1F
`
`orm Selection of Pharmaceutical
`Compounds
`
`Ann W. Newman and G. Patrick Stahly
`SSCI, Inc., West Lafayette, Indiana
`
`I.
`
`INTRODUCTION
`
`The drug development process involves a number of activities which are carried
`out simultaneously, as shown by the oversimplified depiction in Fig. 1. Once a
`molecule is discovered that has desirable biological activity, the process of creat-
`ing a pharmaceutical drug product from this molecule begins. As toxicology and
`efficacy studies are undertaken, methods for manufacture of the active molecule
`and for its delivery in therapeutic doses are sought. Critical to the latter effort is
`finding a form of the active molecule which exhibits appropriate physical proper-
`ties. The form ultimately selected, called the active pharmaceutical ingredient
`(API), or drug substance, must be stable and bioavailable enough to be formulated
`into a drug product, such as a tablet or suspension. This formulation must be
`effective at delivering the active molecule to the targeted biosystem.
`This chapter describes methodology useful in selection of the appropriate
`solid form of a drug substance for inclusion in a drug product. Form selection
`is commonly considered among the primary goals of a preformulation study.
`However, the investigative techniques discussed herein also have application in
`early drug substance and drug product development activities (shown by the cir-
`cled area in Fig. 1).
`Solid form selection involves the preparation and property evaluation of
`many derivatives of an active molecule. Drug substance properties of importance
`in the drug development process may be categorized as shown in Table 1. These
`properties depend on the nature of the drug substance and the final formulation.
`Many bioactive organic molecules contain ionizable groups such as carboxylic
`
`Merck Exhibit 2218, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`Merck Exhibit 2191, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Fig. 1 The drug development process.
`
`acid or amino groups. Reaction of these compounds with acids or bases produce
`salts, which have much different physical properties than the neutral parents. A
`single molecular entity, be it a salt or a neutral molecule, often exists in multiple
`solid forms, each of which exhibits unique physical properties. The properties
`of many such forms need to be evaluated relative to the intended formulation.
`A lyophilized product that will be dissolved and injected needs to be chemically
`stable in the dry state and adequately soluble in the carrier. On the other hand,
`the drug substance in a tablet formulation needs to be processable, chemically
`stable, and physically stable in the dry state, as well as having adequate solubility
`for delivery.
`Form selection activities should be started as early in the development pro-
`cess as material availability allows. Salt selection, including preparation and eval-
`
`Table 1 Some Important Properties of Drug Substances
`
`Bioavailability
`
`Chemical and physical stability
`
`Processibility
`
`Dissolution rate
`Solubility
`Toxicity
`
`Excipient compatibility
`Hygroscopicity
`Oxidative stability
`Photostability
`Thermodynamic stability
`Crystal form
`
`Color
`Compactibility
`Density
`Ease of drying
`Filterability
`Flowability
`Hardness
`Melting point
`Particle size
`
`Merck Exhibit 2218, Page 2
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`Merck Exhibit 2191, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`uation of samples, and polymorph screening can be carried out with as little as
`half a gram of active compound. Results of form selection include information
`that can be used in planning the final step of the manufacturing process (often
`crystallization) as well as information that is critical to formulation development.
`The nature and extent of work to be performed during development can
`be modeled after the draft International Committee on Harmonization (ICH) Q6A
`document on specifications, which can be found on the Food and Drug Adminis-
`tration (FDA) website (www.fda.cder.gov). This document outlines the specifi-
`cations needed for a New Drug Application and contains several decision trees
`to guide the selection of specifications. The Q6A decision tree 4 (Fig. 2) describes
`
`Fig. 2 Flow chart 4 from the ICH Q6A document (www.fda.cder.gov).
`
`Merck Exhibit 2218, Page 3
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`Merck Exhibit 2191, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Fig. 2 Continued
`
`methods for the study of solids for a polymorph screen as well as characterization
`of the drug substance in the drug product. Other decision trees have also been
`reported in the literature (1).
