`JOURNALOF.
`
`
`
`'flELSEVIERr”
`
`Lupin Ex. 1051 (Page 1 of 11)
`Lupin Ex. 1051 (Page 1 of 11)
`
`
`
`© 2002 Elsevier Science B.V. All rights reserved ’1
`
`This journal and the individual contributions contained in it are protected by the copyright of Elsevier Science B.V., and the "
`following terms and conditions apply to their use:
`
`Pho recopying
`Single pl~otozopies of single articles may be made for personal use as allowed by national copyright laws. Permission of the publisi~er
`and payment of a fee is required for all other photocopying, including multiple or systematic copyi~ag, copying for advertising or
`promotional purposes, resale, and all forms of document delivery, Special rates are available for educational institutions that wish
`to make photocopies for non-profit educational classroom use.
`
`Permissions may be sought directly from Elsevier Science via their homepage (http://www.elsevier.eom) by selecting "Customer
`support" and then "Permissions". Alternatively you can send an e-mail to: permissions@elsevier.co.uk, or fax to: (+44) 1[765
`853333.
`
`In the USA, users may clear permissions and make payments through the Co-pyright Clearance Center, Inc., 222 Rosewood Drive,
`Danvers, MA 01923, USA; phone: (+ 1) (978) 7508400, fax: (+ 1)(978) 7504744, and in the UK through the Copyright Licensing
`Agency Rapid Clearance Service (CLARCS), 90 Tottenhaln Court Road, London W1P 0LP, UK; phone: (+44) 207 631 5555; fax:
`(+ 44) 207 631 5500. Other countries may have a local reprographic rights agency for payments.
`
`Derivative Works
`Subscribers may reproduce tables of contents or prepare lists of articles including abstracts for internal circulation within their
`institutions. Permission of the publisher is required for resale or distribution outside the institution.
`
`Permission of the publisher is required for all other derivative works, including compilations and translations.
`
`Electronic Storage or Usage
`Permission of the Publisher is required to store or use electronically any material contained in this journal, including any article or
`part of an article.
`
`Except as outlined above, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any fc rm
`or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the publisher.
`
`Address permissions requests to: Elsevier Science Global Rights Department, at the mail, fax and e-mail addresses noted above.
`
`Notice
`No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability,
`negligence or otherwise, or fl’om any use or operation of any methods, products, instructions or ideas contained in the mate’ial
`herein. Because of rapid advances in the medical sciences, in particular, 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 ~ot
`constitute a guarantee or endorsement of the quality or value of such product or of the clailnS made of it by its manufacturer.
`
`Advertising hfformation. Advertising orders and enquiries can be sent to: USA, Canada and South America: Mr Tino de Carlo, The
`Advertising Department, Elsevier Science Inc., 655 Avenue of the Americas, New York, NY 10010-5107, USA; phone: (+ 1) (212)--i
`633 3815; fax: (+ 1) (212) 633 3820; e-mail: t.decarlo@elsevier.com. Japan: The Advertising Departlnent, Elsevier Scieuce K.K., 9-15
`Higashi-Azabu 1-chorea, Minato-ku, Tokyo 106-0044, Japan; phone: (+81) (3) 556I 5033; fax: (+81) (3) 5561 5047. Europe and
`ROW: Rachel Leveson,Gower, The Advertising Department, Elsevier Science Ltd., The Boulevard, Langford Lane, Kidlington,
`Oxford OX5 1GB, UK; phone.: (+44) (1865) 843565; fax: (+44) (1865) 843976; e-mail: r.leveson-gower@elsevier.co.uk
`
`Orders, claims a~d product enqairies: please contact the Customer Support Departinent at the Regional Sales Office nearest you:
`
`New York: Elsevier Science, PO Box 945, New York, NY 10159-0945, USA; phone: (+1) (212) 6333730 [toll free nulnber for North Amer~:an
`customers: 1-888-4ES-INFO (437~4636)]; fax: (+ 1) (212) 6333680; e-mail: usinfo-f@elsevier.com
`
