international
`journal of
`pharmaceutics
`
`volume 105 no. 3 (9 May 1994)
`completing volume 105
`
`\
`
`'
`
`'
`
`\:•
`
`i'
`
`
`
`Apotex Exhibit 1014.001
`
`

`

`© 1994, Elsevier Science B.V. All rights reserved
`
`ISSN 0378-5173/1994/$7.00
`No part of this publication may be reproduced, stored in a retrieval system, or transinitted in any form or by any means, electronic,
`mechanical, photocopying, recording or otherwise, without the prior written pyrmission of the Publisher, Elsevier Science B.V.,
`Copyright + Permissions Department P.O. Box 521, 1000 AM Amsterdam, The Netherlands.
`No responsibility is assumed by the Publisher for any injury aii.d/or damage to persons or property as a matter of products liability,
`negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material
`herein. Because of the rapid advances in the medical sciences, the Publisher recommends that independent verification of
`diagnoses and drug dosages should be made.
`Although all advertising material is expected to conform to ethical (medical) standards, inclusion in this publication does not
`constitute a guarantee or endorsement of the quality or value of such product or of the claims made of it by its manufacturer.
`Special regulations for authors. Upon accepta11ce of an article by the journal, the author(s) will be asked to transfer copyright of the
`article to the Publisher. This transfer will ensure the widest possible dissemination of information.
`Special regulations for readers in the U.SA. This journal has been registered with the Copyright Clearance Center, Inc. Consent is
`given for copying of articles for personal or internal use, or for the personal use of specific clients. This consent is given on the
`condition that the copier pays through the Center the per-copy fee stated in the code on the first page of each article for copying
`beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. The appropriate fee should be forwarded with a copy of
`the first page of the article to the Copyright Clearance Center, Inc., 27 Congress Street, Salem, MA 01970, U.S.A. If no code
`appears in an article the author has .not given broad consent to copy and permission to copy must be obtained directly from the
`author. The fee indicated on the first page of an article in this issue will apply retroactively to all articles published in this journal,
`regardless of the year of publication. This consent does not extend to other kinds of copying such as for general distribution, resale,
`advertising and promotion purposes, or for creating new collective works. Special written permission must be obtained from the
`Publisher for such copying.
`
`Elsevier Science B.V.
`Fax ( +31-20)5803258. Tel. ( +31-20)5803422
`
`Postal Address:
`
`Courier Service Address:
`
`International Journal of Pharmaceutics
`Elsevier Science B.V.
`P.O. Box 1527
`1000 BM Amsterdam
`The Netherlands
`
`International Journal of Pharmaceutics
`Elsevier Science B.V.
`Molenwerf 1
`1014 AG Amsterdam
`The Netherlands
`
`This journal is in the ADONIS Service, whereby copies of individual articles can be printed out from CD-ROM on request. An
`explanatory leaflet can be obtained by writing to ADONIS B.V., P.O. Box 839, 1000 AV Amsterdam, The Netherlands.
`
`Printed in the Netherlands
`
`This journal is printed on acid-free paper.
`
`Apotex Exhibit 1014.002
`
`

