- Email: [email protected]
14 and cosolvent
Solubilization Behavior of Poorly Soluble Drugs with Combined Use of Gelucire 44/14 and Cosolvent KOHSAKU KAWAKAMI, KYOKO MIYOSHI, YASUO IDA Developmental Research Laboratories, Shionogi & Co., Ltd., 12-4 Sagisu 5-chome, Fukushima-ku, Osaka 553-0002, Japan
Received 7 October 2003; revised 16 January 2004; accepted 18 January 2004
ABSTRACT: Gelucire 44/14 is a surface-active excipient that can solubilize poorly soluble drugs. We investigated its solubilization behavior when coexisting with dimethylacetoamide (DMA) or dimethylsulfoxide (DMSO), both of which are also expected to enhance drug solubility. Gelucire was confirmed to form micelles by surface tension and fluorescence measurements both in water and water/cosolvent mixtures. Light-scattering measurements revealed that DMA and DMSO affect the micellar morphology in a different manner. DMA helped form large structures by being entrapped in the hydrophobic region of the micelles and/or inducing the aggregation. DMSO was likely to be anchored to the interfacial layer and did not induce micelle growth. Two model drugs, phenytoin and indomethacin, were employed to observe the solubilization behavior of poorly soluble drugs in Gelucire/cosolvent mixtures. The solubility of these drugs in the mixtures could be explained very well by using the new solubility model introduced in this article. Addition of cosolvents to the Gelucire solution did not enhance the solubility very much, and thus the combined use of cosolvents with Gelucire offered only little advantage from the viewpoint of solubility. ß 2004 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 93:1471–1479, 2004
Keywords: Gelucire 44/14; dimethylacetoamide; micelle; solubility; log-linear model; two-phase model
INTRODUCTION Solubilization of poorly soluble drugs has been a very important issue in screening studies of new chemical entities as well as in formulation research.1,2 As organic cosolvents (this will be called simply as ‘‘cosolvent’’ hereafter in this article), dimethylsulfoxide (DMSO) and dimethylacetoamide (DMA), have been widely used3–6 because of their large solubilization capacity for poorly soluble drugs and their relatively low toxicity. Gelucire 44/14 has also been recognized as a powerful solubilization agent.7,8 Therefore, the combination of Gelucire and cosolvents is
Correspondence to: Kohsaku Kawakami (Telephone: 81 6 6458 5861; Fax: 81 6 6458 0987; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 93, 1471–1479 (2004) ß 2004 Wiley-Liss, Inc. and the American Pharmacists Association
dimethylsulfoxide;
cosolvent;
commonly employed to solubilize poorly soluble drugs in research of the pharmaceutical industries. The solubilization by cosolvents can be expressed by the log-linear model,9–11 that is, log S ¼ log Sw þ sf ;
ð1Þ
where S and Sw are the solubilities in the cosolvent–water mixture and water, respectively. f is the fraction of the cosolvent. s is the solubilization capacity and can be defined by the following equation using the octanol–water partition coefficient, log P, of the solute. s ¼ slog P þ t
ð2Þ
Here, s and t are constants that depend only on nature of the cosolvents. Because these values have already been provided for ethanol, propylene glycol, polyethylene glycol 400, and glycerol in literature,11 all we need to know to predict the
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 6, JUNE 2004
1471
1472
KAWAKAMI, MIYOSHI, AND IDA
solubility in their water/cosolvent mixtures may be the log P of the solutes. It has also been shown that the s value has a strong correlation with log P for these cosolvents.11 However, this log-linear model is not applicable, if the solute forms solvate or the crystal structure is altered by the addition of the cosolvent.12 The solubilization by surfactants can be described in a simple manner as follows: S ¼ xðCs Ccmc Þ þ pSw
ð3Þ
Here, x, Cs, and Ccmc are the solubilizing capacity of micelles, surfactant concentration, and critical micelle concentration, respectively. p is the coefficient responsible for the change in the bulk solubility, which is affected by pH change, salting out, etc., caused by the surfactant. Prediction of x is difficult, because it greatly depends on the solubilization site of the drugs in micelles.13 If the guest drugs are localized deep in the micelles, x usually increases with increase of the alkyl chain length.13 It should be noted that eq. 3 ignores the growth of micelles with the surfactant concentration and structural change by the guest drug molecules. It is also difficult to predict how the addition of cosolvent affects micelle formation, because its addition significantly change the solution conditions, which affect the interaction between surfactant molecules.14–17 Possible behaviors of the cosolvent molecules are: (1) increase of surfactant solubility and inhibition of micelle formation,14,17 (2) action as a cosurfactant and alteration of micelle characteristics,15 (3) formation of pools inside micelles, (4) decrease of surfactant solubility and enhancement of micelle formation,14 (5) existence in the continuum phase with no effect on micelle formation. This article describes how the addition of DMSO (or DMA) affects the formation of Gelucire 44/14 micelles and the solubility of the drug in the mixture. The effect of using Gelucire in screening studies in the pharmaceutical industries is also discussed.
An increase in the solubility by the surfactant frequently makes a much more significant contribution than the cosolvent. Therefore, the solubility change in the micelles, caused by the addition of cosolvents, may needed to be considered. In the model, fraction of surfactant, cosolvent, and water are defined as fs, fc, and 1 fs fc, respectively, as shown in Figure 1. The notion of the partial molar volume is ignored for simplicity. If the fraction of a of the cosolvent is dissolved in the continuum phase and the remaining cosolvent forms the micelle phase with the surfactant, the fraction of each phase, Fc and Fm, is expressed as Fm ¼ fs þ ð1 aÞfc
ð5Þ
Fc ¼ 1 fs þ ð1 aÞfc :
ð6Þ
Applying the log-linear model to the solubility of drug in the micelle phase, the solubility in the cosolvent, Sc, is written as Sc ¼ Sw expðsÞ ¼ xexpðs0 Þ;
ð7Þ
where s0 is the solubilization capacity in the micelle phase. The total solubility of the drug can be written as 0 S xFm s fc ð1 aÞ safc ¼ exp þ Fc exp : ð8Þ Sw Fm Sw Fc This equation will be referred to as the twophase model hereafter.
MATERIALS AND METHODS Materials Gelucire 44/14 was a gift from Gattefosse´ (Cedex, France). It was melted at 608C and mixed well
A SOLUBILITY MODEL FOR COMBINED USE OF SURFACTANT AND COSOLVENT If the micelles and cosolvents individually contribute to the solubilization behavior, the resultant solubility Ss may be approximated as Ss ¼ xCs þ Sw expðsf Þ:
ð4Þ
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 6, JUNE 2004
Figure 1. Schematic presentation of the two-phase model.