`In this chapter we describe the form selection process. A short review of
`the analytical techniques commonly employed is followed by sections covering
`salt and solid form selection. Form selection should be approached in a planned,
`rational manner, but it is important to realize that not all compounds will allow
`adherence to a single experimental plan. The exercise is a scientific one, and it
`will yield the best results only if carried out with judgment and flexibility.
`
`II. ANALYTICAL TECHNIQUES
`
`A number of analytical techniques are commonly used in form selection studies.
`Various publications (2–4) and books (5,6) describe physical characterization
`
`Merck Exhibit 2218, Page 4
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`Merck Exhibit 2191, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`of solid-state pharmaceuticals. A brief description of common methods will be
`presented in this section.
`
`A. X-Ray Diffraction
`
`Crystalline organic solids are made up of molecules which are packed or ordered
`in a specific arrangement. These molecules are held together by relatively weak
`forces, such as hydrogen bonding and van der Waals interactions. The arrange-
`ment of the molecules is defined by a unit cell, which is the smallest repeating
`unit of a crystal. The unit cell can be divided into planes, as shown in Fig. 3.
`X-ray diffraction techniques used for characterizing pharmaceutical solids
`
`Fig. 3 A packing diagram of unit cells divided into planes.
`
`Merck Exhibit 2218, Page 5
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`Merck Exhibit 2191, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`(1)
`
`(n ⫽ 1, 2, 3, . . .)
`
`include the analysis of single crystals and powders. The electrons surrounding
`the atoms diffract X-rays in a manner described by the Bragg equation:
`nλ ⫽ 2d sin θ
`where
`λ ⫽ X-ray wavelength
`d ⫽ spacing between the diffracting planes
`θ ⫽ diffraction angle
`A schematic of the diffraction phenomenon is given in Fig. 4. X-rays will be
`diffracted at an angle defined as θ. Knowing the diffraction angle and the X-ray
`wavelength, the spacing between the planes can be calculated. Conditions of the
`Bragg equation must be satisfied to achieve constructive interference of the dif-
`fracted X-rays and produce a beam that can be measured by the detector. If the
`conditions of the Bragg equation are not satisfied, diffracted waves interfere de-
`structively, with a net diffracted intensity of zero.
`For single-crystal diffraction, a good-quality single crystal of the sample
`of interest is required. From the angles and intensities of diffracted radiation, the
`structure of the crystal can be elucidated and the positions of the molecules in
`the unit cell can be determined. The result is often displayed graphically as the
`asymmetric unit, which is the smallest part of a crystal structure from which
`the complete structure can be obtained using space-group symmetry operations.
`
`Fig. 4 A schematic representation of X-ray diffraction.
`
`Merck Exhibit 2218, Page 6
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`Merck Exhibit 2191, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`The unit cell parameters, the lengths (a, b, c) as well as the angles (α, β, γ) of
`the unit cell are also determined from the crystal structure. There are seven classes
`of unit cells: triclinic, monoclinic, orthorhomic, tetragonal, hexagonal, rhombo-
`hedral, and cubic. For pharmaceutics, only triclinic (a ≠ b ≠ c, α ≠ β ≠ γ ≠90°),
`monoclinic (a ≠ b ≠ c, α ≠ γ ≠90°, β ⫽ 90°), and orthorhombic (a ≠ b ≠ c,
`α ⫽ β ⫽ γ ⫽ 90°) unit cells are commonly observed.
`The unit cells can be ‘‘packed’’ into a three-dimensional display of the
`crystal lattice. The orientation of the molecules is responsible for various proper-
`ties of the crystalline substance. For example, hydrogen bonding networks may
`provide high stability, and spaces in the structure may allow easy access of small
`molecules to provide hydrated or solvated forms.
`Crystal structures provide important and useful information about solid-
`state pharmaceutical materials. Unfortunately, it is not always possible to grow
`suitable single crystals of a drug substance. In these cases, X-ray diffraction of
`powder samples can be used for comparison of samples.