`

`ELSEVIER
`
`International Journal of Pharmaceutics 105 (1994) 209-217
`
`international
`journal of
`pharmaceutics
`
`Research Papers
`An integrated approach to the selection of optimal salt form
`for a new drug candidate
`
`Kenneth R. Morris, Michael G. Fakes, Ajit B. Thakur, Ann W. Newman, Ambarish K.
`Singh, John J. Venit, Ciro J. Spagnuolo, Abu T.M. Serajuddin *
`Pharmaceutical Development Division, Bristol-Myers Squibb Pharmaceutical Research Institute, New Brunswick, NJ 08903, USA
`
`(Received 25 September 1992; Modified version received 13 October 1993; Accepted 20 October 1993)
`
`Abstract
`
`A general method was developed to select the optimal salt form for BMS-180431, a novel HMG-CoA reductase
`inhibitor and a candidate for oral dosage form development, in an expeditious manner at the onset of the drug
`development process. The physicochemical properties such as hygroscopicity, physical stability of crystal forms at
`different humidity conditions, aqueous solubility, and chemical stability of seven salts, e.g., sodium, potassium,
`calcium, zinc,. magnesium, arginine and lysine, were studied using a multi~tier approach. The progression of studies
`among different tiers was such that the least time-consuming experiments were conducted earlier, thus saving time
`and effort. A ·'go/no go' decision was made after each tier of testing the salts, thus avoiding generation of extensive
`data on all available salt forms. The hygroscopicities of all BMS-180431 salts were evaluated at tier 1 and four salts
`(sodium, potassium, calcium and zinc) were dropped from consideration due to excessive moisture uptake within the
`expected humidity range of pharmaceutical manufacturing plants (30-50% R.H. at ambient temperature). The
`remaining three salts were subjected to the tier 2 evaluation for any change in their crystal structures with respect to
`humidity and the determination of their aqueous solubilities in the gastrointestinal pH range. The magnesium salt
`was dropped from further consideration due to humidity-dependent changes in its crystal structure and low solubility
`in water (3.7 mg/ml at room temperature). Arginine and lysine salts, which were resistant to any change in their
`crystalline structui~s under extremes of humidity conditions (6 and 75% R.H.) and had high aqueous solubilities
`( > 200 mg/ ml), were elevated to tier 3 for the determination of their chemical stability. Based on solid state stability
`of these two salts under accelerated conditions (temperature, humidity, an.d presence of excipients), consideration of
`ease of synthesis, ease of analysis, potential impurities, etc., and input from the marketing group with respect to its
`preference of counter i.on species, the arginine salt was selected for further development. The number of tiers
`necessary to reach a decision on the optimal salt form of a compound may depend on the physicochemical properties
`studied and the number of salts available. This salt selection process can be completed within 4-6 weeks and be
`easily adopted in the drug development program.
`
`Key words: Salt selection; HMG-CoA reductase inhibitor; BMS-180431; Hygroscopicity; Arginine salt; Lysine salt
`
`* Corresponding author.
`
`(
`0378-5173/94/$07.00 © 1994 Elsevier Science B.V. All rights reserved
`SSDI 0378-5173(93)E0327-M
`
`Apotex Exhibit 1014.003
`
`

