Study of the lorazepam: Cyclodextrin inclusion complexes using electrospray ionization mass spectrometry

Article abstract of DOI:10.1016/S0040-4039(01)01181-9

As an example of drug-cyclodextrin interactions in aqueous media, the cyclodextrin complexation of lorazepam was studied by electrospray ionization mass spectroscopy (ES-MS). It was concluded that highly concentrated aqueous hydroxypropylcyclodextrins are more suited for the drug complexation and that the inclusion complex includes one molecule of the drug with two and three cyclodextrin molecules as well as two molecules of the drug with three molecules of hydroxypropylcyclodextrins. It was postulated that in the hydroxycyclodextrin cavity the drug molecule decomposes via the elimination of one molecule of formaldehyde.

Full text of DOI:10.1016/S0040-4039(01)01181-9

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 6077–6082
Study of the lorazepam: cyclodextrin inclusion complexes using
electrospray ionization mass spectrometry
Renata Kobetic´,a Branko S. Jursic,b,* Sidney Bonnette,b Jane S.-C. Tsaia and Salamone J. Salvatorea
aRoche Diagnostics Corporation, R&D Department, 9115 Hague Road, Indianapolis, IN 46250, USA
bDepartment of Chemistry, University of New Orleans, New Orleans, LA 70148, USA
Received 24 May 2001; accepted 8 June 2001
Abstract—As an example of drug–cyclodextrin interactions in aqueous media, the cyclodextrin complexation of lorazepam was
studied by electrospray ionization mass spectroscopy (ES-MS). It was concluded that highly concentrated aqueous hydroxypropyl-
cyclodextrins are more suited for the drug complexation and that the inclusion complex includes one molecule of the drug with
two and three cyclodextrin molecules as well as two molecules of the drug with three molecules of hydroxypropylcyclodextrins.
It was postulated that in the hydroxycyclodextrin cavity the drug molecule decomposes via the elimination of one molecule of
formaldehyde. © 2001 Elsevier Science Ltd. All rights reserved.
An important part of chemical research is to determine
the existence as well as alter the physico-chemical prop-
erties of various drugs in aqueous media. If the drug
molecule has aromatic groups that are complementary
in size to the cyclodextrin cavity, then cyclodextrin is
the molecule of choice to achieve these goals.1 One of
the most commonly used sedatives (particularly for
anxiety and insomnia)2 that has aromatic moieties to
which cyclodextrins can bind are benzodiazepines.3
Therefore, it seems appropriate to explore their physical
properties in aqueous cyclodextrins.
istry is b-cyclodextrin (b-CD). Naturally, its capability
to form inclusion complexes with these drugs is of great
importance, but it is also significant to explore the
influence of the cyclodextrin cavity size on the inclusion
complex formation. This can be accomplished by
broadening the complexation study to cover a- and
g-cyclodextrins (a-CD and g-CD). On the other hand,
b-cyclodextrin has relatively low solubility in water,4
therefore to explore the influence of the concentration
on the formation of the cyclodextrin complexes,
hydroxypropyl derivatives of a-, b-, and g-cyclodextrins
were also included in this study.5 These cyclodextrins
(hydroxypropyl-a-cyclodextrin, HP-a-CD; hydroxy-
propyl-b-cyclodextrin, HP-b-CD; hydroxypropyl-g-
cyclodextrin, HP-g-CD) have both higher solubilities
and binding capabilities than normal cyclodextrins.6 To
explore the influence of polyhydroxyl molecules already
present in natural fluids on the cyclodextrin drug com-
plexation, the study also includes such additives as
glycerol, PEG, and propylene glycol.7
The objective of this work is to explore the possibility
of using the ES-MS technique to better understand the
nature of the benzodiazepine–cyclodextrin inclusion
complex. Lorazepam (L) as one important member of
the benzodiazepine family was selected to study the
capabilities of cyclodextrin (CD) to form detectable
benzodiazepine–cyclodextrin
inclusion
complexes
(Scheme 1). The most widely used cyclodextrin in chem-
H
O
O
N
RO
O
Inclusion
H2O or H2O/Glycerol
OH
+
Complexes
RO
OR
N
Cl
n
R=H or C(CH3)2OH
n=6,7, or 8
Cl
Cyclodextrins (CDs)
Lorazepam (L)
Scheme 1. An equilibrium between lorazepam and cyclodextrins in water media.
