International Union
`of Pure and Applied Chemistry (IUPAC)
`
`Handbook of
`Pharmaceutical Salts
`Properties, Selection, and Use
`
`P. Heinrich Stahl, Camille G. Wermuth (Eds.)
`
`Verlag Helvetica Chimica Acta · Zurich
`
`@ WILEY-VCH
`
`Weinheim · New York · Chichester
`Brisbane · Singapore · Toronto
`
`Merck Exhibit 2042, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Dr. P. Heinrich Stahl
`Lerchenstrasse 28
`D-79104 Freiburg im Breisgau
`
`Prof. Camille G. Wermuth
`Louis Pasteur University, Strasbourg
`Faculty of Pharmacy
`74, route du Rhin
`F-67 400 Illkirch
`
`This book was carerully produced. Newrrhcless, editor and publishers do not warrant the infomiation contained therein
`to be free of errors. Readers are advised to k.eep in mind that statements, data, illustrations, procedural details, or other
`iteri1s may inadvertently be inaccurate.
`
`Published jointly by
`VHCA, Verlag Helvetica Chimica Acta, Ziirich (Swi1zerland)
`WILEY-VCH. Weinbeim (Federal Republic of Germany)
`
`Editorial Directors: Thomas Kolitzus, J)r. M. Volk.an Kisakiirek
`Production Manager: Norben Wolz
`
`Cover Design: Bettina Bank
`
`Lihrary of Congress Card No. applied for.
`
`A CLP catalogue record for this book is available from the British Ubrary.
`
`Di.e Deutsche Bibliothek. - ClP-Cataloguing-in-Publication-Data
`
`A catalogue record for this publication is available from Die Deutsche Bihliothek
`
`JSBN 3-906390-26-8
`
`© Verlag Helvetica Chimica Acta, Postfach, CH-8042 Zilricb, Switzerland, 2002
`
`Printed on acid-free paper.
`
`All rights reserved [including those of translation into other languages). No pan of this book may be rcproduced in any form
`- by photoprinring, microfilm, or any other means - nor transmitted or translated into a machine language without written
`permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such,
`are not to be considered unprotected by law.
`
`Printing:.Kontad Triltsth, Print und DigitaJe Medien, D.-97199 0chsenfuri-Hohcstadt
`Printed in Germany
`
`Merck Exhibit 2042, Page 2
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`1].cAL SALTS:
`
`a,D. :flensch(cid:173)
`~.~g. Heidelberg,
`
`~"Klaassen, 5th
`
`schaftsver-
`
`Chapter 6
`
`Salt-Selection Strategies
`
`by Abu T. M. Serajnddin* and Madhu Pudipeddi
`
`Contents
`
`1. Introduction
`2. Selettion of Chemical Forms of Salt,;
`2. L Feasibility Assessment for Salt Formation
`2.2. Application of pH-Solubility Relationship: Case Histories
`2.2.1. Case History 1: REV5901
`2.:Z.2. Case History 2: GW1818
`2.2.3. Case History 3: Phenytoin
`2.3. Theoretical Modeling of pH-Solubility Relationship
`2.4. Feasibility of Disalt Formation
`2.4.1. Feasibility of Salt Formation for Dibasic Compounds
`2.4.2. Feasibility of Salt Formation for Diprotic Acids
`2.5. Effect of Counter-Ions on Salt Solubility
`2.5. l. Common-Jon Effect on Salt Solubility and Dissolution
`2.5.2. in-situ Screening of Counter-Ion Effects on Salt Solubility
`2.6. Effect of Organic Solvents on Salt Formation
`3. Selection of Physical Form
`3 .1. A Multi-Tier Approach
`4. Salt-Selection 1'iming
`5. Salt-Selection Team
`6. Summary and Cc;mclusions
`REFERENCES
`
`1. Introduction
`
`Because of the introdµction of combinatorial chemistry and high(cid:173)
`throughput screening (HTS) during the past ten years, the pharmaceutical in-
`
`Merck Exhibit 2042, Page 3
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`·.· .. .,_
`
`m~tJ
`stfilii
`at;i.Qi
`b~p
`de'\1,
`rn~
`ter((
`;;;-,:c
`,:.:.'-'
`,.-:-·-~·,'l;l
`
`·cIPel
`'C()!l
`
`136
`
`PHARMACEUTICAL SALTS:
`
`dustry is going through a revolutionary change in the way it bas been discov(cid:173)
`ering and developing drugs [l]. Larger, more lipopbilic, and less water-solu(cid:173)
`ble leads are being selected as a result of the quest for more potent and high(cid:173)
`ly specific molecules. The widespread use of dimethyl sulfoxide·(DMSO) in
`HTS also favors the selection of lipophilic, water-insoluble compounds,
`which are easily solubilized in this solvent. Since some of the attributes of
`newer drug molecules are unfavorable to their development as dosage forms,
`the 'developability' is becoming a critical consideration for the transition of
`a chemical entity from the discovery phase to the development phase [21 [3].
