strategies to address low drug solubility

1521-0081/65/1/315499$25.00 http://dx.doi.org/10.1124/pr.112.005660PHARMACOLOGICAL REVIEWS Pharmacol Rev 65:315499, January 2013Copyright 2013 by The American Society for Pharmacology and Experimental Therapeutics

ASSOCIATE EDITOR: ARTHUR CHRISTOPOULOS

Strategies to Address Low Drug Solubilityin Discovery and Development

Hywel D. Williams, Natalie L. Trevaskis, Susan A. Charman, Ravi M. Shanker, William N. Charman,Colin W. Pouton, and Christopher J. H. Porter

Drug Delivery, Disposition and Dynamics (H.D.W., N.L.T., W.N.C., C.J.H.P.), Centre for Drug Candidate Optimisation (S.A.C.),and Drug Discovery Biology (C.W.P.), Monash Institute of Pharmaceutical Sciences, Monash University, Parkville,

Victoria, Australia; and Pfizer Global Research and Development, Groton Laboratories, Groton, Connecticut (R.M.S.)

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319

A. What Is Low Drug Solubility? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320B. Determinants of Aqueous Solubility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323

1. Ideal Versus Nonideal Solubility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323C. Hydrophobic or Lipophilic Drug Candidates? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324D. Solubility of Electrolytes, Weak Electrolytes, and Nonelectrolytes . . . . . . . . . . . . . . . . . . . . . . . 324E. Solubility and Dissolution Rate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325F. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 326

II. In Vitro Complexities of Working with Poorly Water-Soluble Drugs . . . . . . . . . . . . . . . . . . . . . . . . . 326A. Drug Precipitation, Adsorption, Binding, and Complexation in In Vitro Assays . . . . . . . . . . 326B. Changes to Thermodynamic Activity Resulting from Complexation, Binding,

or Solubilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327III. In Vivo Assessment of Poorly Water-Soluble Compounds. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328

A. Parenteral Administration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3291. Complexities with Parenteral Administration of Poorly Water-Soluble Drugs. . . . . . . . . 3292. Parenteral Formulation Approaches for Poorly Water-Soluble Drugs . . . . . . . . . . . . . . . . . 329

B. Oral Administration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3311. Formulations to Support Drug Discovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3312. Use of Enabling Formulations to Promote Oral Absorption . . . . . . . . . . . . . . . . . . . . . . . . . . 3333. Preclinical Toxicology Formulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3334. Development of Clinical Formulations for Poorly Water-Soluble Drugs . . . . . . . . . . . . . . . 334

IV. Buffers and Salt Formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334A. Solution Behavior of Weak Electrolytes and Their Salts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335

1. Ionic Equilibria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3352. pH Solubility Relationships for Weak Electrolytes and Salts of Weak Electrolytes . . . . 3353. Factors Affecting Salt Formation at pHmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3374. Determinants of Salt Solubility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338

a. pHmax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338b. Choice of Counterion to Maximize Salt Solubility. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338c. The Effect of Common Ions on Salt Solubility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341d. Effect of Organic Solvents on Salt Solubility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342

B. pH adjustment Strategies for Addressing Low Drug Solubility . . . . . . . . . . . . . . . . . . . . . . . . . . 3421. Buffered Systems Used in Parenteral Formulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3422. Impact of Cosolubilizers and Electrolytes on pH-Mediated Solubilization . . . . . . . . . . . . . 343

a. pH Adjustment and Cosolvents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343b. pH Adjustment and Strong Electrolytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344c. pH Adjustment and Surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344

Address correspondence to: Christopher J. H. Porter, Drug Delivery, Disposition and Dynamics, Monash Institute of PharmaceuticalSciences, Monash University, 381 Royal Parade, Parkville, VIC, 3052, Australia. E-mail: [email protected]

dx.doi.org/10.1124/pr.111.005660.

315

by guest on May 10, 2014

pharmrev.aspetjournals.org

Dow

nloaded from

http://pharmrev.aspetjournals.org/

d. pH Adjustment and Cyclodextrins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3443. Effects of Dilution on Drug Solubilization by pH Adjustment . . . . . . . . . . . . . . . . . . . . . . . . 345

a. Methods for Assessing Precipitation Potential for Buffered Parenteral Formulations . 3464. Buffer Systems in Nonparenteral Formulations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346

C. The Use of Salts to Address Low Aqueous Solubility in Parenteral Formulations . . . . . . . . 347D. The Use of Salt Forms to Address Low Aqueous Solubility in Oral Formulations . . . . . . . . 347

1. The Use of Pharmaceutical Salts to Enhance Dissolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347a. Effect of Self-Buffering on Salt Dissolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348b. Effect of pH Changes on Drug Supersaturation and Precipitation. . . . . . . . . . . . . . . . . 349c. Potential for Salt Conversion to Un-Ionized Drug/Hydrates/Other Salt Forms In Situ 350d. Effect of Common Ions on Dissolution in the Gastrointestinal Tract . . . . . . . . . . . . . . 352

2. Physical Properties of Pharmaceutical Salts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3533. Potential Toxicity of Pharmaceutical Counterions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353

E. Feasibility of Salt Formation and Salt Selection Strategies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3541. Salt Formation Feasibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3542. Salt-Screening Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355

F. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356V. Optimization of Crystal Habit: Polymorphism and Cocrystal Formation . . . . . . . . . . . . . . . . . . . . . 357

A. Polymorphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3581. Crystal Packing, Polymorphism, and Phase Transformations . . . . . . . . . . . . . . . . . . . . . . . . 3582. Effect of Polymorphism on Drug Solubility, Dissolution Rate, and Oral Absorption . . . 359

B. Cocrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3601. Cocrystals as a Mechanism of Enhanced Drug Solubility and Dissolution . . . . . . . . . . . . 3602. Solubility Assessment of Cocrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3623. Solubility Advantages of Cocrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3634. Design and Preparation of Cocrystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3655. Recent Examples of Pharmaceutical Cocrystals for Improving Solubility

and Bioavailability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 366C. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367

VI. Cosolvents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367A. Rationale for the Use of Cosolvents for the Solubilization of Poorly Water-Soluble Drugs 367B. Commonly Used Organic Cosolvents in Parenteral Drug Delivery . . . . . . . . . . . . . . . . . . . . . . . 368C. Solubilization by Cosolvents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369

1. Cosolvent Effects on Solubility Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3692. Predicting Solubility in Cosolvent Systems: The Log-Linear Solubility Model. . . . . . . . . 370

D. Impact of Solubilizers and Electrolytes on Cosolvent-Mediated Drug Solubilization . . . . . . 3721. Cosolvents and pH Adjustment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3722. Cosolvents and Surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3723. Cosolvents and Cyclodextrins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3724. Cosolvents and Strong Electrolytes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373

