12 ROLE OF HPLC IN PREFORMULA TION Irina Kazakevich 12.1 INTRODUCTION Preformulation is a bridge between discovery and development where devel- opment scientists participate in selection and optimization of lead compounds. It is very critical at this stage to evaluate the developability of potential drug candidates in order to select new chemical entities and decrease the number of failures during future drug development. On average, only one out of ten new chemical entities (NCE) entering first- in-human testing reaches registration, approval, and marketing stage. The reasons for failures of development compounds include problems with bio- pharmaceutical properties, clinical safety, toxicology, efficacy, cost of goods, and marketing (see Figure 12-1) [1, 2]. The biopharmaceutical properties such as gastrointestinal and plasma solubility, lipophilicity (LogD), permeability, first-pass metabolism, systemic metabolism, protein binding, and in vivo bioavailability are related to the solubility, chemical stability, and permeabil- ity of drug candidates and have to be considered at discovery lead selection before recommendation to the development stage. A major challenge in any drug discovery program is achieving reasonable bioavailability upon oral administration; therefore, any information that high- lights potential problems with cell permeability and absorption is valuable when reviewing structural families as leads for drug discovery. Lipinski et al. [3] have reviewed 2245 compounds selected from the United States Adopted 577 HPLC for Pharmaceutical Scientists, Edited by Yuri Kazakevich and Rosario LoBrutto Copyright © 2007 by John Wiley & Sons, Inc. Name (USAN), International Nonproprietary Name (INN), and World Drug Index (WDI), comparing calculated physical properties and clinical exposure. Four parameters were chosen that were associated with solubility and per- meability, namely, molecular weight, octanol/water partition coefficient, the number of hydrogen bond donors, and the number of hydrogen bond acceptors. It was concluded that compounds are most likely to have poor absorption when molecular weight is >500, the calculated LogP is >5, the number of hydrogen bond donors is >5, and the number of hydrogen bond acceptors is >10. Lipinski has referred to this analysis as “rule of five” because the cutoffs for each of the four parameters were all close to five or a multiple of five. The rule of five can serve as qualitative absorption/permeability predictor. The absorption of drug molecules in the gastrointestinal tract is dependent upon the pK a of the compound and the pH of the gastrointestinal region (Figure 12-2). Almost 63% of all drugs are ionized in aqueous solution and can exist in a neutral or a charged state, depending on the pH of the local environment [4]. Based on the major goal of preformulation—identification of possible failure in future development—numerous studies are performed to fully char- acterize prospective drug candidates. The major analytical technique in each preformulation group is liquid chromatography. Ninety percent of all ana- lytical equipment in preformulation groups are HPLC systems equipped with UV and MS detection systems. HPLC is a fast and reliable method for con- centration and identity determination by UV and/or MS detection, respec- tively.The type of HPLC methods differ based on the specific preformulation tests that will be described below. In the early stage of preformulation, characterization of the drug molecule involves ionization constants and partition coefficient determinations, aqueous and nonaqueous kinetic and equilibrium solubility determination, pH solubil- ity profile, chemical stability assessment, and salt and polymorph screening. Assessment of biopharmaceutics and toxicological screening are also essential 578 ROLE OF HPLC IN PREFORMULATION Figure 12-1. Reasons for attrition from 1991 to 2000. (Reprinted with permission from reference 1.) at this stage. At the later stage of preformulation, after recommendation of NCE to development, the development support from preformulation group involves a more detailed solid-state characterization program, elaborating on moisture sorption, compressibility, melting point, particle size, shape, and surface area assessments, as well as excipient compatibility and prototype for- mulation stability evaluation. Further information on the role of preformulation in drug development process can be found in several excellent monographs [6–8] with the focus on pharmaceutical aspects of process development. 12.2 INITIAL PHYSICOCHEMICAL CHARACTERIZATION (DISCOVERY SUPPORT) During the early discovery stage the medicinal chemists use in vitro activities and fast in vivo small animal studies to discover the best compound to develop. The support from development scientist consists of providing information about LogP, pK a , and LogD for ionizable drugs and aqueous solubility. These physical characteristics can affect the absorption of drug candidate and, there- fore, drug bioaivalability.The requirements for HPLC analysis at this stage are speed and efficiency of the separation. It is critical to mention that at the early stage of discovery, very little information is available about the properties of INITIAL PHYSICOCHEMICAL CHARACTERIZATION (DISCOVERY SUPPORT) 579 Figure 12-2. Physical properties of the gastrointestinal tract. (Reprinted with permis- sion from reference 5.) molecule and only a