SOLUBILIZATION BEHAVIOR OF POORLY SOLUBLE DRUGS
before use. Indomethacin (IDM), phenytoin (PHT), pyrene, and dimethylacetoamide (DMA) were obtained from Wako Pure Chemical Industries (Osaka, Japan). Dimethylsulfoxide (DMSO) was supplied from Nacalai Tesque (Kyoto, Japan). All other reagents were used as supplied. Surface Tension Measurement The surface tension of the Gelucire solutions was measured by the Wilhelmy plate method using Kru¨ss K12 Surface Tensiometer (Kru¨ss, Hamburg, Germany). The sample room of the instrument was maintained at 25 0.28C by circulating temperature-controlled water. Before every measurement, the platinum plate was heated with a burner, followed by rinsing with distilled water and acetone. Consequently, the surface tension value was checked with the surfactant-free solution to confirm the absence of contamination. All the data were read at 3 min after the creation of new surfaces. This time period is too short, if the diffusion of the surfaceactive molecules to the air/liquid surface is very slow and/or the orientation and the rearrangement of the molecules occur at the surface. These effects were confirmed not to be significant for the Gelucire solutions. Evaluation of Polarity Inside Micelles The polarity inside the micelles was investigated using pyrene as a fluorescent probe, because its fluorescent spectrum greatly depends on the polarity of its surroundings.17,18 Its spectrum splits into five sharp peaks, and the most frequently used parameter is the peak ratio of the first peak and the third peak (I1/I3). The ratio increases with the increase of the polarity. In the experiment, pyrene was solubilized at 100 ng/mL in aqueous solution with the aid of a trace amount of methanol and subjected to fluorescent measurement (Hitachi F-4010 Fluorescence Spectrophotometer, Tokyo, Japan). The excitation wavelength was 335 nm and the spectrum was measured from 360 to 400 nm. The first and the third peak were found around 374 and 384 nm, respectively. To determine the critical micellar concentration of Gelucire in each solvent, the I1/I3 values were examined at various concentrations of Gelucire with or without 10 w/v% cosolvents. Next, the I1/I3 values of the 5 w/v% Gelucire solutions were measured at various cosolvent concentrations to observe the polarity change inside the micelles
1473
after addition of the cosolvent. All the measurements were done with the room temperature controlled at around 258C. Light-Scattering Measurement The increase in the micelle size was evaluated by laser light scattering measurements on Coulter N4 Plus equipped with He-Ne laser light (Coulter Corp., FL) at the scattering angle 908. Under the same instrumental condition, the scattering intensity I can be written as 2 2 n 1 6 ; ð9Þ I/R c 2 n þ2 where R, c, and n are the particle radius, particle concentration, and the relative refractive index between the particles and the continuum phase, respectively; 7.5 or 15 w/v% aqueous Gelucire solutions mixed with various amounts of DMSO or DMA were assessed in this experiment. All the measurements were done with the room temperature controlled at around 258C. Solubility Measurement Gelucire solutions were prepared by dissolving an adequate amount in water or the cosolvent solutions. An excess amount of PHT or IDM was loaded in a 1-mL centrifuge tube, to which 1 mL of the solvent was introduced. The tubes were rotated at 50 rpm in a temperature-controlled room at 258C for 24 h, followed by centrifugation at 1000 rpm for 2 min. The supernatants were diluted with acetonitrile and filtrated using syringe filters of 450-nm pore size in preparation for HPLC analysis. PHT concentration was measured using YMC-pack ODS-AM-302 (150 mmL 4.6 mm i.d., YMC Co., Kyoto, Japan) with a flow rate at 1 mL/min. The mobile phase was 0.2 vol % acetic acid/acetonitrile ¼ 65/35. The detection wavelength and the injection volume were 240 nm and 15 mL, respectively. IDM was measured on the same column and the injection volume. The mobile phase and the detection wavelength were 0.1 vol % trifluoroacetic acid/ acetonitrile ¼ 40/60 and 320 nm, respectively.
RESULTS Micelle Formation of Gelucire Solutions Figure 2 shows the surface tension of the Gelucire solution with or without 10 w/v% cosolvents. The JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 6, JUNE 2004
1474
KAWAKAMI, MIYOSHI, AND IDA
Figure 2. Surface tension measurements of the Gelucire solutions. The solvent was water (circles), 10% DMSO aqueous solution (triangles), or 10% DMA aqueous solution (squares).
surface tension of the aqueous solution decreased with increase of the Gelucire concentration. Although this figure indicates the micelle formation of the Gelucire molecules, the critical micellar concentration (cmc) was difficult to determine precisely, most likely because Gelucire consists of many components. Nevertheless, it is clear that addition of the cosolvents shifted the cmc to higher concentrations. This trend was more clearly shown by the fluorescent measurement as can be seen in Figure 3. The polarity around pyrene molecules started to decrease around 2 mg/ mL for the aqueous Gelucire solution, indicating the formation of micelles. The cmc is usually
Figure 3. Fluorescent measurement of the Gelucire solutions. I1/I3 is the intensity ratio of the first and the third fluorescent peaks of pyrene (see text). The symbols are the same as those for Figure 1. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 6, JUNE 2004
taken as the concentration where the intensity ratio stops to decrease.17 However, because the observed intensity ratio reflects averaged polarity between bulk and the micellar phases, the middle point has been used for surfactants with low cmc values (below 1 mM).17 In the case of Gelucire, the cmc value seems to be much lower than this. Comparison with the fluorescent data with the surface tension experiments suggests that the surfactant concentration where the intensity ratio starts to decrease seems to be the best choice for deciding cmc values. The cmc values of Gelucire increased to 8 mg/mL for DMSO solution and 30 mg/mL for DMA solution. These were expected results, because the monomer concentration of Gelucire should increase by addition of the cosolvents. Figure 4 shows the change in the light-scattering intensity of Gelucire solutions with addition of the cosolvents. As can be clearly seen, the scattering intensity increased dramatically with the addition of DMA but not with DMSO. This drastic increase can be explained by the increase of micelle size according to eq. 4. Thus, there are two possible explanations: (1) DMA molecules became entrapped in the micelle core to form swollen micelles, and/or (2) the Gelucire micelles grew or became aggregated by the addition of DMA. Polarity Inside Gelucire Micelles Coexisted with Cosolvents Figure 5 shows the polarity change inside Gelucire micelles caused by the addition of the cosolvents. As can be seen, the polarity increased
Figure 4. Scattered light intensity measurements of the Gelucire solutions at 908. DMSO (open) or DMA (closed) was added to the 15 w/v% (circle) or 7.5 w/v% (triangles) aqueous Gelucire solutions.
SOLUBILIZATION BEHAVIOR OF POORLY SOLUBLE DRUGS
Figure 5. Fluorescent measurement of the 5 w/v% Gelucire solutions with various cosolvent concentrations. I1/I3 is the intensity ratio of the first and the third fluorescent peaks of pyrene (see text). The symbols are the same as those for Figure 1. Error bars represent the standard deviations.
with their addition for both cases, with the extent of the increase being larger for DMA. However, because the I1/I3 values for DMSO and DMA were 1.69 and 1.56 respectively, the polarity change inside the micelles did not seem significant with the addition of either cosolvent. Solubilization of Phenytoin As a poorly soluble model drug, phenytoin (PHT) was employed first. Figure 6 shows the solubility change of PHT on the addition of cosolvents. The increase in the solubility followed the log-linear
Figure 6. Solubility of phenytoin in aqueous DMSO (open) or DMA (closed) solution. The lines were drawn from the best-fit of the log-linear model. Error bars represent the standard deviations. For the DMA solutions, all error bars are hidden in the symbols.