`There is now a greater collaboration between discovery and development sci(cid:173)
`entists in evaluating such developability criteria as solubility, dissolution rate,
`stability, permeability, and so forth, for the selection of optimal-development
`candidates. Since, as mentioned in Ch.apt. 2, salt formation can improve sol(cid:173)
`ubility and dissolution rate of basic and acidic drugs, thus increasing their ab(cid:173)
`sorption rate and bioavailability, we will present in this chapter various strat(cid:173)
`egies for the selection of optimal salt forms for new drug candidates. The
`physicochemical principles to be described in this chapter will also be help(cid:173)
`ful in identifying acidic or basic drug candidates that can form more devel(cid:173)
`opable salts.
`The salt selection should be viewed as a part of the overall objective of
`selecting the 'optimal form' of a drug candidate for development. When one
`refers to the optimal form, it involves both chemical and physical forms. A
`new chemical entity can be an acid, a base, or a neutral species. If it is a neu(cid:173)
`tral species, there are no options for chemical manipulation to make it more
`developable other than possibly preparing prodrugs. On the other hand, if it
`is an acid or a base, one can select the free acid or base form, or, alternative(cid:173)
`ly, one can select a salt form. In the selection of free vs. salt form, questions
`that need to be answered are: Is the acid or base form preferred because of
`biopharmaceutical considerations? ls the salt form more suitable? ls the prep(cid:173)
`aration of stable salt forms feasible? Among various potential salt forms of a
`particular drug candidate, which has the most desirable physicochemical and
`biopharmaceutica1 properties?
`Along with the evaluation of chemical form, the strategy for the selec(cid:173)
`tion of physical form must also be considered. One needs to determine
`whether the compound exists in crystalline or amorphous form, and, if crys(cid:173)
`talline, whether it exhibits polymorphism. One also needs to investigate:
`Does the compound exist in hydrate or solvate form? If so, how is such a
`form affected by temperature and moisture? How stable is a particular form
`in solid state and in solution? The ultimate selection of the 'optimal form'
`of a new drug candidate for development depends on a balance among the
`physicochemical properties of its various available chemical and physical
`forms.
`
`Merck Exhibit 2042, Page 4
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`nrate,
`pment
`~:ve sol(cid:173)
`~itheir ab-
`;,.,
`
`PROPERTIES, SELECTION, AND USE
`
`137
`
`Another critical element of a salt-selection process in any drug-develop(cid:173)
`ment program is the timing. Here, the critical questions are: When does one
`start salt selection? Should a new drug candidate be selected after consider(cid:173)
`ation of its feasibility for salt formation? Or should any such consideration
`be postponed, until the new candidate has been selected and forwarded to the
`development stage? How can the salt selection be integrated in the develop(cid:173)
`ment process such that it does not become a rate-limiting step or does not ex(cid:173)
`tend development time?
`'fhe success of a salt-selection program also depends on how various dis(cid:173)
`ciplines within drug discovery and development interact and collaborate. The
`composition of a salt-selection team and the responsibilities of individual
`team members may have profound effects on time and re.<;ources spent on a
`salt-selection program.
`Based on the above considerations, salt-selection strategies for new drug
`candidates may have the following components:
`
`,) selection of chemical forms of salts,
`ii) selection of physical forms of salt.'!,
`iiz) salt-selection timing,
`iv) composition of salt-selection team
`
`In the present chapter, strategies for the selection of chemical forms of
`salts will be described in detail. Strategies for the selection of physical forms
`will be discussed in less detail, since Chapt. 3 and 7 wil1 also cover several
`aspects of these strategies. Salt-selection timing and composition of salt-se(cid:173)
`lection teams will be discussed only briefly, since no clear picture of how
`these are practiced in various drug companies has emerged yet.
`
`2. Selection of Chemical Forms of Salts
`
`At the outset of any salt-selection program, it is important to determine
`whether a particular acid or base is amenable to salt formation. If the salt for(cid:173)
`mation appears to be feasible, the question then arises is which one of the
`many available counter-ions would be most suitable for the purpose. Some
`of these issues will be addressed in this section.