`Amsterda~n: Elsevier Science, PO Box 21i, 1000 AE Amsterdam, The Netherlands; phone: (-t-31) 20 4853757; fax: (+31) 20 4853432; e-mail:
`nlin re-f @elsevier.hi
`
`Tokyo: Elsevier Science, 9-15 Higashi-Azabu 1-chorea, Minato-ku, Tokyo 106~0044, Japan; phone.: (+81) (3) 5561 5033; fax: (+8i) (3) 5561 5047;
`e-mail: iafo@elsevier.co.jp
`
`Singapore: Elsevier Saience, No. 1 Temasek Avenue, # 17-0l Milienia Tower, Singapore 039192; phone: (+ 65) 434 3727; fax: (+ 65)337 2230’ e-ma I:
`asiain fo@elsevler.co~_n,sg
`
`Rio de Janeiro: Elsevier Science, Rua Sate de Setembro 111/16 Andar, 20050-002 Centre, Rio de Janeiro-RJ, Brazil; phone: (+ 55) (21) 509 5340: lax:
`(+ 55) (21) 507 1991; e-mail: elsevier@campus.com.br [Note (Latin America): for orders, claims and help desk inforination, please contact the Regional
`Sales Office in New York as listed above]
`
`@ The paper used in this publication meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper)
`
`Printed in the United Kingdom
`
`Lupin Ex. 1051 (Page 2 of 11)
`
`
`
`~LSEVIER
`
`International Journal of Pharmaceutics 241 (2002) 253-261
`
`international
`journal of
`pharmaceutics
`
`www.elsevier.coln/locate/ijl~harm
`
`Water dynamics in channel hydrates
`exchange
`
`investigated using H/D
`
`Matti U.A. Ahlqvist, Lynne S. Taylor *
`
`Solid State Analysis, Pharmaceutical and Analytical R&D, AstraZe~eca R&D Mi~h~dal, S-4~1 83 Mi~lndal, Sweden
`
`Received 21 January 2002; accepted 25 April 2002
`
`Abstract
`
`The dynamics and interactions of water with different channel hydrates were studied. Caffeine 4/5-hydrate,
`theophylline monohydrate and sodium cromoglycate were used as model compounds. The hydrogen/deuterium (H/D)
`~xchange of the different hydrates following exposure to deuterium oxide vapour was studied using FT-Raman
`spectroscopy. The aim of the work was to (1) investigate the potential for H!D exchange studies to provide
`information about channel hydrates and (2) correlate the mobility of the water molecules inside lattice channels with
`structural parameters of the specific hydrates. The rate of exchange in the three different compounds was shown to
`vary considerably with caffeine 4/5 hydrate undergoing exchange much more rapidly than either sodium cromoglycate
`cr theophylline monohydrate, with exchange in the latter compound being extremely slow. Based on the known
`crystal structures, it was possible to rationalise the results and to draw conclusions about the mechanism of exchange
`for the model COlnpounds. It was found that the mobility, of the water molemfles in a channel hydrate is very
`dependent on the dimensions of the hydrate channel. Thus H/D-exchange studies rnay provide very useful structural
`and energetic information about channel hydrates. © 2002 Elsevier Science B.V. All rights reserved.
`
`Keywords: Theophylline molmhydrate; CafFeine 4/5 hydrate; Sodium cromoglycate; H!D exchange; Raman spectroscopy; Deuterium
`oxide
`
`t ~
`
`’ Introduction
`
`Pharlnaceutical materials are often capable of
`) incorporating water into the crystal lattice to form
`
`a hydrate. It has been estimated that one third of
`all pharlnaeeutical substances on the market have
`hydrated forms (Stahl, 1980). Water may be in-
`
`* Corresponding author. Tel.: + 46-31-776-1282; fax: + 46-
`31-776-3835
`E-mail address: lynne.taylor@astrazeneea.com (L.S, Taylor),
`
`corporated in the lattice in several ways and the
`properties of the hydrate crystal will depend on
`how the water is bound. It is thus useful to be
`able to divide hydrates into a number of classes
`which reflect structural and/or behavioural differ-
`ences. One such classification system for hydrates
`has been suggested by Morris and Rodriguez-
`Hornendez (Morris and Rodriguez-Hornedo,
`1993) whereby hydrates can be structurally di-
`vided into three different groups, namely isolated
`lattice site hydrates, lattice channel hydrates and
`metal ion co-ordinated hydrates.
`
`0378-5173/02/$ - see fi’ont matter © 2002 Elsevier Science B.V. All
`PII: S0378-5 173(02)00242-9
`
`rights reserved.