1475
Figure 7. Solubility of phenytoin (squares) and indomethacin (circles) in aqueous Gelucire solution with linear fit. Error bars represent the standard deviations.
model (eq. 1), and the solubilization capacity s was determined as 6.7 for DMSO and 11.9 for DMA. Figure 7 shows the solubility of PHT in the aqueous Gelucire solution, which shows good proportionality between Gelucire concentration and solubility. The solubilizing capacity x was determined as 8.1 mg/g. It should be noted that pH value of the 20% Gelucire solution was 5.3, which is well below pKa of PHT, 8.3.19 Therefore, the change in pH due to the addition of the Gelucire did not contribute to the bulk solubility change, that is, p in eq. 3 was near unity. Next the solubility of PHT in the Gelucire solution, which contains various amounts of cosolvents, was evaluated. Figure 8 shows the
Figure 8. Solubility of phenytoin in the Gelucire/ DMSO/water solution with various DMSO concentrations. The Gelucire concentration was 15 w/v% (circle) or 5 w/v% (triangles). The thin lines are the expected values from eq. 4. The thick lines are the best fits of the twophase model. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 6, JUNE 2004
1476
KAWAKAMI, MIYOSHI, AND IDA
Figure 9. Solubility of phenytoin in the Gelucire/ DMA/water solution with various DMA concentrations. The Gelucire concentration was 15 w/v% (circle) or 5 w/v% (triangle). The thin lines are the expected values from eq. 4. The thick lines are the best fits of the two-phase model.
observed PHT solubility in the Gelucire solution combined with various fractions of DMSO. The calculated solubility curves from eq. 4 and the twophase model (eq. 8) are also shown. The experimental solubility was much higher than Ss, indicating some synergy effect between Gelucire and DMSO. The two-phase model could successfully explain the experimental data using a as the only fitting parameter. This trend was the same for the DMA system as shown in Figure 9. The fitting parameters obtained were shown in Table 1.
Figure 10. Solubility of indomethacin in aqueous DMSO (open) or DMA (closed) solution. The lines were drawn from the best-fit of the log-linear model. Error bars represent the standard deviations.
by employing pKa ¼ 4.519 after the addition of Gelucire. This contribution can apparently be ignored in eq. 3, because the increase in the solubility by the micelles was more than two orders of magnitude. The linear relationship between the solubility and the Gelucire concentration was observed as in the case of PHT, and the solubility capacity x, 18.2 mg/g, was twice as large as that for PHT. In this case, Ss could be calculated by Ss ¼ xðCs Ccmc Þ þ Scs ;
Solubilization of Indomethacin Figure 10 shows the solubility of indomethacin (IDM) in DMSO and DMA solutions. Unlike the case of PHT, the addition of a small amount of cosolvent decreased the solubility. Thus, the loglinear model did not work for these cases. The solubility of IDM in the aqueous Gelucire solution is shown in Figure 7. The pH of the saturated aqueous solution of IDM was 5.6, giving p ¼ 0.54
ð10Þ
where Scs is the solubility of IDM in the aqueous cosolvent solution. Figures 11 and 12 show the measured solubility of IDM in the Gelucire– DMSO and Gelucire–DMA solutions, presented together with the calculated solubility curves. As for the cases of PHT, the solubility values of IDM could not be explained by eq. 10 but reproduced by the two-phase model using s as a fitting parameter. The fitting parameters for IDM were also presented in Table 1.
Table 1. Fitting Results of the Solubility Data 5% Gelucire
a s x/Sw
15% Gelucire
PHT-DMA
IDM-DMA
PHT-DMSO
IDM-DMSO
PHT-DMA
IDM-DMA
PHT-DMSO
IDM-DMSO
0.99 11.9 293
0.99 12.5 1049
0.97 6.7 293
0.97 8.0 1049
0.93 11.9 293
0.93 9.4 1049
0.68 6.7 293
0.68 6.6 1049
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 6, JUNE 2004
SOLUBILIZATION BEHAVIOR OF POORLY SOLUBLE DRUGS
Figure 11. Solubility of indomethacin in the Gelucire/DMSO/water solution with various DMSO concentrations. The Gelucire concentration was 15 w/v% (circle) or 5 w/v% (triangles). The thin lines are the expected values from eq. 4. The thick lines are the best fits of the two-phase model.
DISCUSSION Micelle Morphology Change by Cosolvent Addition The surface tension and the fluorescent measurements suggested the micelle formation of Gelucire both in water and the 10% cosolvent solutions. Although the addition of cosolvents increased the monomer concentration, this effect was negligible for the solubilization behavior of the guest molecules, because the Gelucire concentrations studied were more than three orders of magnitude higher than the critical micellar concentrations. The
1477
more important contribution of the cosolvents seemed to be a change in the micellar characteristics. According to the model calculation, the cosolvents incorporated in the micelle phase were below 10% in most cases. Interestingly, the lightscattering study proved that DMSO and DMA affected the micelle morphology in different ways. DMSO might act as an anchor to the interfacial layer of the micelles to decrease the micelle size, because the light-scattering intensity did not increase at all, despite the increased volume fraction of the micelles. On the other hand, DMA seemed to penetrate deeper into the micelle to form the DMA core. However, this was not enough to explain the dramatic increase in the lightscattering intensity, because the incorporated fraction of DMA molecules seemed to be very small according to the model calculation. If we suppose that 10% of the added DMA is entrapped in the Gelucire micelle core, the volume fraction of the micelles becomes approximately 17% after the addition of 20% DMA. This means only a 4% increase in the diameter. If the DMA molecules are also distributed to the interfacial layer, this increase becomes much smaller. The fluorescent measurements did not support the DMA pool formation either, because the I1/I3 value did not increase enough. (It should be noted that pyrene molecules might be located not in the DMA pool, although it was formed, but in the micelle interfacial layer.) These observations imply that the micelle growth and/or the formation of the aggregates might also occur. The different effects of DMSO and DMA were likely to have been caused by the difference in the depth of the penetration in the interfacial layer of the micelles. Applicability of the Solubility Model
Figure 12. Solubility of indomethacin in the Gelucire/DMA/water solution with various DMA concentrations. The Gelucire concentration was 15 w/v% (circle) or 5 w/v% (triangles). The thin lines are the expected values from eq. 4. The thick lines are the best fits of the two-phase model.
