Bioorganic & Medicinal Chemistry Letters 25 (2015) 1561–1567
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters journal homepage: www.elsevier.com/locate/bmcl
Thermodynamic equilibrium solubility measurements in simulated fluids by 96-well plate method in early drug discovery q Sonali S. Bharate a,⇑, Ram A. Vishwakarma b,⇑ a b
Preformulation Laboratory, CSIR—Indian Institute of Integrative Medicine, Canal Road, Jammu 180001, India Medicinal Chemistry Division, CSIR—Indian Institute of Integrative Medicine, Canal Road, Jammu 180001, India
a r t i c l e
i n f o
Article history: Received 20 November 2014 Revised 2 February 2015 Accepted 6 February 2015 Available online 16 February 2015 Keywords: Thermodynamic equilibrium solubility Physiological medium 96-Well plate UV detection Shake flask method QSPR study
a b s t r a c t An early prediction of solubility in physiological media (PBS, SGF and SIF) is useful to predict qualitatively bioavailability and absorption of lead candidates. Despite of the availability of multiple solubility estimation methods, none of the reported method involves simplified fixed protocol for diverse set of compounds. Therefore, a simple and medium-throughput solubility estimation protocol is highly desirable during lead optimization stage. The present work introduces a rapid method for assessment of thermodynamic equilibrium solubility of compounds in aqueous media using 96-well microplate. The developed protocol is straightforward to set up and takes advantage of the sensitivity of UV spectroscopy. The compound, in stock solution in methanol, is introduced in microgram quantities into microplate wells followed by drying at an ambient temperature. Microplates were shaken upon addition of test media and the supernatant was analyzed by UV method. A plot of absorbance versus concentration of a sample provides saturation point, which is thermodynamic equilibrium solubility of a sample. The established protocol was validated using a large panel of commercially available drugs and with conventional miniaturized shake flask method (r2 >0.84). Additionally, the statistically significant QSPR models were established using experimental solubility values of 52 compounds. Ó 2015 Elsevier Ltd. All rights reserved.
Lipinski et al.1 caution that solubility is a decisive and key parameter for drug discovery. It is wise to solve solubility insufficiencies at discovery stage via structural modifications of a lead candidate so that chemistry can move from a pool of poorly soluble, orally inactive compounds towards those with some degree of oral activity. The understanding of physicochemical and solidstate properties of drug substance provides the basis to develop suitable dosage form. However, before this, it is prerequisite to know thermodynamic solubility so that the drug substance and its performance can be optimized.2–7 Drug penetration through lipid membrane depends on lipophilicity of the molecule and thus, may be correlated with partition coefficient value. Absorption of drug depends on the correct balance between these two opposite properties.8 A lack of solubility affects the ability of drug to achieve efficacious and toxicologically relevant exposures in animals. This parameter also affects future developability and formulation efforts for a lead candidate.9,10 The acceptance criteria and solubility classifications for discovery project teams and medicinal chemists propose that 100 lg/mL are considered to be highly soluble.11 This classification range is projected to provide general guidelines on potential solubility issues for human oral absorption. However, this criterion usually is too low for animal dosing in solution formulation. Based on these classifications, compounds with aqueous solubility of >100 lg/mL are unlikely to show solubility related issues in further development. However compounds with solubility between 1 and 100 lg/mL may require formulation development in order to overcome poor absorption issues associated with their low solubility. Compounds with even lower solubility typically represent a real formulation challenge.12,13 In spite of these known facts, solubility problems are often recognized for the first time when an oral solid dosage form or parenteral formulation has to be developed. This late identification of poor solubility is attributable to the extensive use of organic solvents at early stages of drug discovery and even in the first animal experiments. By the time formulation development starts, it is usually no longer practical to chemically change the active compound. Unfortunately, attempts to save drug candidate through suitable formulation are expensive, time-consuming and not always successful. Therefore, solubility of new substances should be determined as early as possible and is a critical step in assessing likely developability of a compound.1
1562
S. S. Bharate, R. A. Vishwakarma / Bioorg. Med. Chem. Lett. 25 (2015) 1561–1567
1
2
3
4
5
6
7
8
9
10
11
12
A B C
Absorbance
D E F G H
Concentration (µg/mL)
Figure 1. Experimental design for determination of thermodynamic aqueous solubility of compounds.