few milligrams of compound is available for characteri- zation. Therefore, it is important to choose the most efficient column and the simplest mobile phase. Also, recommended is the use of more contemporary HPLC systems as UPLC from Waters employing columns with dimensions of 50 × 2.1mm, 1.8-µm particle size and the Fast 1200 system from Agilent with column dimensions of 50 × 4.6mm, 1.8-µm particle size, respectively, to enhance the turnaround time for sample analysis. Other platforms would include using Chromolith Speedrod ® monolithic columns at high flow rates. Also, taking into consideration the short column length, gradient elution should be recommended for all HPLC methods at this stage of drug candidate characterization. The post-run equilibration time is not significant in the case where short columns are used, and dwell volume is improved significantly for a new generation of HPLC systems. Many types of modeling techniques are available in the discovery phase of drug development, from structure activity relationships (SAR) to physiology based pharmacokinetics (PBPK) and pharmacokinetics-/pharmacodynamics (PK/PD) to help choosing some of the lead compounds. Some tests that are carried out by discovery include techniques related to structure determina- tion, metabolism, and permeability: NMR, MS/MS, elemental analysis, PAMPA, CACO-2, and in vitro metabolic stability. Although they are impor- tant as a part of physicochemical molecular characterization under the bio- pharmaceutics umbrella, they will not be discussed here. The reader can find relevant information in numerous monographs [9, 10]. 12.2.1 Ionization Constant, pK a Most potential drug candidates are weak bases or acids. Solubility and many other properties of the drug molecule is dependent on its ionization state. Acids are usually considered to be proton donors and bases are proton accep- tors. Any drug molecule with basic functionality in aqueous media holds the following equilibrium: (12-1) where the ionization equilibrium constant could be expressed as (12-2) It is obvious from the above equilibrium that the ratio of ionic to nonionic form of the drug in the solution is controlled by the proton concentration, which is commonly represented by pH values (negative logarithm of proton concentration). Taking the negative logarithm of expression (12-2), the well- known Henderson–Hasselbalch equation could be obtained: K a = [] ⋅ [] [] + BH BH + BH B H + ↔+ + 580 ROLE OF HPLC IN PREFORMULATION (12-3) T his allows for the estimation of the prevailing drug form at a particular pH. Ionic form of any organic molecule is usually more soluble in aqueous media, while the neutral form is usually more hydrophobic and thus shows an increased affinity for lipids. Variation of the ionization state of the molecule at different pH has typical sigmoidal shape (as shown in Figure 12-3). Corresponding expression for this dependence could be derived from equation (12-2) and the mass balance of the ionic and nonionic form of the drug: (12-4) If one assumes quantity q equal to 100, then concentration of B or BH + forms will numerically be equal to the percentage of corresponding form in the solu- tion and solving equation (12-3) with expression (12-4) one will get the expres- sion for BH + concentration expressed as a percent of ionized form (12-5) The inflection point of this curve corresponds to the point where pH = pK a , and it is a common way for the determination of the drug pK a values. Several different techniques are usually employed for pK a determination. They were described in detail by Comer [11]. BH ppH ppH + − () − () [] = ⋅ + 100 10 110 K K a a q = [] + [] + BBH ppH BH B K a =+ [] [] + log INITIAL PHYSICOCHEMICAL CHARACTERIZATION (DISCOVERY SUPPORT) 581 Figure 12-3. Dependence of the relative amount (in the form of a percent) of proto- nated form on the pH of aqueous media. In practice the most common technique to determine pK a value is by employing potentiometric titration based on the detection of the variations of either the conductivity or current at fixed applied potential at various pH values .The automated potentiometric titration system well known as a GLpK a or PCA200 from Sirius Analytical [12] is considered to be a good approach for pK a determination with water-soluble drugs at pH 2–8 for the new drug candidates when the amount of drug substance is limited. For poorly water- soluble compounds it is advised to use GlpK a with D-Pass or Sirius Profiler SGA as a pH/UV method for determination of compounds that have inher- ently lower concentration in the solution media. HPLC is another convenient method for measurement of the NCE pK a values. As was shown by Melander and Horvath [13], the retention of any ionizable analyte closely resembles the curve shown in Figure 12-3. Chro- matographic determination of the pK a could be accurately performed with very limited amount of sample. Fast HPLC method with optimum analyte retention is suitable for this purpose, but the influence of the organic mobile- phase modifier on the mobile phase pH and analyte pK a should be accounted for in order to provide the accurate calculation of the respective pK a value. Detailed discussion of the HPLC-based methods for the pK a determination is given in Chapter 4. In the case of sufficient drug supply the old-fashioned solubility method can be used for pK a determination based on the different equilibrium solubility at different pH