L., 2010. Interaction of arginine with proteins and the mechanism by which it inhibits aggregation. J. Phys. Chem. B 114, 13426−13438. http://dx.doi.org/10.1021/jp108399g

[48] Hirano, A., Kameda, T., Arakawa, T., Shiraki, K. 2010. Arginine-assisted solubilization system for drug substances: Solubility experiment and simulation. J. Phys. Chem. B 114, 13455–13462.

http://dx.doi.org/10.1021/jp101909a

[49] Shah, D., Li, J., Shaikh, A.R., Rajagopalan, R., 2012. Arginine–aromatic interactions and their effects on arginine-induced solubilization of aromatic solutes and suppression of protein aggregation. Biotechnol. Prog. 28, 223-231. http://dx.doi.org/10.1002/btpr.710

[50] Marsac, P. J., Konno, H., Taylor, L. S., 2006. A comparison of the physical stability of amorphous felodipine and nifedipine systems. Pharm. Res. 23, 2306−2316. http://dx.doi.org/10.1007/s11095- 006-9047-9

[51] Marsac, P. J., Li, T., Taylor, L. S., 2009. Estimation of drug−polymer miscibility and solubility in amorphous solid dispersions using experimentally determined interaction parameters. Pharm. Res.

26, 139−151. http://dx.doi.org/10.1007/s11095-008-9721-1

[52] Ghebremeskel, A.N., Vemavarapu, C., Lodaya, M., 2007. Use of surfactants as plasticizers in preparing solid dispersions of poorly soluble API: Selection of polymer–surfactant combinations using solubility parameters and testing the processability. Int. J. Pharm. 328, 119-129.

http://dx.doi.org/10.1016/j.ijpharm.2006.08.010

28

[53] Izutsu, K., Fujimaki, Y., Kuwabara, A., Aoyagi N., 2005. Effect of counterions on the physical properties of L-arginine in frozen solutions and freeze-dried solids. Int. J. Pharm. 301, 161-169.

http://dx.doi.org/10.1016/j.ijpharm.2005.05.019

[54] Gordon, M., Taylor, J. S., 1952. Ideal copolymers and the second order transition of synthetic rubbers 1. Non-crystalline co-polymers. J. Appl.Chem. 26, 493−500.

http://dx.doi.org/10.1002/jctb.5010020901

[55] Gupta, P., Thilaqavathi, R., Chakraborti, A. K., Bansal, A. K. 2005. Role of molecular interaction in stability of celecoxib−PVP amorphous systems. Mol. Pharm. 2, 384−391.

http://dx.doi.org/10.1021/mp050004g

[56] Collier, J.W., Shah, R.B., Gupta, A., Sayeed, V., Habib, M.J., Khan, M.A., 2010. Influence of formulation and processing factors on stability of levothyroxine sodium pentahydrate. AAPS PharmSciTech, 11, 818–825. http://dx.doi.org/10.1208/s12249-010-9434-8

[57] Craye, G., Löbmann, K., Grohganz, H., Rades, T., Laitinen, R., 2015. Characterization of

Amorphous and Co-Amorphous Simvastatin Formulations Prepared by Spray Drying. Molecules, 20, 21532-21548. http://dx.doi.org/10.3390/molecules201219784

[58] Cao, J., Ding, L., Hu, W., Chen, X., Chen, X., Fang, Y., 2014. Ternary system based on fluorophore- surfactant assemblies--Cu²⁺ for highly sensitive and selective detection of arginine in aqueous solution. Langmuir, 30, 15364-15372. http://dx.doi.org/10.1021/la5039798

[59] Dominguez, H., 2017. Interaction of the interleukin 8 protein with a sodium dodecyl sulfate micelle: A computer simulation study. J. Mol. Model., 23, 210. https://doi.org/10.1007/s00894-017- 3386-9

[60] Penfold, J., Thomas, R.K., Li, P., 2016. Anionic surfactant - Biogenic amine interactions: The role of surfactant headgroup geometry. J. Colloid Interface Sci. 15, 213-219.

http://dx.doi.org/10.1016/j.jcis.2015.12.030

29

[61] Yano, K., Masaoka, Y., Kataoka, M., Sakuma, S., Yamashita, S., 2010. Mechanisms of membrane transport of poorly soluble drugs: Role of micelles in oral absorption processes. J. Pharm. Sci. 99, 1336–1345. http://dx.doi.org/10.1002/jps.21919

[62] Kragh-Hansen, U., le Maire, M., Møller, J.V.,1998. The mechanism of detergent solubilization of liposomes and protein-containing membranes. Biophys. J. 75, 2932-2946.

http://dx.doi.org/10.1016/S0006-3495(98)77735-5

[63] Kamei, N., Khafagy, E., Hirose, J., Takeda-Morishita, M., 2017. Potential of single cationic amino acid molecule “Arginine” for stimulating oral absorption of insulin. Int. J. Pharm. 521, 179-183.