There exists variety of approaches for solubility determination at different phases of drug discovery and development including in silico approaches and experimental solubility (kinetic or thermodynamic). Measuring thermodynamic equilibrium solubility is considered as the ‘gold standard’.14,15 Kinetic solubility measurement involves predissolution of compound in a co-solvent [generally in DMSO (dimethyl sulfoxide)] followed by the precipitation after dilution in a suitable solution.16 However, the use of cosolvents leads to overestimation of solubility values.17–20 Thermodynamic solubility is helpful in diagnosing in vivo results as well as it serves as a guide for formulation development and for regulatory submissions. Thermodynamic equilibrium solubility experiments involve addition of an excess of solid compound to dissolution medium. The equilibrium concentration of a compound in solution is determined at the end of dissolution process. This is often considered as the ‘true’ solubility of compound and is generally measured using shake-flask method. This method is labourintensive and when performed in test tubes or vials, requires a large amount of compound. Thus, solubility experiment can be ongoing for several days due to long equilibration times.21,22 Among all available literature methods for solubility determination of compounds, most of them are for kinetic solubility.19,23–25 For measuring thermodynamic aqueous solubility, shake flask method is routinely used approach, which requires large amount of sample, as it uses high volume of dissolution medium. Therefore, the shake-flask approach is not suitable in early drug discovery lead optimization stage. Although miniaturized shake flask methods have also been established,12,26 still these protocols require significantly large amount of sample (100 mg), which is practically not amenable to medicinal chemistry lead optimization. Furthermore, during early development stage, large numbers of compounds/NCEs (new chemical entity) are synthesized with very small quantities (0.84). Chemical structures of data set used for this study are shown in Figures S1–S3.
Table 1 Experimental solubility data of model drugs in water, PBS, SGF and SIF Model compounds
Aceclofenac Albendazole Amphotericin B Anastrozole Aplysinopsin Ascorbic acid Aspirin Berberine HCl Bromocryptolepine Budesonide Carbamazepine Cetirizine Chlorzoxazone Ciprofloxacin Clarithromycin Clotrimazole Colchicine Colistin sulfate Crocin Curcumin Diclofenac sodium Docetaxel Embelin Fexofenadine HCl Fluconazole Flurbiprofen sodium Fluticasone propionate Garcinol Glycyrrhetinic acid Indomethacin Isogarcinol Isoniazid Khellin Loratadine Mangiferin Meloxicam N-Acetylcysteine Nifedipine Nimodipine Omeprazole Paracetamol Piperine Pyrazinamide Rifampicin Roxithromycin Simvastatin Streptomycin Tamoxifen Telmisartan Terbinafine HCl Tinidazole Valsartan a b
BCS class
II IV IV III na I I na na II II I/III II III IV II I IV na na II II na III III II II na na II na III na II na II na II II II I na III II IV II na II II I II II
Detection wavelength (nm)
270 290 320 260 390 260 280 340 280 250 280 240 280 280 250 270 350 250 430 430 280 230 290 230 280 280 270 280 250 260 280 270 330 250 260 360 280 280 360 300 280 340 270 330 230 240 250 280 300 280 310 260
Solubility by 96-well plate protocol, (mol/L)
log Sa
Solubility by shake flask method, Sa (mol/L)
log
Waterb (lit.)
