`
`www.elsevier.com/locate/addr
`
`Salt formation to improve drug solubility ☆
`Abu T.M. Serajuddin ⁎
`
`Science, Technology and Outsourcing Section, Novartis Pharmaceuticals Corporation, One Health Plaza, East Hanover, NJ 07936, USA
`
`Received 23 April 2007; accepted 10 May 2007
`Available online 29 May 2007
`
`Abstract
`
`Salt formation is the most common and effective method of increasing solubility and dissolution rates of acidic and basic drugs. In this article,
`physicochemical principles of salt solubility are presented, with special reference to the influence of pH–solubility profiles of acidic and basic
`drugs on salt formation and dissolution. Non-ideality of salt solubility due to self-association in solution is also discussed. Whether certain acidic
`or basic drugs would form salts and, if salts are formed, how easily they would dissociate back into their free acid or base forms depend on
`interrelationships of several factors, such as S0 (intrinsic solubility), pH, pKa, Ksp (solubility product) and pHmax (pH of maximum solubility). The
`interrelationships of these factors are elaborated and their influence on salt screening and the selection of optimal salt forms for development are
`discussed. Factors influencing salt dissolution under various pH conditions, and especially in reactive media and in presence of excess common
`ions, are discussed, with practical reference to the development of solid dosage forms.
`© 2007 Elsevier B.V. All rights reserved.
`
`Keywords: Salt; solubility; pH–solubility profile; Common-ion effect; Self-association; Dissolution rate; Salt selection; Counterion; Microenvironmental pH
`
`Contents
`
`1.
`2.
`
`3.
`
`4.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`Introduction .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`Principles of salt formation and salt solubility .
`pH–solubility interrelationship of free base and its salt .
`.
`.
`2.1.
`pH–solubility interrelationship of free acid and its salt .
`.
`.
`2.2.
`.
`.
`2.3.
`Effect of counterion on salt solubility .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`Effects of solubility, pKa and Ksp on pHmax .
`.
`.
`.
`.
`.
`.
`2.4.
`2.5. Deviation of pH–solubility interrelationship from ideality .
`Structure–solubility relationships .
`2.6.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`2.7.
`Effect of organic solvent on salt solubility .
`.
`.
`.
`.
`.
`.
`.
`.
`Principles of salt dissolution .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`3.1. General solubility–dissolution rate relationships .
`.
`.
`.
`.
`.
`3.2. Dissolution in reactive media .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`3.3.
`Common-ion effect on dissolution of salts .
`.
`.
`.
`.
`.
`.
`.
`.
`Solubility considerations in salt screening .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`4.1.
`Identification of chemical form .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`4.2. Determination of salt solubility .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`4.3.
`Recent trends in salt forms .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`604
`604
`604
`605
`606
`606
`607
`608
`609
`609
`609
`610
`611
`611
`612
`613
`613
`
`☆ This review is part of the Advanced Drug Delivery Reviews theme issue on “Drug solubility: How to measure it, how to improve it”.
`⁎ Tel.: +1 862 778 3995; fax: +1 973 781 7329.
`E-mail address: abu.serajuddin@novartis.com.
`
`0169-409X/$ - see front matter © 2007 Elsevier B.V. All rights reserved.
`doi:10.1016/j.addr.2007.05.010
`
`Merck Exhibit 2193, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`604
`
`A.T.M. Serajuddin / Advanced Drug Delivery Reviews 59 (2007) 603–616
`
`5.
`
`Practical considerations of salt solubility in dosage form design .
`5.1.
`Liquid formulations .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`5.2. Microenvironmental pH of salts in solid dosage forms .
`.
`6. Conclusions and outlook .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`References .
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`.
`.
`.
`.
`.