`
`2.1. Feasibility Assessment for Salt Formation
`
`No predictive procedure to determine whether a particular acidic or basic
`drug would form a salt with a particular counter-ion has been reported in the
`literature. Anderson and Flora {4] reported that successful salt formation gen-
`
`Merck Exhibit 2042, Page 5
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`138
`
`PHARMACEUTICAL SALTS:
`
`erally requires that the pKa value of a conjugate acid should be·smaller than the
`pKa value of the conjugate base to ensure sufficient proton transfer from the
`acidic to the basic species. Thus, relatively stronger acids like HBr, HCJ.,
`H2S04, or one of the sulfonic acids (pKa < 2.0) would be suitable for the prep(cid:173)
`aration of salts of weakly basic amines having pKa < 4. Other investigators also
`provided similar but rather general guidelines for the selection of counter-ions.
`Wells [5], and also Tong and Whitesell [6] recommended that, for the prepara(cid:173)
`tion of salt forms of a basic drug, the pKa of the acid used should be at least
`2 pH units lower than the pKa of the drug. Although these are valuable guide(cid:173)
`lines, a more predictive method for assessing the feasibility of salt formation
`would be necessary to minimize trials and errors in a salt-selection program.
`As described in Chapt. 2, the pH-solubility interrelationship and the lo(cid:173)
`cation of pHmax in the pH scale play critical roles in determining which salt,
`if any, can be synthesized for a particular free acid or base. Dittert et al. [7]
`reported as early as in 1964, although not for the specific purpose of salt se(cid:173)
`lection, that whether a basic drug would exist as the free base or as a salt un(cid:173)
`der certain pH conditions can be determined by studying its solubility vs. pH
`relationship. Later, Kramer and Flynn [8] demonstrated that the pH-solubil(cid:173)
`ity relationship of a basic drug could be expressed by two independent cur(cid:173)
`ves, and the point where the two curves intersected was the pHmax• the pH of
`maximum solubility. This is shown in Fig. I, and the relevant equations are
`given below:
`
`At pH> pHmax:
`
`Sr= [BH+1 + [Bls
`= [B]s. (1 + [H30+J/Ka)
`
`(1)
`
`(2)
`
`At pH< pH,nax:
`
`ST= [BH+] 5 + [B]
`= [BH+]s · (1 + Kj[H30~)
`In both Eqns. I and 2, ST is the total or equilibrium solubility under a partic(cid:173)
`ular pH condition, [BJ and [BW] ate concentrations of free and protonated
`species of the base, respectively, and the subscript s represents the concen(cid:173)
`tration in equilibrium with the solid phase. Fig. I essentially illustrates that
`a salt would not be formed in an aqueous medium, unless the pH of the sat(cid:173)
`urated solution of a basic drug is not lowered below the pHmax, and any salt
`formed would be reconverted to its free base form, if the pH of a saturated
`salt solution is raised above the P~ax· In other words, solid phases that re(cid:173)
`main in equilibrium with solutions at pH below and above pHmax are a salt
`and- the free base, respectively.
`Similar pH-solubility relationship also exists for acidic drugs [9] [ I 0]. As
`illustrated in Fig. 2, for a rnonoprotic acid, the free acid would be the equi(cid:173)
`librium species at a pH below the pHmax• and a salt would be formed only if
`the pH is raised above the PRmax by using suitable counter-ions. The relevant
`
`Merck Exhibit 2042, Page 6
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`PROPERTIES, SELECTION, AND USE
`
`139
`
`ST= [BH1s +[B]
`= [BH"'ls{1 + Kai[H301}
`
`I
`I
`l
`
`Solid Phase:
`Salt
`
`Solid Phase:
`Base
`
`pH
`
`Fig. 1. Schematic representatii:m of the pH-solubility profde of a monobasic compound, show(cid:173)
`ing that solubilities of base and salt can b-e expressed by two independent curves cornspond(cid:173)
`ing to two i.ndependen.t equations. The point where the two curves intersect is the PH.nax-
`
`Sr= [AH].+ [A-]
`= [AHi. (1 + Kj[H3O+})
`
`Sr= [A-Js+ [AH]
`= [Al. ( 1 + [H3O+y Ka)
`
`Solid Phase:
`Acid
`
`Solid Phase:
`Salt
`
`pH
`
`Fig. 2. pH-Solubility profile analogous to Fig. I of a monoprotic acid
`
`equations are given below:
`
`At pH< pHmax.:
`
`ST= [AH],+ [A-]
`= [AH]s · (1 + Kal[H30+])
`ST= [A-] 5 + [AH]
`= {KJs · (1 + [H30']/KJ
`
`(3)
`
`(4)
`
`Merck Exhibit 2042, Page 7
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`140
`
`PHARMACEUTICAL SALTS:
`
`In Eqns. 3 and 4, ST is again the total solubility under a particular pH condi(cid:173)
`tion, [AH] and [ A-] are concentrations of free and ionized species of the acid,
`respectively, and the subscript s represents the concentration in equilibrium
`with the solid phase.