ABBREVIATIONS: ABT-229; 8,9-anhydro-40-deoxy-39-N-desmethyl-39-N-ethylerythromycin B-6,9-hemiaceta; AMG 517, N-(4-[6-(4-trifluoromethyl-phenyl)-pyrimidin-4-yloxy]-benzothiazol-2-yl)-acetamide; AUC, area under the curve; BCRP, breast cancerresistant protein; BCS, Biopharma-ceutical Classification System; CD, cyclodextrin; CMC, critical micelle concentration; CNT, classical nucleation theory; CRA13,naphthalen-1-yl(4-(pentyloxy)naphthalen-1-yl)methanone; DG, diglyceride; DMA, dimethylacetamide; DMSO, dimethyl sulfoxide; DSC,differential scanning calorimetry; FA, fatty acid; FABP, fatty acid-binding protein; FaSSGF, fasted-state simulated gastric fluid;FaSSIF, fasted-state simulated intestinal fluid; FATP, fatty acid transport protein; FeSSIF, fed-state simulated intestinal fluid; FTIR,Fourier transform infrared spectroscopy; GI, gastrointestinal; HDL, high-density lipoprotein; HPC, hydroxypropyl cellulose; HLB,hydrophilic-lipophilic balance; HPMC, hydroxypropyl methylcellulose; HPH, high-pressure homogenization; HPMCAS, hydroxypropylmethylcellulose acetate succinate; K-832, 2-benzyl-5-(4-chlorophenyl)-6-[4-(methylthio)phenyl]-2H-pyridazin-3-one; L-883555, N-cyclopropyl-1-{3-[6-(1-hydroxy-1-methylethyl)-1-oxidopyridin-3-yl]phenyl}-1,4-dihydro-[1,8]naphthyridin-4-one 3-carboxamide; LBF, lipid-based for-mulations; LC, long chain; LCQ789, 5-(4-chlorophenyl)-1-phenyl-6-(4-(pyrazin-2-yl)phenyl)-1H-pyrazolo[3,4-d]pyrimidin-4(5H)-one;LDL, low-density lipoprotein; LFCS, Lipid Formulation Classification System; LPC, lysophosphatidylcholine; MC, medium chain; mdror MDR, multi-drug resistant; MG, monoglyceride; MRP, multiresistance protein; NMR, nuclear magnetic resonance; NSC-639829, N-[4-(5-bromo-2-pyrimidyloxy)-3-methylphenyl]-(dimemethylamino)-benzoylphenylurea; OZ209, cis-adamantane-2-spiro-39-89-(aminomethyl)-19,29,49-trioxaspiro[4.5]decane mesylate; PEG, polyethylene glycol; PG-300995, 2-(2-thiophenyl)-4-azabenzoimidazole; P-gp, P-glycoprotein; PL,phospholipid; PPI, polymer precipitation inhibitor; PVP, polyvinylpyrrolidone; RPR200765, {t-2-[4-(4-fluoro-phenyl)-5-pyridin-4-yl-1H-imidazol-2-yl]-5-methyl-[1,3]dioxan-r-5-yl}-morpholin-4-yl-methanone;; SCF, super critical fluids; SEDDS, self-emulsifying drug-delivery systems; SD,solid dispersion; SGF, simulated gastric fluid; SLN, solid lipid nanoparticle; TG, triglyceride; TPGS, D-a-tocopheryl polyethylene glycol succinate;TRL, triglyceride-rich lipoprotein; UWL, unstirred water layer; XRPD, X-ray powder diffraction.

316 Williams et al.

E. Effects of Dilution on Drug Solubilization by Cosolvents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374F. Potential Pharmacological Properties of Cosolvents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374

1. In Vitro Assessment of Cosolvent Toxicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3742. Lytic Effects of Commonly Used Cosolvents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375

G. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375VII. Surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375

A. Rationale for the Use of Surfactants to Enhance the Solubilization of PoorlyWater-Soluble Drugs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375

B. Commonly Used Nonionic Surfactants in Drug Delivery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377C. Solubilization by Surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378

1. Micelle Formation and Drug Solubilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3782. Quantification of Micellar Solubilization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3793. Effect of Surfactant Type and Structure on Drug Solubilization . . . . . . . . . . . . . . . . . . . . . . 3804. Effect of Temperature on Drug Solubilization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 380

D. Impact of Cosolubilizers and Electrolytes on Surfactant-Mediated Drug Solubilization . . . 3811. Surfactants and Cosolvents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3812. Surfactants and pH Adjustment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3813. Surfactant Mixtures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3824. Surfactants and Cyclodextrins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3835. Surfactants and Strong and Weak Electrolytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383

E. Effects of Dilution on Drug Solubilization by Surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383F. Potential Pharmacological Properties of Nonionic Surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . 384

1. Toxicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3842. Pharmacokinetic Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 386

G. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387VIII. Cyclodextrins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387

A. Rationale for the Use of Cyclodextrins in the Solubilization of Poorly Water-Soluble Drugs. . 387B. Drug-Cyclodextrin Inclusion Complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 389

1. Binding Equilibria between a Drug and Cyclodextrin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3892. Measurements of Drug-Cyclodextrin Binding Constants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3903. Secondary Equilibria: Micelles and Cyclodextrin Aggregate Formation . . . . . . . . . . . . . . . 392

C. Drug Factors Affecting Complexation by Cyclodextrins. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3931. Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3932. Polarity and Charge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3933. Presence of Counterions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395

D. Impact of Cosolubilizers and Excipients on Cyclodextrin-Mediated Drug Solubilization . . 3951. Cyclodextrins and Water-Soluble Polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3952. Cyclodextrins and Cosolvents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3963. Cyclodextrins and Surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 396

E. In Vivo Utility of Drug-Cyclodextrin Complexes in Parenteral Formulations. . . . . . . . . . . . . 3961. Effect of Cyclodextrins on Drug Pharmacokinetics after Intravenous Administration . 3962. Effects of Dilution on Drug Solubilization by Cyclodextrins . . . . . . . . . . . . . . . . . . . . . . . . . . 3973. Potential Toxicological Properties of Parenterally Administered Cyclodextrins. . . . . . . . 397

F. In Vivo Utility of Drug-Cyclodextrin Complexes in Oral (Nonparenteral) Formulations . . 3981. Effect of Cyclodextrins on Oral Drug Bioavailability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3982. Oral Administration of Physical Mixtures Versus Isolated

Drug-Cyclodextrin Complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3993. Using Cyclodextrins to Stabilize Amorphous Drug . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3994. Cyclodextrin Complexation as a Limitation to Oral Bioavailability . . . . . . . . . . . . . . . . . . . 4005. Potential Toxicological Properties of Orally Administered Cyclodextrins. . . . . . . . . . . . . . 4016. Effect of Cyclodextrins on Membrane Transport and Drug Metabolism . . . . . . . . . . . . . . . 402

G. Manufacture of Cyclodextrin Formulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4021. Parenteral Formulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4022. Solid Drug-Cyclodextrin Complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402

H. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402IX. Particle Size Reduction Strategies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403

A. Particle Size Effects on Dissolution, Solubility, and In Vivo Performance . . . . . . . . . . . . . . . . 404

Strategies for Low Drug Solubility 317

1. Effect of Particle Size on Dissolution Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4042. Effect of Particle Size on Saturated Solubility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4053. Effect of Particle Size and Shape on the Diffusional Layer Thickness . . . . . . . . . . . . . . . . 405

B. Common Methods to Reduce Particle Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4061. Top-Down Particle Size Reduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 406

a. Pearl Milling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 406b. High-Pressure Homogenization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407

2. Bottom-up Nanoparticle Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408a. Controlled Crystallization Via Solvent Shift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409b. Precipitation after Solvent Evaporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 410

C. Oral and Parental Delivery of Formulations Containing Nanosized Drug Particles. . . . . . . 4111. Oral Delivery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4112. Nanosuspensions for Parenteral Delivery of Poorly Water-Soluble Drugs . . . . . . . . . . . . . 411

D. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415X. Solid Dispersions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 416

A. Classification of Solid Dispersions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4171. Nontraditional Solid Dispersions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419

B. Mechanisms by which Solid Dispersions Enhance Dissolution Rate andOral Bioavailability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4191. Reduced Particle Size and Enhanced Drug Wetting and Solubilization . . . . . . . . . . . . . . . 4192. Administration of Drug in the Amorphous Form. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4213. Maintenance of Supersaturation after Drug Dissolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421

a. The Impact of Dissolution Rate and Supersaturation Ratio onPrecipitation Inhibition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423

b. Mechanisms by which Polymers Maintain Supersaturation in Solution . . . . . . . . . . . 423c. Screening for Potential Polymer Precipitation Inhibitors. . . . . . . . . . . . . . . . . . . . . . . . . . 425

C. Recent Examples of Bioavailability Enhancement Using Solid Dispersions . . . . . . . . . . . . . . 426D. Role of the Carrier and the Drug-Carrier Ratio in Dictating Drug Release Kinetics

from Solid Dispersions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427E. Common Methods to Manufacture Solid Dispersions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 429

1. Melting/Fusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4292. Solvent Evaporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4293. Solvent Evaporation Using Supercritical Fluids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4304. Electrospinning and Microwave Irradiation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 431

F. Physical Stability of Amorphous Solid Dispersions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4311. Structural Changes at the Glass Transition Temperature. . . . . . . . . . . . . . . . . . . . . . . . . . . . 4312. Glass-Forming Potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4323. Polymer Effects on the Glass Transition Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4334. Moisture Effects on Glass Transition Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 434