`

`210
`
`1. Introduction
`
`KR. Morris et al./ International Journal of Pharmaceutics 105 (1994) 209-217
`,!
`
`Berge et al. (1977) reviewed various advan(cid:173)
`tages of using salt forms of drugs in pharmaceuti(cid:173)
`cal formulations, which include improved dissolu(cid:173)
`tion rate and bioavailability of poorly water-solu(cid:173)
`ble compounds. For some drugs, preparation of
`stable salts may not be feasible, and free acid or
`base forms may be preferred (Serajuddin et al.,
`1986). In selecting the optimum chemical form of
`a new drug candidate, one must, therefore, take
`into consideration all physicochemical properties
`which would influence its physical and chemical
`stability, processability under manufacturing con(cid:173)
`ditions, dissolution rate, and bioavailability. Such
`a selection of chemical form must be done at the
`initial stage of drug development. Changing the
`chemical form in the middle of a developmental
`program may require repeating most of the bio(cid:173)
`logical, toxicological, formulation, and stability
`tests performed. On the other hand, continuing
`the development of a nonoptimal chemical form
`may lead to increased developmental and produc(cid:173)
`tion costs and even product failure.
`Although the importance of using the optimal
`salt form of a compound in dosage form design is
`well-recognized (Berge et al., 1977; Hirsch et al.,
`1978), there is no generally accepted procedure
`of selecting such a form during the drug develop(cid:173)
`ment process. More often than not, the medicinal
`chemists select salt forms on a practical. basis,
`such as previous experience with the salt type,
`ease of synthesis, percent yield, etc. (Berge et al.,
`1977). It is, therefore, desirable that a procedure
`be developed for the selectfon of salt or other
`chemical form of a drug candidate expeditiously
`at the outset of the developmental program. We
`have developed an integrated approach which
`was successfully applied to the selectfon of the
`optimal salt forms of several compounds. Its ap(cid:173)
`plication in the selection of the salt form of
`BMS-180431, a new HMG-CoA reductase in(cid:173)
`hibitor which is a candidate for the development
`as a solid dosage form, is d~scribed in this paper.
`
`2. Development of salt selection strategy
`
`Gould (1986) described a salt selection process
`based on melting point, solubility, stability, wetta(cid:173)
`bility, etc., of various salt forms. However, in the
`absence of clear go/ no go decisions at any par(cid:173)
`ticular stage of the salt selection process, this
`approach would lead to the< generation of exten(cid:173)
`sive physicochemical data on, all salt forms syn(cid:173)
`thesized. Gould concluded that "the balance re(cid:173)
`quired in assessing the correct salt form to
`progress into drug development makes it a diffi(cid:173)
`cult semiempirical exercise." A more rational ap(cid:173)
`proach is, therefore, required to select the appro(cid:173)
`priate salt form expeditiously during drug devel(cid:173)
`opment. In the present method the physicochemi(cid:173)
`cal tests were conducted at different tiers and a
`go/ no go decision was made after each tier of
`testing the salts, thus avoiding generation of ex(cid:173)
`tensive data on each salt form synthesized. The
`studies were planned such that the least time(cid:173)
`consuming experiments which could still give a
`go/ no go decision were conducted at tier 1.
`Progressively more time-consuming and labor-in(cid:173)
`tensive experiments were conducted at tier 2, tier
`3, etc. In this way, many different salt forms
`could be screened with the minimum of experi(cid:173)
`mental effort.
`Based on the review of literature (Berge et al.,
`1977; Hirsch et el., 1978; Gould, 1986; Serajuddin
`et al., 1986) and our experience in product devel(cid:173)
`opment, we identified low hygroscopicity, in(cid:173)
`tegrity of crystal form at different storage condi(cid:173)
`tions, aqueous solubility, and chemical stability as
`primary criteria for the selection of BMS-180431
`salts, and set limits for the acceptability of these
`criteria. All salt forms of the compound which
`were found to be crystalline were tested at tier 1
`for their hygroscopicity. A high degree of mois(cid:173)
`ture sorption or desorption by the salts under
`expected ambient humidity conditions of pharma(cid:173)
`ceutical manufacturing plants may create han(cid:173)
`dling and manufacturing difficulties, such as
`change in potency of the drug substance, change
`
`Apotex Exhibit 1014.004
`
`

`

`K.R. Morris et al./ International Journal of Pharmaceutics 105 (1994) 209-217
`
`211
`
`in the true density, tariation in flow behavior, etc.
`There may be batch-to-batch variability in the
`potency of dosage forms if care is not taken to
`ensure that the bulk drug substance maintained
`its declared potency prior to batching. The change
`in moisture content may also affect the physical
`and chemical stability of salts. Therefore, at the
`end of tier 1, all salt forms with excessive mois(cid:173)
`ture sorption/ desorption behavior were dropped
`from further consideration.
`The salts which were considered to have ac(cid:173)
`ceptable hygroscopicity were then screened in
`tier 2 for changes in crystal structure under exc
`tremes of humidity conditions by using combina(cid:173)
`fions of powder X-ray diffraction and thermal
`analysis techniques. This would indicate any
`propensity for pseudopolymorphic and solution(cid:173)
`mediated polymorphic changes which might occur
`during manufacturing or accelerated stability test(cid:173)
`ing of the bulk material or the solid dosage form.
`At this stage, the salts were also screened for
`their aqueous solubilities to determine if there is
`any potential dissolution and bioavailability prob(cid:173)
`lems and whether the formulation of a solution
`dosage form, if required, is feasible. The go/ no
`go decision would depend on the consideration of
`both the physical stability of crystalline structure
`at different humidity conditions as well as the
`solubility. The criteria for the selection of salts at
`tier 2 may depend on the judgment of the drug
`development scientists in consideration of the
`type of dosage form and the expected dose of the
`compound. A salt with lower solubility which can
`· still provide good dissolution rate in the judgment
`of a formulation scientist could be selected over a
`salt which is highly soluble but prone to crys(cid:173)
`talline changes. On the other hand, if the solubil(cid:173)
`ity is not acceptable .in consideration of the disso(cid:173)
`lution rate or if a: solution with high drug concen(cid:173)
`tration· is required for oral or parenteral use,
`another salt with some propensity for changes in
`crystal properties under extremes of humidity may
`be considered.
`Finally, at tier 3, the selected salts )Vere sub(cid:173)
`jected to accelerated thermal stability and photo(cid:173)
`stability screening. Since the stability testing of
`salts required much time and effort, placing this
`at tier 3 limited the number of salts on which
`)
`
`such tests were conducted and avoided genera(cid:173)
`tion of unnecessary data with other salt forms.
`Compatibility screening with selected excipients
`may also be conducted at this stage.
`In the above scheme, the number of salt forms
`available and the physicochemical properties con(cid:173)
`sidered important for the bulk drug substance as
`well for the expected dosage forms will dictate
`how many tiers would be necessary to select a salt
`form. There may also be rare situations where all
`salts progressed from a lower tier to a higher one
`ai:'e unacceptable for development. For example,
`the solubility of all salts at tier 2 may be unac(cid:173)
`ceptable or chemical stability of all the salts at
`tier 3 may be poor. If this happens, additional salt
`forms or free acids/bases should be considered
`prior to reevaluating any salt that was dropped at
`an earlier tier. Also, the criteria of progression
`from a lower tier to the next higher one may
`depend on the physicochemical properties of the
`available salts. If, for example, all salts are found
`to be highly hygroscopic, it might be necessary to
`progress some of them to a higher tier, keeping in
`mind that, if selected, they might require special
`manufacturing and storage conditions.
`F
`
`OH
`
`"'""'• COOR
`OH
`
`. I
`
`·~
`
`F
`
`3. Experimental
`
`3.1. Mat~ripJs
`
`The following salts of BMS-180431 were used
`during salt selection: sodium, potassium, calcium,
`zinc, magnesium, arginine, and lysine. A few other
`salts were also prepared; however, they were
`found to be noncrystalline and, therefore, not
`considered for salt selection.
`
`Apotex Exhibit 1014.005
`
`