`Solubilities of salts, as described by Eqns. 2 and 4, can be influenced by
`excess counter-ions present in solution. However, counter-ions influence sol(cid:173)
`ubilities through solubility products only after salts are formed and, therefore,
`might not adversely affect the feasibility of salt formation. The issue of sol(cid:173)
`ubility product on the salt-selection strategy will be discussed in a later sec(cid:173)
`tion of this chapter.
`Serajuddin and co-workers [10-14], and numerous other authors [8] [9]
`[15-18] confirmed the application of the above-mentioned pH-solubility re(cid:173)
`lationships in determining under which pH conditions salts of particular acid(cid:173)
`ic and basic drugs can be formed.
`
`2.2. Application of pH-Solubility Relationship: Case Histories
`
`The application of pH-solubility relationships in determining the feasibil(cid:173)
`ity of salt formation can be explained by a few case histories.
`
`2.2.1. Case History I: REV5901
`
`To determine the feasibility of salt formation for REV-5901 (Fig. 3), a
`base with the pKa value 3.7, Serajuddin et al. [14) determined its pH-solu(cid:173)
`bility profile as shown in Fig. 4. An identical profile was obtained, when
`either the free base or the hydrochloride salt was used as; the starting solid
`phase. The pHmax of the compound was I, indicating, a<; mentioned above, a
`salt fonn would exist only at pH below 1.0. Indeed, only two salts, a hydro(cid:173)
`chloride salt and a sulfate salt, could be prepared for REV-5901, since only
`strong acids like HCI and H2S04 could lower the pH of a saturated solution
`below the pHmax of 1. A salt formation with relatively weaker acids like phos(cid:173)
`phoric acid, acetic acid, lactic acid, tartaric acid, and so forth, would not be
`feasible, since such acids would be unable to lower the pH below 1.0. Thus,
`just from the pH-solubility relationship, one can narrow down the type and
`the number of salts that can be prepared, saving much efforts and resources
`that could otherwise be wasted in attempting to synthesize many different salt
`forms. Based on Fig. 4, rne would even question the suitability of hydro(cid:173)
`chloride and sulfate salts for development, because such salts would be con(cid:173)
`verted to the free base fonn, when the microenvironmental pH in presence
`of moisture rises above 1.0. It was, indeed, observed that both of these salts
`
`Merck Exhibit 2042, Page 8
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`;e}MTS:
`;:.::,'.·,
`Ji'::
`~ondi-
`t¢facid,
`lbrium
`
`:edby
`f~sol-
`t,;<•"
`·"·efore,
`)fSol(cid:173)
`!ftr sec(cid:173)
`f}v
`;8l[9]
`!!#ity re'(cid:173)
`ffi' acid-
`
`~ibil-
`
`)lg. 3), a
`iH~solu-
`when
`I!b solid
`~e,a
`1 !ydro(cid:173)
`:.
`; only
`solution
`;· phos-
`' 1ot be
`~:Thus,
`"'-e and
`,
`,urces
`)fe[lt salt
`fhydro-
`:con(cid:173)
`:m;sence
`ese salts
`
`PROPERTIES, SELECTION,AND USE
`
`141
`
`pK,, =3.6
`S0 = 0,002 mg/ml
`
`Fig. 3. Chemical stn,cture of REV5901, the compound used in Case History 1
`
`1.0
`
`0.08
`
`,........
`
`0.8
`
`E ---0)
`.s
`0.6
`~
`:0 0.4
`::,
`0
`U)
`
`W ch 0.06
`
`.§.
`0.04
`~
`:0 :::,
`0
`<J)
`
`0.02
`
`[fil
`
`6.
`
`3
`
`5
`
`6
`
`4
`pH
`
`0.2
`
`00
`
`1
`
`2
`
`4
`
`5
`
`6
`
`3
`pH
`
`Fig. 4. pH-Solubility profile of REV5901 (A), where triangles represent the solubility obtained
`with. the free base and circles the data obta.ined with the hydrochloride salt. In the insert (B),
`the solubility is shown on an expanded scale. Either HCI OT NaOH was used to adjust pH.