We proposed a simple solubility model for combined use of the surfactant and the cosolvents. In the study of PHT, rather good agreement was obtained between the theoretical and experimental results, although a was the only fitting parameter in the calculation. The obtained a values, 0.97 for DMSO and 0.99 for DMA, were quite reasonable because of their miscible characteristics with water. Therefore, the two-phase model seemed to work well, if the log-linear model is applicable to the solubility of the drug in the cosolvent–water mixture. The log-linear model could not explain the IDM solubility in the cosolvent–water mixture. Although this phenomenon was not studied JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 6, JUNE 2004
1478
KAWAKAMI, MIYOSHI, AND IDA
further, this might be caused by a preferred interaction between cosolvent and water molecules. DMSO has been known to form 1:2 complex with water molecules when the DMSO concentration is below 10%,20,21 and DMA may act in a similar way. This interaction may decrease the IDM solubility. In the fitting procedure, we employed a values obtained from the PHT study. Thus, the only fitting parameter was s, which could not be calculated in the solubility measurements. In this case, s includes the deviation of the solubility from the log-linear model as well as the solubilization capacity of the cosolvent. It should also be mentioned that we obtained almost the same a and s values, if a was also considered as a fitting parameter. The two-phase model was confirmed to be a good approximation regardless of the solubility profile of the drugs to the water–cosolvent mixture. If the solubility profile follows the log-linear model, like PHT, the partition of the cosolvent between the micelle phase and the continuum phases, a, is the only fitting parameter to explain the solubility in the surfactant–cosolvent–water mixture. Although the determination of s and x in the preliminary experiments is recommended, similar results could be obtained by the best-fit analysis without knowing those values. Even though the solubility profile does not obey the log-linear model, like IDM, the same analysis was possible. In this study, only Gelucire 44/14 was employed as surfactant. We are planning to extend this analysis to other surfactants, cosolvents, and drugs. Practical Implications for Using Gelucire/Cosolvent Mixtures in Screening Studies The combined use of Gelucire and cosolvents is not a new idea in the pharmaceutical industry for solubilizing poorly soluble drugs, although little has been discussed in literature. The use of cosolvent facilitates the preparation procedure of the solution, because, in many cases, new candidate drugs are supplied as DMSO solution in the screening process. Our important finding here is that the addition of cosolvents did not contribute much to improving the solubility. For example, in the IDM–DMSO system, the increase in the solubility by the addition of 15% DMSO to a 15% Gelucire solution was only 1.2 times. The solubilization capacity of cosolvents for poorly soluble drugs is frequently much lower than that of micelles. In those cases, the effect of cosolvents on the micellar characteristics is much more JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 6, JUNE 2004
important than their direct contribution to the drug solubility.
CONCLUSIONS The solubility of poorly soluble drugs in Gelucire solution containing DMA or DMSO was investigated, and a two-phase model was introduced to explain the solubility data. Gelucire was confirmed to form micelles both in water and water/ cosolvent mixtures. Light-scattering measurements revealed that DMA induce large micellar aggregates by being entrapped in the micelles and/or inducing aggregation. However, DMSO was likely to be anchored to the interfacial layer and did not induce micelle growth. Two poorly soluble drugs, phenytoin and indomethacin, were employed as the model drugs. Their solubility in the mixtures could be well explained by using our two-phase model, which requires only one fitting parameter. Addition of the cosolvents to the Gelucire solution had only a little advantage from the viewpoint of improving solubility.
REFERENCES 1. Bittner B, Mountfield RJ. 2002. Intravenous administration of poorly soluble new drug entities in early drug discovery: The potential impact of formulation on pharmacokinetic parameters. Curr Opin Drug Dis Dev 5:59–71. 2. Bittner B, Mountfield RJ. 2002. Formulations and related activities for the oral administration of poorly water-soluble compounds in early discovery animal studies. Pharm Ind 64:800–807. 3. Seethala R, Fernandes PB, editors. 2001. Handbook of drug screening. New York: Marcel Dekker. 4. Pan L, Ho Q, Tsutsui K, Takahashi L. 2001. Comparison of chromatographic and spectroscopic methods used to rank compounds for aqueous solubility. J Pharm Sci 90:521–529. 5. Kerns EH. 2001. High throughput physicochemical profiling for drug discovery. J Pharm Sci 90:1838– 1858. 6. Krishna G, Chen KJ, Lin CC, Nomeir AA. 2001. Permeability of lipophilic compounds in drug discovery using in-vitro human absorption model, Caco-2. Int J Pharm 222:77–89. 7. Roussin P, Laforeˆt JP. 1997. Gelucire1 44/14: A high-performance system to enhance bioavailability of poorly water soluble drugs. B T Gattefosse´ 90:51–58. 8. Damian F, Blaton N, Naesens L, Balzarini J, Kinget R, Augustijns P, Van den Mooter G. 2000.
SOLUBILIZATION BEHAVIOR OF POORLY SOLUBLE DRUGS
9.
10.
11.
12.
13.
14.
Physicochemical characaterization of solid dispersions of the antiviral agent UC-781 with polyethylene glycol 6000 and Gelucire 44/14. Eur J Pharm Sci 10:311–322. Rubino JT, Yalkowsky SH. 1987. Cosolvency and deviations from log-linear solubilization. Pharm Res 4:231–236. Grant DJW, Brittain HG. 1995. Solubility of pharmaceutical solids. In: Brittain HG, editor. Physical characterization of pharmaceutical solids. New York: Marcel Dekker. pp 321–386. Millard JW, Alvarez-Nu´n˜ez, Yalkowsky SH. 2002. Solubilization by cosolvents. Establishing useful constants for the log-linear model. Int J Pharm 245:153–166. Bustamante P, Navarro J, Romero S, Escalera B. 2002. Thermodynamic origin of the solubility profiles of drugs showing one or two maxima against the polarity of aqueous and nonaqueous mixtures: Niflumic acid and caffeine. J Pharm Sci 91:874– 883. Florence AT, Attwood D. 1998. Physicochemical principles of pharmacy, 3rd ed. New York: Palgrave. pp 199–251. Almgren M, Swarup S, Lo¨froth JE. 1985. Effect of formamide and other organic polar solvents on the
15.
16. 17.
18.
19.
20.
21.