PBS
SGF
SIF
Water
PBS
SGF
SIF
4.24 6.42 (6.31) 4.66 (4.09) 6.47 4.50 3.34 2.80 (2.78) 3.67 (3.27) 4.19 5.03 (5.33) 4.07 (4.29) 3.99 3.63 (3.75) 3.62 6.17 (7.35) 5.00 3.43 6.10 3.81 5.87 (6.78) 3.33 (2.54) 5.27 (6.06) 5.77 (5.99) 3.83 (3.55) 6.49 (6.48) 3.48 6.7 (6.99) 5.78 5.97 4.25 (4.20) 6.07 3.23 4.11 (4.01) 4.49 (6.06) 4.32 4.94 (5.64) 4.61 4.64 (5.76) 5.32 (5.24) 3.64 (3.84) 2.70 4.15 (4.85) 2.24 4.31 6.22 4.72 6.07 4.97 5.41 3.61 3.49 4.34 (4.38)
3.65 5.72 4.66 6.47 4.25 3.07 2.30 4.67 4.09 5.03 4.47 3.98 3.63 4.22 6.17 4.94 3.70 6.40 3.81 5.87 3.90 5.61 5.77 3.83 6.49 3.79 6.7 5.78 5.67 4.25 6.07 3.23 4.11 4.98 4.32 4.45 4.91 4.64 6.62 3.64 3.00 4.55 2.27 4.01 5.92 4.72 6.07 6.39 5.11 6.52 3.22 3.74
5.55 3.82 4.87 5.77 4.10 3.94 2.48 4.47 4.59 5.03 4.07 3.76 3.33 3.92 6.17 5.24 4.30 6.40 3.81 5.87 5.81 6.21 5.77 4.43 5.79 4.48 6.00 6.08 5.97 5.85 6.07 3.23 4.51 5.28 4.72 5.85 5.21 4.94 5.02 3.36 3.00 4.36 2.25 4.01 6.22 4.72 6.07 6.39 4.41 3.61 3.79 4.74
3.65 5.72 5.36 5.77 6.41 4.94 5.56 3.40 6.49 5.03 4.28 5.36 3.33 3.92 5.87 6.54 3.70 5.80 3.81 5.87 4.60 5.91 5.77 6.03 6.49 5.39 6.00 5.18 5.97 6.38 6.07 3.54 4.51 5.88 4.72 4.64 5.21 5.84 5.02 4.24 5.48 4.15 2.10 4.01 6.22 5.62 6.07 6.39 5.71 5.82 3.49 3.74
4.21 7.31 4.64 6.83 nd nd 2.80 nd nd 5.23 4.40 nd 3.84 nd nd nd nd nd nd nd 3.75 nd nd 3.85 6.27 nd 7.71 nd nd nd nd nd 4.00 4.52 nd 4.96 nd 4.55 5.39 nd 2.87 nd nd 4.21 nd 4.60 nd nd 4.53 3.64 nd 4.24
3.59 4.89 4.58 6.98 nd nd 2.79 nd nd 5.27 4.48 nd 3.78 nd nd nd nd nd nd nd 3.84 nd nd 4.17 6.33 nd 8.00 nd nd nd nd nd 4.14 4.91 nd 4.66 nd 4.44 6.40 nd 2.93 nd nd 4.14 nd 4.61 nd nd 5.19 6.17 nd 3.91
8.06 3.69 4.95 6.75 nd nd 2.84 nd nd 5.07 4.34 nd 3.47 nd nd nd nd nd nd nd 6.74 nd nd 4.33 6.07 nd 6.70 nd nd nd nd nd 4.01 3.84 nd 7.39 nd 4.87 4.97 nd 2.91 nd nd 4.02 nd 4.59 nd nd 4.39 3.81 nd 4.52
3.54 6.12 5.42 6.16 nd nd 5.43 nd nd 5.24 4.85 nd 3.62 nd nd nd nd nd nd nd 3.94 nd nd 6.43 6.05 nd 6.39 nd nd nd nd nd 4.23 5.88 nd 4.54 nd 6.24 5.00 nd 5.48 nd nd 4.14 nd 4.94 nd nd 5.64 6.21 nd 3.88
Average of three determinations. References for literature values are provided in the supporting information file; nd, not determined; na, not available.