`
`did not have acceptable properties for development, and the free base form
`of REV-5901 was ultimately selected.
`
`2.2.2. Case History 2: GW1818
`
`Tong and Whitesell [6] studied the feasibility of salt formation of a basic
`drug GW1818, which bad the pKa: value of 8.0and the intrinsic free base sol(cid:173)
`ubility of 0.0044 mg/ml. For this compound, the pHmax• was ca. 5, and, as a
`result, the formation of stable salts with both strong and weak counter-ions,
`such as hydrochloride, methanesulfonate, phosphate, and succinate, was fea(cid:173)
`sible. This is because all of these counter;..ions coµld loy.,er the pH below 5.
`
`Merck Exhibit 2042, Page 9
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`cl~j
`tein
`eairl
`Itil~
`;~
`3if~
`
`rt1'£. ;
`__t' . . c~ ~
`
`. ing
`. c:'01
`Spt
`··•.···•<lt
`
`142
`
`PHARMACEUTICAL SALTS:
`
`2.2.3. Case History 3: Phenytoin
`
`The feasibility of salt formation for an acidic drug can be illustrated by
`the pH-solubility profile of phenytoin (Fig. 5), a compound with a pKa value
`of 8.4 and the intrinsic free acid solubility of 0.02 mg/ml at 37 °C [19]. The
`sodium salt is the commercially available salt form for phenytoin, and there
`are numerous reports in the literature demonstrating that the free acid form
`of phenytoin precipitate out of salt solutions depending on pH. There is also
`the propensity for the conversion of salt to free acid in solid dosage form. It
`is apparent from Fig. 5 that the salt formation for phenytoin is feasible only
`with strong alkalis like NaOH because it can raise the pH above the pHmax
`value of 11. Since relatively weaker bases like Mg(OHh, Ca(OHh, etc., and
`the commonly used amine bases like arginine, lysine, etc., would not raise
`the pH of an aqueous solution above 11, they will not form salts with phen~
`ytoin. Fig. 5 also indicates that any salt formed would be converted to the
`free acid if the microenvironmental pH were below 11. If, unlike phenytoin,
`the pHmax of an acid were, for example, around 8, there would be a much
`better option for salt formation, because the pH could be raised above 8 by
`using a larger selection of alkalis and bases.
`
`- 140
`E 120
`--O'I 100
`E.
`60
`>-
`40
`~ 20
`:0
`10
`::i
`6
`0
`en
`4-
`2
`1
`0.6
`0.4
`0.2
`0.1
`0.06
`0.04
`0.02
`
`pKa = 8.4
`So = 0.04 mg/ml
`
`4 }\
`I 0 Free Acid
`
`A Sodium Salt
`
`0
`
`2
`
`4
`
`6
`
`pH
`
`8
`
`10
`
`12
`
`14
`
`Fig. 5. pH-Solubility pro.file of phenytoin at 37°C indicating pHmax at JJ. Identical profiles
`were obtained when either free acid or the sodium salt of the drug substance was used. pH was
`adjusted using either NaOH or HCl.
`
`Merck Exhibit 2042, Page 10
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`,AL'SALTS:
`
`PROPERTlES; SELECTION, AND USE
`
`143
`
`ated by
`a value
`a19]. The
`d there
`id fottn
`~e is also
`m. It
`
`2.3. Theoretical Modeling. of pff-Solubilify Relationship
`
`In the case histories-mentioned above, pH.;...solubility profiles, which were
`determined experimentally, have been used to identify the pHmax and to de(cid:173)
`termine the feasibility of salt fonnation, However; at drug discovery· and·
`early development stages, when the supply of: drug substances is limited, it
`might not be practical to determine pH-solubility profiles experimentally. In
`such a situation, the pH-solubiiityrelationships corresponding toEqns, 1 and
`3 for basic and acidic drugs,. respectively, can be generated theoretically; if
`pKa and S0 values are available, 1'.hen, an· estimate of pHmax values can be
`made by assuming celiain values. for• salt solubilities.