`X-ray powder diffraction (XRPD) is the analysis of a powder sample. The
`typical output is a plot of intensity versus the diffraction angle (2θ). Such a plot
`can be considered a fingerprint of the crystal structure, and is useful for determi-
`nation of crystallographic sameness of samples by pattern comparison. A crystal-
`line material will exhibit peaks indicative of reflections from specific atomic
`planes. The patterns are representative of the structure, but do not give positional
`information about the atoms in the molecule. One peak will be exhibited for all
`repeating planes with the same spacing. An amorphous sample, on the other hand,
`will exhibit a broad hump in the pattern called an amorphous halo, as shown in
`Fig. 5.
`
`Fig. 5 The XRPD pattern exhibited by an amorphous material.
`
`Merck Exhibit 2218, Page 7
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`Merck Exhibit 2191, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`XRPD is dependent on a random orientation of the particles during analysis
`to obtain a representative powder pattern. The sample, as well as sample prepara-
`tion, can greatly effect the resulting pattern. Large particles or certain particle
`morphologies, such as needles or plates, can result in preferred orientation. Pre-
`ferred orientation is the tendency of crystals to pack against each other with some
`degree of order and it can affect relative peak intensities, but not peak positions,
`in XRPD patterns. If a powder is packed into an XRPD sample holder and the
`surface is smoothed with a microscope slide or similar device, crystals at the
`surface can become aligned so that a nonstatistical arrangement of crystal faces
`is presented to the X-ray beam. The result is that some reflections are artificially
`intensified and others are artificially weakened. One way to determine if preferred
`orientation is causing relative peak intensity changes is to grind and reanalyze
`samples. Grinding reduces particle size and disrupts the crystal habit, both of
`which tend to minimize preferred orientation effects. However, grinding can
`cause crystal form changes, so care must be taken to interpret the patterns with
`this in mind. The effects of preferred orientation can be profound, as illustrated
`by the XRPD patterns shown in Fig. 6.
`
`Fig. 6 XRPD patterns of the same sample before (top) and after (bottom) grinding. The
`polymorphic form of the sample was not changed by grinding.
`
`Merck Exhibit 2218, Page 8
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`Merck Exhibit 2191, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Qualitative analysis of powder patterns can be used to determine if multiple
`samples are the same crystal form or if multiple crystal forms have been pro-
`duced. Mixtures of samples can also be evaluated. When mixtures are obtained,
`XRPD can also be used in a quantitative mode to calculate the amount of each
`phase present.
`
`B. Thermal Methods
`
`Thermal methods of analysis discussed in this section are differential scanning
`calorimetry (DSC), thermogravimetry (TG), and hot-stage microscopy (HSM).
`All three methods provide information upon heating the sample. Heating can be
`static or dynamic in nature, depending on the information required.
`Differential scanning calorimetry monitors the energy required to maintain
`the sample and a reference at the same temperature as they are heated. A plot
`of heat flow (W/g or J/g) versus temperature is obtained. A thermal transition
`which absorbs heat (melting, volatilization) is called endothermic. If heat is re-
`leased during a thermal transition (crystallization, degradation), it is called exo-
`thermic. The area under a DSC peak is directly proportional to the heat absorbed
`or released and integration of the peak results in the heat of transition.
`Samples are loaded into pans for DSC analysis. Pan configuration (open,
`crimped, hermetically sealed, hermetically sealed with a pinhole, etc.) and scan
`rate can result in variations in position and intensity of the thermal events. These
`variations can be used to gain further information about the sample as well as
`other crystal forms.
`The observance of thermal transitions by DSC is insufficient to fully char-
`acterize the behavior of a substance on heating. It is not known if an endothermic
`transition observed in the DSC is a volatilization or a melt without corroborating
`information, such as TG or HSM data. It is important to understand the origin
`of the DSC transitions to fully characterize the system and understand the rela-
`tionship between various solid forms.
`Thermogravimetry measures the weight change of a sample as a function
`of temperature. A total volatile content of the sample is obtained, but no informa-
`tion on the identity of the evolved gas is provided. The evolved gas must be
`identified by other methods, such as gas chromatography, Karl Fisher titration
`(specifically to measure water), TG–mass spectroscopy, or TG–infrared spectros-
`copy. The temperature of the volatilization and the presence of steps in the TG
`curve can provide information on how tightly water or solvent is held in the
`lattice. If the temperature of the TG volatilization is similar to an endothermic
`peak in the DSC, the DSC peak is likely due or partially due to volatilization.