G. Factors Affecting Drug Crystallization from Solid Dispersions. . . . . . . . . . . . . . . . . . . . . . . . . . . 4341. Drug-Polymer Intermolecular Association . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4352. Inhibition of Crystal Nucleation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4353. Molecular Mobility, Strength, and Fragility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4364. Limitations to the Use of Thermal Analysis to Estimate Molecular Mobility. . . . . . . . . . 4395. Drug-Polymer Miscibility and Relationship to Drug Loading . . . . . . . . . . . . . . . . . . . . . . . . . 439

H. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 444XI. Lipid-Based Formulations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445

A. Mechanisms of Bioavailability Enhancement by Lipid-Based Formulations . . . . . . . . . . . . . . 4461. Stimulation of Intestinal Lipid Absorption and Transport Pathways . . . . . . . . . . . . . . . . . 4462. Enhanced Drug Dissolution and Solubilization in the Intestinal Lumen . . . . . . . . . . . . . . 4503. Enhanced Intestinal Permeability and Inhibition of Intestinal Efflux and

First-Pass Metabolism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 452a. Correlation of In Vitro Effects of Lipid-Based Formulation Excipients on

Permeability with Changes to In Vivo Exposure?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4574. Promotion of Lymphatic Drug Transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 459

B. Design and Formulation of Lipid-Based Formulations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4611. Lipid-Based Fomulations for Parenteral Administration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 461

318 Williams et al.

2. Lipid-Based Formulations for Oral Administration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4633. The Lipid Formulation Classification System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4644. Selection of Lipid-Based Formulation Excipients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 465

a. Solvent Capacity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 466b. Mutual Miscibility. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 466c. Toxicity/Irritancy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 466d. Purity and Chemical Complexity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467e. Capsule Compatibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467f. Polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 468

C. Assessment of Lipid-Based Formulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4681. In Vitro Dispersion Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4682. In Vitro Digestion Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4683. In Vivo Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 470

D. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471XII. Emerging Strategies for Improving the Aqueous Solubility of Poorly Water-Soluble Drugs . . . 472

A. Drug Adsorption to Microporous Adsorbents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 472B. Solid Lipid Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 473

XIII. Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 474Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 475References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 476

AbstractDrugs with low water solubility are pre-disposed to low and variable oral bioavailability and,therefore, to variability in clinical response. Despitesignificant efforts to design in acceptable developabil-ity properties (including aqueous solubility) during leadoptimization, approximately 40% of currently marketedcompounds and most current drug development candi-dates remain poorly water-soluble. The fact that somanydrug candidates of this type are advanced into develop-ment and clinical assessment is testament to an increas-ingly sophisticated understanding of the approachesthat can be taken to promote apparent solubility in thegastrointestinal tract and to support drug exposure afteroral administration. Here we provide a detailed com-mentary on the major challenges to the progression ofa poorly water-soluble lead or development candidate

and review the approaches and strategies that can betaken to facilitate compound progression. In particular,we address the fundamental principles that underpin theuse of strategies, including pH adjustment and salt-formselection, polymorphs, cocrystals, cosolvents, surfac-tants, cyclodextrins, particle size reduction, amorphoussolid dispersions, and lipid-based formulations. In eachcase, the theoretical basis for utility is described alongwith a detailed review of recent advances in the field.The article provides an integrated and contemporarydiscussion of current approaches to solubility anddissolution enhancement but has been deliberatelystructured as a series of stand-alone sections to allowalso directed access to a specific technology (e.g., soliddispersions, lipid-based formulations, or salt forms)where required.

I. Introduction

High hydrophobicity and intrinsically low water solu-bility are increasingly common characteristics of hits,leads, development candidates, and ultimately marketeddrugs. Many hypotheses have been put forward as towhy these trends have emerged, and the true explana-tion is clearly multifaceted. The application of combi-natorial chemistries to generate large chemical librariesand the common application of high-throughput screen-ing modalities, often in nonaqueous media (or mixed-solvent media), have probably played a role. The desirefor increased potency, coupled with the realization thatreceptor binding is mediated, at least in part, byhydrophobic interactions, further magnifies the likelihoodthat drug candidates will have limited aqueous solubility.Finally, the quest to explore unprecedented drug targets,some of which are associated with intracellular signalingpathways, lipid processing architecture, or highly lipo-philic endogenous ligands, only amplifies the requirement

for highly lipophilic, poorly water-soluble drug candidatesto access and interact with the target.

These drivers ultimately bias the identification ofpoorly water-soluble hits during early drug screens.Poor water solubility is a significant risk factor in loworal absorption because drug molecules must, in mostcases, be in solution to be absorbed, and oral bioavail-ability is usually a required characteristic in a targetproduct profile of an orally administered medicine. Assuch, medicinal chemistry strategies during lead opti-mization typically seek to modify physicochemical pro-perties (including solubility) such that drug leads havemore developable characteristics. Many decision gates,or idealized character panels, are used to identify andreject drug candidates with inappropriate developabilityproperties and, subsequently, to synthetically modifystructures to improve physicochemical characteristics.Perhaps the best known of these is Chris Lipinskis ruleof 5 (Lipinski et al., 1997), but there are many others.


In all cases, however, at least moderate water solubilityis usually a focus.Nonetheless, even with contemporary medicinal

chemistry programs and increasingly sophisticatedlead optimization strategies, it is apparent that forsome targets, reducing lipophilicity and increasingwater solubility will result in an unacceptable reductionin potency. In spite of attempts to circumvent solubilityproblems, approximately 40% of currently marketeddrugs (Fig. 1) and up to 75% of compounds currentlyunder development have been suggested to be poorlywater-soluble (Di et al., 2009, 2011). Furthermore, theproblems of low water solubility do not seem to bediminishing and may well be increasing (Takagi et al.,2006). Low water solubility therefore continues to bea challenge to successful drug development.This review seeks to provide an overview of the

strategies that may be taken to address the problemsof low solubility during drug discovery and develop-ment and includes a comprehensive review of formu-lation approaches to support the clinical developmentof oral and parenteral drug products for poorly water-soluble drugs. In addition, even when lead optimiza-tion is successful in increasing aqueous solubility, earlypreclinical studies are still required with less thanoptimal leads to provide the data sets to allow informedprogression or rejection; therefore, we also address thestrategies that might be used when dealing with thecomplexities of low solubility during the discoveryphase. In the former case, at least for oral drug products,the market typically dictates the need for traditionalsolid dosage forms (e.g., capsules, tablets). In the latter,where studies are usually conducted in small animals(rodents), liquid formulations are often used to alloworal dosing, and therefore a slightly different ap-proach must be taken.

This review is structured to provide an initial andrelatively brief introduction to the determinants of lowwater solubility to provide the theoretical basis forthe approaches that might be taken to address solu-bility challenges. We subsequently present summarysections that outline potential strategies for addressingsolubility issues, first in vitro and second in vivo, withthe latter section addressing both parenteral and oraladministration. Subsequently, we provide a compre-hensive review of the technologies that can be used topromote solubility or dissolution. These latter sectionsare necessarily dense and are intentionally separatedfrom the higher-level strategy summaries provided inthe introduction. The technology overviews providea reference source for the summaries that precede them.A schematic representation of each of these formulationapproaches is provided in Fig. 2 and is intended tohighlight the variety of proven strategies available tothose working with poorly water-soluble drugs.

A. What Is Low Drug Solubility?

Although low water solubility of drug candidatespresents varied and significant challenges throughoutdrug discovery and development, the greatest concernis generally the risk of reduced and variable absorp-tion after oral administration. The value at whichlimited solubility begins to impact absorption isdifficult to state definitively since it is dependent ona number of other system variables, including drugpermeation, dose, and the environment presentwithin the gastrointestinal (GI) tract. To understandthese variables, and therefore, the factors that impactthe required solubility for a drug candidate, it isinstructive first to appreciate the drivers of flux across anabsorptive membrane since these in turn will determinewhether the drug dose can be absorbed over the timescaleavailable.