1479
micelle formation of sodium dodecyl sulfate. J Phys Chem 89:4621–4626. Aveyard R, Binks BP, Cooper P, Fletcher PDI. 1990. Incorporation of hydrocarbons into surfactant monolayers. Adv Colloid Interface Sci 33:59–77. Israelachvili JN. 1991. Intermolecular and surface forces, 2nd ed. New York: Academic Press, Chapt 16. Aguiar J, Carpena P, Molina-Bolı´var JA, Ruiz CC. 2003. On the determination of the critical micelle concentration by the pyrene 1:3 ratio method. J Colloid Interface Sci 258:116–122. Kawakami K, Nishihara Y, Hirano K. 1999. Liposome/emulsion transition induced by a-tocopheryl acetate. Langmuir 15:7454–7460. Florence AT, Attwood D. 1998. Physicochemical principles of pharmacy, 3rd ed. New York: Palgrave. pp 84–85. Shin DN, Wijnen JW, Engberts JBFN, Wakisaka A. 2001. On the origin of microheterogeneity: A mass spectroscopic study of dimethyl sulfoxide-water binary mixture. J Phys Chem B 105:6759–6762. Shin DN, Wijnen JW, Engberts JBFN, Wakisaka A. 2002. On the origin of microheterogeneity: A mass spectroscopic study of acetonitrile-water and dimethyl sulfoxide-water binary mixtures (part 2). J Phys Chem B 106:6014–6020.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 6, JUNE 2004
Received 7 October 2003; revised 16 January 2004; accepted 18 January 2004
ABSTRACT: Gelucire 44/14 is a surface-active excipient that can solubilize poorly soluble drugs. We investigated its solubilization behavior when coexisting with dimethylacetoamide (DMA) or dimethylsulfoxide (DMSO), both of which are also expected to enhance drug solubility. Gelucire was confirmed to form micelles by surface tension and fluorescence measurements both in water and water/cosolvent mixtures. Light-scattering measurements revealed that DMA and DMSO affect the micellar morphology in a different manner. DMA helped form large structures by being entrapped in the hydrophobic region of the micelles and/or inducing the aggregation. DMSO was likely to be anchored to the interfacial layer and did not induce micelle growth. Two model drugs, phenytoin and indomethacin, were employed to observe the solubilization behavior of poorly soluble drugs in Gelucire/cosolvent mixtures. The solubility of these drugs in the mixtures could be explained very well by using the new solubility model introduced in this article. Addition of cosolvents to the Gelucire solution did not enhance the solubility very much, and thus the combined use of cosolvents with Gelucire offered only little advantage from the viewpoint of solubility. ß 2004 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 93:1471–1479, 2004
Keywords: Gelucire 44/14; dimethylacetoamide; micelle; solubility; log-linear model; two-phase model
INTRODUCTION Solubilization of poorly soluble drugs has been a very important issue in screening studies of new chemical entities as well as in formulation research.1,2 As organic cosolvents (this will be called simply as ‘‘cosolvent’’ hereafter in this article), dimethylsulfoxide (DMSO) and dimethylacetoamide (DMA), have been widely used3–6 because of their large solubilization capacity for poorly soluble drugs and their relatively low toxicity. Gelucire 44/14 has also been recognized as a powerful solubilization agent.7,8 Therefore, the combination of Gelucire and cosolvents is
Correspondence to: Kohsaku Kawakami (Telephone: 81 6 6458 5861; Fax: 81 6 6458 0987; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 93, 1471–1479 (2004) ß 2004 Wiley-Liss, Inc. and the American Pharmacists Association
dimethylsulfoxide;
cosolvent;
commonly employed to solubilize poorly soluble drugs in research of the pharmaceutical industries. The solubilization by cosolvents can be expressed by the log-linear model,9–11 that is, log S ¼ log Sw þ sf ;
ð1Þ
where S and Sw are the solubilities in the cosolvent–water mixture and water, respectively. f is the fraction of the cosolvent. s is the solubilization capacity and can be defined by the following equation using the octanol–water partition coefficient, log P, of the solute. s ¼ slog P þ t
ð2Þ
Here, s and t are constants that depend only on nature of the cosolvents. Because these values have already been provided for ethanol, propylene glycol, polyethylene glycol 400, and glycerol in literature,11 all we need to know to predict the
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 6, JUNE 2004
1471
1472
KAWAKAMI, MIYOSHI, AND IDA
solubility in their water/cosolvent mixtures may be the log P of the solutes. It has also been shown that the s value has a strong correlation with log P for these cosolvents.11 However, this log-linear model is not applicable, if the solute forms solvate or the crystal structure is altered by the addition of the cosolvent.12 The solubilization by surfactants can be described in a simple manner as follows: S ¼ xðCs Ccmc Þ þ pSw
ð3Þ
Here, x, Cs, and Ccmc are the solubilizing capacity of micelles, surfactant concentration, and critical micelle concentration, respectively. p is the coefficient responsible for the change in the bulk solubility, which is affected by pH change, salting out, etc., caused by the surfactant. Prediction of x is difficult, because it greatly depends on the solubilization site of the drugs in micelles.13 If the guest drugs are localized deep in the micelles, x usually increases with increase of the alkyl chain length.13 It should be noted that eq. 3 ignores the growth of micelles with the surfactant concentration and structural change by the guest drug molecules. It is also difficult to predict how the addition of cosolvent affects micelle formation, because its addition significantly change the solution conditions, which affect the interaction between surfactant molecules.14–17 Possible behaviors of the cosolvent molecules are: (1) increase of surfactant solubility and inhibition of micelle formation,14,17 (2) action as a cosurfactant and alteration of micelle characteristics,15 (3) formation of pools inside micelles, (4) decrease of surfactant solubility and enhancement of micelle formation,14 (5) existence in the continuum phase with no effect on micelle formation. This article describes how the addition of DMSO (or DMA) affects the formation of Gelucire 44/14 micelles and the solubility of the drug in the mixture. The effect of using Gelucire in screening studies in the pharmaceutical industries is also discussed.
An increase in the solubility by the surfactant frequently makes a much more significant contribution than the cosolvent. Therefore, the solubility change in the micelles, caused by the addition of cosolvents, may needed to be considered. In the model, fraction of surfactant, cosolvent, and water are defined as fs, fc, and 1 fs fc, respectively, as shown in Figure 1. The notion of the partial molar volume is ignored for simplicity. If the fraction of a of the cosolvent is dissolved in the continuum phase and the remaining cosolvent forms the micelle phase with the surfactant, the fraction of each phase, Fc and Fm, is expressed as Fm ¼ fs þ ð1 aÞfc
ð5Þ
Fc ¼ 1 fs þ ð1 aÞfc :
ð6Þ
Applying the log-linear model to the solubility of drug in the micelle phase, the solubility in the cosolvent, Sc, is written as Sc ¼ Sw expðsÞ ¼ xexpðs0 Þ;
ð7Þ
where s0 is the solubilization capacity in the micelle phase. The total solubility of the drug can be written as 0 S xFm s fc ð1 aÞ safc ¼ exp þ Fc exp : ð8Þ Sw Fm Sw Fc This equation will be referred to as the twophase model hereafter.
MATERIALS AND METHODS Materials Gelucire 44/14 was a gift from Gattefosse´ (Cedex, France). It was melted at 608C and mixed well
A SOLUBILITY MODEL FOR COMBINED USE OF SURFACTANT AND COSOLVENT If the micelles and cosolvents individually contribute to the solubilization behavior, the resultant solubility Ss may be approximated as Ss ¼ xCs þ Sw expðsf Þ:
ð4Þ
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 6, JUNE 2004
Figure 1. Schematic presentation of the two-phase model.