values. This method is very precise, but time- and drug- consuming, and is described in detail in reference 6. Drug substance often contains several ionizable groups that may signifi- cantly complicate experimental measurement of the pK a . All different types of pK a determination methods are essentially based on the measurement of the titration curve. If the pK a values of several ionizable groups in the mole- cule are within 2 pH units from each other, experimental measurement become very tedious. Recent advancements in the molecular computational methods and developments of physicochemical databases for a large number of known compounds allow computer-based prediction of the pK a values on the basis of known physicochemical correlations and fast computer screening of known values for related or structurally similar compounds from the data- base. Detailed discussion of these programs is given in Chapter 10. 12.2.2 Partition and Distribution Coefficients One of the most important physicochemical parameters associated with oral absorption, central nervous system (CNS) penetration, and other pharmaco- kinetic parameters is lipophilicity of organic compounds, which determines distribution of a molecule between the aqueous and the lipid environments. The lipophilicity in the form of LogP was included in Lipinski’s rule of five as one of the major characteristics of drug-like organic molecules. It was stated that LogP should be not more than five for drug candidates to have a good 582 ROLE OF HPLC IN PREFORMULATION oral absorption property. In Table 12-1, some LogP values for various types of dosage forms are given. T he partition coefficient itself is a constant and is defined as the ratio of concentration of compound in aqueous phase to the concentration in an immiscible solvent, as the neutral molecule. In practical terms the neutral mol- ecule exists for bases >2 pH units above pK a and for acids >2 pH units below pK a . In practice, log P will vary according to the conditions under which it is measured and the choice of partitioning solvent. LogP is the logarithm of dis- tribution coefficient at a pH where analyte is in its neutral state. This is not a constant and will vary according to the protogenic nature of the molecule. The choice of partition solvent has been a subject of debate. Different type of solvents have been used for the determination of partitioning coefficient [14], but the majority of the data are generated using water–n-octanol parti- tioning. Octanol was chosen as a simple model of a phospholipid membrane. However, it has shown serious shortcomings in predicting blood–brain barrier or skin penetration. Other solvents such as chloroform, cyclohexane, and propylene glycol dipelargonate (PGDP) have been used for modeling biolog- ical membranes. Octanol is a hydrogen-bonding solvent, and thus it shows certain specificity in its ability to dissolve some components. For example, K 0 w for phenol in hexane is only 0.11 while in octanol it is equal to 29.5. There were several attempts to rationalize solvent effects using solubility parameters [15], dielec- tric constant [16], and others,but none appear to be consistent.n-Octanol gives the most consistent results with other physicochemical properties and drug absorption in gastrointestinal tract. The classical measurement of LogP is the shake flask method [17].A known amount of drug is dissolved in a flask containing both octanol phase and aqueous buffer at controlled pH to ensure the existence of only nonionic form (at least two units from the drug pK a ). The flask is shaken to equilibrate the sample between two phases. There must be no undissolved substance present in both phases. After the system reaches its equilibrium, which is time- and temperature-dependent, the concentration of drug is analyzed by HPLC in both phases. Partitioning coefficient is calculated as (12-6) K c c w w 0 0 = INITIAL PHYSICOCHEMICAL CHARACTERIZATION (DISCOVERY SUPPORT) 583 TABLE 12-1. Preferable Dosing Form for Different LogP Regions LogP Dosing Form Low <0 Injectable Medium 0–3 Oral High 3–4 Transdermal Very high 4–7 Toxic buildup in fatty acids This method allows for the accurate determination of K 0 w only within the − 1000 to +1000 region or approximately within six orders of magnitude span. T hese experiments could be complicated by solubility and equilibration kinet- ics and the properties of a substance. For example, if a studied compound has a property of nonionic surfactant, it will be mainly accumulated at the water–organic interface, and shaking of this two-phase system will create a stable emulsion difficult for analytical sampling. The ultracentrifugation at speed of 14,000rpm for 15–20min can be enough in most cases to separate two phases. Actual equilibration of the system is tested by several measure- ments of the equilibrium concentration at different time intervals. Because of the wide range of partitioning coefficient values, in most cases the decimal logarithm of K 0 w is used, and it is denoted as LogP: (12-7) The biggest challenge for the use of HPLC in the LogP measurement is the determination of the drug concentration in the octanol phase. If the octanol solution is being injected onto the reversed-phase column, it can modify the stationary phase, shift the analyte retention, and lead to an incorrect mea- surement due to the retention shift. To avoid this problem the dilution in the corresponding mobile phase is