`
`. 614
`. 614
`. 614
`. 614
`. 615
`
`1. Introduction
`
`Salts of acidic and basic drugs have, in general, higher solu-
`bilities than their corresponding acid or base forms. Salt formation
`to increase aqueous solubility is the most preferred approach for
`the development of liquid formulations for parenteral adminis-
`tration [1]. For solid dosage forms, Nelson [2,3] demonstrated as
`early as in 1950s that dissolution rates of salt forms of several
`weakly acidic compounds under gastrointestinal (GI) pH con-
`ditions were much higher than those of their respective free acid
`forms. He attributed the higher dissolution rate of a salt to its
`higher solubility (relative to the free acid form) in the aqueous
`diffusion layer surrounding the solid. Pronounced differences
`were observed in rates and extents of absorption of novobiocin [4]
`and tolbutamide [5] as compared to their respective sodium salts.
`Monkhouse and coworkers [6,7] reviewed physicochemical and
`biopharmaceutical advantages of salts over their free acid or base
`forms. The interest in salt formation has grown greatly over the
`past half a century and, in recent years, it has become the most
`commonly applied technique of increasing solubility and dis-
`solution rate in drug product development.
`The primary reason for the increased interest in salt formation
`is that with the progress in medicinal chemistry and, especially
`due to the recent introduction of combinatorial chemistry and
`high-throughput screening in identifying new chemical entities
`(NCE) [8,9], the solubility of new drug molecules has decreased
`sharply [10]. While a value of less than 20 μg/mL for the solu-
`bility of a NCE was practically unheard of until the 1980s, the
`situation has changed so much that in the present day drug
`candidates with intrinsic solubilities (solubility of neutral or
`unionized form) of less than 1 μg/mL are very common [11].
`Lipinski [12] reported that 31.2% of a group of 2246 compounds
`synthesized in academic laboratories between 1987 and 1994 had
`solubility equal to or less than 20 μg/mL. According to the recent
`experience of the present author, approximately one-third of new
`compounds synthesized in medicinal chemistry laboratories have
`an aqueous solubility less than 10 μg/mL, another one-third have
`a solubility from 10 to 100 μg/mL, and the solubility of the
`remaining third is N100 μg/mL. With such a predominance of
`poorly water-soluble compounds, careful attention must be paid
`to identification and selection of optimal salt forms for de-
`velopment. In certain cases, salt formation may not be feasible due
`to physical and chemical properties of NCEs. In other cases, even
`though salts can be synthesized, they may not serve the purpose of
`enhancing dissolution rate and bioavailability. It is important that
`the reasons behind such situations are understood.
`Despite major advantages of the use of salts, only limited
`attention has been paid historically to the selection of optimal
`salt forms for pharmaceutical product development [6,13]. At
`
`the beginning of drug development programs, salts were often
`selected based on ease of synthesis, ease of crystallization, cost
`of raw material, etc., and no systematic studies to evaluate their
`physicochemical properties, such as physical and chemical
`stability, processability into dosage forms, solubility and dis-
`solution rate at different pH conditions, etc., were conducted. If
`a salt was later found to be suboptimal for the desired formu-
`lation or if problems developed, it was often difficult to change
`the salt form without delaying the drug development program,
`since it required repeating most of the biological, toxicological,
`formulation and stability tests that had already been performed
`[14]. For most practical purposes, identification and selection of
`salt forms of NCEs still remain a trial and error process.
`One major objective of the present article is to review the basic
`principles of salt formation and how salts influence solubility and
`dissolution rate in a comprehensive manner, such that they can be
`easily applied to the development of drug substances as well as
`dosage forms. Efforts will be made to indicate the application of
`such principles in screening various salt candidates for a NCE,
`identification of optimal salt form, and ultimately formulation of
`dosage forms using the selected salt. Wherever possible, advan-
`tages and disadvantages of salt forms relative to their respective
`free acid or base forms will be presented.