`It is usually necessary to utilize multiple techniques to determine if more than
`one thermal event is responsible for a given DSC peak.
`
`Merck Exhibit 2218, Page 9
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`Merck Exhibit 2191, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Hot-stage microscopy is a technique that supplements DSC and TG. Events
`observed by DSC and/or TG can be readily characterized by HSM. Melting, gas
`evolution, and solid–solid transformations can be visualized, providing the most
`straightforward means of identifying thermal events. Many polymorphic systems
`have been investigated using only these thermal methods, as illustrated by the
`publications of Kuhnert-Brandsta¨tter (7). Details of the methodologies used in
`hot-stage microscopy have also been reviewed (8).
`Thermal analysis can be used to determine the melting points, recrystal-
`lizations, solid-state transformations, decompositions, and volatile contents of
`pharmaceutical materials. DSC can also be used to analyze mixtures quantita-
`tively.
`
`C. Vibrational Spectroscopy
`
`Common methods used to characterize drugs and excipients are infrared (IR) and
`Raman spectroscopy. These techniques are sensitive to the structure, conforma-
`tion, and environment of organic compounds. Because of this sensitivity, they
`are useful characterization tools for pharmaceutical crystal forms. Qualitative as
`well as quantitative analysis can be performed with both techniques.
`Infrared spectroscopy is based on the conversion of IR radiation into molec-
`ular vibrations. For a vibration to be IR-active, it must involve a changing molec-
`ular dipole (asymmetric mode). For example, vibration of a dipolar carbonyl
`group is detectable by IR spectroscopy. Whereas IR has been traditionally used
`as an aid in structure elucidation, vibrational changes also serve as probes of
`intermolecular interactions in solid materials.
`Sampling techniques for IR include pellets, mulls, and diffuse reflectance.
`Diffuse reflectance is the best choice for crystal form determination, due to the
`minimal sample manipulation required. Mulls can also be used for form identifi-
`cation, but peaks due to the suspension medium may interfere with the peaks of
`interest.
`Raman spectroscopy is based on the inelastic scattering of laser radia-
`tion with loss of vibrational energy by a sample. A vibrational mode is Raman-
`active when there is a change in the polarizability during the vibration. Symmetric
`modes tend to be Raman-active. For example, vibrations about bonds be-
`tween the same atom, such as in alkynes, can be observed by Raman spectros-
`copy.
`
`Small amounts of samples can be analyzed by Raman spectroscopy and a
`variety of sample holders are available, ranging from stainless steel holders to
`glass NMR tubes. The samples are analyzed neat, eliminating the need for sample
`preparation procedures that may induce solid form changes. Since a laser is used,
`only a small portion of the sample is in the beam during analysis.
`
`Merck Exhibit 2218, Page 10
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`Merck Exhibit 2191, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`D. Nuclear Magnetic Resonance (NMR) Spectroscopy
`
`NMR spectroscopy probes atomic environments based on the different resonance
`frequencies exhibited by nuclei in a strong magnetic field. Many different nuclei
`are observable by the NMR technique, but those of hydrogen and carbon atoms
`are most frequently studied. NMR spectroscopy of solutions is commonly used
`for structure elucidation. However, solid-state NMR measurements are extremely
`useful for characterizing the crystal forms of pharmaceutical solids.
`Nuclei that are typically analyzed with this technique include those of 13C,
`31P, 15N, 25Mg, and 23Na. Different crystal structures of a compound can result in
`perturbation of the chemical environment of each nucleus, resulting in a unique
`spectrum for each form. Once resonances have been assigned to specific atoms
`of the molecule, information on the nature of the polymorphic variations can be
`obtained. This can be useful early in drug development, when the single-crystal
`structure may not be available. Long data acquisition times are common with
`solid-state NMR, so it is often not considered for routine analysis of samples.