recommended. Also, when LogP is more than four, the concentration of drug in water phase is very small, causing a detec- tion problem with UV detection. This becomes even more troublesome if the compound of interest has a weak UV chromophore. The use of MS detection and proper ionization mode is recommended to increase the sensitivity. Direct HPLC experiment can be used for estimation of LogP, but this tech- nique is valid only for neutral molecules or for ionized molecules analyzed in their neutral state [18]. The following is a brief description of this method. Compounds with known LogP is injected onto C18 hydrophobic column, and the respective retention factors are used to create a calibration curve. The estimation of LogP for unknown compounds can be made on the basis of this calibration curve. This method is straightforward, but requires the previous knowledge of pK a values for ionizable compounds to avoid the possible ion- ization that will lead to incorrect determination of values of LogP. Recently, an automated isocratic liquid chromatography system, dedicated to the measurement of LogP,Profiler LDA, was introduced into the market by Sirius- Analytical, Ltd. There were numerous attempts to use the retention time of compound in correlation with its distribution properties in RP HPLC [19, 20]. The retention factor was used to calculate a distribution coefficient between stationary phase and mobile phase. In case of Sirius Profiler LDA automated system, a set of molecules with known LogP values was used to calibrate the system and convert the chromatographic retention time into octanol/water partition coefficients. The system could cover the LogP range from −1 to 5.5 by choosing between three different methods and different column lengths LogP = () log K w 0 584 ROLE OF HPLC IN PREFORMULATION ranging from 1 to 25cm, but was recently removed from the market.The well- known automated pH titrator from Sirius , GlpK a , can be used as well to deter- mine the octanol/water partition coefficient. The measurement is based on a two-phase acid/base titration in a mixture of water/octanol [21]. Partition coefficient discussed above represents oil/water equilibrium dis- tribution of only neutral forms of a substance. The distribution at different pH is described by LogD, which is the logarithm of the ratio of the concentrations of all forms of analyte in oil and water phases at particular pH. Logarithm of distribution coefficient at pH 7.4 is often used to estimate the lipophilicity of a drug at the pH of blood plasma. As follows from the definition, the distribution coefficient is dependent on the pH. It is usually assumed that in the oil-phase drug molecule could exist in only nonionic form; thus the distribution coefficient, D 0 w , for basic drug B could be written as (12-8) If LogP and pK a for a studied drug is known, then it is possible to express D 0 w as a function of pH of aqueous phase through these values using equations (12-3) and (12-6)–(12-8). Resulting expression is (12-9) Figure 12-4 represents the comparison of the pH dependencies of ionic form of a basic drug with LogD. Log D pH LogP Log ppH w K a 0 110 () () =−+ [] − D B w 0 = [] [] + [] oil water + water BBH INITIAL PHYSICOCHEMICAL CHARACTERIZATION (DISCOVERY SUPPORT) 585 LogD Figure 12-4. Normalized dependence of the protonated form of the base (solid) and its LogD dependence on the aqueous pH (dashed). At high pH, the neutral form of a drug (basic compound) has a distribu- tion coefficient equal to its partitioning coefficient. With the decrease of the pH of the aqueous phase, the degree of drug ionization increases, thus increas- ing its total concentration in the aqueous phase. As the pH decreases, the ionic equilibrium is shifted toward the protonated form of a drug, which con- tinually increases its concentration in the aqueous phase and decreases its content in oil phase. There is no plateau region in the LogD curve at low pH for basic compounds (Figure 12-4). On the other hand, for acidic compounds, there is a plateau region in the LogD curve at low pH (pHs below the pK a ); and then as the pH increases, the more ionic equilibrium is shifted toward the ionized form of the acid, which continually increases its concentration in the aqueous phase and decreases its content in the oil phase. This results in the absence of plateau in the LogD curve at high pH (pH > pK a ) for acidic compounds. These are only the theoretical dependencies; real behavior of actual mole- cule usually is significantly altered due to different types of intermolecular interactions. Molecular solvation, association, hydrogen bonding, and counte- rions all have a significant effect on drug ionization constant and partitioning and distribution coefficients. Detailed and comprehensive discussion of these effects could be found in the book by Avdeef [22]. 