Assuming appropriate chemical and metabolic sta-bility and an absence of transporter or carrier-mediatedprocesses, flux (F) across an absorptive membrane is theproduct of the concentration gradient and the perme-ability across the membrane and is described as follows:

F5D K A

hCm 2C0 1

where Cm is the drug concentration immediatelyadjacent to the membrane, C0 is the drug concentrationon the abluminal side of the membrane, K is the par-tition coefficient between the aqueous solution overlay-ing the membrane and the membrane itself, h is thewidth of the diffusion layer, D is the diffusion coefficientof the drug in the membrane, and A is the surface area.Assuming that the drug concentration on the abluminalside of the absorptive membrane is low relative to theconcentration at the absorptive site (i.e., Cm .. C0), eq.1 can be simplified to the following:

Fig. 1. A comparison of the distribution of solubilities for the top 200 oraldrug products in the United States (US), Great Britain (GB), Spain, andJapan and from the World Health Organization (WHO) Essential DrugList. Very soluble drugs: over 1000 mg/ml; freely soluble drugs: 1001000mg/ml; soluble drugs: 33100 mg/ml; sparingly soluble drugs: 1033 mg/ml;slightly soluble drugs: 110 mg/ml; very slightly soluble drugs: 0.11 mg/ml;practically insoluble drugs: ,0.1 mg/ml. Adapted from Takagi et al. (2006).

320 Williams et al.

F5D K A

hCm 2

The values for D, K, and h are typically fixed fora particular system and are used to define thepermeability coefficient (P) where

P5D Kh

3

In the absence of supersaturation, the maximumconcentration that can be attained at the surface of anabsorptive membrane is equivalent to the equilibriumsolubility (Cs) of the drug, and therefore the maximumflux (per unit area) (F) is the product of solubility andpermeability:

F95P Cs 4Appreciation of this relationship illustrates that

knowledge of the solubility alone is insufficient toanticipate whether solubility will limit flux (or absorp-tion) since flux is also a function of permeability. Tosome extent, therefore, low solubility can be offset byhigh permeability; similarly, if permeability is low, therequirements for solubility to generate appropriateflux increase.A well recognized approach applied in early drug

discovery for estimating the required solubility andpermeability needed to achieve good oral absorption isthe concept of a maximum absorbable dose (MAD),

originally derived by Johnson and Swindell (1996) andfurther applied by Curatolo (1998) and Lipinski (2000):

MAD 5 Cs kabs SIWV SITT 5

where Cs is the solubility (mg/ml) at pH 6.5 (represent-ing the pH of the small intestine); kabs is the rateconstant (h21) for intestinal absorption (which isrelated to the permeability); SIWV is the smallintestinal water volume (in milliliters), which istypically assumed to be ;250 ml (the volume of fluidassumed to be present in the fasted GI tract aftera glass of water has been drunk when takingmedication orally); and SITT is the small intestinaltransit time (min) of ;270 min (4.5 h). Rearrangingthis relationship provides an expression for thenecessary or target solubility for a given dose and kabs(or permeability) and provides an initial indication asto whether solubility is likely to limit oral absorption.This concept is shown graphically in Fig. 3, the datafrom which are taken from a seminal review thatshows the theoretical required solubility to providegood oral absorption for drugs with projected dosesranging from 0.1 to 10 mg/kg and permeabilitiesranging from low to high. At one end of the spectrum,highly potent drugs for which the dose is low and themembrane permeability is high have relatively lowsolubility requirements to achieve good oral absorp-tion. At the other extreme, low-potency drugs for whichthe dose is high and the permeability is low need

Fig. 2. Schematic diagram illustrating the common strategies currently used (and discussed in this review) to address low drug solubility in drugdiscovery and development.


considerably higher solubility for good oral absorption(by several orders of magnitude in this example). Asa broad initial estimation, and assuming moderatepotency (1.0 mg/kg) and moderate permeability (kabs 51.0 h21), this approach indicates that where aqueoussolubilities are ,50 mg/ml, problems associated withlow water solubility might be anticipated. It is clear,however, that the required solubility to support drugabsorption must be evaluated in light of both thepotency (or dose) and the permeability characteristics.It is also clear that at stages during the developmentpathway, in particular during preclinical toxicitytesting, exposure at doses considerably in excess ofthe predicted clinical dose will be required, magnifyingthe need for solubility support.A further application of the solubility-permeability

relationship to oral drug absorption is the Biophar-maceutics Classification System (BCS) (Fig. 4),originally developed by Amidon et al. (1995), withsubsequent variations by others (Wu and Benet, 2005;Butler and Dressman, 2010; Chen et al., 2011). Theprinciples of the BCS are well described elsewhere(Amidon et al., 1995; Yu et al., 2002; Dahan et al., 2009),but in brief, the BCS allows classification of drugmolecules as a function of their solubility and

permeability properties. Originally proposed to providea scientific basis for biowaivers based on a correlation ofin vitro drug dissolution and in vivo drug absorption,this classification system has found much broaderapplicability across many areas of drug discovery anddevelopment. According to the BCS, class I moleculesare those having both high solubility and high perme-ability (and therefore likely few problems with oralabsorption); class II compounds are those that have lowsolubility and high permeability (where solubility is theprimary limitation to absorption); class III compoundshave high solubility but low permeability (whereabsorption is limited by membrane permeation andnot solubility); and class IV compounds are those inwhich both poor solubility and poor permeability limitdrug absorption. The focus of the current review istherefore BCS class II compounds, which often exhibitsolubility-limited absorption. Class IV compounds arealso relevant, although they have additional prob-lems associated with low permeability.

Drug dose is also an important factor in the BCSbecause highly soluble drugs are defined as those inwhich the highest dose will dissolve in 250 ml overthe pH range of the GI tract (i.e., pH 16.8). Severalexamples of dose, solubility, and volume requirements(taken from Amidon et al., 1995) are shown in Table 1.As with the MAD, these calculations are not designed tobe definitive but rather illustrate that low-dose com-pounds, such as digoxin, can have good absorption andbioavailability, even when GI solubility is low, whereasthe absorption of high-dose compounds, such as griseo-fulvin, is more often low, variable, and highly formula-tion dependent as a result of their solubility limitations.

A complication of the BCS definition of highsolubility is that the highest strength dose must besoluble in 250 ml of water at all pH values that mightbe encountered in the GI tract. Therefore, drugs maybe classified as class II even though they have goodsolubility at one end of this pH range. For example,many weak acids have low solubility at pH 1 and arestrictly classified as BCS class II compounds but arequite soluble at intestinal pH (pH 67) and in manycases do not exhibit solubility-limited absorption. Assuch, a BCS class II designation does not alwaysdictate that solubility will be a limitation to absorption;rather, compounds in this BCS class are more likely tobe solubility limited than are those in class I. Indeed,

Fig. 3. The relationship between the projected dose and the requiredaqueous solubility for low, medium, and high permeability compounds.Permeability is indicated by the magnitude of the absorption rateconstant (kabs), where low permeability, kabs 5 0.1; medium 5 1.0, high 510 h21. From this analysis, moderately permeable compounds (kabs 51.0 h21), with projected potencies of 1.0 mg/kg, require aqueoussolubilities of ;52 mg/ml. Adapted from Lipinski (2000).

TABLE 1Examples of dose, solubility, and volume requirements for a range of poorly water-soluble drugs

Drug Dose Solubility Volume Required forComplete Solubilization Absorption

mg mg/ml ml

Piroxicam 20 0.007 2857 Low, variableDigoxin 0.5 0.024 21 GoodGriseofulvin 500 0.015 33,333 Low, variableChlorthiazide 500 0.78 636 Reasonable

From Amidon et al. (1995).

322 Williams et al.

a more practical description of BCS class II may bedrugs that do not have high solubility rather thanthose that have low solubility since the classificationsystem specifically identifies compounds that fall intoclass I and all those with solubilities below this fallinherently into class II (or class IV if permeability isalso low) (Fig. 4).The preceding discussion serves to illustrate further

that low solubility is a somewhat arbitrary conceptwhen assessing the likelihood that solubility will limitdrug absorption and that additional knowledge of thelikely dose and membrane permeability is inevitablyrequired to put a solubility value into an appropriatecontext.