`However, it is usually a very sensitive technique, and sample preparation is mini-
`mal. NMR spectroscopy can be used either qualitatively or quantitatively, and
`can provide structural data, such as the identity of solvents bound in a crystal.
`
`E. Moisture Sorption/Desorption and Hygroscopicity
`
`Hygroscopicity and the formation of hydrated crystal forms can be investigated
`by means of moisture sorption/desorption methods. Sample analysis may be car-
`ried out using automated equipment or by periodic weighing of samples kept
`over saturated salt solutions providing various relative humidities (RHs). In either
`case, water taken in or released by a sample is detected as a change in sample
`weight. If a material readily loses water of hydration at low relative humidity,
`the stability of the hydrate may need to be investigated further. If a material
`readily gains moisture at ambient or high relative humidity, hygroscopicity stud-
`ies will be needed to determine if a change in crystal form is associated with the
`water uptake. This is done by characterizing material equilibrated under various
`relative humidity conditions using techniques suitable for detection of crystal
`form, such as XRPD, TG, DSC, and IR spectroscopy. Changes in water content
`and crystal form may lead to definition of specific handling conditions under
`which a change in form will not occur.
`
`F. Summary
`
`Only a brief description of selected techniques for solid-state characterization
`has been given above. Many other techniques are available. It is imperative that
`
`Merck Exhibit 2218, Page 11
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`Merck Exhibit 2191, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`a multidisciplinary approach be applied to the characterization solids; no single
`analytical method can provide all the information necessary to understand the
`nature and properties of solid pharmaceutical compounds.
`
`III. SALT SELECTION
`
`A. Factors Guiding Salt Selection
`
`Salt selection is a critical part of the drug development process because selection
`of an appropriate salt can significantly reduce time to market. Changing salts in
`the middle of a development program may require repeating most of the biologi-
`cal, toxicological, formulation, and stability studies performed initially. However,
`continuing the development of a nonoptimal salt may lead to increased develop-
`mental and production costs, even product failure. Selection of the correct salt
`early in the development process will avoid these problems and facilitate down-
`stream development activities. In addition, salts that exhibit advantageous proper-
`ties are usually patentable as new chemical entities.
`Salts are used to alter the physical, chemical, biological, and economic
`properties of a drug substance. The change in crystal structure accomplished by
`forming a salt can lead to greatly improved properties. The advantages of using
`salt forms in pharmaceutical formulations have been extensively reviewed (9).
`A variety of factors can guide the salt selection process and a partial list of consid-
`erations is given in Table 2.
`A change in the solubility of a drug substance is often a major reason for
`choosing a salt. In many cases, substances containing free acid or base groups
`have poor aqueous solubility which saltification of these groups can improve,
`leading ultimately to greater bioavailability. Increasing the solubility of a weak
`acid–base drug substance by forming a variety of salts has been reported for
`
`Table 2 Factors Guiding the Salt Selection
`Process
`
`Bioavailability
`Chemical stability
`Crystallinity
`Dissolution rate
`Cost
`Handling properties
`Hygroscopicity
`Melting point
`Intended formulation
`
`pH of salt solutions
`Physical stability
`Processing properties
`Purity
`Solubility
`Taste
`Toxicity
`Wettability
`Yield
`
`Merck Exhibit 2218, Page 12
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`Merck Exhibit 2191, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Fig. 7 The structure of RS-82856.
`
`RS-82856 (Fig. 7) (10). Five salts (chloride, hydrogen sulfate, phosphate, sodium,
`and potassium) exhibited significantly higher solubility and dissolution rates than
`the parent drug. Based on a variety of physical parameters (solubility, dissolution
`rate, melting point, hygroscopicity, and chemical stability), the hydrogen sulfate
`form was recommended for development. A bioavailability study in dogs compar-
`ing the parent drug and the hydrogen sulfate salt resulted in the salt being ab-
`sorbed approximately two to three times more efficiently than the parent drug.
`The solubility and dissolution data were good indicators of the bioavailability of
`this material.