SOLUBILIZATION BEHAVIOR OF POORLY SOLUBLE DRUGS
before use. Indomethacin (IDM), phenytoin (PHT), pyrene, and dimethylacetoamide (DMA) were obtained from Wako Pure Chemical Industries (Osaka, Japan). Dimethylsulfoxide (DMSO) was supplied from Nacalai Tesque (Kyoto, Japan). All other reagents were used as supplied. Surface Tension Measurement The surface tension of the Gelucire solutions was measured by the Wilhelmy plate method using Kru¨ss K12 Surface Tensiometer (Kru¨ss, Hamburg, Germany). The sample room of the instrument was maintained at 25 0.28C by circulating temperature-controlled water. Before every measurement, the platinum plate was heated with a burner, followed by rinsing with distilled water and acetone. Consequently, the surface tension value was checked with the surfactant-free solution to confirm the absence of contamination. All the data were read at 3 min after the creation of new surfaces. This time period is too short, if the diffusion of the surfaceactive molecules to the air/liquid surface is very slow and/or the orientation and the rearrangement of the molecules occur at the surface. These effects were confirmed not to be significant for the Gelucire solutions. Evaluation of Polarity Inside Micelles The polarity inside the micelles was investigated using pyrene as a fluorescent probe, because its fluorescent spectrum greatly depends on the polarity of its surroundings.17,18 Its spectrum splits into five sharp peaks, and the most frequently used parameter is the peak ratio of the first peak and the third peak (I1/I3). The ratio increases with the increase of the polarity. In the experiment, pyrene was solubilized at 100 ng/mL in aqueous solution with the aid of a trace amount of methanol and subjected to fluorescent measurement (Hitachi F-4010 Fluorescence Spectrophotometer, Tokyo, Japan). The excitation wavelength was 335 nm and the spectrum was measured from 360 to 400 nm. The first and the third peak were found around 374 and 384 nm, respectively. To determine the critical micellar concentration of Gelucire in each solvent, the I1/I3 values were examined at various concentrations of Gelucire with or without 10 w/v% cosolvents. Next, the I1/I3 values of the 5 w/v% Gelucire solutions were measured at various cosolvent concentrations to observe the polarity change inside the micelles
1473
after addition of the cosolvent. All the measurements were done with the room temperature controlled at around 258C. Light-Scattering Measurement The increase in the micelle size was evaluated by laser light scattering measurements on Coulter N4 Plus equipped with He-Ne laser light (Coulter Corp., FL) at the scattering angle 908. Under the same instrumental condition, the scattering intensity I can be written as 2 2 n 1 6 ; ð9Þ I/R c 2 n þ2 where R, c, and n are the particle radius, particle concentration, and the relative refractive index between the particles and the continuum phase, respectively; 7.5 or 15 w/v% aqueous Gelucire solutions mixed with various amounts of DMSO or DMA were assessed in this experiment. All the measurements were done with the room temperature controlled at around 258C. Solubility Measurement Gelucire solutions were prepared by dissolving an adequate amount in water or the cosolvent solutions. An excess amount of PHT or IDM was loaded in a 1-mL centrifuge tube, to which 1 mL of the solvent was introduced. The tubes were rotated at 50 rpm in a temperature-controlled room at 258C for 24 h, followed by centrifugation at 1000 rpm for 2 min. The supernatants were diluted with acetonitrile and filtrated using syringe filters of 450-nm pore size in preparation for HPLC analysis. PHT concentration was measured using YMC-pack ODS-AM-302 (150 mmL 4.6 mm i.d., YMC Co., Kyoto, Japan) with a flow rate at 1 mL/min. The mobile phase was 0.2 vol % acetic acid/acetonitrile ¼ 65/35. The detection wavelength and the injection volume were 240 nm and 15 mL, respectively. IDM was measured on the same column and the injection volume. The mobile phase and the detection wavelength were 0.1 vol % trifluoroacetic acid/ acetonitrile ¼ 40/60 and 320 nm, respectively.
RESULTS Micelle Formation of Gelucire Solutions Figure 2 shows the surface tension of the Gelucire solution with or without 10 w/v% cosolvents. The JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 6, JUNE 2004
1474
KAWAKAMI, MIYOSHI, AND IDA
Figure 2. Surface tension measurements of the Gelucire solutions. The solvent was water (circles), 10% DMSO aqueous solution (triangles), or 10% DMA aqueous solution (squares).
surface tension of the aqueous solution decreased with increase of the Gelucire concentration. Although this figure indicates the micelle formation of the Gelucire molecules, the critical micellar concentration (cmc) was difficult to determine precisely, most likely because Gelucire consists of many components. Nevertheless, it is clear that addition of the cosolvents shifted the cmc to higher concentrations. This trend was more clearly shown by the fluorescent measurement as can be seen in Figure 3. The polarity around pyrene molecules started to decrease around 2 mg/ mL for the aqueous Gelucire solution, indicating the formation of micelles. The cmc is usually
Figure 3. Fluorescent measurement of the Gelucire solutions. I1/I3 is the intensity ratio of the first and the third fluorescent peaks of pyrene (see text). The symbols are the same as those for Figure 1. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 6, JUNE 2004
taken as the concentration where the intensity ratio stops to decrease.17 However, because the observed intensity ratio reflects averaged polarity between bulk and the micellar phases, the middle point has been used for surfactants with low cmc values (below 1 mM).17 In the case of Gelucire, the cmc value seems to be much lower than this. Comparison with the fluorescent data with the surface tension experiments suggests that the surfactant concentration where the intensity ratio starts to decrease seems to be the best choice for deciding cmc values. The cmc values of Gelucire increased to 8 mg/mL for DMSO solution and 30 mg/mL for DMA solution. These were expected results, because the monomer concentration of Gelucire should increase by addition of the cosolvents. Figure 4 shows the change in the light-scattering intensity of Gelucire solutions with addition of the cosolvents. As can be clearly seen, the scattering intensity increased dramatically with the addition of DMA but not with DMSO. This drastic increase can be explained by the increase of micelle size according to eq. 4. Thus, there are two possible explanations: (1) DMA molecules became entrapped in the micelle core to form swollen micelles, and/or (2) the Gelucire micelles grew or became aggregated by the addition of DMA. Polarity Inside Gelucire Micelles Coexisted with Cosolvents Figure 5 shows the polarity change inside Gelucire micelles caused by the addition of the cosolvents. As can be seen, the polarity increased
Figure 4. Scattered light intensity measurements of the Gelucire solutions at 908. DMSO (open) or DMA (closed) was added to the 15 w/v% (circle) or 7.5 w/v% (triangles) aqueous Gelucire solutions.
SOLUBILIZATION BEHAVIOR OF POORLY SOLUBLE DRUGS
Figure 5. Fluorescent measurement of the 5 w/v% Gelucire solutions with various cosolvent concentrations. I1/I3 is the intensity ratio of the first and the third fluorescent peaks of pyrene (see text). The symbols are the same as those for Figure 1. Error bars represent the standard deviations.
with their addition for both cases, with the extent of the increase being larger for DMA. However, because the I1/I3 values for DMSO and DMA were 1.69 and 1.56 respectively, the polarity change inside the micelles did not seem significant with the addition of either cosolvent. Solubilization of Phenytoin As a poorly soluble model drug, phenytoin (PHT) was employed first. Figure 6 shows the solubility change of PHT on the addition of cosolvents. The increase in the solubility followed the log-linear
Figure 6. Solubility of phenytoin in aqueous DMSO (open) or DMA (closed) solution. The lines were drawn from the best-fit of the log-linear model. Error bars represent the standard deviations. For the DMA solutions, all error bars are hidden in the symbols.