B. Determinants of Aqueous Solubility

A detailed description of the thermodynamic deter-minants of drug solubility in aqueous media is beyondthe scope of the current discussion; for more informa-tion, the interested reader is directed to the followingreferences: Grant and Higuchi (1990), Yalkowsky (1999),and Murdande et al. (2010a). An overview of the broadphysicochemical drivers of drug solubility is war-ranted, however, since these drivers underpin theapproaches that can be taken to enhance solubility.Simplistically, the potential for a drug (solute) to passinto solution in an aqueous fluid (solvent) is dictated bythree separate events, as shown in Fig. 5. First, solutemolecules must be abstracted from the solid state,a process that involves breaking solute-solute bonds.The strength of solute-solute attractive forces in dif-ferent solids varies significantly and is typically higherfor electrolytes than for nonelectrolytes (since ionicattractive forces in the solid state are stronger thannonionic forces), for crystalline solids compared withamorphous materials, and for planar nonelectrolytes(which pack more effectively in the solid state)

compared with nonplanar molecules. The melting pointprovides a reasonable indication of the strength of in-termolecular solute-solute interactions in the solidmaterial. Second, a void must be created in the solventthat is sufficient to accommodate the abstracted solutemolecule. Since intermolecular forces in the liquidstate are much lower than those in the solid state, theenergy required to create a void in the solvent is lowand is usually ignored when assessing energy changesduring dissolution. Finally, the solute molecule isinserted into the solvent void. For molecules with someaffinity for a polar solvent such as water, this process isenergetically favorable and therefore drives drug solu-bility. This last concept is the rationale behind the oftenquoted maxim like dissolves like. This is largely true,but it captures only half the story since it ignores theimpact of changes to solid-state properties on solubility.Since the energy transitions associated with changesto solvent structure are small, it is apparent that thereare two primary determinants of drug solubility: 1) theenergy required to overcome the strength of intermo-lecular forces in the solute solid state and 2) the energygenerated on the interaction of solute and solventmolecules in solution (solvation)

For an analogous structural series, solubility istherefore generally lower for molecules that havehigher melting points (stronger attractive forces) andlower affinity for water (poor solvation).

1. Ideal Versus Nonideal Solubility. Solubilitytheory defines an idealized condition or an ideal solu-tion, where the intermolecular forces between soluteand solvent are equivalent to those between solute andsolute and those between solvent and solvent. Underthese circumstances, the mixing of solute and solventmolecules in the liquid state results in no net energy

Fig. 4. Diagrammatic representation of the BCS, which classifies drugsaccording to their permeability and solubility properties. Compounds aredefined as high solubility when the quantity of drug that is present inthe highest strength immediate release dose form is soluble in 250 ml ofwater across the likely range of gastrointestinal pH (1.27.5). Highlypermeable compounds are defined as those that are .90% absorbedor where permeability as assessed by in vitro or in vivo methods isequivalent to or higher than that of a reference compound that is 90%absorbed. Adapted from Amidon et al. (1995).

Fig. 5. Three essential steps are required for a solute drug molecule to bedisplaced from a solid particle and to enter solution. Step 1: A singlesolute molecule is removed from the crystal lattice; energy is required inthis step to overcome solute-solute interactions in the solid state. Step 2:A void is created within the solvent to accommodate the solute molecule.Although this step also requires energy, it is likely to be considerablylower than the energy required in step 1. Step 3: The solute moleculeinserts into the solvent, forming solute-solvent interactions. Simplisti-cally, if the energy released from the solute-solvent interactions (i.e., step3) is greater than the energy required for steps 1 and 2, solubility isfavored.


change. Where this is the case, solubility is dependenton the strength of the solute crystal lattice and may bedefined by

logX52DHf

2:303R

Tm 2TTmT

6

where X is the ideal mole fraction solubility of solute,DHf is the enthalpy of fusion of solute, Tm is solutemelting point, T is the absolute temperature, and R isthe gas constant.In reality, ideal solutions are highly unusual, and

solution conditions close to ideality exist pharmaceu-tically only in solutions of highly lipophilic drugs innonaqueous (lipidic) solutions. In contrast, in aqueoussolution, the differences between solute-solute inter-actions and solute-solvent interactions are highly sig-nificant, leading to nonideal solution behavior andmuch lower solubility than would otherwise be pre-dicted. In this case, the difference between ideal andnonideal behavior is captured by a correction factortermed the activity coefficient (g), where

logX52DHf

2:303R

Tm 2TTmT

2 log g 7

In turn, the activity coefficient is a function of themolar volume of the solvent (Vs), the volume fraction ofthe solute (w1), and the difference in the solubilityparameters for the solvent (d1) and solute (d2):

log g 5Vsw21

2:303RTd1 2 d22 8

The solubility parameters provide a measure of thecohesive forces in either solute or solvent, and from eq. 8,it is evident that smaller differences between thesolubility parameters of the solvent and solute give riseto lower activity coefficients and, therefore, an increasein solubility toward ideality (i.e., d1 5 d2) (Rubino, 2002).The solubility equation (eq. 7) captures the determi-

nants of solubility as dictated by changes to the solid-state properties of a drug (and usually manifest bychanges to melting point) and the activity coefficient(reflecting differences between the solubility parame-ters), and it reiterates the importance of solute-soluteinteractions in the solid state and the need to minimizedifferences between solute and solvent properties tomaximize solvation (i.e., like dissolves like). Theseprinciples underpin all the approaches to solubilityenhancement that are subsequently described in thisreview, as each of these approaches either change solid-state properties or change the nature of the interactionbetween drug (solute) and solvent molecules.

C. Hydrophobic or Lipophilic Drug Candidates?

An understanding of the primary drivers of drugsolubility allows an important distinction to be made

between poorly water-soluble drugs that are limited bysolid-state properties (e.g., the strength of the crystallattice), those that are limited by solvation (i.e., solute-solvent interactions in solution), and those that arelimited by both. In practice, most compounds fall intotwo categories since almost all poorly water-solublecompounds have limited affinity for water (i.e., they arehydrophobic) and are therefore solvation-hydrationlimited. Where compounds are hydrophobic and alsoshow strong intermolecular forces in the solid state,they are typically poorly soluble in both aqueous andnonaqueous solvents. In contrast, hydrophobic mole-cules, in which solubility is not limited by solid-stateproperties, often show varying degrees of solubility innonaqueous solvents such as lipids (since the molecularproperties that reduce hydration in aqueous media oftenpromote solvation in nonaqueous media). The lattercompounds are therefore both hydrophobic and lipo-philic, with the former simply being hydrophobic. Thedifference between these two types of molecules can beillustrated using the analogies of brick dust for hydro-phobic molecules with poor solubility in all solvents andgrease balls for compounds that are hydrophobic andlipophilic and show reasonable solubility in lipids (Stellaand Nti-Addae, 2007). This distinction is importantsince the formulation options available for eitherhydrophobic or lipophilic compounds differ considerably.Simplistically, the increased lipid solubility of lipophilicdrug candidates allows access to liquid surfactants andlipid-based delivery technologies that can be filled intosoft gelatin capsules (or sealed hard gelatin capsules),whereas the lack of solubility for hydrophobic moleculesin almost all vehicles precludes formulation in anythingother than modified solid dosage forms.

D. Solubility of Electrolytes, Weak Electrolytes,and Nonelectrolytes

The solvation properties of drug molecules, andtherefore a significant part of the driving force for drugsolubility in aqueous media, are highly dependent on theextent of ionization. The presence of a charged functionalgroup provides an opportunity for favorable ion-dipoleinteractions with polar solvents such as water, whichdirectly enhance hydration and water solubility. Strongelectrolytes (such as NaCl) that completely dissociate inwater are generally highly water soluble. This is notalways the case, however, as solubility remains a com-posite function of the energy required to break thecrystal lattice versus the energy liberated on hydrationof the ions formed. In some extreme cases, for example,an inorganic salt such as AgCl, the crystal lattice energyis sufficiently high that aqueous solubility remains low.As a strong electrolyte, therefore, dissociation of AgClmolecules in solution is complete, but as a poorly water-soluble strong electrolyte, solubility is limited.