1564
S. S. Bharate, R. A. Vishwakarma / Bioorg. Med. Chem. Lett. 25 (2015) 1561–1567
A compound in discovery phase, where aqueous solubility looks poor, it is relevant to measure solubility in bio-relevant media namely phosphate buffer saline pH 7.4 (PBS), simulated gastric fluid pH 1.2 (SGF) and simulated intestinal fluid pH 6.8 (SIF) which predicts behavior of drug candidate in the body.24,27,28,36–38 The solubility in water provides direction for preparing pre-clinical (animal) dosing; and PBS is a standard for some investigators who prefer this media for dosing, which is certainly valuable if IV dosing is required. SGF and SIF provide solubility at the extremes of GIT, which is valuable information. Many compounds contain ionizable groups and therefore measuring solubility at different pH levels (pH 1–8) is more useful than measuring it at single pH. Determining solubility at pH gradient of gastrointestinal tract provides more complete understanding of absorption in GIT.39 To more accurately reflect the physiologic conditions of GIT, the use of biorelevant test media viz. SGF with pepsin and SIF with pancreatin is more realistic.40–42 Although SGF and SIF (with digestive enzymes) are not exactly similar to in vivo digestion, their results are very close to the results of in vivo digestion of the mammalian system.43 Compound has to dissolve into gastric fluid and/or intestinal fluid so that it can cross lumen and achieve desired systemic exposure. It is noteworthy to mention that experimental solubility values in simulated fluids have never been reported in the literature for such large dataset. For most of the compounds used in this study, literature solubility in SGF/SIF is not available. Therefore, using our established and validated 96-well plate protocol, solubility of 52 compounds was measured in PBS, SGF and SIF (Table 1).30,44,45 Structural features of the compound have a great impact on solubility of compound in SGF (pH 1.2) and SIF (pH 6.8). The presence of ionizable functional groups in a structure makes it more soluble in respective media. This has been depicted by the representative structures in Figure 3. Solubility of compounds is generally represented as log S, where S is the concentration of compound in mol/L for a saturated aqueous solution in equilibrium. About 85% of drugs have ( )log S values between 1 and 5 and virtually none have values above 6. Values below 1 are not problematic because they are associated with highly polar molecules that may have low membrane permeability in the absence of active transport. It is evident that for most of the compounds, with ( )log S value between 1 and 5, the polarity needs to be compromised to get reasonable aqueous solubility with the hydrophobicity required for acceptable membrane passage.46
S
N
Cl
NH
N H O Albendazole
O N
O
Aceclofenac
Cl NH OH
Fexofenadine
OH
O
O
N
N
O
Cl HO
OH
N H Aplysinopsin
NH
O
N HN
Validation of 96-well plate protocol with miniaturized shake flask method. Experimental solubility of 26 commercially available compounds in water, PBS, SGF and SIF was determined by shake flask method. It was observed that, 96-well plate solubility results correlated well with the data obtained from shake-flask method. The correlation coefficient r2 obtained for 96-well plate and shake flask method in water, PBS, SGF and SIF were found to be 0.855, 0.915, 0.843 and 0.849, respectively (Fig. 4). Relative overestimation by 96-well method was most significant for the compounds with solubility values lower than 5 lg/mL (for example ibuprofen, albendazole, fluticasone propionate; denoted by red points in Fig. 4). However for compounds having low aqueous solubility (
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters journal homepage: www.elsevier.com/locate/bmcl
Thermodynamic equilibrium solubility measurements in simulated fluids by 96-well plate method in early drug discovery q Sonali S. Bharate a,⇑, Ram A. Vishwakarma b,⇑ a b
Preformulation Laboratory, CSIR—Indian Institute of Integrative Medicine, Canal Road, Jammu 180001, India Medicinal Chemistry Division, CSIR—Indian Institute of Integrative Medicine, Canal Road, Jammu 180001, India
a r t i c l e
i n f o
Article history: Received 20 November 2014 Revised 2 February 2015 Accepted 6 February 2015 Available online 16 February 2015 Keywords: Thermodynamic equilibrium solubility Physiological medium 96-Well plate UV detection Shake flask method QSPR study
a b s t r a c t An early prediction of solubility in physiological media (PBS, SGF and SIF) is useful to predict qualitatively bioavailability and absorption of lead candidates. Despite of the availability of multiple solubility estimation methods, none of the reported method involves simplified fixed protocol for diverse set of compounds. Therefore, a simple and medium-throughput solubility estimation protocol is highly desirable during lead optimization stage. The present work introduces a rapid method for assessment of thermodynamic equilibrium solubility of compounds in aqueous media using 96-well microplate. The developed protocol is straightforward to set up and takes advantage of the sensitivity of UV spectroscopy. The compound, in stock solution in methanol, is introduced in microgram quantities into microplate wells followed by drying at an ambient temperature. Microplates were shaken upon addition of test media and the supernatant was analyzed by UV method. A plot of absorbance versus concentration of a sample provides saturation point, which is thermodynamic equilibrium solubility of a sample. The established protocol was validated using a large panel of commercially available drugs and with conventional miniaturized shake flask method (r2 >0.84). Additionally, the statistically significant QSPR models were established using experimental solubility values of 52 compounds. Ó 2015 Elsevier Ltd. All rights reserved.