`For some drugs, preparation of stable salts may not be feasible, or free acid
`or free base forms may be preferred. A reported example compares the free base
`and hydrochloride salt of the poorly water-soluble drug, α-pentyl-3-(2-quino-
`linylmethoxy)benzenemethanol, known as REV 5901 (Fig. 8) (11). For this drug
`substance, lower solubility of the chloride salt, along with equivalent dissolution
`rates, resulted in the free base being chosen for development.
`It should be noted that a salt usually exhibits a higher melting point than
`the free acid or base, which can result in greater stability and easier processing.
`However, there is often a relationship between melting point and aqueous solubil-
`ity. Gould, in his study of the salts of basic drugs, concluded that ‘‘ideal solubility
`of a drug in all solvents decreases by an order of magnitude with an increase of
`100°C in its melting point’’ (12). An example of this phenomenon is the antima-
`larial drug α-(2-piperidyl)-3,6-bis(trifluoromethyl)-9-phenanthrenemethanol hy-
`
`Fig. 8 The structure of REV 5901.
`
`Merck Exhibit 2218, Page 13
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`Merck Exhibit 2191, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Fig. 9 The
`methanol.
`
`structure of α-(2-piperidyl)-3,6-bis(trifluoromethyl)-9-phenanthrene-
`
`drochloride (Fig. 9). The melting point and solubility data are shown in Table 3
`(13). Overall, a substantial decrease in solubility was observed with the increase
`in melting point of the salts. It should also be noted that the solubility of salts
`can be affected not only by changing the lattice energy (melting point), but also
`by enhancing water–drug interactions. The study of chlorhexidine (Fig. 10)
`showed that the solubility of this drug was significantly enhanced by increasing
`the number of hydroxyl groups on the conjugate acid (14).
`The melting point of a drug substance salt can be greatly influenced by the
`counterion. For UK47880 (Fig. 11), a relationship was observed between the
`melting points of the salts and the corresponding conjugate acid (12). Salts pre-
`
`Table 3 Melting-Point and Solubility Data for
`α-(2-piperidyl)-3,6-bis(trifluoromethyl)-9-phenanthrenemethanol
`hydrochloride (13)
`
`Salt form
`
`Free base
`DL-Lactate
`L-Lactate
`2-Hydroxyethane sulfonate
`Sulfate
`Mesylate
`Hydrochloride
`
`Melting point
`of salt (°C)
`
`215
`172
`192
`251
`270
`290
`331
`
`Aqueous
`solubility
`(mg/mL)
`
`7.5
`1850
`925
`620
`20
`300
`13
`
`Merck Exhibit 2218, Page 14
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`Merck Exhibit 2191, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Fig. 10 The structure of chlorhexidine.
`
`Fig. 11 The structure of UK47880.
`
`pared from high-melting aromatic acids exhibited higher melting points, whereas
`salts prepared from low-melting flexible aliphatic acids yielded oils. In the case
`of epinephrine (Fig. 12), the effect of hydrogen bonding on the melting points
`of the salts was apparent (12). Small acids prone to form hydrogen bonds (ma-
`lonic and maleic) resulted in higher-melting salts. The bitartrate and fumarate
`salts were found to be lower-melting due to their size and possibly unfavorable
`symmetry, respectively.
`A salt can also provide improved chemical stability compared to the parent
`drug substance. An example of this was reported for xilobam, whose structure
`is shown in Fig. 13 (15). In order to protect xilobam from the effects of high
`temperature and humidities without decreasing the dissolution rate, three arylsul-
`fonic acid salts (tosylate, 1-napsylate, and 2-napsylate), as well as the saccharate
`salt, were prepared. The 1-napsylate was found to be the most chemically stable
`form at 70°C and 74% RH after 7 days. Dissolution data from compressed tablets
`
`Fig. 12 The structure of epinephrine.
`
`Merck Exhibit 2218, Page 15
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`Merck Exhibit 2191, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Fig. 13 The structure of xilobam.
`
`showed that the 1-napsylate salt released xilobam at a faster rate than the free
`base. This work demonstrated that a strong acid with an aryl group protected the
`easily hydrolyzed base from the effects of high temperature and humidity.