`

`212
`
`K.R. Morris et al./ International Journal of Pharmaceutics 105 (1994) 209-217
`
`Computer
`
`Fig. 1. Schematic diagram of the moisture sorption-desorption
`apparatus.
`
`3.2. Moisture sorption studies
`
`The rate and extent of moisture sorption at
`different humidity conditions were determined by
`using a Cahn Digital Recording Balance fitted
`with a system to maintain and monitor specific
`relative humidity conditions. The system consists
`of a tubular glass chamber surrounding the pan
`of the balance, a flask for the humidity-control(cid:173)
`ling salt solution, a peristaltic pump to circulate
`the saturated air, and an in-line chilled mirror
`dew point meter for the determination of relative
`humidity within the system. Saturated aqueous
`solutions of MgCl 2 , Mg(N0 3) 2 and NaCl were
`used at room temperature to maintain 33, 52 and
`75% R.H. conditions, respectively. Anhydrous
`CaS04 (Drierite®, Hammond, OH) provided the
`6% relative humidity condition. Once assembled,
`a closed system is formed wherein the desired
`relative humidity condition can be readily at(cid:173)
`tained. The sample can be placed on the pan of
`the balance through the access port with practi(cid:173)
`cally no perturbation of the humidity inside the
`chamber. Once a sample ( ~ 10 mg) reaches equi(cid:173)
`librium at a particular relative humidity, the envi(cid:173)
`ronment surrounding the sample ma:y be changed
`to a different humidity condition by changing the
`flask containing the salt solution. The atmo-
`
`sphere within the glass chamber attains the new
`humidity condition within 5 min of such a change,
`and the sample then reaches equilibrium with
`this new atmospheric condition. Highly repro(cid:173)
`ducible moisture sorption data were obtained by
`this method. In a separate study, when multiple
`samples (15 mg each) from the same batch of a
`drug substance were equilibrated at different hu(cid:173)
`midity conditions, the final weights at each hu(cid:173)
`midity condition were within ± 0.02 mg. This
`indicated that for a 2% moisture uptake, the
`precision of the experiment would be ± 6.6%,
`while for a 10% moisture uptake it would be
`within ± 1.5%. 1 Also, when the equilibrium mois(cid:173)
`ture contents of a sample measured at different
`humidity conditions are compared, the influence
`of any artifacts such as the adsorption of mois(cid:173)
`ture to the sample pan is also minimal; for exam(cid:173)
`ple, in a moisture uptake run without any sample
`on the pan, the weight gain between 6 and 75%
`RH was < 0.01 mg. A schematic diagram of the
`moisture sorption/ desorption apparatus is given
`in Fig. 1.
`
`3.3. Determination of moisture content
`
`The moisture contents of samples exposed to
`different relative humidity conditions were deter(cid:173)
`mined primarily by thermal gravimetric analysis
`(TGA) using a DuPont 2000 Thermal Analyst
`system .. However, since the weight loss in the
`TGA represents both water and any other volatile
`material, initial moisture contents of all samples
`as well as moisture contents of specific samples
`after equilibration at different humidity condi(cid:173)
`tions were confirmed by coulometric titration us(cid:173)
`ing a Brinkman 684 Karl-Fischer Coulometer.
`The TGA was performed by weighing accurately
`about 10 mg of sample on an open sample pan
`and then heating the sample at a rate of l0°C per
`min. During the experiment, dry nitrogen was
`purged over the sample at a rate of 40 cm3 per
`min. For coulometric analysis, sample sizes were
`selected to yield 1-3 mg of water. The accuracy
`of the system was ensured by titrating 2 mg of
`water (10 µ,I of a 20% w /v solution of water in
`methanol) where the results varied between 97.5
`and 102 % of theoretical.
`
`Apotex Exhibit 1014.006
`
`