http://dx.doi.org/10.1016/j.ijpharm.2017.01.066

[64] Robinson, A.D., Sun, S., Poyton, M.F., Johnson, G.A., Pellois, J.P., Jungwirth, P., Vazdar, M., Cremer, P.S., 2016. Polyarginine interacts more strongly and cooperatively than polylysine with phospholipid bilayers. J. Phys. Chem. B 120, 9287–9296. http://dx.doi.org/10.1021/acs.jpcb.6b05604

[65] Avdeef, A., Kansy, M., Bendels, S., Tsinman, K., 2008. Absorption-excipient-pH classification gradient maps: sparingly soluble drugs and the pH partition hypothesis. Eur. J. Pharm. Sci., 33, 29-41.

http://dx.doi.org/10.1016/j.ejps.2007.09.009

[66] Avdeef, A., Bendels, S., Tsinman, O., Tsinman, K., Kansy, M., 2007. Solubility-excipient

classification gradient maps. Pharm. Res., 24, 530-545. http://dx.doi.org/10.1007/s11095-006-9169- 0

[67] Laitinen, R., Löbmann, K., Grohganz, H., Priemel, P., Strachan, C.J., Rades, T. Supersaturating drug delivery systems: the potential of coamorphous drug formulations. Int. J. Pharm., 532, 1-12.

http://dx.doi.org/10.1016/j.ijpharm.2017.08.123

30

Table 1. Glass transition (Tg), recrystallization (Trc) and melting (Tm) temperature values (average ± standard deviation) for the amorphous materials prepared by cryomilling.

Material Tg (°C) Trc (°C) Tm (°C)

GBC^a 71.9 ± 0.7 127.1 ± 0.3 167.0 ± 0.6

SLS^b ND ND 197.7±0.2

GBC-SLSl 51.5±2.7; 72.5±1.3 122.1±0.2 156.2±0.5

GBC-SLSh 54.4±1.0; 68.9±1.3 120.9±0.1 150.1±0.4

GBC-SER^a 70.1 ± 1.3 113.8 ± 1.4 186.9 ± 2.6

GBC-SER-SLSl 54.0±0.3; 70.8±0.6 113.8±0.3 153.9±0.2; 193.5±5.5 GBC-SER-SLSh 53.1±4.3; 67.8±0.3 104.8±0.9; 124.4±0.4 148.8±0.3; 164.2±0.1

GBC-ARG 62.8±1.78 ND ND^c

GBC-ARG-SLSl 55.3±0.5 ND^d 176.8±1.8^d

GBC-ARG-SLSh 61.2±0.1 ND^d 162.9±3.7^d

aFrom Laitinen et al., 2014 ND=not detected

b loss of water at 101.3±0.3

cA large endotherm was observed at T>160°C which may be a combination of GBC/ARG melting and thermal degradation

dA rise in the baseline was observed which may be due to recrystallization event leading to melting

31

Table 2. Area under curve (AUC, µg*h) values for dissolved and permeated GBC during the 6h dynamic dissolution/permeation test.

GBC GBC-SLSl GBC-SLSh GBC-SER GBC-SER-SLSl GBC-SER-SLSh GBC-ARG GBC-ARG-SLSl GBC-ARG-SLSh Crystalline

Dissolution 6615±272 7857±824 7122±1436 4014±165 7282±245 7786±1643 8519±1488 13639±1287 14195±1419 Permeation 2576±257 1940±357 3176±1342 2518±892 2254±482 4313±1554 3099±235 5068±2034 4492±1046

Amorphous

Dissolution 43147±16197 45097±10825 21357±5243 24299±7070 23813±8479 25179±3768 486354±82833 493179±27929 368416±50084 Permeation 9212±2399 21164±4423 8459±928 4246±613 15648±8297 6423±507 66187±12333 52886±14751 64480±33245

AUC increase physical mixtures vs. crystalline GBC

Dissolution - 1.2^a 1.1 0.6^b 1.1^a 1.2 1.3 2.1^a 2.1^a

Permeation - 0.8 1.2 1.0 0.9 1.7 1.2 2.0 1.7^a

AUC increase co-amorphous mixtures vs. crystalline GBC

Dissolution - 6.8^a 3.2^a 3.7^a 3.6^a 3.8^a 74^a 75^a 56^a

Permeation - 8.2^a 3.2^a 1.6^a 6.1^a 2.5^a 26^a 21^a 25^a

AUC increase amorphous vs. corresponding crystalline formulation

Dissolution 6.5^a 5.7^a 3.0^a 6.1^a 3.3^a 3.2^a 57^a 36^a 26^a

Permeation 3.6^a 11^a 2.7^a 1.7^a 6.9^a 1.5^a 21^a 10^a 14^a

AUC increase co-amorphous mixtures vs. amorphous GBC

Dissolution - 1.0 0.5^b 0.6^b 0.6^b 0.6^b 11.3^a 11.4^a 8.5^a

Permeation - 2.3^a 0.9 0.5^b 1.7 0.7 7.2^a 5.7^a 7.0^a

aStatistically significantly higher

bStatistically significantly lower

32 Figure legends

Fig. 1. Chemical structures of glibenclamide (GBC), serine (SER), arginine (ARG) and sodium lauryl sulfate (SLS).