`One particular issue with the use of salts in drug development
`is that, while salts are usually prepared from organic solvents,
`they are destined to encounter aqueous environment (water,
`humidity) during dosage form development and, in case of an
`orally administered tablet or capsule, at the time of dissolution in
`GI fluid. Therefore, a perfectly good salt isolated from an or-
`ganic solvent may not behave well in an aqueous environment
`due to low solubility, conversion to free acid or base forms, poor
`stability, etc., thus limiting its use in dosage forms.
`
`2. Principles of salt formation and salt solubility
`
`The aqueous solubility of an acidic or basic drug as a function
`of pH dictates whether the compound will form suitable salts or
`not and, if salts are formed, what some of their physicochemical
`properties might be [15]. pH–solubility interrelationships also
`dictate what counterions would be necessary to form salts, how
`easily the salts may dissociate into their free acid or base forms,
`what their dissolution behavior would be under different GI pH
`conditions, and whether solubility and dissolution rate of salts
`would be influenced by common ions [15,16].
`
`2.1. pH–solubility interrelationship of free base and its salt
`
`Kramer and Flynn [17] demonstrated that the pH–solubility
`profile of a basic drug may be expressed by two independent
`
`Merck Exhibit 2193, Page 2
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`Fig. 2. pH–solubility profiles of haloperidol determined by using methane-
`sulfonic (mesylic) (□), hydrochloric (○) and phosphoric (▵) acids (reproduced
`from Ref. [24] with permission).
`
`only one point, can both the free base and salt coexist as solids.
`If the pH of a saturated solution with excess solid free base is
`lowered from above the pHmax to below the pHmax, the solid
`phase will convert to the salt, and it is important to note here
`that the pH will not drop below pHmax until enough acid is
`added to convert the entire excess free solid base into salt. The
`reverse is true for the conversion of a salt to the free base; no
`free base will precipitate out until the pH is raised above the
`pHmax.
`There are numerous reports in the literature confirming
`interrelationships of solubilities of bases and their salt forms
`as per the schematics in Fig. 1 [17,18–23]. An example of
`typical pH–solubility profiles is given in Fig. 2, where solu-
`bilities of haloperidol and its methanesulfonate (mesylate),
`hydrochloride and phosphate salts as a function of pH are
`shown [24].
`
`2.2. pH–solubility interrelationship of free acid and its salt
`
`Fig. 3 shows a schematic diagram for the pH–solubility
`interrelationship of a free acid and its salt form. The free acid
`would be the equilibrium species at a pH below pHmax, and it
`would convert to a salt only if it is equilibrated with a solution at
`a pH above pHmax by adding a sufficient quantity of an alkali or
`
`where BH+ and B represent, respectively, protonated (salt) and
`free base forms of the compound. When the aqueous medium at
`a given pH is saturated with the free base, the total solubility
`(ST) at that pH may be expressed as follows:
`
`ð
`ST; base pH N pH max
`
`s
`
`
`s 1 þ H3O
`Ka
`
`
`
`þ
`
`ð3Þ
`
`A.T.M. Serajuddin / Advanced Drug Delivery Reviews 59 (2007) 603–616
`
`605
`
`Fig. 1. Schematic representation of the pH–solubility profile of a basic drug
`indicating that the solubilities may be expressed by two independent curves and
`that the point where two curves meet is the pHmax (reproduced from Ref. [15]
`with permission).
`
`curves, one where the free base is the saturation or equilibrium
`species and the other where the salt is the equilibrium species.