In reality, most drugs are organic materials that areeither nonelectrolytes (which do not dissociate to form

324 Williams et al.

ionic species in water) or weak electrolytes thatdissociate only partially in water such that both un-ionized solute and the dissociated ions are present insolution.The solubility of weak electrolytes is highly de-

pendent on the degree of ionization (dissociation) sincethe affinity of the ionized species for water is markedlyhigher than that of the un-ionized species. The degreeof dissociation is in turn dependent on the pKa of theweak electrolyte and the pH of the solution into whichit is dissolving. Simplistically, at pH values above thepKa of a weak acid and below the pKa of a weak base,solubility increases significantly as a result of ioniza-tion. For weak acids and bases with a single ionizablegroup, solubility increases by a factor of 10 for every pHunit away from the pKa (although this trend does notcontinue indefinitely and solubility is ultimatelylimited by the solubility product of the salts that areformed in situ with available counterion; see SectionIV). Nonetheless, optimization of solution pH is aneffective means by which solubility can be enhancedand is commonly used to enhance the solubilityproperties of solution formulations for weak electro-lytes. For solid dosage forms, the principles of pH-dependent solubility may be manipulated by theisolation of a drug (or drug candidate) as an appropri-ate salt form. The use of pH and salt selection toenhance solubility and the dissolution rate is describedin more detail in Section IV. For nonelectrolytes,solubility behavior is not complicated by the effects ofionization and remains a function of hydration and thestrength of the crystal lattice.

E. Solubility and Dissolution Rate

The solubility of a drug in aqueous solution isa fundamentally important property that affects notonly the potential for drug absorption after oraladministration and the ability to administer the drugparenterally but also the ease of manipulation andtesting in the laboratory and during manufacture.However, drug solubility is an equilibrium measureand the rate at which solid drug or drug in a formula-tion passes into solution (i.e., the dissolution rate) isalso critically important when the time available fordissolution is limited. This rate is particularly relevantafter oral administration since intestinal transit timeis finite and the rate of drug dissolution must sig-nificantly exceed the rate of transit for absorption to bemaximized. For example, the absorption of a drug withreasonable solubility may still be poor if the rate ofdissolution is low since the solubility limit may neverbe reached during the intestinal transit time. Simi-larly, even where the rate of dissolution is relativelyrapid, if the equilibrium solubility is low, the quantityof drug available in solution for absorption is unlikelyto support rates of flux across the intestine that aresufficient to absorb the entire drug dose in the time

available. For different drugs, and under differentcircumstances, either solubility or dissolution rate (orboth) may be the limiting feature.

The process of drug dissolution from the solid state issummarized in Fig. 6. An unstirred water layer ispresent on the surface of every dissolving solid andprovides the barrier to drug equilibration with the wellstirred bulk solution. The dissolution rate is thereforedefined by the rate at which drug diffuses across theunstirred water layer, and the equations that describedrug dissolution (i.e., the Noyes Whitney Equation, eq.9) are analogous to simple diffusion equations. The rateof transfer across the unstirred layer is a function of theconcentration gradient across the unstirred layer, thewidth of the diffusion layer (h), the surface area ofcontact of the solid with the dissolution fluid (A), andthe diffusion rate of the drug in water (D). Theconcentration gradient in turn is a function of themaximum drug concentration at the surface of thedissolving solid (drug solubility Cs) and the concentra-tion in the well-stirred bulk (C) (Noyes and Whitney,1897):

dcxdt

5D Ah

Cs 2C 9

If we assume sink conditions where the concentra-tion of drug in bulk solution (C) is low relative to theconcentration on the surface of the dissolving solid (Cs),then this relationship collapses to

dcxdt

5D Ah

Cs 10

Fig. 6. Graphic depicting the process of drug dissolution from a soliddrug particle. An unstirred water layer of width (h) is present on thesurface of the dissolving solid. A concentration gradient is establishedacross the unstirred water layer that drives dissolution and results fromthe difference in drug concentration between that on the surface of thedissolving solid (usually assumed to be the equilibrium solubility of thedrug, Cs) and the concentration in the well stirred bulk (C).


For most small molecules, the diffusion rate constantin water (D) is relatively high and manipulation ofdrug structure does not typically affect D to the extentthat it has a significant impact on dissolution rate.Similarly, whereas the width of the diffusion layer canbe altered by agitation or stirring in vitro, this aspectcannot be easily manipulated in vivo.The major determinants of in vivo drug dissolution

rate are therefore solubility and surface area. Sincesolubility is a function of the strength of the crystallattice and the affinity of the solute (drug) for theaqueous environment, three major strategies can bedefined to increase the solubility or dissolution rate(realizing that the dependence of dissolution onsolubility dictates that increases in solubility inher-ently increase the dissolution rate):

Reducing intermolecular forces in the solid state(increases solubility and dissolution rate)

Increasing the strength of solute-solvent interac-tions in solution (increases solubility and dissolu-tion rate)

Increasing the surface area available for dissolu-tion (increases dissolution rate, potential to mod-erately increase solubility at very small particlesizes (,1 mm)

F. Summary

Low aqueous solubility and reduced dissolutionrates are a common property of many new drugcandidates, and these properties create a number ofchallenges during drug discovery and development.An understanding of the determinants of solubilityand dissolution provides a framework from whichapproaches to enhance solubilization may be devel-oped. In subsequent sections of this review, we firstaddress the complexities of working with poorlywater-soluble drugs in vitro and subsequently sum-marize the approaches that can be taken to assist inthe development of both parenteral and oral formula-tions. The main body of the review follows andprovides a detailed account of the technologicalapproaches that can be taken to support the de-velopment of effective formulations for poorly water-soluble drugs. Comment is made as to the manydifferent approaches that might be taken during leadoptimization and preclinical development and alsothose strategies that are also appropriate for exten-sion into clinical development and ultimately to themarket. To constrain the scope of this review, syntheticmedicinal chemistry approaches to solubility manipula-tion are not addressed and the discussion is limited toapproaches that do not result in the generation ofa fundamentally new chemical entity. For more in-formation on approaches to solubility manipulation viastructural modification, the interested reader is directedto the following reviews: Fleisher et al. (1996), Stella and

Nti-Addae (2007), Di et al. (2009), Keseru and Makara(2009), and Ishikawa and Hashimoto (2011).

II. In Vitro Complexities of Working with PoorlyWater-Soluble Drugs

Despite the drive to identify drug-like leads withacceptable physicochemical properties, lead series areoften plagued by poor aqueous solubility. The basis forthis trend was discussed earlier in this introduction,but regardless of the source, working with inherentlypoorly water-soluble compounds creates a number ofchallenges throughout drug discovery, beginning withprimary activity screens and progressing through tosecondary in vitro assays and in vivo assessment.

In many in vitro assays, there is little scope for im-proving solubility given the poor tolerability of many invitro biologic test systems for solubilizing components.In this instance, the focus must be on understandingthe consequences of simple solution preparation andmanipulation (e.g., dilution), appreciating the poten-tially confounding effects of compound precipitation onassay results, and grasping the impact of commonlyused buffer components in dictating free compoundconcentrations in solution. The sections that followhighlight some of the common problems associated withthe in vitro testing of poorly water-soluble compoundsand, where available, approaches that can be taken toovercome, or at least minimize, these issues.

A. Drug Precipitation, Adsorption, Binding, andComplexation in In Vitro Assays

Beyond the realms of chemical synthesis, most invitro evaluations of compound performance are con-ducted using aqueous-based biologic buffers and re-lated media. The range of in vitro testing protocols is,of course, broad but might include binding or displace-ment assays, enzyme inhibition studies, activity screen-ing in cell culture, traditional organ-bath pharmacology,assessment of uptake and transport in cell culturesystems, or excised tissues and metabolic stabilitystudies using microsomes or hepatocytes. In all cases,an accurate knowledge of the concentration of drug insolution is required as it is critical to the determinationof the experimental endpoint. For example, most invitro assays of drug activity are based on a concentra-tion-response relationship, with the endpoint beingsome measure of the inhibitory or effective concentra-tion (e.g., IC50 or EC50). In other assays, changes to drugconcentrations in solution during the assay are used asan indicator of cellular uptake or transport, metabolism,or binding, all of which rely on a known concentration ofcompound in solution.