Lipinski et al.1 caution that solubility is a decisive and key parameter for drug discovery. It is wise to solve solubility insufficiencies at discovery stage via structural modifications of a lead candidate so that chemistry can move from a pool of poorly soluble, orally inactive compounds towards those with some degree of oral activity. The understanding of physicochemical and solidstate properties of drug substance provides the basis to develop suitable dosage form. However, before this, it is prerequisite to know thermodynamic solubility so that the drug substance and its performance can be optimized.2–7 Drug penetration through lipid membrane depends on lipophilicity of the molecule and thus, may be correlated with partition coefficient value. Absorption of drug depends on the correct balance between these two opposite properties.8 A lack of solubility affects the ability of drug to achieve efficacious and toxicologically relevant exposures in animals. This parameter also affects future developability and formulation efforts for a lead candidate.9,10 The acceptance criteria and solubility classifications for discovery project teams and medicinal chemists propose that 100 lg/mL are considered to be highly soluble.11 This classification range is projected to provide general guidelines on potential solubility issues for human oral absorption. However, this criterion usually is too low for animal dosing in solution formulation. Based on these classifications, compounds with aqueous solubility of >100 lg/mL are unlikely to show solubility related issues in further development. However compounds with solubility between 1 and 100 lg/mL may require formulation development in order to overcome poor absorption issues associated with their low solubility. Compounds with even lower solubility typically represent a real formulation challenge.12,13 In spite of these known facts, solubility problems are often recognized for the first time when an oral solid dosage form or parenteral formulation has to be developed. This late identification of poor solubility is attributable to the extensive use of organic solvents at early stages of drug discovery and even in the first animal experiments. By the time formulation development starts, it is usually no longer practical to chemically change the active compound. Unfortunately, attempts to save drug candidate through suitable formulation are expensive, time-consuming and not always successful. Therefore, solubility of new substances should be determined as early as possible and is a critical step in assessing likely developability of a compound.1
1562
S. S. Bharate, R. A. Vishwakarma / Bioorg. Med. Chem. Lett. 25 (2015) 1561–1567
1
2
3
4
5
6
7
8
9
10
11
12
A B C
Absorbance
D E F G H
Concentration (µg/mL)
Figure 1. Experimental design for determination of thermodynamic aqueous solubility of compounds.
There exists variety of approaches for solubility determination at different phases of drug discovery and development including in silico approaches and experimental solubility (kinetic or thermodynamic). Measuring thermodynamic equilibrium solubility is considered as the ‘gold standard’.14,15 Kinetic solubility measurement involves predissolution of compound in a co-solvent [generally in DMSO (dimethyl sulfoxide)] followed by the precipitation after dilution in a suitable solution.16 However, the use of cosolvents leads to overestimation of solubility values.17–20 Thermodynamic solubility is helpful in diagnosing in vivo results as well as it serves as a guide for formulation development and for regulatory submissions. Thermodynamic equilibrium solubility experiments involve addition of an excess of solid compound to dissolution medium. The equilibrium concentration of a compound in solution is determined at the end of dissolution process. This is often considered as the ‘true’ solubility of compound and is generally measured using shake-flask method. This method is labourintensive and when performed in test tubes or vials, requires a large amount of compound. Thus, solubility experiment can be ongoing for several days due to long equilibration times.21,22 Among all available literature methods for solubility determination of compounds, most of them are for kinetic solubility.19,23–25 For measuring thermodynamic aqueous solubility, shake flask method is routinely used approach, which requires large amount of sample, as it uses high volume of dissolution medium. Therefore, the shake-flask approach is not suitable in early drug discovery lead optimization stage. Although miniaturized shake flask methods have also been established,12,26 still these protocols require significantly large amount of sample (100 mg), which is practically not amenable to medicinal chemistry lead optimization. Furthermore, during early development stage, large numbers of compounds/NCEs (new chemical entity) are synthesized with very small quantities (0.84). Chemical structures of data set used for this study are shown in Figures S1–S3.