`

`KR. Morris et al./ International Journal of Pharmaceutics 105 (1994) 209-217
`
`213
`
`3.4. Differential scanning calorimetric (DSC) anal(cid:173)
`ysis
`
`The energetics of the moisture-solid interac(cid:173)
`tion in samples, e.g.,. hydrate formation, was de(cid:173)
`termined by DSC analysis. A sample sealed in an
`aluminum pan with a pin hole was heated at a
`rate of 5 or 10°C per min from room temperature
`to about 200°C using a DuPont 2000 Thermal
`Analyst system.
`
`3.5. Powder X-ray diffraction (XRD) analysis
`
`Powder X-ray diffraction patterns of the sam(cid:173)
`ples were collected using a Phillips APD 3720
`powder diffraction system with a vertical go(cid:173)
`niometer in the 8 /28 geometry. The X~ray gen(cid:173)
`erator (model XRG 3100) was operated at 45 kV
`with a copper radiation source. A scintillation
`detector was used to scan the range between 2
`and 32° 28. A sample was packed in a 1.5 cm X 1.0
`cm sample holder with a thickness of 2 mm and
`its initial powder pattern was determined. The
`sample holder was then stored overnight in a
`desiccator with a particular relative humidity, and
`the powder X-ray diffraction analysis was then
`repeated. The top/surface of the powder bed was
`fully exposed to the atmosphere, and, since the
`thickness of the powder bed was small, there was
`no barrier to the diffusion of moisture into the
`bed. By this procedure, it is also possible to
`determine the change in crystal structure when a
`sample equilibrated. at one relative humidity is
`reequilibrated at a different humidity condition.
`Since the samples were not disturbed after the
`initial packing, the parameters such as size and
`orientation of particles remained unchanged.
`
`3. 6. Determination of solubility
`
`The solubilities of various salts under simu(cid:173)
`lated gastric and intestinal pH conditions were
`determined by equilibrating excess of solid mate(cid:173)
`rial with 0.01 M HCl and water at room tempera(cid:173)
`ture using a Burrell® wrist action shaker. The
`solids were equilibrated 'with water for 24 h;
`however, since the preliminary studies showed
`that the compound is relatively unstable in· acidic
`
`media due to the formation of its lactone form
`( ~ 5% degradation at pH 2 in 6 h at room
`temperature) shaking with 0.01 M HCl was per(cid:173)
`formed for 2 h only. The pH values of the solu(cid:173)
`tions were recorded prior to their filtration
`through 0.45 µ,m pore size Millipore filters. The
`solutions were assayed for drug concentrations
`using high-pressure
`liquid chromatography
`(HPLC).
`
`3. 7. Determination of solid-state stability
`
`Accurately weighed samples of salts ( ~ 10 mg
`each) were stored at 40 and 50°C in closed 4 cm3
`glass vials and at 40°C/75% R.H. in open glass
`vials. For photostability studies, samples stored in
`closed clear glass vials were exposed to 900 foot(cid:173)
`candle fluorescent light; the vials stored under a
`similar condition with aluminum foil wrappers
`around them served as controls. The stability
`samples were assayed at different intervals by
`HPLC.
`
`3.8. Screening of drug-excipient compatibility
`
`Drug-excipient compatibility screening studies
`were performed on arginine and lysine salts of
`BMS-180431 using a procedure reported earlier
`(Serajuddin et al., 1991). Ternary or quaternary
`mixtures of the drug and excipients (drug: excipi(cid:173)
`ent, 1 : 7) were weighed into 4 cm3 glass vials,
`approx. 20% water was added to the contents of
`vials (70 µ,l water to 320 mg powder in a vial), the
`vials were sealed tightly, and stored at 50°C.
`Samples stored at room temperature without
`added water served as controls. Each drug-excipic
`ent mixture consisted of 40 mg of drug, 250 mg of
`a diluent (tricalcium phosphate, mannitol or mi(cid:173)
`crocrystalline cellulose), and 30 mg of a third
`component which was a disintegrant, binder or
`lubricant. Duplicate samples of each mixture were
`withdrawn, examined visually for any physical
`change, and analyzed chemically using HPLC.
`
`3.9. HPLC analysis
`
`The samples were dissolved in a 40 : 60 v /v
`mixture of acetonitrile and water, and analyzed
`by HPLC using a 4.6 mm x 250 mm reversed
`
`Apotex Exhibit 1014.007
`
`