Fig.2. X-ray diffractograms of a) the crystalline starting materials glibenclamide (GBC), arginine (ARG), serine (SER) and sodium layryl sulfate (SLS). The diffractogram of SER is cut off due to the very high intensity of the diffraction peak at 2theta of 23° and b) 1. cryomilled GBC-SLSl, 2. GBC-SLSh, 3.

GBC-SER-SLSl, 4. GBC-SER-SLSh, 5. GBC-ARG, 6. GBC-ARG-SLSl and 7. GBC-ARG-SLSh.

Fig.3. Fourier-transform infrared spectra of a) crystalline and amorphous glibenclamide (GBC), crystalline sodium lauryl sulfate (SLS) and amorphous GBC-SLSl and GBC-SLSh; b) crystalline serine (SER) and GBC-SER, GBC-SER-SLSl and GBC-SER-SLSh mixtures and c) crystalline and amorphous arginine (ARG, prepared by spray-drying [30]) and amorphous GBC-ARG, GBC-ARG-SLSl and GBC- ARG-SLSh mixtures.

Fig.4. a) X-ray diffractograms of: GBC-ARG stored at 1. 4 °C/0%, 2. 40 °C/0% and 3. ambient/60% RH for 18 months, GBC-ARG-SLSl stored at 4. 4 °C/0%, 5. 40 °C/0% and 6. ambient/60% RH for 11 months and GBC-ARG-SLSh stored at 7. 4 °C/0%, 8. 40 °C/0% and 9. ambient/60% RH for 11 months and FTIR-spectra after storage under different conditions for 11 months for b) GBC-ARG; c) GBC- ARG-SLSl and d) GBC-ARG-SLSh.

Fig. 5. Total cumulative amounts of glibenclamide (GBC, µg) dissolved in the donor cell in a dynamic dissolution/permeation test at pH 7.2 from a) crystalline formulations (mean ± sd, n=3); and b) amorphous formulations (mean ± sd, n=6). Note the axis break on the y-axis.

Fig.6. Total cumulative amounts of glibenclamide (GBC, µg) permeated to the acceptor cell in a dynamic dissolution/permeation test at pH 7.2 in the case of a) crystalline formulations (mean ± sd, n=3); and b) amorphous formulations (mean ± sd, n=6).

33

Fig. 7. Area under curve (AUC, mean) values for permeation as a function of mean AUC-values for the dissolution for the crystalline and amorphous formulations. Note the axis break on the x-axis.

Supplementary material for “Permeability of glibenclamide through a PAMPA membrane: the effect of co-amorphization”

Table 1. Composition of the prepared formulations of glibenclamide (GBC), arginine (ARG), serine (SER) and sodium lauryl sulfate (SLS). All listed formulations were prepared as physical mixtures (PM) and as cryomilled formulations (CM).

Formulation GBC (mmol) Amino acid (mmol) SLS (mmol)

GBC-SLSl 0.96 - 0.08

GBC-SLSh 0.92 - 0.16

GBC-ARG 0.75 0.75 -

GBC-ARG-SLSl 0.72 0.72 0.06

GBC-ARG-SLSh 0.70 0.70 0.11

GBC-SER 0.83 0.83 -

GBC-SER-SLSl 0.80 0.80 0.07

GBC-SER-SLSh 0.78 0.78 0.12

Figure 1. X-ray diffractograms of a) from bottom to top: GBC-SLSl stored at 4 °C/0%, 40 °C/0% and ambient/60% RH and GBC-SLSh stored at 4 °C/0%, 40 °C/0% and ambient/60% RH for 11 months; b) from bottom to top: GBC-SER-SLSl stored at 4 °C/0%, 40 °C/0% for 11 months and ambient/60% RH for two months and GBC-SER-SLSh stored at 4 °C/0%, 40 °C/0% for six months and ambient/60% RH for two months.

b) a)

Figure 2. FTIR spectra of a) amorphous GBC-SLS combinations stored for 11 months except for six months under ambient/60% RH –condition; b) GBC-SER-SLSl stored for 11 months and GBC-SER-SLSh stored for six months and for two months under ambient/60% RH –condition.

a)

b)

b) a)

Figure 3. GBC concentrations present in the donor cell (µg/ml) and the cumulative amounts of glibenclamide (GBC, µg) permeated to the acceptor cell in the case of a) crystalline formulations (mean ± sd, n=3); and b) amorphous formulations (mean ± sd, n=6). The dashed lines indicate the equilibrium solubility value of GBC in pH 7.2 buffer.

b)