`Essentially, the following equilibrium exists when a basic com-
`pound or its salt is dissolved in water:
`
`þ þ H2O fKa B þ H3O
`
`þ
`
`BH
`
`or
`
`
`½Ka ¼ B½ H3O
`
`þ
`½
`
`þ
`BH
`
`
`
`ð1Þ
`
`ð2Þ
`
` ¼ B½
`
`þ BH½
`
`Þ ¼ B½
`
`þ
`
`
`
`¼ B½
`s 1 þ 10pKa pH
`where the subscript “s” represents the saturation species. On the
`other hand, when the salt
`is the saturation species,
`the
`equilibrium solubility at a particular pH may be expressed by:
`
`ð
`ST; salt pH bpH max
`
`Þ ¼ BH½
`
`þ
`¼ BH
`½
`
`þ
`
`s
`
`
`½þ B½ ¼ BH
`
`
`
`s 1 þ Ka
`
`þ
`
`
`þ
`H3O
`s 1 þ 10pH pKa
`
`ð4Þ
`
`
`
`The two independent curves mentioned above may be
`obtained by varying hydrogen ion concentrations (or pH)
`in Eqs. (3) and (4), and the point where the curves intersect
`the pH of maximum solubility. This is
`is called pHmax,
`shown schematically in Fig. 1, where the solubility profile at
`a pH higher than the pHmax is represented by Eq. (3), while
`Eq. (4) represents the solubility profile below pHmax. If
`the solid phase that is in equilibrium with a solution is ana-
`lyzed, it would be the free base at pH N pHmax and the salt
`at pH b pHmax. Only at pHmax, which theoretically represents
`
`Fig. 3. Schematic representation of the pH–solubility profile of an acidic drug
`indicating that the solubility may be expressed by two independent curves and
`that the point where the two curves meet is the pHmax (reproduced from Ref. [15]
`with permission).
`
`Merck Exhibit 2193, Page 3
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`606
`
`A.T.M. Serajuddin / Advanced Drug Delivery Reviews 59 (2007) 603–616
`
`organic counterion. The relevant equations below and above
`pHmax are given below [25]:
`
`
`s 1 þ Ka
`½
`H3O
`
`ð
`ST; acid pHbpH max
`
`Þ ¼ AH½
`¼ AH½
`
`s
`
`
`
` ¼ AH½
`þ A ½
`
`
`
`
`s 1 þ 10pH pKa
`
`ð
`ST; salt pH N pH max
`
` ¼ A ½
`
`þ AH½
`
`Þ ¼ A ½
`
`
`
` ½
`
`s 1 þ 10pKa pH
`¼ A
`
`s
`
`
`
`
`þ
`ð5Þ
`
`
`
`
`½
`
`s 1 þ H3O
`
`Ka
`ð6Þ
`
`As indicated in Fig. 3, the solid phase in equilibrium with a
`saturated solution at pH b pHmax is the free acid and the solid
`phase at pH N pHmax is the salt; only at pHmax, both forms coexist.
`Interconversion from the salt to the free acid form or vice versa
`may occur if the pH shifts from one side of the pHmax to the
`other. There are numerous reports in the literature indicating that
`Eqs. (5) and (6) are, in general, followed for solubilities of free
`acids and their salts, respectively [22,23,26–29]. In all cases, salts
`had higher solubilities than their corresponding free acids,
`although solubilities of different salt forms of a particular acid
`could vary.
`
`2.3. Effect of counterion on salt solubility
`
`Salt-forming agents used to prepare salts, such as acids to
`form salts of basic drugs and bases to form salts of acidic drugs,
`exert influences on salt solubility by exerting common-ion
`effects in solution. This may be seen in Fig. 2, where solubilities
`of methanesulfonate, hydrochloride and phosphate salts of
`haloperidol decreased gradually at pH below 2.5. This is due to
`the common-ion effect since the acids used to lower pH
`generated excess counterions.