`

`... --
`
`214
`
`KR. Morris et al./ International Journal of Pharmaceutics 105 (1994) 209-217
`
`CD
`
`20
`18
`~ 16 m-
`a; a 14
`a: a; 12
`l!!:i=
`.a~ 10
`.!!!C 8
`Oo
`::!E-
`
`6
`
`(a)
`
`33% RH
`
`52% RH
`
`Potassium
`Sodium
`
`Calcium
`75% RH
`
`- - - - -Z inc
`
`+-Change in RH
`
`4
`
`2 0
`
`2
`
`3
`Time, hours
`
`4
`
`5
`
`6
`
`I -
`
`18
`16
`CD
`~ 14
`~~ 12
`a: a; 10
`I!!:;=
`=~ 8
`.!!!C 6
`Oo
`::!E-
`4
`'#.
`
`2
`
`(b)
`
`33% RH
`
`52% RH
`
`75% RH
`
`Magnesium
`
`- - - - - -Lys ine
`
`i=:==:::::::;;:::::=--<--- Arginine
`t
`'Change In RH
`Change In RH
`2
`
`4
`
`0 0
`
`Table 1
`Moisture contents of various BMS-180431 salts initially and
`after equilibration at different humidity conditions
`% moisture relative to dry weight
`Salt
`Initial 33%R.H. 52% R.H.
`4.4
`4.4
`Sodium
`16.0
`Potassium
`9.5
`10.6
`17.1
`Calcium
`5.0
`10.4
`13.4
`4.1
`2.9
`Zinc
`6.5
`Magnesium 13.8
`13.5
`14.0
`Arginine
`2.8
`3.0
`3.2
`Lysine
`3.2
`3.6
`4.0
`
`75% R.H.
`ND
`ND
`16.1
`8.3
`14.9
`3.8
`6.0
`
`ND, not determined.
`
`R.H., as determined by recording weight gain or
`loss using a Cahn balance, are shown in Fig. 2.
`The initial moisture content varied from salt to
`salt. Also, when multiple lots of the same salt
`were received, the initial moisture content varied
`depending on the extent of drying of the samples
`(not shown in figure). As shown in Fig. 2, the
`moisture sorption rates were very rapid; the sam(cid:173)
`ples reached equilibrium in < 10 min when ex(cid:173)
`posed to 33, 52 or 75% R.H. Initial moisture
`contents as well as equilibrium moisture contents
`of various salts under these three different hu(cid:173)
`midity conditions are tabulated in Table 1.
`The energetics of the interaction of water with
`these salts were examined by DSC. The loss of
`moisture from the metal salts (sodium, potassium,
`
`5
`
`6
`
`3
`Time, hours
`Fig. 2. Moisture sorption by (a) potassium, sodium, calcium
`and zinc salts, and (b) magnesium, lysine and arginine salts of
`BMS-180431 at 33, 52 and 75% R.H. The initial moisture
`contents were determined separately by TGA. The % R.H.
`was changed to the next higher value after the sample reached
`equilibrium moisture content. at the lower % R.H. The typical
`changes in % R.H. are indicated by arrows.
`
`phase column (YMC-C18 column, S-5; YMC, Inc.,
`Morris Plains, NJ), a Waters autoinjector and an
`Applied Biosystems Spectroflow 783 detector.
`The mobile phase consisted of a · 40 : 15 : 45 v /v
`mixture of acetonitrile, methanol and a phospho(cid:173)
`ric acid solution (0.1 % H 3P0 4 in water), and had
`a final pH of 2.5. The flow rate of. the mobile
`phase was 1 ml/ min and the run time after each
`injection was 20 min. The sample volume was 20
`,ul, and the detection was by ultraviolet light
`absorbance at a wavelength of 296 nm.
`
`~-
`
`~ 0.3
`0 u:: - 0.1
`
`ra
`Q) :c
`
`4. Results and discussion
`
`4.1. Moisture sorption (tier I)
`
`Moisture sorption curves for various salts when
`exposed to 33% R.H., followed by 52 and 75%
`
`Temperature, °C
`
`Fig. 3. DSC scans of BMS-180431 calcium salt indicating its
`dehydration behavior. Scan A shows the sample as received
`for testing and scan B is for the sample equilibrated at 75%
`R.H.
`
`Apotex Exhibit 1014.008
`
`