`Fig. 6shows pH-solubility profiles of a basic compound generated theo(cid:173)
`retically according to Eqn. 1 by using a fixed pK., value of 8.0 and· various
`S0 values ranging from 0.0001 to lO mglml. In this case, solubilities of salt
`forms corresponding to various theoretical curves are unknown. Any partic(cid:173)
`ular value for the salt solubility can be assumed for the purpose of estimat(cid:173)
`ing pHmax values. If, for example, a salt solubility of 20, mg/ml is assumed
`corresponding to each curve in Fig. 5, the estimated pHmax values• corre(cid:173)
`sponding to Si> values of0.0001, 0.001, 0.0l, 0.1, 1, and lOmg/ml would be
`3, 4, 5, 6, 7,. and 8, respectively. The pHmax values would not differ much
`even if the solubility of salt form somewhat differs, because, as mentioned in
`Chapt. 2, for a ten~fold difference in salt solubility, the PHmax differs by one
`umt only. Thus, from the theoretical analysis of pH-solubility relationships
`in Fig. 6, it may be concluded that the salt formation of a base with the pKa
`value of 8.0 might be feasible with most commonly used acid~ when S0 val-
`
`50
`
`40
`
`=' E
`0) 30
`.s
`~ ~ = 0.0001
`:a 20
`::,
`0
`Cl).
`
`10
`
`0
`0
`
`2
`
`4
`
`pH
`
`6
`
`8
`
`10
`
`Fig. 6. Theoretical pH-solubility ptofil.es dem01Mtrating the effect of intrinsic solubility (S0)
`ranging from 0.0001 to JO mg/ml of a basic dn,g with the pKQ = 8.0
`
`Merck Exhibit 2042, Page 11
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`144
`
`PHARMACEUTICAL SALTS:
`
`ues a.e ca. 0.01 mg/ml and higher. This is because pHmax values in these cas'"
`es would be around 5 and higher. For S0 values of 0.001 and 0.0001 mg/ml,
`however, relatively stronger acids would be required to form salts because
`pHma:x values would be ca. 4 and ca. 3, respectively.
`The theoretical analysis in Fig. 6 will change if the pKa value of a basic
`drug is lower. As shown in Chapt. 2, there is a direct relationship between
`pKa and pHmax; the. JlHmax decreases by 1 for each unit decrease in the pKa
`value. Thus, if the µKa value in Fig. 6 would be 4.0 instead of 8.0, the pHmax
`would be 3 and lower for S0 values of l mg/ml and lower. In such a situa(cid:173)
`tion, the possibility of salt formation becomes limited, because only relative(cid:173)
`ly stronger acids like HCl, methanesulfonic acid, ethanesulfonic acid, etc.,
`can lower the pH of saturated solutions below 3. The salt formation may not
`at all be feasible if the S0 is below 0.01 mg/ml because the Pfl.nax in this case
`would be less than l.
`A confumation of the validity of above theoretical analysis may be ob(cid:173)
`tained from the work of La.kkaraju et al. [20], where the authors studied
`pH-solubility relationships of two structurally similar compounds, avitriptan
`and BMS-181885 (Fig. 7). The compounds were dibasic in nature, each of
`them with pKa values of 8.0 and 3.6. However, the S0 values of the com(cid:173)
`pounds differed; they were 0.006 and 0.0007 mg/ml for avitriptan and BMS-
`181885, respectively. Because of this difference in S0 values, the pHmax val(cid:173)
`ues, due to the effect of the higher pKa (8.0), were ca. 5 for avitriptan and
`ca. 4 for BMS-181885. As a consequence, salts with many different counter(cid:173)
`ions, including acetate, lactate, succinate, and tartrate, could be synthesized
`for avitriptan. But, with BMS-181885, it was not possible to lower the pH of
`a saturated solution below 4 by using acetic acid, lactic acid, succinic acid,
`
`0 HCJJ:O
`
`3
`
`0 ,, ,,,o
`HN-S,.,
`H3C
`
`✓
`
`~NH
`
`H
`
`CH3
`/ ' \ Qt=\
`N N~ 1;N
`N_!J
`\___J
`
`Avitriptan
`pKa,1 = 8.0; pK.,2 = 3.6
`So= 0.006 mg/ml
`
`BMS-181885
`pKa.1 = 8.0; PKa.2= 3.6
`So= 0.0007 mg/ml
`
`Fig. 7. Chemical structures of avitripran and BMS-181885
`
`Merck Exhibit 2042, Page 12
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`;lCALSALTS:
`
`IL
`,::;'th these cas-
`
`between
`thepKa
`epHmax
`'a situa(cid:173)
`reJative-
`
`PROPERTIES, SELECTION, AND USE
`
`145
`
`or tartaric acid, and, therefore, the salt formation with any of these counter(cid:173)