1475
Figure 7. Solubility of phenytoin (squares) and indomethacin (circles) in aqueous Gelucire solution with linear fit. Error bars represent the standard deviations.
model (eq. 1), and the solubilization capacity s was determined as 6.7 for DMSO and 11.9 for DMA. Figure 7 shows the solubility of PHT in the aqueous Gelucire solution, which shows good proportionality between Gelucire concentration and solubility. The solubilizing capacity x was determined as 8.1 mg/g. It should be noted that pH value of the 20% Gelucire solution was 5.3, which is well below pKa of PHT, 8.3.19 Therefore, the change in pH due to the addition of the Gelucire did not contribute to the bulk solubility change, that is, p in eq. 3 was near unity. Next the solubility of PHT in the Gelucire solution, which contains various amounts of cosolvents, was evaluated. Figure 8 shows the
Figure 8. Solubility of phenytoin in the Gelucire/ DMSO/water solution with various DMSO concentrations. The Gelucire concentration was 15 w/v% (circle) or 5 w/v% (triangles). The thin lines are the expected values from eq. 4. The thick lines are the best fits of the twophase model. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 6, JUNE 2004
1476
KAWAKAMI, MIYOSHI, AND IDA
Figure 9. Solubility of phenytoin in the Gelucire/ DMA/water solution with various DMA concentrations. The Gelucire concentration was 15 w/v% (circle) or 5 w/v% (triangle). The thin lines are the expected values from eq. 4. The thick lines are the best fits of the two-phase model.
observed PHT solubility in the Gelucire solution combined with various fractions of DMSO. The calculated solubility curves from eq. 4 and the twophase model (eq. 8) are also shown. The experimental solubility was much higher than Ss, indicating some synergy effect between Gelucire and DMSO. The two-phase model could successfully explain the experimental data using a as the only fitting parameter. This trend was the same for the DMA system as shown in Figure 9. The fitting parameters obtained were shown in Table 1.
Figure 10. Solubility of indomethacin in aqueous DMSO (open) or DMA (closed) solution. The lines were drawn from the best-fit of the log-linear model. Error bars represent the standard deviations.
by employing pKa ¼ 4.519 after the addition of Gelucire. This contribution can apparently be ignored in eq. 3, because the increase in the solubility by the micelles was more than two orders of magnitude. The linear relationship between the solubility and the Gelucire concentration was observed as in the case of PHT, and the solubility capacity x, 18.2 mg/g, was twice as large as that for PHT. In this case, Ss could be calculated by Ss ¼ xðCs Ccmc Þ þ Scs ;
Solubilization of Indomethacin Figure 10 shows the solubility of indomethacin (IDM) in DMSO and DMA solutions. Unlike the case of PHT, the addition of a small amount of cosolvent decreased the solubility. Thus, the loglinear model did not work for these cases. The solubility of IDM in the aqueous Gelucire solution is shown in Figure 7. The pH of the saturated aqueous solution of IDM was 5.6, giving p ¼ 0.54
ð10Þ
where Scs is the solubility of IDM in the aqueous cosolvent solution. Figures 11 and 12 show the measured solubility of IDM in the Gelucire– DMSO and Gelucire–DMA solutions, presented together with the calculated solubility curves. As for the cases of PHT, the solubility values of IDM could not be explained by eq. 10 but reproduced by the two-phase model using s as a fitting parameter. The fitting parameters for IDM were also presented in Table 1.
Table 1. Fitting Results of the Solubility Data 5% Gelucire
a s x/Sw
15% Gelucire
PHT-DMA
IDM-DMA
PHT-DMSO
IDM-DMSO
PHT-DMA
IDM-DMA
PHT-DMSO
IDM-DMSO
0.99 11.9 293
0.99 12.5 1049
0.97 6.7 293
0.97 8.0 1049
0.93 11.9 293
0.93 9.4 1049
0.68 6.7 293
0.68 6.6 1049
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 6, JUNE 2004
SOLUBILIZATION BEHAVIOR OF POORLY SOLUBLE DRUGS
Figure 11. Solubility of indomethacin in the Gelucire/DMSO/water solution with various DMSO concentrations. The Gelucire concentration was 15 w/v% (circle) or 5 w/v% (triangles). The thin lines are the expected values from eq. 4. The thick lines are the best fits of the two-phase model.
DISCUSSION Micelle Morphology Change by Cosolvent Addition The surface tension and the fluorescent measurements suggested the micelle formation of Gelucire both in water and the 10% cosolvent solutions. Although the addition of cosolvents increased the monomer concentration, this effect was negligible for the solubilization behavior of the guest molecules, because the Gelucire concentrations studied were more than three orders of magnitude higher than the critical micellar concentrations. The
1477
more important contribution of the cosolvents seemed to be a change in the micellar characteristics. According to the model calculation, the cosolvents incorporated in the micelle phase were below 10% in most cases. Interestingly, the lightscattering study proved that DMSO and DMA affected the micelle morphology in different ways. DMSO might act as an anchor to the interfacial layer of the micelles to decrease the micelle size, because the light-scattering intensity did not increase at all, despite the increased volume fraction of the micelles. On the other hand, DMA seemed to penetrate deeper into the micelle to form the DMA core. However, this was not enough to explain the dramatic increase in the lightscattering intensity, because the incorporated fraction of DMA molecules seemed to be very small according to the model calculation. If we suppose that 10% of the added DMA is entrapped in the Gelucire micelle core, the volume fraction of the micelles becomes approximately 17% after the addition of 20% DMA. This means only a 4% increase in the diameter. If the DMA molecules are also distributed to the interfacial layer, this increase becomes much smaller. The fluorescent measurements did not support the DMA pool formation either, because the I1/I3 value did not increase enough. (It should be noted that pyrene molecules might be located not in the DMA pool, although it was formed, but in the micelle interfacial layer.) These observations imply that the micelle growth and/or the formation of the aggregates might also occur. The different effects of DMSO and DMA were likely to have been caused by the difference in the depth of the penetration in the interfacial layer of the micelles. Applicability of the Solubility Model
Figure 12. Solubility of indomethacin in the Gelucire/DMA/water solution with various DMA concentrations. The Gelucire concentration was 15 w/v% (circle) or 5 w/v% (triangles). The thin lines are the expected values from eq. 4. The thick lines are the best fits of the two-phase model.