`
`Lupin Ex. 1051 (Page 3 of 11)
`
`
`
`254
`
`M.U.A. Ah/qvist, L.S. Taylor/h~ten~ational Journal qf Pharmaeeutie~ 241 (2002) 253 26!
`
`The present work focuses on understanding
`water dynamics and interactions in lattice channe!
`hydrates and to some extent metal ion co-ordi-
`nated hydrates. Lattice channel hydrates comprise
`an interesting class of hydrates because small
`differences in structural parameters may have a
`large effect on the stability of the hydrate (Pettier
`and Byrn, 1982; Byrn et al., 1999). The water
`molecules in lattice channel hydrates can be either
`localised (hydrogen bonded to specific groups in
`the parent molecule) or disordered. Some lattice
`channel hydrates may also expand or contract to
`host more or less water molecules as the ambient
`humidity is altered and these are termed expanded
`lattice channel hydrates (Morris and Rodriguez-
`Hornedo, 1993).
`Hydrates are commonly investigated by hydra-
`tion/dehydration experiments. However, since hy-
`dration/dehydration often leads to a change of
`phase, alternative methods for studying hydrates
`are of interest. Hydrogen/deuterium (H!D) ex-
`change has been used in a number of studies to
`investigate interactions of water with cyclodex-
`trins by lnonitoring the exchange of crystal water
`and hydroxyl groups following exposure to deu-
`terium oxide vapour (Steiner et al., 1995; Amado
`and Ribeiro-Claro, 1997; Moreira da Silva et al.,
`1997). More recently, this method has been used
`to probe amorphous and crystalline sugars
`(Ahlqvist and Taylor, 2002). The purpose of the
`present study was to (1) investigate the potential
`for H!D exchange experilnents to provide infor-
`mation about channel hydrates and (2) correlate
`
`the mobility of the water molecules inside lattice
`channels with structural parameters of the specific .
`hydrates.
`Theophylline monohydrate and caffeine 4/5-by,
`drate were chosen as model compounds for thi
`investigation. They can both form lattice channel
`hydrates and are essentially isomorphous with!
`very similar molecular structures (Sutor, 1958a,b).
`Despite the silnilarities, the stability of their hy-
`drated forms are very different. While caffeine
`effloresces under ambient conditions, theophylline
`must be heated to 35-50 °C or stored at low
`relative humidities (RH) before dehydration takes
`place (Byrn et al., 1999). Sodium cromoglycate "
`was also selected as a model compound, since il is
`capable of forming non-stoichiometric hydrates.
`Thus, when sodium cromoglycate is exposed to
`high RH, the crystals will expand to hold more
`water and at low RH, the converse occurs and the
`crystals contract. In addition, sodium cromogly- "
`cate contains water molecules in different chemi-
`cal environments. For example, at 75% FH
`sodium cromoglycate contains seven molecules of
`water per molecule of drug, four of the water ,.
`molecules are co-ordinated to the sodium ions
`and the other three water molecules are located in
`a lattice channel and held by hydrogen bonding to
`the parent molecule (Cox et al., 1971; Stephenson
`and Diseroad, 2000). The molecular structures ~f
`theophylline, caffeine and sodium cromoglycate-~
`are shown in Fig. 1. Raman spectroscopy wrs
`used to monitor H/D exchange in these model
`systems. .
`
`Theophylline
`
`0
`Sodium cromoglycate
`
`Caffeine
`
`Fig. 1, Molecular structures of theophylline, sodiuln cromoglycate and caffeine. The asterisk on theophylline and caffeine marks file
`nitrogen that accepts a hydrogen bond fl’om the water molecules in the laydrate structure,
`
`Lupin Ex. 1051 (Page 4 of 11)
`
`
`
`M.U.A, Ahlqoist, L,S, Taylor/International Journal of Pharmaceutics 241 (2002) 253-261
`
`255
`
`2. Materials and methods
`
`2. I. Materials
`
`:Anhydrous theophylline, anhydrous caffeine
`and sodium cromoglycate were purchased from
`Sigma Chemical Company (St Louis, MO).
`Theophylline m0nohydrate was obtained as
`needle like crystals by crystallisation from water.