Table 1 Experimental solubility data of model drugs in water, PBS, SGF and SIF Model compounds
Aceclofenac Albendazole Amphotericin B Anastrozole Aplysinopsin Ascorbic acid Aspirin Berberine HCl Bromocryptolepine Budesonide Carbamazepine Cetirizine Chlorzoxazone Ciprofloxacin Clarithromycin Clotrimazole Colchicine Colistin sulfate Crocin Curcumin Diclofenac sodium Docetaxel Embelin Fexofenadine HCl Fluconazole Flurbiprofen sodium Fluticasone propionate Garcinol Glycyrrhetinic acid Indomethacin Isogarcinol Isoniazid Khellin Loratadine Mangiferin Meloxicam N-Acetylcysteine Nifedipine Nimodipine Omeprazole Paracetamol Piperine Pyrazinamide Rifampicin Roxithromycin Simvastatin Streptomycin Tamoxifen Telmisartan Terbinafine HCl Tinidazole Valsartan a b
BCS class
II IV IV III na I I na na II II I/III II III IV II I IV na na II II na III III II II na na II na III na II na II na II II II I na III II IV II na II II I II II
Detection wavelength (nm)
270 290 320 260 390 260 280 340 280 250 280 240 280 280 250 270 350 250 430 430 280 230 290 230 280 280 270 280 250 260 280 270 330 250 260 360 280 280 360 300 280 340 270 330 230 240 250 280 300 280 310 260
Solubility by 96-well plate protocol, (mol/L)
log Sa
Solubility by shake flask method, Sa (mol/L)
log
Waterb (lit.)
PBS
SGF
SIF
Water
PBS
SGF
SIF
4.24 6.42 (6.31) 4.66 (4.09) 6.47 4.50 3.34 2.80 (2.78) 3.67 (3.27) 4.19 5.03 (5.33) 4.07 (4.29) 3.99 3.63 (3.75) 3.62 6.17 (7.35) 5.00 3.43 6.10 3.81 5.87 (6.78) 3.33 (2.54) 5.27 (6.06) 5.77 (5.99) 3.83 (3.55) 6.49 (6.48) 3.48 6.7 (6.99) 5.78 5.97 4.25 (4.20) 6.07 3.23 4.11 (4.01) 4.49 (6.06) 4.32 4.94 (5.64) 4.61 4.64 (5.76) 5.32 (5.24) 3.64 (3.84) 2.70 4.15 (4.85) 2.24 4.31 6.22 4.72 6.07 4.97 5.41 3.61 3.49 4.34 (4.38)
3.65 5.72 4.66 6.47 4.25 3.07 2.30 4.67 4.09 5.03 4.47 3.98 3.63 4.22 6.17 4.94 3.70 6.40 3.81 5.87 3.90 5.61 5.77 3.83 6.49 3.79 6.7 5.78 5.67 4.25 6.07 3.23 4.11 4.98 4.32 4.45 4.91 4.64 6.62 3.64 3.00 4.55 2.27 4.01 5.92 4.72 6.07 6.39 5.11 6.52 3.22 3.74
5.55 3.82 4.87 5.77 4.10 3.94 2.48 4.47 4.59 5.03 4.07 3.76 3.33 3.92 6.17 5.24 4.30 6.40 3.81 5.87 5.81 6.21 5.77 4.43 5.79 4.48 6.00 6.08 5.97 5.85 6.07 3.23 4.51 5.28 4.72 5.85 5.21 4.94 5.02 3.36 3.00 4.36 2.25 4.01 6.22 4.72 6.07 6.39 4.41 3.61 3.79 4.74
3.65 5.72 5.36 5.77 6.41 4.94 5.56 3.40 6.49 5.03 4.28 5.36 3.33 3.92 5.87 6.54 3.70 5.80 3.81 5.87 4.60 5.91 5.77 6.03 6.49 5.39 6.00 5.18 5.97 6.38 6.07 3.54 4.51 5.88 4.72 4.64 5.21 5.84 5.02 4.24 5.48 4.15 2.10 4.01 6.22 5.62 6.07 6.39 5.71 5.82 3.49 3.74
4.21 7.31 4.64 6.83 nd nd 2.80 nd nd 5.23 4.40 nd 3.84 nd nd nd nd nd nd nd 3.75 nd nd 3.85 6.27 nd 7.71 nd nd nd nd nd 4.00 4.52 nd 4.96 nd 4.55 5.39 nd 2.87 nd nd 4.21 nd 4.60 nd nd 4.53 3.64 nd 4.24
3.59 4.89 4.58 6.98 nd nd 2.79 nd nd 5.27 4.48 nd 3.78 nd nd nd nd nd nd nd 3.84 nd nd 4.17 6.33 nd 8.00 nd nd nd nd nd 4.14 4.91 nd 4.66 nd 4.44 6.40 nd 2.93 nd nd 4.14 nd 4.61 nd nd 5.19 6.17 nd 3.91
8.06 3.69 4.95 6.75 nd nd 2.84 nd nd 5.07 4.34 nd 3.47 nd nd nd nd nd nd nd 6.74 nd nd 4.33 6.07 nd 6.70 nd nd nd nd nd 4.01 3.84 nd 7.39 nd 4.87 4.97 nd 2.91 nd nd 4.02 nd 4.59 nd nd 4.39 3.81 nd 4.52
3.54 6.12 5.42 6.16 nd nd 5.43 nd nd 5.24 4.85 nd 3.62 nd nd nd nd nd nd nd 3.94 nd nd 6.43 6.05 nd 6.39 nd nd nd nd nd 4.23 5.88 nd 4.54 nd 6.24 5.00 nd 5.48 nd nd 4.14 nd 4.94 nd nd 5.64 6.21 nd 3.88
Average of three determinations. References for literature values are provided in the supporting information file; nd, not determined; na, not available.