* Corresponding author.
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)01181-9

6078
R. Kobetic´ et al. / Tetrahedron Letters 42 (2001) 6077–6082
In all our ES-MS studies with HP-CD 1:220 and 2:321
lorazepam–cyclodextrin inclusion complexes were
observed as demonstrated on the examples of
lorazepam (L) inclusion complexes with the HP-a-CD
(Fig. 1). Although the concentration of cyclodextrin is
1000 times higher than the drug concentration, the
instrument detector sensitivity is tuned to lorazepam
and in this way a better noise to signal ratio for
inclusion complexes was observed. The free guest
molecule signal (L+Na=344) is very small in compari-
son with the HP-a-CD inclusion complexes, indicating
that practically the entire drug is complexed by the
cyclodextrin. It appears that the complex with two
HP-a-CD and one L is the most dominant inclusion
complex, although there are signals of drug complexed
three HP-a-CDs (Fig. 1).
The formation of the inclusion complexes between benzo-
diazepines and cyclodextrins is known,8 but the inclu-
sion complex ‘behavior’ in aqueous solution is still not
completely understood. The earlier inclusion complexa-
tion studies were based on solubility studies using UV–
Vis, CD, IR, and DSC techniques.9 Recent
complexation studies were performed in mixed or
organic solvents10 rather than in water since most of the
benzodiazepines have very low solubility in water. In
many instances additives such as glycerol help to dis-
solve the drugs, although water is the media in which
cyclodextrin application is utilized the most.11 In this
paper only the results obtained from electrospray ion-
ization mass spectroscopy (ES-MS) will be presented12
although in some instances our extensive NMR studies
of the benzodiazepines will be used for comparison.
Contrary to a traditional spectroscopic approach, such
as is the case with NMR13 spectroscopic study, ES-MS
spectroscopy allows us to detect and explore all kinds
of different molecular associates between the guest and
the host molecules.14 ‘Soft’ ionization methods have
been employed to study the non-covalent interactions
in the gas-phase successfully.15 Recently, Bartlett16 and
others17 showed that ES-MS is a very useful soft tech-
nique for the study of molecular complexes.
In the ES-MS spectra of lorazepam inclusion complexes
with HP-CDs (HP-a-CD, HP-b-CD, and HP-g-CD) the
observed m/z is 30 mass units below the expected
molecular mass. This is clearly not the case when a-CD
or g-CD as host molecules were used (Table 1). The
influence of the hydroxypropyl moiety on the cyclodex-
trin ring does not only increase the solubility in water
and the cyclodextrin binding capability, but it also
alters the chemical stability of the drug. It is obvious
that the presence of the hydroxypropyl group catalyzed
the elimination of the molecular unit with m/z 30. The
first postulate is that lorazepam and HP-a-CD form
acetal A through a reaction of the hemiacetal hydroxy
group of HP-CD with the hydroxyl group of lorazepam
(Scheme 2). In this way a molecule of water is elimi-
nated (m/z=18). According to our ES-MS data even if
this acetal is formed it decomposes and its signals
cannot be observed.
We were challenged in our studies by the fact that our
guest lorazepam (and the majority of the other benzo-
diazepines as well) has very poor water solubility (ꢀ0.1
mM). Fortunately, our NMR studies of the lorazepam–
cyclodextrin inclusion complexes reveal that to observe
a substantial change in the chemical shift of the aro-
matic portion of the spectra, the required molar ratio
L:CD is around 1:100. The concentration ratio used in
the experiments are also determined by the fact that
signals of the inclusion complexes should be sufficiently
strong to be clearly separated from the spectra base-
line.18 For practical purposes, the concentration of
cyclodextrins was as high as 100 mM,19 while the
lorazepam concentration was kept at a constant 0.1
mM.