`T,~e sample was sieved and the 125-250 ~trn
`fraction was retained. The weight loss on heat-
`i~g during thermogravimetric analysis (TGA)
`was 9.4% (theoretical weight loss for the mono-
`hydrate is 9.1%) and X-ray powder diffraction
`(XRPD) pattern of the obtained hydrate was in
`good agreement to that reported previously
`(Wang, 1974). Theophylline monodeuteriate was
`obtained from a deuterium oxide solution and
`stored at ~-,44% RH (D~O). The sample lost
`9.9% weight on heating during TGA (theoretical
`~fight loss for the molaodeuteriate is 10.0%) and
`XRPD confirmed that the monodeuteriate and
`t!ae monohydrate had the same crystal structure.
`Caffeine 4/5-hydrate was prepared by crystalli-
`sation from a hot (80 °C) aqueous solution. The
`solution was slowly cooled to room temperature
`and the needle-like crystals were harvested by
`filtration. The wet crystals were dried at 44%
`RH for 7 days and then stored at 75% RH. The
`weight loss on heating during TGA was 6.8%
`(theoretical weight loss for the 4/5-hydrate is
`6.9%). The XRPD pattern was coincident with
`that reported previously (Wang and Jaw, 1979).
`Microscopy of the material showed an approxi-
`mate particle size of 100 ~tm. Caffeine 4i5-deu-
`teriate was obtained by crystallisation from a
`1 ot D~O solution. The solvent was relnoved un-
`der a nitrogen purge until damp crystals re-
`mained, final drying was performed at 75% RH
`(D~O). The weight loss on heating during TGA
`was 7.4% (theoretical weight loss for the 4/5-
`eeuteriate is 7.5%). XRPD confirmed that the
`crystal structure of the deuteriated sample was
`iJentical to that of the hydrated sample.
`Sodiuln cromoglycate was equilibrated at 75%
`RH which corresponds to a water content of
`7-7.8 water molecules per molecule of drug
`(Cox et al., 1971). Microscopy confirmed a par-
`
`ticle size of around 100 ~tlTl_ for the bladed crys-
`tals.
`
`2.2. Materials characterisation
`
`A Mettler Toledo TGA 850 system was used
`for the TGA naeasurements. The heating rate
`was 10 °C!rain and open almn_ina pans were
`used.
`XRPD analysis was performed using low
`background sample holders and a Siemens
`DS000 X-ray diffractometer.
`Qualitative analysis of particle size of the sam-
`ples was performed on an Olympus BX50 micro-
`scope fitted with a Sony Exwave HAD digital
`video camera and Linksys software for Windows
`v.1.85.
`
`2,3, Spectroscopic monitoring of the
`HiD-exchange
`
`A Perkin-Ehner System 2000 FTIR with a
`Raman accessory, equipped with an InGaAs de-
`tector was used for collecting Raman spectra of
`theophylline.
`Raman spectra of caffeine and sodium cromo-
`glycate were collected with a Bio-Rad FTS 575C
`FTIR instrument with a Raman accessory,
`equipped with a liquid nitrogen cooled Ge
`detector.
`For both instruments, a Nd:Yag laser (1064
`nln) was used for excitation. The laser power on
`the sample was typically 500 mW and the spec-
`tral resolution was 4 era-1. Scattered radiation
`was collected at an angle of 180° and the Stokes
`radiation is reported. Indene was used as a stan-
`dard to monitor the wavenumber accuracy.-
`Spectra were coilected as a function of time with
`a minimuln of 1064 scans for each time-point.
`Samples were analysed in triplicate in rotated
`sample vials where a constant humidity environ-
`ment was provided by a saturated salt solution
`prepared with D~O. These hygrostats have been
`described in detail previously (Ahlqvist and Tay-
`lor, 2002). The experimental variation was better
`than _+ 10% and the mean of the data are pre-
`sented.
`
`Lupin Ex. 1051 (Page 5 of 11)
`
`
`
`256
`
`M. U,A. Ahlq~ist L.S, Taylor/hzlernational Journal qf Pl~armaceutics 241 (2002) 253-261
`
`CH-stretch
`
`OH-stretch
`overlapping
`with CH- ,
`
`F=73 h
`T=3h
`T=0h
`
`3600 3400
`
`3200
`
`3000 2800 2600 2400 2200 2000
`Raman Shift (cm4)
`
`Fig. 2. Ralnan spect~a of caffeine 4/5 hydrate as a [’unction of time of exposure to deuterium oxide vapour ( ~ 75 %RH.D20). With
`time the peak due to OH stretch (3400-3100 era- I) decreases and the peak due to OD stretch (2600-2300 cm- ~) increases. A1
`spectra are norlnalised to the CH stretchiug peak.