`The common-ion effect may be explained by the following
`equilibrium that exists below pHmax for the salt of a basic drug:
`ð
`Þ
`½
`
`
`þ X ½
`
`ð7Þ
`f BH
`
`þ
`
`BH
`
`
`
`X
`
`solid
`
`þ
`
`s
`
`where (BH+X−)solid denotes undissolved solid salt that is in
`equilibrium with solution, [BH+]s is the salt solubility, and [X−]
`is the counterion concentration. The apparent solubility product
`(Ksp′ ) can be derived from Eq. (7) as follows:
`
`K Vsp ¼ BH½
`
` ½
`
`s X
`p
`ffiffiffiffiffiffiffi
`In the absence of excess counterion, [BH+]s = [X−], and there-
`K Vsp
`. Under such a condition, the solubility of a
`fore, solubility =
`salt remains unchanged as seen in the flat region of salt solubility in
`Fig. 2. On the other hand, if a significant amount of excess
`counterion is used either to lower pH or as a formulation adjuvant
`in dosage form (e.g., in adjusting ionic strength, tonicity, etc.), a
`major decrease in solubility may be observed, according to:
`½
`
`
`¼ Ksp= X ½
`
`þ
`BH
`
`ð8Þ
`
`ð9Þ
`
`þ
`
`s
`
`Streng et al. [30] studied the combined effect of the addition
`of NaCl and HCl on aqueous solubility of the HCl salt of a basic
`drug; the solubility in the relatively flat region of the pH–
`solubility profile (pH 3 to 6) decreased by a factor of 3 when
`0.05 M NaCl was added to the solution, while at pH below 3 the
`solubility further decreased due to the effect of Cl− ion as-
`sociated with HCl added to adjust pH. Similarly, in developing a
`liquid formulation for the sodium salt of an acidic drug,
`Serajuddin et al. [31] observed a decrease in solubility from
`7.8 mg/mL to 1.1 mg/mL with the addition of 0.1 M NaCl to
`adjust the ionic strength of solution. The common-ion effect
`also has a major influence on solubility and dissolution rates of
`salts in the GI tract, where the solubility of HCl salts are
`particularly sensitive to the presence of chloride ion [24].
`The overall impact of counterions on salt solubility depends
`on the magnitude of Ksp value. According to Eq. (9), for an
`equal change in [X−],
`the common-ion effect will be less
`pronounced in a salt of higher Ksp (i.e. higher solubility) than in
`a salt with lower Ksp (i.e. lower solubility). For example, the
`aqueous solubility of
`tiaramide HCl at 37 °C remained
`practically constant around 200 mg/mL (∼0.5 M) during the
`lowering of pH from 4.0 to 1.6 by the addition of HCl, since, as
`compared to the drug concentration, changes in the chloride ion
`concentration during the pH adjustment were negligible.
`Further, the solubility of tiaramide HCl decreased by just 25%
`to ∼150 mg/mL at pH 1. In contrast, there are numerous reports
`in the literature indicating drastic common-ion effects on salts
`having relatively low aqueous solubilities [18,21]; three such
`examples demonstrating major impacts of maleate and chloride
`ions on solubilities of a maleate salt [32] and two hydrochloride
`salts [33,34], respectively, at low pH are shown in Fig. 4. Since,
`as mentioned earlier, most compounds currently synthesized in
`drug discovery laboratories have poor aqueous solubilities and,
`as a consequence, their salt forms are also found to have
`relatively low aqueous solubilities, an investigation of potential
`impacts of common ions is critically important in salt selection
`and during dosage form development.
`
`2.4. Effects of solubility, pKa and Ksp on pHmax
`
`The concept of pHmax is an important one in the physical
`chemistry of salts. It is apparent from Figs. 1 and 3 that pHmax
`plays a major role in determining whether a salt would be
`formed or not, and, in case it is formed, whether it would remain
`‘as is’ or would convert to the corresponding free acid or base
`form. As mentioned previously, it is only at the pHmax that both
`forms could coexist. Therefore, at the pHmax, both Eqs. (3) and
`(4) can be valid for the solubility of a basic drug and, similarly,
`both Eqs. (5) and (6) can be valid for the solubility of an acidic
`drug. Bogardus and Blackwood [18] proposed that, for a basic
`drug, the saturation solubilities of free base and its salt form
`may be set equal at pHmax, and solving the relevant equations
`for pHmax, they derived the following relationship:
`
`pH max ¼ pKa þ log
`
`sffiffiffiffiffiffiffi
`p
`B½
`
`Ksp
`
`ð10Þ
`
`Merck Exhibit 2193, Page 4
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`A.T.M. Serajuddin / Advanced Drug Delivery Reviews 59 (2007) 603–616
`
`607
`
`It is evident from Fig. 5 that a stronger basicity (higher pKa), a
`higher intrinsic solubility and a lower salt solubility will favor salt
`formation for a basic drug by increasing pHmax. Analogous
`relationships may also be derived for the salt formation of an
`acidic drug where an increase in S0 and decreases in pKa and salt
`solubility will decrease pHmax and, therefore, favor salt formation.