Further evidence that the acetal A is an unlikely
product formed in aqueous L and HP-a-CD comes
from our chromatographic study. Both reverse phase
column and reverse phase TLC chromatographic sepa-
rations of the water solution made from lorazepam (0.1
mM) and HP-a-CD (100 mM) were performed. Only
Figure 1. The ES-MS spectra for the aqueous lorazepam (0.1 mM) and HP-a-CD (100 mM).

R. Kobetic´ et al. / Tetrahedron Letters 42 (2001) 6077–6082
6079
Table 1. Major peaks and host–guest ratio between
lorazepam (L) as guest and cyclodextrins (a-CD, g-CD,
HP-a-CD, HP-b-CD, and HP-g-CD)22 as host molecules
obtained in water
The influence of the concentration of cyclodextrin on
the formation of a higher order of inclusion complexes,
as well as a relatively low affinity of benzodiazepines to
form cyclodextrin inclusion complexes is perfectly
1
demonstrated in Fig. 2. From our H NMR studies we
Cyclodextrin
(host)
Host:guest
-ratio
Major peaks observed, m/z
know that there is practically no change in the chemical
shift for aromatic signals of 0.1 mM lorazepam if the
cyclodextrin concentration is below 10 mM. Further-
more, disproportion of the guest and host proton sig-
nals in the NMR spectra of this aqueous solution
renders the NMR method a difficult experimental tech-
nique. When HP-a-CD is used in only 10-fold excess
(Fig. 2) there is no ES-MS detectable cyclodextrin
complexation of lorazepam. Only the free drug, the free
HP-a-CD, and the self-complexed HP-a-CD signals
were observed (Fig. 2).
a-CD
g-CD
1:1
2:1
2:1
1319 [a-CD+L+Na]+
1157 [2a-CD+L+2Na]2+
2937 [2g-CD+L+Na]+
1480 [2g-CD+L+2Na]2+
2130 [3g-CD+L+2Na]2+
2260–2898, cluster of 12 peaks 58
units apart
3:1
2:1
HP-a-CD
HP-b-CD
HP-g-CD
[2HP-a-CD+L-H2CO+Na]+
1773 (1802) to 2092, cluster of 12
peaks 29 units apart
3:2
2:1
3:2
2:1
3:2
[3HP-a-CD+2L-2H2O+2Na]2+
2580–3226, cluster of 12 peaks 58
units apart
In all highly concentrated aqueous cyclodextrins, the
ES-MS signals that indicate double complexation of the
drug was observed (Table 1).25 The drug tends to be
complexed with at least two hydroxypropylcyclodex-
trins, while the formation of the complex with three
hydroxypropyl cyclodextrins and two lorazepam
molecules is also strong. The size of the cavity seems to
not have an influence on the intensity of the 3:2 com-
plex. This can be explained by the fact that the hydroxy-
propyl group can stabilize the cyclodextrin inclusion
complex in two ways. One is by shielding a large
portion of the drug’s benzene moiety from the bulk of
water because it cannot fully fit into the HP-a-CD
cavity. Another is by surrounding the drug’s benzene
moiety to make it fit better into large HP-g-CD cavity.
This, of course, cannot be accomplished in a-CD, b-
CD, and g-CD; therefore, it is reasonable to expect that
the ES-MS intensity of signals for drug complexation
with these cyclodextrins is not as strong.
[2HP-b-CD+L-H2CO+Na]+
2015–2379, cluster of 12 peaks 29
units apart
[3HP-b-CD+2L-2H2CO+2Na]+
2852–3491, cluster of 12 peaks 58
units apart
[2HP-g-CD+L-H2CO+Na]+
2257–2581, cluster of 12 peaks 29
units apart
[3HP-g-CD+2L-2H2CO+2Na]2+
UV-detectable molecules in the solution do not have
attached cyclodextrin moieties.23 In two separate experi-
ments, aqueous solutions of HP-a-CD and lorazepam
were stirred at room temperature (first experiment) or
refluxed overnight (second experiment). In both cases
we were not able to detect the presence of lorazepam,
instead we detected two dominant products. At room
temperature the dominant product seems to be hetero-
cycle I (Scheme 2) and at higher temperatures ketone
P.24
One can argue that polyhydroxyl molecules can actu-
ally help the formation of cyclodextrin complexed
lorazepam in a similar manner as covalently attached
hydroxypropyl groups do in HP-CDs. According to our
1H NMR spectroscopic studies, the presence of polyhy-
H
N
O
O
H
N
CH3
CH3
O
N
Cl
O
O
Cl
α-CD
FW=973
CH3
CH3
-H2O
N
Cl
H
N
FW=1334
CH3
O
O
HO
O
Cl
A
H3C
OH
- H2CO
-HP-α-CD
- H2CO
FW=2307
N
Cl
+
Cl
H
NH2
O
L
N
O
α-CD+(CH3)2CO
+2 H2O
-CO2
-NH3
FW=321
N
FW=1031
Cl
Cl
Cl
Cl
FW=291
P
I
Scheme 2. Possible pathways for decomposition of lorazepam in aqueous HP-CDs.