`ions was not feasible for this compound. However, BMS-181885 formed
`salts with stronger acids like H3PO4 and HCL
`lt should be noted here that self-association of drug molecules in solu(cid:173)
`tions may sometimes lead to deviations in pH-solubility profiles predicted
`from p.K,. and solubility. Nevertheless, the theoretical modeling can still ser(cid:173)
`ve ~ a useful method of predicting the fea<;ibility of salt formation because
`a pHmax value may be estimated within a reasonable range. Also, the self-as(cid:173)
`sociation often shifts pHmax in favor of salt formation,
`
`2.4. Feasibility of Disalt Formation
`
`Eqns. 1-4 are applicable to compounds with only one pK,. value, and,
`therefore, the discus.~ion in this chapter has so far focused primarily around
`the feasjbility of salt formation for bases with one protonatable moiety and
`acids with one ionizable species. Such compounds can form only mono-salts
`(e.g., mono-hydrochloride, mono-sodium, etc.). In addition, a compound may
`have both basic and acidic moieties. Such a compound can also be classified
`as one forming a mono-salt, because only one of these groups can be used at
`one time for ,i;alt formation. In contra<;t, drugs can also be polybasic or poly(cid:173)
`protic, which might.be able to form poly-salts. Examples of disalts; such as
`dihydrochloride, disodium, etc., are common in the literature. Some of the
`questions that arise for compounds with multiple basic moieties or multiple
`acidic moieties are: Should mono- or poly-salt be synthesized for such com(cid:173)
`pounds? Is the formation of poly-salt feasible? If the synthesis of both forms
`of salts is feasible, which one is preferred for a particular drug candidate?
`Some of these issues are addressed below.
`
`2.4.1. Feasibility of Salt Formation for Dibasic Compounds
`
`Serajuddin and co-workers [20} [21] have demonstrated that the feasibil(cid:173)
`ity of salt formation for a dibasic compound can also be predicted from its
`pH-solubility relationship. As illustrated schematically in Fig. 8, the solubil(cid:173)
`ity of a free base increases with a decrease in pH, and, after the first pHi11ax
`(or p8=.x 1) is reached, a mono-salt might be formed. The solubility of the
`mono-salt formed then increases because of the protonation of the second ba-
`. sic moiety, thus reaching pHmax, 2 . Below pHmax. 2, a disalt could be formed.
`Depending on pH and counter-ions used to prepare salts, there could be three
`distinct solid phases (free base, mono-salt, and disalt) in equilibrium with
`aqueous solutions. The equations corresponding to solubilities of these three
`
`Merck Exhibit 2042, Page 13
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`146
`
`PHARMACEUTICAL SALTS:
`
`S,.= [BH2 ... ] 5 + [B+]
`= [BH2++], (1 + K0,:/[Hp+])
`
`Solid Phase: Disalt
`
`I
`! Solid Phase: i
`! Mono-Salt I
`pH
`
`S,.= [BH+J+ [BJ.
`=[BJ. (1 + [Hp+VK..,1)
`
`Solid Phase: Base
`
`Fig. 8. A schematic representation of the pH-solubility profile of a dibasic compound, show(cid:173)
`ing that solubilities of the brise and its mono- and disalt forms can be expressed by three in(cid:173)
`dependent curves corresponding to three independent equations. The profile indicates that two
`pHmax values, pHmax, 1 and pHmux, 2, may exist for such a compound.
`
`phases are given below:
`
`At pH< pH max. 2:
`
`At pH> pHmax, 1:
`
`ST = [BH'] + [B ]8
`= [BJ.· (1 + [H3O+)/Ka. 1)
`At pH< pHmax, 1 and> pHmax, 2:
`ST= [BH2] + [BH1s
`= [BH+Js · (l + [H3O+]/Ka, 2)
`ST= [BH2Js + [B+)
`= [BH:zls · (1 + Ka.. z/[H3O+])
`For the sake of simplicity, no consideration of the common-ion effect and the
`solubility product was made in deriving the above equations. It should also
`be mentioned here that distinct regions in the pH-solubility profile corre(cid:173)
`sponding to mono- and disalt forms may not be obtained if pKa and/or P8mu
`values of the compound are not far apart (ca. 2 units). If two pHmax values,
`are indistinguishable, only the disalt may be isolated in pure form.