`

`.. if.
`
`KR. Morris et al./ International Journal of Pharmaceutics 105 (1994) 209-217
`
`215
`
`calcium, zinc and magnesium) occurred in two
`stages, which 'i~ exemplified by the DSC scan of
`the calcium salt as shown in Fig. 3A and B where
`two dehydration endotherms at 59 and 97°C were
`observed. The ~ndotherm at 59°C disappeared
`when the calcium salt was equilibrated at 6%
`R.H., indicating a lower energy of salt-water asso(cid:173)
`ciation than the endotherm at 97°C which repre(cid:173)
`sents more tightly bound, possibly cation-associ(cid:173)
`ated, water. The presence of a dehydration en(cid:173)
`dotherm for a metal salt at a relatively low tem(cid:173)
`temperature
`perature (for example, 30-70°C
`range for the calcium salt in Fig. 3B) also sug(cid:173)
`gests the possibility that the loss of moisture may
`occur when heat is applied or developed during
`processing, e.g., -drying of bulk drug substance,
`compression of powders or granules in the tablet
`press, etc. In contrast, the DSC scans of the
`amino acid salts as received (Fig. 4) did not
`indicate the presence of any hydrate. The DSC
`scans remained essentially unchanged after the
`exposure of samples to humidity conditions be(cid:173)
`tween 33 and 75% R.H., indicating that there was
`also no hydrate formation during this humidity
`challenge test.
`Additionally, Table 1 shows that arginine and
`lysine salts have very low fluctuations in moisture
`content (0.4% or lower) within the range of 33-
`52% R.H, which is prevalent under most manu(cid:173)
`facturing conditions. Although the magnesium salt
`had high initial moisture content, it was relatively
`nonhygroscopic between 33 and 75% R.H. For
`
`0.5
`
`0.0
`
`-0.5
`
`-1.0
`
`-1.5
`
`-2.0
`
`.J2l
`3::
`"i
`0
`iI
`Ri
`Q)
`:I:
`
`r
`I
`/e'--.... v
`
`I _ _,,
`"
`
`2
`
`5
`-
`
`-
`
`25
`
`50
`
`75
`
`100 125 150 175. 206 225 250
`
`Temperature, °C
`Fig. 4. DSC scans of arginine (A) and lysine (B) salts of
`BMS-180431 as received for testing.
`
`75% RH
`
`0.0
`
`5.0
`
`10.0
`
`15.0
`
`Degrees, 28
`
`Fig. 5. Powder X-ray diffraction patterns of BMS-180431
`magnesium salt as received for testing and after equilibration
`at 6 and 75% R.H.
`
`other salts, the fluctuation in moisture content
`between 33 and 52% R.H. was > 3%. Based on
`these considerations, arginine, lysine and magne(cid:173)
`sium salts were selected for the next stage (tier 2)
`of evaluation.
`
`4.2. Crystal structure and solubility (tier 2)
`
`Arginine, lysine, and magnesium salts selected
`because of their low hygroscopicity were further
`evaluated for any change in crystal structure un(cid:173)
`der expected extremes of humidity conditions
`during storage or accelerated stability testing. This
`is important because changes in crystal form in
`the bulk material may affect the physical (and/ or
`chemical) stability and performance of solid
`dosage forms and must be controlled. For exam(cid:173)
`ple, Yamaoka et al. (1982) observed cracking of
`tablets when the drug substances present changed
`from an anhydrate to a hydrate under high(cid:173)
`humidity conditions.
`The powder X-ray diffraction patterns of argi(cid:173)
`.. nine andJ.y:sine salts remained unchanged within
`the humidity range of 6-75% R.H. However, the
`magnesium salt showed changes in its powder
`XRD patterns which could be attributed to a
`change in the crystal form (Fig. 5). The sample
`stored at 6% R.H. loses intensity in the higher
`angle peaks as well as the major reflections at 11°
`28, and gains intensity in the peak at 11.5° 28.
`
`Apotex Exhibit 1014.009
`
`