`
`3. Results
`
`3.1. HID exchange in caJfeine
`
`Fig. 2 shows Raman spectra of caffeine 4/5-hy-
`drate over the 3600-2000 cm-~ wavenumber re-
`gion following exposure to deuterium oxide
`vapour (~ 75% RH) for different time periods.
`The peaks in this region arise from three group
`frequencies. At around 3300-3000 cm- ~, the very
`broad, weak peak arises from OH stretching
`modes of the crystal water molecules in caffeine
`4/5-hydrate. Overlapping with the main part of
`the OH peak are the much stronger CH stretching
`peaks (3150-2800 cm - ~) of the caffeine molecule.
`The bands in the frequency range from 2650 to
`2250 cm-~ are due to stretching modes of D20
`molecules tliat have replaced the water of
`crystallisation.
`It is apparent fi:om Fig. 2 that the spectrum of
`caffeine 4/5-hydrate changes on exposure to deu-
`teriuln oxide vapour. The peak arising from OD
`vibrations evolves with time. The simultaneous
`~lecrease of the OH peak is not as easy to detect
`
`since it overlaps with the CH peaks and is very
`weak even in the fully hydrated material. How-
`ever, careful examination of the spectra shows a
`decrease in OH intensity, indicating that water
`molecules in the hydrate are replaced by deu
`terium oxide.
`The change in the peak area of the OD peak
`(Io>) as a function of time was used to monitor
`the exchange process. The peak area was nor-
`malised against the area of the CH stretching
`peak at 2980-2900 ClTI-1 to eliminate experimen-
`tal variations. Caffeine 4/5 D20 was used as the
`reference to calculate the extent of exchange. The
`assumption was made that the OH peak wiI1
`decrease at the same rate as the OD peak increase.
`This assumption is reasonable since the hydrate
`can only host a certain number of water
`molecules, thus the uptake of a D~O requires the
`loss of a H20 molecule.
`Fig. 3 shows the increase in Io~ as a function of
`time. It can be seen that the exchange is quite
`rapid with 50% exchange after 175 rain. More-
`over, all of the water molecules are able to ex
`change with D20. Since all the water lnolecules in
`
`Lupin Ex. 1051 (Page 6 of 11)
`
`
`
`M.U.A. Ahtqvist, L.S. Taylor !h~ternatfonal Journal of Pharmaceutics 241 (2002) 253-26I
`
`257
`
`caffeine 4/5-hydrate are located within the same
`channel and caffeine does not have any exchange-
`able hydrogens, the complete exchange is as ex-
`pected. This can also be confirmed by the
`observation that the Raman spectrum of caffeine
`4!5-hydrate exposed to deuterium vapour for 4
`clays is virtually identical to the spectrum of caf-
`feine that has been crystallised from D~O (Fig. 4).
`The spectra of caffeine 4/5 H~O and caffeine
`4/:i D~O were also compared. As noted previ-
`ously, the caffeine molecule is unaltered on crys-
`taltisation from DzO, the crystal structure is the
`same as for the hydrate and thus the only differ-
`once is the solvent molecule. It can be seen from
`Fig. 5a that the two spectra are virtually identical
`and replacement of the water molecules with deu-
`telmm oxide causes very little perturbation of the
`caffeine molecule. Minor differences were, how-
`
`2dO0 26’00 ’ 24’00
`Wavenumber (6m-1)
`
`Fig. 4, Comparison of" the 3400-2100 cm ~region of the
`Raman spectra for caffeine 4/5 D20 and caffeine 4/5 H~O
`after full equilibration with deuterium oxide vapour (~ 75
`%RH.D20).
`
`ever, observed in two spectral regions as shown in
`Fig. 5b. In caffeine 4/5 D~O there is an increase in
`peak intensity at 1493 Cln~ and a whilst the peak
`seen at 646 cm- t shifts to 648 cm-~. These peaks
`most likely arise from modes associated with the
`nitrogen atom which is involved in hydrogen
`bonding with water!deuterium oxide.
`
`Theophylline monohydrate was also found to
`undergo H/D-exchange on exposure to deuterium
`oxide vapour. However, as can be seen from Fig.
`3, the H!D-exchange in theophylline monohydrate
`is much slower than in caffeine 4/5-hydrate. Fig.