12.2.3 Solubility and Solubilization Aqueous solubility is one of the most important physicochemical properties of a new drug candidate because it affects both drug absorption and dosage form development. Only a drug in solution can be absorbed by the gastroin- testinal track. The rate of dissolution and the intestinal permeability of the drug molecules are dependent on the aqueous solubility—that is, the higher the solubility, the faster the rate of dissolution. An excellent monograph describing the theory of solubility and solubility behavior of organic com- pounds was written by Grant and Higuchi [23]. For additional information on solubility, the reader can be referred to references 24–27. Solubility is expressed as the concentration of a substance in a saturated solution at a defined temperature. The US Pharmacopeia (USP) gives the solubility definitions shown in Table 12-2. Solubility measurements are generally carried out in the early stages of drug development because it affects drug bioavailability evaluation; in many cases, solubility-limited absorption has been reported. Only a compound that is in solution is available to cross the gastrointestinal membrane. The solubil- ity measurements in aqueous buffered systems at different pHs are used to mimic gastrointestinal human or animal fluids. Solubility determination in DMSO is very important at the early stages of lead candidate selection because of the increasing use of 10mM DMSO solution as a stock solution for biological testing for very slightly soluble lead candidates [29]. In general, 586 ROLE OF HPLC IN PREFORMULATION [...]... identification of specific class for the drug candidate is critical for future development of dosage forms Different platforms are used for solubility measurements: UV; HPLC with UV detection; or HPLC with MS detection UV spectrophotometry is the simplest and fastest method, unfortunately with limited applicability In most cases the drug substance available for the study in the preformulation stage is not... SELECTION For ionic drugs the salt form can be considered as an alternative to increase the solubility Drug substance usually is more soluble in aqueous media in its ionic form Low solubility of the neutral form of the drug substance suggests the necessity to formulate it in the form of salt The reader is referred to reference 46 for more information about the properties, selection, and use of salt forms for. .. low solubility compared to original crystal form [51].The product was recalled from market and was reformulated It was a rare example of a dramatic effect of the existence of multiple crystal forms of a commercial pharmaceutical and showed the importance of polymorphic screening for all type of pharmaceutical dosage forms When the existence of polymorphism for new chemical entity is identified, the property... and fast HPLC to determine the assay concentration for each polymorph at the different temperatures are the same as for solubility determination However, for stability evaluation of the different polymorphs a stability-indicating HPLC method should be used 12.6 PREFORMULATION LATE STAGE (DEVELOPMENT SUPPORT) After a new chemical entity has been selected to move forward to development, the preformulation... source of information to formulation and analytical scientists regarding the properties of the recommended drug molecules The development preformulation support provides the additional testing of prototype formulation and excipient compatibility samples as well as guidance for salt form selection and polymorphs screening The major role of this unit is to bridge discovery and development stages HPLC coupled... mechanism To achieve this goal, the best approach is to perform forced degradation studies at the preformulation stage of drug development with most viable candidates, which may include the free base or acid and several corresponding salt forms The FDA and ICH guidance provides very little information about strategies and principles for conducting forced degradation studies, including problems with poorly... salt counterions are shown in Table 12-6 Salt form selection is mainly covered by solid-state charactezation methods, and HPLC is only used to determine the solubility and solid/solution stability of different salt forms The requirements for HPLC method development is the same as for solubility/stability determination described previously, and the same HPLC method may be applied 12.5 POLYMORPHISM Polymorphism... described model to perform drug–excipient compatibility testing prior to Phase I to eliminate potential future issues related to drug instability in final formulation Because the drug–excipient compatibility testing is conducted at an early drug development stage when a fully validated HPLC method is not available, the same GLP HPLC method as for forced degradation studies can be used for this test as well... thermodynamically unstable forms during preformulation stage of drug development is important Typical methods used for solid-state characterization of polymorphism are DSC, POLYMORPHISM 595 FT/IR, microscopy, and X-ray powder diffraction [49, 50] HPLC is used to evaluate chemical stability of different polymorphic forms as well as for solubility determination, and this parameter is very critical for drug development,... 5 A Avdeef, Absorption and Drug Development, Wiley Interscience, New York, 2003 6 J I Wells, Pharmaceutical Preformulation, Ellis Horwood Chichester, UK, 1988 7 J T Carstensen, Pharmaceutical Preformulation, Technomic Publishing Company, Lancaster, PA, 1998 8 M Gibson (ed.), Pharmaceutical Preformulation and Formulation, Interpharm/ CRC Press, Boca Raton, Florida, 2004, pp 585 9 H van de Waterbeemd, . specific class for the drug candidate is critical for future development of dosage forms. Different platforms are used for solubility measurements: UV; HPLC with UV. H + ↔+ + 580 ROLE OF HPLC IN PREFORMULATION (12-3) T his allows for the estimation of the prevailing drug form at a particular pH. Ionic form of any organic