6080
R. Kobetic´ et al. / Tetrahedron Letters 42 (2001) 6077–6082
Figure 2. The ES-MS spectra for the aqueous lorazepam (0.1 mM) and HP-a-CD (1 mM).
droxyl molecules such as glycerol increase the solubility
of lorazepam in water, but at the same time decrease
References
1. (a) Bender, M. L.; Komiyama, M. Cyclodextrin Chemistry;
Springer-Verlag: Berlin, 1978; (b) Saenger, W. In Inclusion
Compounds; Atwood, J. L.; Davies, J. E.; MacNicol, D. D.,
Eds. Structural Aspects of Cyclodextrins and their Inclu-
sion Complexes; Academic Press: New York, 1984; Vol. 2,
Chapter 8, pp. 231–259; (c) Szejtli, J. In Inclusion Com-
pounds; Atwood, J. L.; Davies, J. E.; MacNicol, D. D., Eds.
Industrial Applications of Cyclodextrins; Academic Press:
New York, 1984; Vol. 3, Chapter 11, pp. 331–390; (d)
Szejtli, J. Chem. Rev. 1998, 98, 1743–1754; (e) Hedges, A.
R. Chem. Rev. 1998, 98, 2035–2044; (f) Uekama, K.;
Hirayama, F.; Irie, T. Chem. Rev. 1998, 98, 2044–2076.
2. Salzman C. Task Force on Benzodiazepine Dependency,
American Psychiatric Association; Benzodiazepine depen-
dence, toxicity, and abuse: a task force report of the
American Psychiatric Association; American Psychiatric
Association: Washington, DC, 1990.
3. For a general review about benzodiazepines see: Burger’s
Medical Chemistry and Drug Discovery: Therapeutical
Agents, 5th ed.; Wolff, M. E., Ed.; Wiley: New York, 1997;
Vols. II–V.
4. Solubility of b-cyclodextrin in water is 1.85 g/100 g of
water, which makes maximal concentration of ꢀ16 mM
for aqueous b-cyclodextrin. For comparison of physical
properties of various cyclodextrins see: Szejtli, J. Cyclodex-
trin Technology; Kluwer Academic: Dordrecht, 1988.
5. All cyclodextrins are obtained from Aldrich Chemical
Company and used in this study without further
purification.
the solubility of the cyclodextrins. It also appears that
there is no substantial difference on the cyclodextrin
lorazepam binding constant, which is estimated to be
between 10−4 and 10−3 mol−1 L. However, 100 mM
concentration of the cyclodextrin with glycerol is hard
to maintain homogeneous. The ES-MS spectroscopic
1
studies fully agree with the H NMR finding. There is
some difference of the ion distribution. For instance, no
double charged CD species can be observed when glyc-
erol is present. Strong signals for the sodium cation
with glycerol are present in the ES-MS spectra (m/z 115
and 190).26
In summary, noncovalent lorazepam–cyclodextrin
inclusion complexes could be studied using ‘soft’ ES-
MS techniques. The formation of cyclodextrin inclusion
complexes occurs and can be easily detected with the
ES-MS technique in highly concentrated aqueous
cyclodextrins. To achieve that, a host–guest molar ratio
of 1000:1 in 100 mM aqueous cyclodextrin is required.