`3b shows that the exchange is still proceeding
`after 60 000 min (42 days) and is nowhere close to
`completion, showing only around 39% exchange
`relative to the deuterated theophylline D~_O refer-
`ence. Even after more than 6 months, exchange
`was not complete (it had reached approximately
`40%, data not shown).
`The HiD-exchange in theophylline monohy-
`drate is more complicated than for caffeine 4/5-
`hydrate since the former also has an NH group
`that can potentially exchange a hydrogen for a
`deuterium. Unfortunately, the .OD and ND peaks
`overlap so this information cannot be directly
`gained from a comparison of the spectra of the
`solvated samples. In order to see if the NH group
`had undergone any exchange, the sample of
`
`Lupin Ex. 1051 (Page 7 of 11)
`
`~ Caffeine 4/5 hydrate
`....... Sodium ¢romoglyoate
`--. Theophylline monohydrate
`
`3.2. HiD-exchange in theophy[line
`
`1000
`
`2000
`
`3000
`
`4000
`
`Time (minutes)
`
`1
`
`0,8
`
`o.s
`
`0.2
`
`0
`
`1
`
`5"o.~ i
`
`0.2
`
`-- Caffeine 4t5 hydrate,
`.. -" "
`flf" ....... Sodium cromoglycate
`--. Yheophylline monohydrate
`
`0 ,
`0
`10000
`
`(h)
`
`20000 30000 40000
`
`50000
`
`60000
`
`Time (minutes)
`Fi;~. 3. Plot of the relative Raman intensity of the OD stretch-
`ing peak (Rel. Io~) as a function of time of exposure to
`deuterium oxide vapour (75 %RH.D20) for ca[’t~ine 4/5 hy-
`dmte, theophylline monohydrate and sodimn eromoglycate.
`Fig, 3a shows the first 5000 rain and Fig. 3b shows the
`exchange over 60 000 rain.
`
`
`
`258
`
`M.U.A. Ahtqvisl, L,S. Taylor iInterizatiottal Journal of Pharmaceutics 241 (2002) 253-26J
`
`theophylline which had been exposed to D20 was
`dried and compared with a dried sample of fully
`deuterated theophylline. This comparison indi-
`cated that there was a small ND peak in the
`spectrum of the dried exchanged sample (data not
`shown), indicating that some NH groups have
`undergone exchange. However, the intensity is
`much lower than for the fully deuterated sample,
`and from the relative intensities of the ND peak
`in the two samples, it can be estimated that only
`around 13% of the NH groups have exchanged.
`In other spectral regions, the exchanged sample
`was more similar to anhydrous theophyiline than
`to deuterated anhydrous theophyltine (the pres-
`ence of a ND group changes a number of peaks in
`the fingerprint region of the spectruna), confirming
`that the NH group is largely unexchangeable.
`
`3,3. H!D-exchange in sodium cromogIycate
`
`Sodium cromoglycate will take up or give off
`water molecules to equilibrate with the ambient
`atlnosphere in a non-stoichiometric manner (Cox
`et al., 1971; Chen et al., 1999). To enable a true
`H/D-exchange to take place, the samples had to
`be pre-equilibrated in 75% RH.H:O prior to being
`exposed to ~ 75% RH’D20. At 75% RH, sodium
`cromoglycate hosts approximately 7 water
`molecules per drug molecule (Cox et el., 1971).
`Upon exposure to deuterium oxide, the water
`molecules in sodium cromoglycate were found to
`exchange with deuterium oxide. As can be seen
`
`fi’om Fig. 3, the rate of the exchange process in
`sodium cromoglycate is slower than that in caf-
`feine 4/5-hydrate but faster than for theophylline
`monohydrate with a 50% H/D-conversion after
`750 rain (12.5 h).
`Fig. 3 indicates that exchange does not reach
`completion as compared with the reference sam.
`ple of sodimaa cromoglycate crystallised from D~0
`atad equilibrated at 75% RH.(D~O) atlnosphere.
`Since all of the water lnolecules are found in a
`channel, it seems like!y that this incomplete
`change may result from the inaccessibility of the
`hydroxyl group of the cromoglycate molecu’e.