We proposed a simple solubility model for combined use of the surfactant and the cosolvents. In the study of PHT, rather good agreement was obtained between the theoretical and experimental results, although a was the only fitting parameter in the calculation. The obtained a values, 0.97 for DMSO and 0.99 for DMA, were quite reasonable because of their miscible characteristics with water. Therefore, the two-phase model seemed to work well, if the log-linear model is applicable to the solubility of the drug in the cosolvent–water mixture. The log-linear model could not explain the IDM solubility in the cosolvent–water mixture. Although this phenomenon was not studied JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 6, JUNE 2004
1478
KAWAKAMI, MIYOSHI, AND IDA
further, this might be caused by a preferred interaction between cosolvent and water molecules. DMSO has been known to form 1:2 complex with water molecules when the DMSO concentration is below 10%,20,21 and DMA may act in a similar way. This interaction may decrease the IDM solubility. In the fitting procedure, we employed a values obtained from the PHT study. Thus, the only fitting parameter was s, which could not be calculated in the solubility measurements. In this case, s includes the deviation of the solubility from the log-linear model as well as the solubilization capacity of the cosolvent. It should also be mentioned that we obtained almost the same a and s values, if a was also considered as a fitting parameter. The two-phase model was confirmed to be a good approximation regardless of the solubility profile of the drugs to the water–cosolvent mixture. If the solubility profile follows the log-linear model, like PHT, the partition of the cosolvent between the micelle phase and the continuum phases, a, is the only fitting parameter to explain the solubility in the surfactant–cosolvent–water mixture. Although the determination of s and x in the preliminary experiments is recommended, similar results could be obtained by the best-fit analysis without knowing those values. Even though the solubility profile does not obey the log-linear model, like IDM, the same analysis was possible. In this study, only Gelucire 44/14 was employed as surfactant. We are planning to extend this analysis to other surfactants, cosolvents, and drugs. Practical Implications for Using Gelucire/Cosolvent Mixtures in Screening Studies The combined use of Gelucire and cosolvents is not a new idea in the pharmaceutical industry for solubilizing poorly soluble drugs, although little has been discussed in literature. The use of cosolvent facilitates the preparation procedure of the solution, because, in many cases, new candidate drugs are supplied as DMSO solution in the screening process. Our important finding here is that the addition of cosolvents did not contribute much to improving the solubility. For example, in the IDM–DMSO system, the increase in the solubility by the addition of 15% DMSO to a 15% Gelucire solution was only 1.2 times. The solubilization capacity of cosolvents for poorly soluble drugs is frequently much lower than that of micelles. In those cases, the effect of cosolvents on the micellar characteristics is much more JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 6, JUNE 2004
important than their direct contribution to the drug solubility.
CONCLUSIONS The solubility of poorly soluble drugs in Gelucire solution containing DMA or DMSO was investigated, and a two-phase model was introduced to explain the solubility data. Gelucire was confirmed to form micelles both in water and water/ cosolvent mixtures. Light-scattering measurements revealed that DMA induce large micellar aggregates by being entrapped in the micelles and/or inducing aggregation. However, DMSO was likely to be anchored to the interfacial layer and did not induce micelle growth. Two poorly soluble drugs, phenytoin and indomethacin, were employed as the model drugs. Their solubility in the mixtures could be well explained by using our two-phase model, which requires only one fitting parameter. Addition of the cosolvents to the Gelucire solution had only a little advantage from the viewpoint of improving solubility.
REFERENCES 1. Bittner B, Mountfield RJ. 2002. Intravenous administration of poorly soluble new drug entities in early drug discovery: The potential impact of formulation on pharmacokinetic parameters. Curr Opin Drug Dis Dev 5:59–71. 2. Bittner B, Mountfield RJ. 2002. Formulations and related activities for the oral administration of poorly water-soluble compounds in early discovery animal studies. Pharm Ind 64:800–807. 3. Seethala R, Fernandes PB, editors. 2001. Handbook of drug screening. New York: Marcel Dekker. 4. Pan L, Ho Q, Tsutsui K, Takahashi L. 2001. Comparison of chromatographic and spectroscopic methods used to rank compounds for aqueous solubility. J Pharm Sci 90:521–529. 5. Kerns EH. 2001. High throughput physicochemical profiling for drug discovery. J Pharm Sci 90:1838– 1858. 6. Krishna G, Chen KJ, Lin CC, Nomeir AA. 2001. Permeability of lipophilic compounds in drug discovery using in-vitro human absorption model, Caco-2. Int J Pharm 222:77–89. 7. Roussin P, Laforeˆt JP. 1997. Gelucire1 44/14: A high-performance system to enhance bioavailability of poorly water soluble drugs. B T Gattefosse´ 90:51–58. 8. Damian F, Blaton N, Naesens L, Balzarini J, Kinget R, Augustijns P, Van den Mooter G. 2000.
SOLUBILIZATION BEHAVIOR OF POORLY SOLUBLE DRUGS
9.
10.
11.
12.
13.
14.
Physicochemical characaterization of solid dispersions of the antiviral agent UC-781 with polyethylene glycol 6000 and Gelucire 44/14. Eur J Pharm Sci 10:311–322. Rubino JT, Yalkowsky SH. 1987. Cosolvency and deviations from log-linear solubilization. Pharm Res 4:231–236. Grant DJW, Brittain HG. 1995. Solubility of pharmaceutical solids. In: Brittain HG, editor. Physical characterization of pharmaceutical solids. New York: Marcel Dekker. pp 321–386. Millard JW, Alvarez-Nu´n˜ez, Yalkowsky SH. 2002. Solubilization by cosolvents. Establishing useful constants for the log-linear model. Int J Pharm 245:153–166. Bustamante P, Navarro J, Romero S, Escalera B. 2002. Thermodynamic origin of the solubility profiles of drugs showing one or two maxima against the polarity of aqueous and nonaqueous mixtures: Niflumic acid and caffeine. J Pharm Sci 91:874– 883. Florence AT, Attwood D. 1998. Physicochemical principles of pharmacy, 3rd ed. New York: Palgrave. pp 199–251. Almgren M, Swarup S, Lo¨froth JE. 1985. Effect of formamide and other organic polar solvents on the
15.
16. 17.
18.
19.
20.
21.
1479
micelle formation of sodium dodecyl sulfate. J Phys Chem 89:4621–4626. Aveyard R, Binks BP, Cooper P, Fletcher PDI. 1990. Incorporation of hydrocarbons into surfactant monolayers. Adv Colloid Interface Sci 33:59–77. Israelachvili JN. 1991. Intermolecular and surface forces, 2nd ed. New York: Academic Press, Chapt 16. Aguiar J, Carpena P, Molina-Bolı´var JA, Ruiz CC. 2003. On the determination of the critical micelle concentration by the pyrene 1:3 ratio method. J Colloid Interface Sci 258:116–122. Kawakami K, Nishihara Y, Hirano K. 1999. Liposome/emulsion transition induced by a-tocopheryl acetate. Langmuir 15:7454–7460. Florence AT, Attwood D. 1998. Physicochemical principles of pharmacy, 3rd ed. New York: Palgrave. pp 84–85. Shin DN, Wijnen JW, Engberts JBFN, Wakisaka A. 2001. On the origin of microheterogeneity: A mass spectroscopic study of dimethyl sulfoxide-water binary mixture. J Phys Chem B 105:6759–6762. Shin DN, Wijnen JW, Engberts JBFN, Wakisaka A. 2002. On the origin of microheterogeneity: A mass spectroscopic study of acetonitrile-water and dimethyl sulfoxide-water binary mixtures (part 2). J Phys Chem B 106:6014–6020.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 93, NO. 6, JUNE 2004