`
`2.5. Deviation of pH–solubility interrelationship from ideality
`
`Organic compounds often undergo self-association in solu-
`tion because of their amphiphilic nature [35,36]. Indeed, bile
`salts are great examples of how organic compounds exhibit
`surface activity and undergo self-association in aqueous solu-
`tions because of their amphiphilic properties [37]. It has been
`reported that salt forms of many drug molecules undergo similar
`aggregation in solution [38–41]. Because of self-aggregation,
`activities of saturated solutions of many salts and even non-salts
`are lower than their measured concentrations in solution, re-
`sulting in non-ideal pH–solubility behavior. An example of
`
`Fig. 4. Typical pH–solubility profiles of poorly water-soluble basic drugs:
`(a) the solubility profile of a compound with intrinsic solubility (S0) of 2 μg/mL
`and a pKa of 6.3, for which a pHmax of ∼3.4 and a common-ion effect below
`pHmax were observed when the pH was lowered using maleic acid (reproduced
`with permission from Ref. [32]); (b) solubility profile of a compound with S0 of
`3.4 μg/mL and pKa of 5.7, for which pHmax of 3.2 and common-ion effect below
`pHmax were observed when the pH was lowered using HCl; and (c) the solubility
`profile of a base having S0 of b0.0001 μg/mL (below detection limit) and
`estimated pKa in the range of 5.5 to 6.0, for which the pH was adjusted by HCl
`and the hydrochloride salt did not have acceptable properties for further
`development due to low pHmax (∼1.5), low salt solubility (0.1 mg/mL at pHmax)
`and strong common-ion effect (reproduced with permission from Ref. [34]).
`
`Pudipeddi et al. [16] depicted the influence of S0 (or [B]s),
`pKa and Ksp on pHmax, according to Eq. (10), by using Fig. 5,
`where:
`
`a) an increase in pKa by one unit increases the pHmax by one
`unit;
`b) an increase in intrinsic solubility, S0, of the base by one order
`of magnitude increases pHmax by one unit; and
`c) a decrease in salt solubility (Ksp) by one order of magnitude
`increases pHmax by one unit.
`
`Fig. 5. Effects of (a) pKa, (b) S0 and (c) Ksp on pHmax (reproduced from Ref. [16]
`with permission).
`
`Merck Exhibit 2193, Page 5
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`608
`
`A.T.M. Serajuddin / Advanced Drug Delivery Reviews 59 (2007) 603–616
`
`increase in activity coefficients of salts in solution [18,21,22]. In
`other words, solubilities of HCl salts decreased not only due
`to common-ion effect, but also due to salting out effects that
`decreased self-association [42].
`Bergstrom et al. [43] reported that experimentally determined
`solubility profiles of 25 amine drugs differed substantially from
`theoretical pH–solubility profiles generated by using the
`Henderson–Hasselbach (HH) equation. The HH equation used
`is essentially the same as Eq. (3) described earlier for the pH–
`solubility profile of a base. This equation would not describe the
`solubility of a salt form and, therefore, a deviation in solubility
`profile at pH b pHmax was expected. Some of the deviations
`observed at pH N pHmax could be due to self-association of
`dissolved species, especially at or near pHmax. The accuracy of a
`theoretical pH–solubility profile also depends on accuracies of
`intrinsic solubility and pKa values used in Eq. (3). The work of
`Bergstrom et al. [43] highlights the importance of accurate S0
`and pKa values; otherwise, the theoretical profiles generated may
`fail to appropriately predict the experimental behavior.