`A choice of salts can also expand the formulation options for a material.
`The antimalarial agent α-(2-piperidyl)-3,6-bis(trifluoromethyl)-9-phenanthrene-
`methanol hydrochloride (Fig. 9) exhibited poor solubility, was delivered as an
`oral formulation, and required a single dosing of 750 mg (13). Seven salts and
`the free base were evaluated. The lactate salt was found to be 200 times as soluble
`as the hydrochloride salt (Table 3). This enhanced solubility would make it possi-
`ble to reduce the oral dose to achieve the same therapeutic response as well as
`develop a parenteral formulation for the treatment of malaria. However, the case
`of lidocaine hydrochloride (Fig. 14) demonstrates that a compound limited to
`parenteral and topical formulations can be expanded to oral administration by
`changing to a salt form with acceptable physical properties (16). The hydrochlo-
`ride salt was hygroscopic, difficult to prepare, and hard to handle. Six salts were
`evaluated for salt formation, solubility, and hygroscopicity. Other salts, such
`as phosphate, exhibited properties acceptable for dry pharmaceutical dosage
`forms.
`Many other examples can be found in the literature that demonstrate the
`applicability of examining a number of salts to obtain the necessary properties
`needed for development and marketing of the drug substance. Excellent reviews
`on salts (9,12) discuss many of the issues involved in targeting salt forms of drug
`substances.
`
`Fig. 14 The structure of lidocaine.
`
`Merck Exhibit 2218, Page 16
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`Merck Exhibit 2191, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`B. Counterions
`
`Salt formation involves proton transfer from an acid to a base. In theory, any
`compound that exhibits acidic or basic characteristics can form salts. The major
`consideration is the relative acidity and/or basicity of the chemical species in-
`volved. To form a salt, the pKa of the acidic partner must be less than the pKa
`of the conjugate acid of the basic partner. These pKa values need to be about
`two units apart for total proton transfer to occur, otherwise an equilibrium mixture
`of all components (acid, base, and salt) is likely to result. Even so, equilibrium
`mixtures of this type can often be used to prepare salts if a driving force is present,
`such as the crystallization of the salt from solution.
`Another consideration is the toxicity of the counterion. A large number of
`anions and cations are available for pharmaceutical compounds, and tabulations
`of those approved by the FDA have appeared in the literature (9,12). An expanded
`but not comprehensive list of acceptable ions is presented in Table 4. In general,
`ions related to normal metabolic chemicals or present in food or drink are usually
`regarded as suitable candidates for preparing salts.
`Target salts are chosen by considering a number of factors. The structure
`and pKa of the drug substance are important values to determine initially. Avail-
`able literature on structurally related compounds can result in excellent leads for
`target salts. The chemical stability of the drug substance, especially as related to
`pH stability, will also play a role. The ease of large-scale preparation of the salt,
`as well as the cost of the counterion and processing, will need to be considered
`to determine if the salt is a feasible choice. The type of drug product and antici-
`pated loading of the drug substance in the drug product can also influence the
`choice of salts. For high drug loadings, a large, bulky counterion, which adds
`substantial mass to the loading, may not be the best choice. Anions that irritate
`the gastrointestinal tract should be avoided for certain drugs, such as anti-
`inflammatories. The relative acid/base strength of the resulting salt and the ten-
`dency to disproportionate should be considered when using basic excipients in
`a formulation.
`A common salt choice for basic drug substances is the hydrochloride, be-
`cause of its availability. However, a number of issues also need to be considered
`when using this salt. Reports have shown that hydrochloride salts do not always
`increase the solubility of poorly soluble basic drugs (1,13,17,18), due to the com-
`mon-ion effect. The presence of chloride ions in the gastric fluid can result in a
`lower solubility for the hydrochloride salt.
`The hydrochloride salt is often a stronger acid than is needed for many
`drug substances, which can result in low pH values for the aqueous solutions.
`This can lead to limitations in parenteral formulations or processing. The highly
`polar nature of hydrochloride salts can also l