Although quite different, the ES-MS and NMR studies
complement each other quite well. We also confirmed
the advantage of hydroxypropylcyclodextrins to form
detectable inclusion complexes with lorazepam and
eliminate the formaldehyde molecule from lorazepam.
The typical complexation ratio is 2:1 and 3:2 for
hydroxypropylcyclodextrin–lorazepam inclusion com-
plexes.
6. For instance, dichlofenac-hydroxypropyl-b-cyclodextrin
complex under commercial name ‘Voltaren’ has been used
for lysis of red blood cells. Solution of pure diclofenac
(10.77 mM) in aqueous hydroxypropyl-b-cyclodextrin
(61.88 mM) is four times more potent than a 40.67 mM
solution of Voltaren: Reer, O.; Bock, T. K.; Muller, B. W.
J. Pharm. Sci. 1994, 1345.
Acknowledgements
The authors thank the Roche Diagnostic Corporation
for the financial support.

R. Kobetic´ et al. / Tetrahedron Letters 42 (2001) 6077–6082
6081
7. In some instances the solubility of the guest molecule was
increased up to eight times in the presence of polyhy-
droxyl additives: Aboutaleb, A. E.; Rahman, A. A.;
Ismail, S. Bull. Pharm. Sci., Assiut University 1985, 8,
47–69.
four-, penta-, and hexahydroxypropyl-a-cyclodextrins.
Except for mono- and hexahydroxypropyl-a-cyclodextrin
all the others can have many isomers are a result of
attaching hydroxypropyl to the different primary
hydroxyl groups of the glucose unit of the cyclodextrin
ring. For instance, due to cyclodextrin asymmetry there
are six dihydroxypropyl-a-cyclodextrin (1,2-, 1,3-, 1,4-,
1,5, and 1,6-. Pairs 1,2- and 1,6- as well as 1,3- and 1,5-
are not equivalent. In the ES-MS all six dihydroxypropyl-
a-cyclodextrin isomers generate one set of signals). The
same is applicable for other hydroxypropylcyclodextrins.
Therefore, in the case of HP-a-CD six ES-MS 1:1, 36 1:2,
and 108 1:3 ES-MS detectable inclusion complexes will be
formed in the solution. Of course the most intense will
also be the one that has the most isomers. To generate
reasonably strong signals of these inclusion complexes the
lorazepam–cyclodextrin ratio is required to be as high as
possible.
8. Uekama, K.; Narisaw, S.; Hirayama, F.; Otagiri, M. Int.
J. Pharm. 1983, 16, 327–338.
9. (a) Uekama, K.; et al. Intl. J. Pharm. 1983, 16, 3327–
3328; (b) Anderson, F. M.; Bundgaard, H. Arch. Pharm.
Chem. Sci. Ed. 1982, 10, 80–87; (c) Abdel-Rahman, A.
A.; et al. J. Pharm. Belg. 1994, 49, 23–32; (d) Sanghavi,
N. M.; et al. Drug Dev. Ind. Pharm. 1993, 19 (6), 701–
712; (e) Keipert, S.; Hildebrand, S. Pharmazie 1989, 44,
H12; (f) Kata, M. et al. Proceedings of the Eight Interna-
tional Symposium on Cyclodextrins 435–438; Szejtli, J.;
Szente, L., Eds.; Kluwer Academic: Netherlands, 1996.
10. Shapiro, Yu. E.; Gorbatyuk, V. Ya.; Yakubovskaya, L.
N.; Andronati, K. S.; Andronati, S. A. Zh. Strukt. Khim.
1996, 37, 876–890.
11. Bioorganic Chemistry: Carbohydrates; Hecht, S. M., Ed.;
Oxford University Press, 1999; Chapter 12, pp. 458–488.
12. The ES-MS spectra were acquired with a sector instru-
ment with a mass of charge (m/z) range of 5000. A
Micromass Autospec M mass spectrometer with an elec-
trospray source was used. The ES-MS parameters (i.e.
pressure, temp., dielectric capillary distance and the
voltage on the needle, etc.) were kept constant in each
series of solutions. A flow rate of 10 mL/min was applied
using volumes of 100 mL of sample solutions with the
constant concentration of lorazepam being 10−4 M and
cyclodextrin concentration variations from 10−1 to 10−4
M.