`
`2.6. Structure–solubility relationships
`
`As shown earlier in Fig. 2, aqueous solubilities of halo-
`peridol salts differed depending on salt-forming agents used.
`There are numerous other reports in the literature presenting
`such results for both acidic and basic drugs [25,30,44–48]. For
`example, Streng et al. [30] reported that solubilities of ter-
`fenadine salts formed with phosphoric acid, hydrochloric acid,
`methanesulfonic acid and lactic acid showed up to 10-fold
`differences, ranging from 0.5 mg/mL to 5 mg/mL. It may be
`noted that acidities and structures of acids used to form the salts
`in this study differed greatly. On the other hand, Serajuddin [33]
`reported for a basic drug, avitriptan, that solubilities of salts
`could be similar if structurally similar acids are used (Table 1).
`Although avitriptan is a dibasic compound with pKa values of
`8.0 and 3.6 [15], it formed mono-salts with all acids tested,
`except for HCl, which was also able to form a dihydrochloride
`salt. The solubility differed significantly only for the hydro-
`chloride salts. The results obtained by O'Connor and Corrigan
`[29] for diclofenac salts with a series of structurally similar
`amines, however, differed from the observation of Serajuddin
`[33] that similar counterions might provide similar solubilities;
`their values differed by a factor as much as N100.
`Anderson and Flora [13]
`indicated that no predictive
`structure–solubility relationships for pharmaceutical salts have
`yet been established. They, however, suggested that
`for
`
`Table 1
`Aqueous solubility of mono-salts of avitriptan containing various counterions
`
`Acid used
`
`pKa of acid
`Solubility (mg/mL at 25 °C)
`−6.1
`3.4 (9.0) a
`HCl
`−1.2
`Methanesulfonic acid
`16.3
`Tartaric acid
`3.0, 4.4
`14.7
`Lactic acid
`3.9
`15.2
`Succinic acid
`4.2, 5.6
`16.1
`Acetic acid
`4.8
`16.5
`a Value in parentheses is for the dihydrochloride salt.
`
`Fig. 6. pH–solubility profile of papaverine hyrochloride determined by using HCl
`or NaOH, as necessary, indicating supersaturation around pHmax and common-ion
`effect at low pH. The broken line shows the theoretical solubility profile determined
`based on S0 and pKa, and the positive deviation of experimental profile indicates
`self-association (reproduced from Ref. [19] with permission).
`
`such non-ideality is shown in Fig. 6, in which the saturation
`solubility of papaverine hydrochloride at or below pHmax
`exhibits higher values than the theoretical profile. There are also
`numerous examples where metastable, supersaturated solutions
`are formed at or near pHmax [13,19–22]. Supersaturated solu-
`tions are formed more often when free base or free acid is used
`as the starting material in the phase solubility study due to
`‘kinetic barriers’ in the transformation of free species to the salt
`form. Such kinetic barriers may also be present in the con-
`version of salts to free species, but they are relatively less
`pronounced. Serajuddin and Mufson [20] reported that
`the
`solubility of papaverine free base at 37 °C increased gradually
`when the pH was decreased by adding HCl to a suspension until
`the pH reached 4. Then, at an almost constant pH of 4 ± 0.1, the
`solubility increased from b10 mg/mL to N120 mg/mL; the
`solubility kept on increasing with the addition of HCl as long as
`solid base was present to dissolve into solution. Although it was
`known that the aqueous solubility of papaverine hydrochloride
`was only 40 mg/mL, there was no sign of precipitation of the
`supersaturated solution to the salt from even at a concentration
`above 120 mg/mL, and it was only after nucleation of the
`supersaturated solution by adding a few crystals of papaverine
`HCl that the precipitation of the salt ensued and the solubility
`dropped to the level of the salt form. With a similar addition of
`HCl to a phenazopyridine aqueous suspension, the solubility of
`the free base at pHmax was observed to reach at least three times
`higher than that of phenazopyridine HCl before precipitation of
`the salt form ensued [21,22]. Such a supersaturation and the
`deviation from the ideal pH–solubility relationship were
`attributed to self-association of drug in solution.