1564
S. S. Bharate, R. A. Vishwakarma / Bioorg. Med. Chem. Lett. 25 (2015) 1561–1567
A compound in discovery phase, where aqueous solubility looks poor, it is relevant to measure solubility in bio-relevant media namely phosphate buffer saline pH 7.4 (PBS), simulated gastric fluid pH 1.2 (SGF) and simulated intestinal fluid pH 6.8 (SIF) which predicts behavior of drug candidate in the body.24,27,28,36–38 The solubility in water provides direction for preparing pre-clinical (animal) dosing; and PBS is a standard for some investigators who prefer this media for dosing, which is certainly valuable if IV dosing is required. SGF and SIF provide solubility at the extremes of GIT, which is valuable information. Many compounds contain ionizable groups and therefore measuring solubility at different pH levels (pH 1–8) is more useful than measuring it at single pH. Determining solubility at pH gradient of gastrointestinal tract provides more complete understanding of absorption in GIT.39 To more accurately reflect the physiologic conditions of GIT, the use of biorelevant test media viz. SGF with pepsin and SIF with pancreatin is more realistic.40–42 Although SGF and SIF (with digestive enzymes) are not exactly similar to in vivo digestion, their results are very close to the results of in vivo digestion of the mammalian system.43 Compound has to dissolve into gastric fluid and/or intestinal fluid so that it can cross lumen and achieve desired systemic exposure. It is noteworthy to mention that experimental solubility values in simulated fluids have never been reported in the literature for such large dataset. For most of the compounds used in this study, literature solubility in SGF/SIF is not available. Therefore, using our established and validated 96-well plate protocol, solubility of 52 compounds was measured in PBS, SGF and SIF (Table 1).30,44,45 Structural features of the compound have a great impact on solubility of compound in SGF (pH 1.2) and SIF (pH 6.8). The presence of ionizable functional groups in a structure makes it more soluble in respective media. This has been depicted by the representative structures in Figure 3. Solubility of compounds is generally represented as log S, where S is the concentration of compound in mol/L for a saturated aqueous solution in equilibrium. About 85% of drugs have ( )log S values between 1 and 5 and virtually none have values above 6. Values below 1 are not problematic because they are associated with highly polar molecules that may have low membrane permeability in the absence of active transport. It is evident that for most of the compounds, with ( )log S value between 1 and 5, the polarity needs to be compromised to get reasonable aqueous solubility with the hydrophobicity required for acceptable membrane passage.46
S
N
Cl
NH
N H O Albendazole
O N
O
Aceclofenac
Cl NH OH
Fexofenadine
OH
O
O
N
N
O
Cl HO
OH
N H Aplysinopsin
NH
O
N HN
Validation of 96-well plate protocol with miniaturized shake flask method. Experimental solubility of 26 commercially available compounds in water, PBS, SGF and SIF was determined by shake flask method. It was observed that, 96-well plate solubility results correlated well with the data obtained from shake-flask method. The correlation coefficient r2 obtained for 96-well plate and shake flask method in water, PBS, SGF and SIF were found to be 0.855, 0.915, 0.843 and 0.849, respectively (Fig. 4). Relative overestimation by 96-well method was most significant for the compounds with solubility values lower than 5 lg/mL (for example ibuprofen, albendazole, fluticasone propionate; denoted by red points in Fig. 4). However for compounds having low aqueous solubility (