`

`216
`
`KR. Morris et al./ International Journal of Pharmaceutics 105 (1994) 209-217
`
`The 75% R.H. pattern completely loses the 11.5°
`28 reflection while maintaining the 11° 2fJ peak.
`As mentioned earlier, solubilities of the salts
`of BMS-180431 were also studied at this stage.
`While the two amino acid salts were freely solu(cid:173)
`ble in water(> 200 mg/ml as free acid content),
`the magnesium salt was only slightly soluble (3.7
`mg/ml as free acid content). The pH values of
`the saturated solutions of these salts in water
`were between 6.8 and 7.6. All three salts, how(cid:173)
`ever, exhibited comparable solubilities under
`acidic pH conditions due to the conversion of
`BMS-180431 to the free acid (pKa = 4.4); the
`solubilities at pH 2.5-2.9 varied between 0.61 and
`0.67 mg/ml. These results indicate that these
`salts wou1d have adequate solubilities for their
`complete dissolution of the expected dose of 25
`mg under gastrointestinal pH conditions if formu(cid:173)
`lated as solid dosage forms. However, the limited
`solubility of the magnesium salt in water could
`present difficulties in the preparation of drug
`solutions for oral and parenteral use. Due to the
`change in crystal form and the limited solubility
`of magnesium salt, it was dropped from further
`consideration.
`
`4.3. Solid-state stability and drug-excipient compat(cid:173)
`ibility (tier 3)
`
`The solid-state stability evaluation of arginine
`and lysine salts for 4 weeks at a temperature as
`high as 50°C and at 40°C/75% R.H. did not show
`any significant degradation. Both salts, however,
`showed some susceptibility to light; the potencies
`of the arginine salt after exposure to 900 foot(cid:173)
`candle light for 1 and 2 weeks were 94.7 and
`87.9%, respectively, while those of the lysine salt
`stored under the same condition were 97.4 and
`95.0%, respectively. This difference in the light
`stability of arginine and lysine salts may not be of
`any practical significance because both salts would
`require protection from light.
`A short-term compatibility screen with some
`excipients commonly used in solid dosage forms,
`namely, tricalcium phosphate, mannitol, micro(cid:173)
`crystalline cellulose, croscarmelose sodium,
`cross-linked polyvinylpyrrolidone, and magnesium
`stearate, was conducted to further evaluate the
`
`arginine and lysine salts. Lactose was not selected
`as one of the excipients due to expected incom(cid:173)
`patibility of amino acid counter ions with lactose.
`The stabilities of the salts in the presence of
`various excipients were comparable; the loss in
`potency after