In most drug discovery settings, moderate- to high-throughput assay formats (i.e., 96-, 384-, or 1536-wellplates) are used, and compounds are introduced intoaqueous biologic media via dilution of concentrated

326 Williams et al.

stock solutions prepared in water miscible organicsolvents, the most common being dimethyl sulfoxide(DMSO). DMSO is an excellent solvent for many poorlysoluble compounds, including those with diversechemical structures, and allows for the preparation ofhighly concentrated stock solutions for subsequentdilution for most compounds. However, compoundprecipitation after dilution of concentrated cosolventsolutions is common (see Section VI) and can lead tovariable responses depending on how the dilution isconducted and the composition of the final assaymedia. This in turn can lead to a high degree ofvariability and inconsistent results between assayswhere these variables may be different. Differentbiologic assays also vary in their tolerance to the finalDMSO (or other cosolvent) concentrations, making itdifficult to standardize experimental conditions anddilution procedures across different assay formats.For many in vitro assays (with the exception of high-

throughput assays specifically designed to assesssolubility), it may be difficult, if not impossible, todetect the presence of finely precipitated material ondilution into aqueous media, and measuring the finalconcentration of drug may be impractical. Workingwith compounds with inherently low aqueous solubilitiesgenerally limits the available range of drug concentra-tions that can be used to define concentration-dependentprocesses. Under these circumstances, complete bindingor inhibition profiles may be difficult to define since thesolubility limit is reached before saturation of bindingsites or approach to maximal inhibition.Furthermore, the physicochemical properties that

predispose compounds to low aqueous solubility canalso lead to an association of drug molecule withhydrophobic environments and surfaces. In fact, thisphenomenon is often a driving force for biologic potencysince partitioning into and across cell membranes andinteraction with cellular and molecular targets arethermodynamically favored compared with residencewithin an aqueous solution. However, these character-istics also lead to an inherent propensity for non-specific adsorption to surfaces such as tubes, pipettetips, filters, syringes, multiwell plates and cellularsupports. Under these circumstances, a decrease indrug concentration in solution may be reflective ofnonspecific adsorption rather than binding, uptake, ortransport, artificially reducing the concentration ofdrug in solution and leading to an erroneous endpointdetermination if concentrations are assumed on thebasis of only the dilution factor. Clearly, there areadvantages to obtaining measured concentrations toprovide confidence in the experimental results whenpractical and also to provide an assurance of massbalance, which is a critical control measurement in anymass transport experiment.The adsorption of drug to filter membranes requires

special mention since separation of free drug from

bound, complexed, or precipitated drug is a commonprocedure during the conduct of many in vitro assays.Assessment of the potential for adsorption duringfiltration is an important validation step but isparticularly critical for poorly water-soluble drugs.Where significant adsorption is evident, and unavoid-able, then it may be necessary to avoid the process offiltration. Equilibrium dialysis may provide an alter-native under such circumstances, but the potential foradsorption to the dialysis membrane is also high. Afinal approach is to use ultracentrifugation to separatelarger drug complexes from free drug in solution, butthese measurements are time consuming, requirespecialized equipment and a significant density differ-ence between species, and ultimately still carry the riskof drug adsorption to centrifuge tubes or plates.

Several different approaches can be taken to over-come issues of adsorption. The first is the choice of tube,filter, culture flask, multiwell plate, filter or pipette tip,as many manufacturers supply materials with modifiedsurface properties to reduce nonspecific adsorption. Itis important to appreciate the consequences of usingdifferent surfaces as those specifically designed to, forexample, promote cell adhesion by making them morehydrophobic, will also increase the likelihood of non-specific drug adsorption. In contrast, more hydrophilicsurfaces will generally reduce adsorption by reducingthe thermodynamic favorability of drug leaving thelargely aqueous solution environment. Another ap-proach involves pre-exposing potential adsorptionsites to drug solution with a view to saturating adsorptionbefore conduct of the experiment. Finally, alteration ofthe solution properties can reduce adsorption by in-creasing drug affinity for the bulk solution and reducingthe effective partition coefficient between solution anddrug adsorption sites. Most commonly, this is achievedvia the inclusion of small quantities of a cosolvent(depending on the sensitivity of the particular assay) orby manipulating solution pH to increase solute-solventinteractions in solution. The use of pH and cosolvents toenhance solubility is described in more detail inSections IV and VI, respectively.

Adsorption may also be reduced via the addition ofa complexation or solubilization agent such as a sur-factant (see Section VII), cyclodextrin (see SectionVIII), or protein. Although these approaches are oftenhighly effective, they introduce a further complexity,namely, changes to the chemical potential or thermo-dynamic activity of free (unbound) drug in solution asdiscussed in the following section.

B. Changes to Thermodynamic Activity Resultingfrom Complexation, Binding, or Solubilization

The effective concentration of a species in solution ismost accurately defined by its thermodynamic activity(a), which is related to the concentration (C) and theactivity coefficient (g):


a 5 g C 11A detailed evaluation of thermodynamic activity and

chemical potential is beyond the scope of this review,but in simple terms, the effective concentration orthermodynamic activity can be considered as the con-centration of drug in solution that is unconstrained byinteraction with any other molecular species and istherefore available to exert its effect, regardless ofwhether the effect is to bind to a receptor or to diffuseacross a membrane. In concentrated solutions, forexample, the close molecular proximity of drug moleculesto each other may promote solute-solute interactions(i.e., enhance intramolecular association), thereby re-ducing thermodynamic activity. Under these circum-stances, the activity coefficient is less than unity, andthe effective concentration or activity is less than themeasured concentration. In typical dilute solutions, thedegree of dilution is expected to reduce solute-soluteintermolecular interactions in solution such that eachmolecule effectively acts independently, and underthese circumstances, the activity coefficient is unityand activity is equal to concentration. This allows thedefinition of equilibrium constants, for example, indilute solution using drug concentration instead of thethermodynamically correct use of activities.The relevance of this discussion to the in vitro

testing of poorly water-soluble drugs is that manystrategies that promote drug solubility in aqueoussolution and assay media can change the thermody-namic activity of the drug in solution. This is mostreadily illustrated by considering the impact of theaddition of a surfactant to an aqueous solution of drug.Surfactants are amphiphilic molecules, and in aqueoussolutions, they can exist as either monomers or asmicellar structures. As surfactant concentrations areincreased above the critical micelle concentration(CMC), the concentration of monomeric surfactantremains constant while the concentration of surfactantpresent as micelles in solution increases. Surfactantmicelles typically enhance drug solubility by providinga hydrophobic environment in the micellar core tosolubilize poorly water-soluble drugs as they havea greater affinity for this hydrophobic environment thanfor the surrounding aqueous environment (see SectionVII). However, under this circumstance, the thermody-namic activity of the drug is much lower than the totalconcentration as the activity coefficient of the drug islower than unity. In the case of drug solubilized ina surfactant micelle, drug can be considered as existingin equilibrium between solubilized drug within themicelle and free drug in an intermicellar phase or bulksolution phase. Here, solubilized drug is highly con-strained by the surrounding micellar structure, and theeffective concentration (i.e., the thermodynamic activity)is more accurately represented by the free (nonmicellar)concentration of drug.

In addition to the use of surfactants, other commoncircumstances where the thermodynamic activity ofa compound might be reduced include the introductionof a complexation agent, such as a cyclodextrin, or theaddition of plasma or serum proteins to cell-basedassay media. For the latter situation, the so-calledserum shift phenomenon is widely recognized where invitro activity changes in response to changing concen-trations of serum present in the media. For a morecomprehensive review of the effect of protein bindingon the pharmacological activity of drugs and the com-plexities of interpreting static (in vitro) versus dynamic(in vivo) situations, the reader is referred to theexcellent review by Smith et al. (2010). In each ofthese cases, the total concentration of drug is signifi-cantly higher than the effective or free concentration,and the situation is further confounded by the difficultyin accurately quantifying free drug concentrations. Incontrast, solubility enhancement through pH manipula-tion or cosolvency will typically have limited impact onthermodynamic activity of drug in solution.

III. In Vivo Assessment of PoorlyWater-Soluble Compounds

In vivo assessment of new drug candidates beginswith early in vivo efficacy studies in animal models andcontinues through pharmacokinetic and dose-limitingtoxicology studies in various preclinical species andultimately into human clinical trials. Preclinicalpharmacokinetic studies are heavily used to supporthuman dose and pharmacokinetic predictions and,along with results from toxicology studies, are used toselect a safe starting dose in humans. For in vivoevaluations in animals, there are multiple approachesthat can be used to facilitate i.v. administration and toimprove exposure after oral dosing by the use ofenabling formulations. However, there is always a riskthat early incorporation of an enabling formulationapproach may shift the focus away from structuraloptimization. Even in circumstances where lead opti-mization strategies are successful in identifying drugcandidates with improved solubility properties, it islikely that at some stage during drug discovery therewill be the need to assess less soluble early leads forintrinsic activity and proof of concept efficacy to justifycompound or series progression.