`Avitriptan (Fig. 7), a dibasic compound, was used as the test compound
`for salt formation. As shown in Fig. 9, protonation of the piperazine N-atom · "
`and the pyrimidine N-atom was responsible for pKa values of 8.0 and 3.6, re0
`spectively, for the compound. By using HCI to adjust pH, it was established
`that the compound could have two pHmax values, one at pH 5 (P~ax. 1) arrd
`the other at pH ca. 2 (pHmax, 2). This is shown in Fig. JO. It is evident frorn
`
`Merck Exhibit 2042, Page 14
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`"'TICAL SALTS:
`
`PRO)"ERTIFS, SELECTION, AND USE
`
`147
`
`0 ~.,,o
`HS(N-S%"
`H3C
`
`P~.1 =8.0
`
`H
`
`'
`
`CH,
`\ { \ o'h
`N---{
`),l
`N
`\____j NJ
`
`0
`. \\ ,::;O
`
`H3C
`H~N-S
`
`/.K"a.2 = 3.6
`
`H
`
`,CH,
`0
`
`-
`,--------,
`+
`•
`'
`'
`N N~N-H
`\__/ NJ
`
`Fig. 9. Protonation of avitripum cone.sponding to its two pK0 values
`
`35 .----------on.As;:::::======::;-,
`
`30 -t 25
`
`I
`.§. 20 Q
`'?
`£ 15
`:0
`:::J 0 10
`
`(j)
`
`I
`I
`
`5
`
`0
`
`1
`
`~0.6
`
`:0 "§ 0.4
`
`Cl)
`H
`p max,i
`
`0.2
`
`6
`
`7
`
`8
`pH
`
`9
`
`,o
`
`HCI
`pl(,,= -6.1
`
`3
`
`5
`
`pH
`
`7
`
`9
`
`11
`
`Fig. 10. pH-Solubility profile of avitriptan at 25 °C where HCI was used to adjust pH, indi(cid:173)
`cating two pH,,,a:r: values. Solubility profile at pH above 5 is shown in the inset.
`
`Fig. JO that both mono- and dihydroch1oride salts can possibly be prepared
`for avitriptan; the monohydroehloride sa]t would be the equilibrium species
`at pH between 2 and 5, and the dihydrochloride salt would be the equilibri(cid:173)
`um species at pH below 2. Among variol!s aci<is used by Lakkaraju et al. [20]
`to form salts with avitriptan, only HCl could lower the pH of a s~turated av-
`
`Merck Exhibit 2042, Page 15
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`148
`
`PHARMACEUTICAL SALTS:
`
`80
`
`~
`
`E 60
`..._
`.s
`0)
`>, 40
`:t::
`:0
`:J
`0
`Cl) 20
`
`0
`1
`
`0
`I
`0
`I
`0
`Q
`9
`c)
`6
`Cb_ --o ___
`
`\
`
`COOH
`I
`HC-OH
`I
`HG-OH
`I
`COOH
`
`pK~., = 3.0
`PK,.,2= 4.4
`
`3
`
`5
`
`pH
`
`7
`
`9
`
`11
`
`Fig. 11. pH-Solubility profile of avitriptan at 25 °C where tartaric acid was used to adjust pH,
`indicating the pl"esence of only one pHmux value
`
`itriptan solution below 2 and thus could form a disalt. The pH of an avitrip(cid:173)
`tan solution could be lowered below 5 (pHmax, 1) but not below 2 (pHmax. 2),
`when methanesulfonic acid, acetic acid, lactic acid, tartaric acid, and succin(cid:173)
`ic acid were used, indicating that these acids would form only mono-salts
`with avitriptan. As a typical example, the pH-solubility profile of avitriptan
`in presence of tartaric acid is shown in Fig. 11, where the pH could not be
`lowered below 2.5 by adding excess amount of tartaric acid. In agreement
`with these pH-solubility considerations, the salt-selection program of avitrip(cid:173)
`tan yielded mono-salts for all counter-ions used except for hydrochloride, al(cid:173)
`though the existence of two bai;ic moieties in the molecule intuitively sug(cid:173)
`gested that attempts for the synthesis of disalts using various counter-ions
`should be made. Thus, conducting a feasibility analysis based on pH-solu(cid:173)
`bility relationships can save considerable time and efforts in a salt-synthesis
`program.
`
`2.4.2. Feasibility of Salt Formation for Diprotic Acids
`
`Equations analogous to Eqns. 5 - 7 above can also be derived for acids
`with two ionizable groups in order to study the feasibility of mono- or disalt
`formation. For such a compound, the solubility of the acid initially increas(cid:173)
`es with an increase in pH due to the ionization of the first ionizable group
`(i.e., the stronger ionizable group with lower pKa value). At a ce1tain pH, the
`firs