`This was checked by comparing the spectrum of
`dried sample of sodium cromoglycate, exposed for
`six months to D~O vapour, with that of a dried
`sample of sodium cromoglycate re-crystallised
`fi’om D~O (data not shown). The sample that was
`exposed to D20 vapour had a small OH peak
`whilst the sample crystallised from D~O had
`This indicates that the OH group in sodium cro-
`moglycate is not exchanged on exposure to
`vapour. This was also confirmed by looking at
`otl~er areas of the spectrum where the dried six
`month sample was virtually identical to that of’
`the dried hydrate, but different fi’om that of the
`dried deuterated salnple which showed a number
`of differences due to the presence of the OD
`hydroxyl fnnction. Thus our results suggest th’~t
`although all the water is exchanged, the hydroxyl
`group is unexchangeable and this group accounts
`for the residual OH intensity.
`
`............. = Caffeine, 415-D~O
`-- = Caffeine, 415-H~O
`
`Caffeine, 4f5,DzO
`
`Caffeine, 4/5-H
`
`3400
`(a)
`
`3000
`
`2600 2200
`1800
`1400
`Wavenumber(cm-1)
`
`1510
`(b)
`Wavenumber (crn‘t)
`Fig. 5. Comparisons of the Ramall spectra of caffeine 4/5 D20 and caffeine 4/5 H20. (a) shows the entire spectra while (b) sho~s
`the regions in which differences are seen (1510-i450 cm- 1, 664-630 cm- ~).
`
`1000
`
`600
`
`t,~90 ’ 1,~70 ’ t4~50
`
`T -
`’~6
`~4 ’ 6~6 ’ 6~8 ’ 6;10’ 632
`
`Lupin Ex. 1051 (Page 8 of 11)
`
`
`
`M.U.A. Ahlqoist, L.S. Taylor/h~ternationaf Join’hal of Pharmaceulics 241 (2002) 255-261
`
`259
`
`4. Discussion
`
`~b
`
`!1 channel hydrates, the water molecules are
`adjacent to other water molecules in the neigh-
`bo,tring unit cell and a channel of water molecules
`is formed along one of the crystallographic axis
`(Morris and Rodriguez-Hornedo, 1993). The
`three compounds examined in this study can all
`be classified as channel hydrates. Results from
`thi:, study have shown that crystal hydrate water
`can be replaced by DzO molecules on exposure to
`Dr) vapour for all three compounds although the
`kinetics and extent of exchange were found to
`vary considerably for the different compounds.
`The single crystal structures of caffeine 4/5 hy-
`drate, theophylline monohydrate and sodium cro-
`moglyclate have been published and were used to
`aid interpretation of the results from this study.
`Despite the fact that theophylline monohydrate
`and caffeine 4/5-hydrate are nearly isomorphous
`--(Byrn et al., 1999), the rate of the HiD exchange
`is very different for the two hydrates. In addition,
`the hydrates of these two compounds exhibit dif-
`fering physical stabilities. Thus caffeine forms a
`non-stoichiometric hydrate whereby the water
`molecule readily effloresces through a channel in
`the a-crystallographic direction (space group P2~/
`e). There is a single hydrogen bond between the
`caffeine molecule and water, between the imina-
`zole nitrogen atom marked in Fig. 1 and the
`hydrogen of a water molecule. The water
`molecules form a zigzag chain which has vacan-
`cies as a result of the non-stoichiometry leading to
`’localised disorder of the water lnolecules (Gerdil
`an~t Marsh, 1960). In contrast theophylline forms
`a stoichiolnetric hydrate which is stable to dehy-
`dration at room temperature. Each theophylline
`molecule hydrogen bonds to a water molecule
`through the nitrogen atom marked in Fig. 1 and
`~the water molecules form a zigzag chain. Unlike
`caffeine 4/5 hydrate, the water molecules are lo-
`caAsed. The strength of the hydrogen bonds be-
`tween drug molecule and water are similar in
`m~:gnitude in the two compounds, suggesting that
`other structural features are important in influenc-
`ing the observed differences in the rate and extent
`of H!D exchange. Perrier and Byrn have high-
`lighted the importance of the cross sectional area
`
`3.3A
`
`Fig. 6. Crystal structural of caffeine 4/5 l~ydrate (looking along
`the a-crystallographic axis), space group P21ic. The contours
`indicate the profile of the van der Waal radii of the atoms that
`are adjacent to the hydrate water channeI. The water
`molecules are drawn as van der Waal representations flOlll two
`perspectives and are shown next to caffeine for size compari-
`sons. Tl~e distances, X and Y, are the shortest distances
`between opposite atoms (in three dim~nsioas) forming the
`channel.