`The aggregation of drug in solution does not occur only at the
`pHmax. Ledwidge and Corrigan [22] demonstrated that salts of
`both basic and acidic drugs may self-associate and exhibit
`surface activity at a wide range of pH. For hydrochloride salts,
`the addition of an excess of chloride ion during the pH ad-
`justment showed a decrease in apparent Ksp values, which was
`attributed to a decrease in self-association and the consequent
`
`Merck Exhibit 2193, Page 6
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`A.T.M. Serajuddin / Advanced Drug Delivery Reviews 59 (2007) 603–616
`
`609
`
`understanding the effect of a series of salt-forming counterions
`on aqueous solubility, contributions of the counterions to crystal
`lattice energies and solvation energies should be considered. In
`dissolving a salt in water, the molar free energy of solution,
`▵Gsoln, may be represented by
`ð11Þ
`DGsoln ¼ DGcation þ DGanion DGlattice
`where ▵Gcation and ▵Ganion are molar free energies of hydration
`of the salt cationic and anionic species, respectively, and ▵Glattice
`is the crystal lattice free energy. Thus, the overall effect of
`counterions on salt solubility will depend on whether hydration
`energies or lattice energies are most sensitive to changes in salt
`structure. By analyzing data for alkali and alkaline earth metal
`salts of several carboxylic acids, the authors suggested that a
`quantitative trend exists where the solubility increases with an
`increase in anion or cation charge and decreases with an increase
`in ionic radius. A general trend was also observed where solubility
`of various ammonium salts of an acidic drug, flurbiprofen,
`increased with a decrease in melting point, indicating that crystal
`lattice energy plays an important role in salt solubility.
`Several other investigations also attempted to establish
`relationships between salt forms and their aqueous solubilities
`[28,29,44–46]. However, no general predictive relationships
`could be obtained. For example, it was reported that solubilities
`of diclofenac salts with several structurally related primary
`amines varied by a factor of as much as 100, and they did not
`show dependence on any one parameter, but on a combination
`of factors like salt crystal
`lattice and pH of saturated salt
`solution [29].
`
`2.7. Effect of organic solvent on salt solubility
`
`The pH–solubility principles described in this section are
`applicable to aqueous solutions. However, as reported by
`Wermuth and Stahl [49], pharmaceutical salts are usually
`synthesized either from organic solvents or from organic–water
`cosolvents. No systematic studies on the solubility of acidic and
`basic drugs in such solvents as a function of added counterions to
`form salts have been reported in the literature.
`Organic solvents are used not only in the preparation of salts;
`they are also used in parenteral and other liquid dosage forms.
`Organic solvents may influence the solubility of a drug candidate
`by (a) increasing solubility of unionized species (S0), (b) de-
`creasing its protonation or ionization, and (c) decreasing solubility
`of the salt form [15]. Thus, as reported by Kramer and Flynn [17],
`an increase in S0 in an organic cosolvent will increase pHmax for a
`basic drug, thus favoring salt formation. This is, of course,
`assuming that the pKa value of the compound would remain
`unchanged. However, the presence of organic solvents may
`adversely affect drug ionization by suppressing the dielectric
`constant [50]. For example, Albert and Serjeant [51] reported that
`the pKa of an acid increased by ∼1 and that of a base decreased by
`∼0.5 in a 60:40 methanol–water mixture. As a consequence, any
`positive influence of organic solvents on salt formation may partly
`or fully be negated by the decreased ionization. The decreased salt
`solubility in the presence of organic solvents may influence drug
`
`produc