A complication of poor aqueous solubility in the invivo assessment of compounds throughout drug dis-covery and development is that compound supply isfrequently limited (particularly in early discovery), andmaterial that is available most likely will not have thefinal, or optimal, solid-state properties. As compoundsprogress through discovery and into development, thesolid-state properties will almost certainly change;crystal forms will become better defined, particle sizereduction may be introduced, and salt forms will likely

328 Williams et al.

become available for compounds with ionizable func-tional groups. The timing at which these changes occurduring development will impact early preclinical data,and this needs to be appreciated and factored intostudy design and data interpretation.

A. Parenteral Administration

1. Complexities with Parenteral Administration ofPoorly Water-Soluble Drugs. In comparison with thelarge number of drugs intended for oral administration,parenteral formulations constitute a more limited pro-portion of marketed products. However, the generationof useful parenteral formulations remains a key compo-nent of drug discovery and drug development for almostall drug molecules (regardless of the intended route ofclinical administration) since data obtained after i.v.administration is necessary to generate fundamentalpharmacokinetic parameters (volume of distribution,clearance, half-life, absolute bioavailability) to supportcompound optimization and progression, and such dataare also useful to support early stage pharmacodynamicassessment of drugs intended for acute administration.Low aqueous solubility significantly complicates the

development of i.v. formulations since, in almost allcases, simple solution formulations are required. Thefollowing section describes several different approachesthat can be used as i.v. solution formulations, dependingon the physicochemical properties of the drug and theintended use (e.g., animal or humans). Depending onthe formulation strategy adopted, increases in solubilityof several orders of magnitude are possible. The risksassociated with i.v. administration of poorly water-soluble drugs include, first, the potential for drugprecipitation on administration and, second, the poten-tial for formulation components to cause pain, phlebitis,inflammation, hemolysis, or unacceptable toxicology.Precipitation is more likely for formulations where pHadjustment or cosolvents form the basis for drug solu-bilization since both are likely to be altered significantlyon dilution (see Sections IV and VI). In contrast, sur-factant solubilization, complexation, and i.v. emulsionsare far less sensitive to dilution (see Sections VII, VIII,and XI). Where dilution-mediated precipitation ispossible, slow administration into large veins, wherethe blood flow is higher (and therefore the supply ofplasma proteins and lipoproteins to provide in vivobinding is higher) is preferred, and this can also reducethe possibility of pain on injection, hemolysis, andphlebitis since the irritant component is maximallydiluted. For studies in rats and dogs, this can beaccomplished by implanting a dosing cannula into thejugular vein (or another large vein) with administrationover a period of 5 to 10 min for rats and up to 30 min orlonger in dogs. In contrast, administration into smallveins in the extremities (where blood flow is muchslower), such as the tail vein in rodent models, and rapidbolus injection, should be avoided with poorly water-

soluble drugs, if possible, because of the potential forprecipitation. Addition of small quantities of surfactants(,0.5%) or polymers may assist in preventing pre-cipitation at the injection site.

Drug administration via other parenteral routes,including s.c., i.m., and i.p. administration, is onoccasion warranted, particularly in proof-of-conceptbiology studies where there is evidence of poor oralexposure resulting from solubility limitations. Withnon-i.v. parenteral routes, the potential for the routeand the formulation vehicle to have an effect (eitherdesirable or undesirable) on the absorption profile isgreater, and depot effects are often seen. For theseadministration routes, solution formulations are stillpreferred since the volume of fluid at the injection siteis low, and, therefore, dissolution of suspension for-mulations may be limited. The lack of formulationdilution at the injection site dictates that the range ofexcipients and conditions to promote drug solubiliza-tion is even more limited than it is after i.v. admini-stration. Ideally, non-i.v. parenteral formulationsshould be as close as possible to physiologic pH, andconcentrations of cosolvents should be minimized.

2. Parenteral Formulation Approaches for PoorlyWater-Soluble Drugs. The most common approachesfor the development of parenteral formulations forpoorly water-soluble drugs are summarized in thefollowing discussion, and strategies for rapid identifi-cation of the most appropriate approach are discussed.The acceptable limits for each of these will depend onmany factors including the species (animal or human),dose volume, rate of administration (e.g., infusion orbolus), parenteral route (i.v., i.m., s.c., i.p.), and durationof treatment (Gad et al., 2006; Li and Zhao, 2007). Asummary of parenteral formulation approaches is givenin Table 2, and these are briefly described as follows.

pH adjustment (see Section IV) is a powerfulmechanism by which the solubility of weak acids andweak bases may be enhanced, but the approach islimited by the pKa of the compound and the acceptablepH range of the formulation. In this regard, acidicsolutions are generally more readily tolerated than arebasic solutions, and working pH ranges of 29 (Li andZhao, 2007) or 49 (Lee et al., 2003) have beensuggested. The pH of parenteral solutions can bemanipulated via the addition of small quantities ofstrong acids or bases, such as HCl or NaOH but is moreusefully achieved via the use of a buffer, generally withas low a buffer concentration as possible while stillmaintaining the desired pH. Common buffer systemsused in parenteral formulations include phosphate andbicarbonate at neutral or slightly basic pH and citrate,lactate, tartrate, acetate, and maleate for more acidicsolutions (Flynn, 1980; Kipp, 2007).

Cosolvents (see Section VI) are also commonly usedto promote solubility in parenteral formulations, andethanol, propylene glycol (PG), low-molecular-weight


polyethylene glycol (PEG 300 or 400), dimethylaceta-mide (DMA), and DMSO have been used. Combina-tions of cosolvents provide benefits since cosolvency isadditive, whereas toxicity is often dependent on theparticular solvent.In many cases, judicious manipulation of solution pH

in combination with cosolvents is sufficient to providesolubilization and is typically viewed as the first-lineapproach. For example, in a review of 317 discoverycompounds, Lee et al., (2003) reported that .90% ofcompounds could be formulated using pH and cosol-vents and that .60% of these could be formulatedusing cosolvent levels of less than 50% v/v.However, where pH adjustment and cosolvents fail

to provide the required solubility, alternative approachesmust be used. Second-tier approaches include the use ofcyclodextrins, surfactants, and lipid emulsions and,ultimately, more complex formulations such as lip-osomes, microemulsions, or nanosuspensions. Formore complex systems, however, the potential effect,either positive or negative, on systemic clearance anddistribution (and also absorption for non-i.v. paren-teral formulations) should be carefully considered.Surfactants (see Section VII) can also be used to

enhance drug solubility via micellar solubilization.Surfactants are generally limited to those that arenonionic because of their more favorable safety profileand may include polysorbates (e.g., Tween 80), Cremo-phors (including Cremophor EL and RH40), SolutolHS-15 (BASF Corporation, Washington, DC), vitamin E

TPGS, and poloxamers (Pluronics, including PluronicF68; BASF Corporation). Whereas surfactants providefor relatively robust solubilization, a complexity is thegrowing realization that many surfactants can affecttransport processes, such as cellular efflux, and mayalso result in significant adverse effects, includinghypersensitivity reactions (Gelderblom et al., 2001; tenTije et al., 2003). Cremophor EL, in particular, seems topromote an immune reaction in some species (mostnotably dogs) when administered parenterally. Surfactantsbring additional complexities to data interpretation and,where possible, might usefully be avoided in favor of pHchanges, cosolvents, or the use of cyclodextrins.

Complexation agents (see Section VIII), such ascyclodextrins, provide a mechanism for enhancingsolubility for many drugs and have been increasinglyused as modified cyclodextrins, such as hydroxypropyl-b-cyclodextrin (HP-b-CD) and sulfobutylether-b-cyclo-dextrin (SBE-b-CD), have become more cost effective.HP-b-CD and SBE-b-CD have higher aqueous

strategies to address low drug solubility

Documents