`
`of the water channel in relationship to the dehy~
`dration behaviour of hydrates (Perrier and Byrn,
`1982), and it is suggested that the size of the
`hydrate water channels in theophylline and caf-
`feine will be an important factor affecting the
`H/D-exchange rate.
`Figs. 6 and 7 show structural diagrams of caf-
`feine 4/5-hydrate, looking along the a-crystallo-
`
`~b
`
`Fig. 7. Crystal structure of theophylline monohydrate (looking
`along the c-crystallographic axis) showing only the molecules
`that are in direct contact with the hydrate water channel.
`Other details are as for Fig. 6
`
`Lupin Ex. 1051 (Page 9 of 11)
`
`
`
`260
`
`M. U,A. Ahlqvisl, L,S, Taylor/Interncttior~al Jour1~al qf Pharmaeeuties 241 (2002) 253-261
`
`graphic axis and theophylline monohydrate, look-
`ing along the c-crystailographic axis. The van der
`Waal radii of the atoms surrounding the hydrate
`water channel are represented as a contour, the
`parent molecules are represented as sticks. The
`water molecules are drawn next to the parent
`molecules with van der Waal radii for size
`comparison.
`The size of the hydrate water channels were
`measured from crystallographic data (Sutor,
`1958a,b) ilnported from Cambridge Crystallo-
`graphic Database using WebLab ViewerPro 3.7
`software. The size of the hydrate channels were
`estimated by measuring the two shortest distances
`(roughly perpendicular) across the channel. The
`distances were measured froln the edge of the van
`der Waal radius of one drug atom to the edge of
`the van der Waal radius of the opposite drug
`atoln (see Figs. 6 and 7). The size of the water
`molecule was determined in a silnilar wa3~ by
`measuring three distances between edges of the
`van der Waal radii of the atoms in the molecule
`(see Figs. 6 and 7).
`By comparing the size of a water molecule with
`the measured size of the caffeine 4!5-hydrate wa-
`ter channel, it seems likely that a water molecule
`can enter the hydrate channel in any orientation.
`This means that all water molecules that approach.
`the channel entrance in caffeine 4/5-hydrate may
`penetrate the hydrate and diffuse into the
`molecule. It can also be assumed by reference to
`Fig. 6 that the size of the channel may allow for
`one water molecule to pass by another water
`molecule inside the channel. This allows for diffu-
`sion of intact water molecules along the channel
`and the replacement of water molecules by D~O
`molecules. Moreover, the vacancies in the hydrate
`chain discussed above will facilitate the movement
`of water within the channel.
`In contrast, the hydrate channel in theophylline
`monohydrate is considerably smaller than that
`found in caffeine 4/5 hydrate and is approxi-
`mately the same size as a water molecule as shown
`by Fig. 7. The packing of theophylline molecules
`around the hydrate water molecules is thus much
`closer than in caffeine 4/5 hydrate. This structural
`difference is the most likely explanation for the
`extremely low rate and incomplete extent of H/D
`
`exchange for theophylline. There is, however,
`around 40% exchange in theophylline monohy.
`drate as compared with the deuterated form. It is
`likely that some exchange of water molecules
`the channels can take place at the ends of the
`crystals (where the hydrate channel commences)
`which are in contact with D20 atmosphere and
`further exchange may also occur in disordered
`regions of the crystals. However, given the strue.
`ture of theophylline lnonohydrate as discussed
`above, it is unlikely, that the majority of the
`H/D-exchange occurs by diffusion of intact Dz0
`molecules. Instead the exchange mechanism
`be proton chain transfer which would provide an’
`explanation for the extremely slow exchange pro-
`cess in theophylline monohydrate. Proton chain
`transfer involves disruption of the water covalent
`bonds and is, therefore, more energetically de-
`manding and slower than the exchange of intact
`water molecules. Moreira da Silva et al. postu-
`lated that part of the H/D-exchange process in
`[3-cyclodextrin occurs by proton chain transfer
`(Moreira da Silva et al., 1997; Steiner et al., 1995).
`Their assumption that the proton chain transfer is
`an essentially slower process than the diffusion of
`intact water molecules is given further support by
`the work done on c,-cyclodextrin by Amado m~d
`R