13. On the NMR time scale the recorded NMR chemical
shifts for guest molecule represent an average chemical
shift resulting from fast equilibrium that involve free
guest molecules as well as guest molecules involved in
various cyclodextrin inclusion complexes.
14. For more details about electro-spray ionization mass
spectroscopy, see: Electrospray Ionization Mass Spec-
trometry: Fundamentals, Instrumentation, and Applica-
tions; Cole, R. B., Ed.; Wiley-Interscience: New York,
1997.
15. Przybylski, M.; Glocker, M. O. Angew. Chem., Int. Ed.
Engl. 1996, 35, 806–826 and references cited therein.
16. Bartlett, M. G.; Srinivasan, K. Rapid Commun. Mass
Spectrom. 2000, 14, 624–632.
17. (a) Blanschke, G.; et al. Chirality 1999, 11, 635–644; (b)
Selva, A.; et al. J. Mass Spectrom. 1998, 33, 729–734; (c)
Selva, A.; et al. J. Mass Spectrom. 1996, 31, 1364–1370;
(d) Bakhitar, R.; Hop, C. E. C. A. Rapid Commun. Mass
Spectrom. 1997, 11, 1478–1481; (e) Gallagher, R. T.; et al.
Intl. J. Mass Spectrom. Ion Process. 1997, 165/166, 523–
531; (f) Sawada, M. Mass Spectrom. Rev. 1997, 16,
73–90; (g) Lamcharfi, E. J. Mass Spectrom. 1996, 31,
982–986.
19. Considering our reasoning mentioned above (variety of
HP-CD inclusion complexes with lorazepam) it would be
optimal to prepare a 0.1 mM concentration of lorazepam
in 1000 mM (1 M) concentration of HP-CD. Although
these cyclodextrins will allow us to prepare this highly
concentrated solution the cyclodextrin crystallizes from
the solution after several hours. Further problems appear
with carrying out the ES-MS experiment. Crystallization
of the cyclodextrin occurs in the injection needle and
injection chamber must be cleaned after every ES-MS
recording even when the concentration of the cyclodex-
trin is 0.1 mM. With these difficulties associated with
experimental studies, high concentration cyclodextrin
solutions such as 1000 mM (1 M) were not possible and
a majority of our studies were performed with 0.01 M
cyclodextrin concentration (1:100 guest–host ratio), but
to obtain the better spectra presented in this paper the
concentration of cyclodextrin was 0.1 M.
20. The signal values for the L:HP-CD 1:2 complex are
observed with the elimination of formaldehyde. This
reaction also occurs under simulated laboratory condi-
tions and a paper of synthetic formaldehyde elimination
from diazepams will be published separately. For exam-
ples of elimination of formaldehyde from b-oxo alcohols,
see: Olsen, S.; Aalrust, E.; Blom, H. Chem. Ber. 1957, 90,
1389–1398; Olsen, S.; Aalrust, E.; Blom, H. Chem. Ber.
1957, 90, 765–771. The reaction is very similar to the
elimination of carbon monoxide from barbituric acids.
For this reaction, see: Wulff, G.; Clarkson, G. Carbo-
hydr. Res. 1994, 257, 81–95; Jursic, B. S. Tetrahedron
Lett. 2000, 41, 5325–5328. The basic ES-MS signal at
m/z=2260 corresponds to [2a-CD+L-HCHO+Na]+=[2×
973+321-30+23]. After these ES-MS signals 10 addition
signals for various HP-a-CD isomers in the complex
separated by 58 units follow the 2a-CD+1L complex
signal.
21. The values for this complex start with [3a-CD+2L-
2H2CO+2Na]2+=[3×973+2×321−2×30+2×23]2+=3547.
This amounts to m/z=1803.5, which corresponds to our
first signal at 1773. This signal is followed by 12 signals
separated by m/z=29, a half molecular weight of one
hydroxypropyl unit.
22. We were not able to detect the b-CD inclusion complex
with lorazepam due to fact that b-CD has relatively low
water solubility.
18. When the lorazepam–cyclodextrin concentration ratio is
1:100 or 1:1000 practically all lorazepam is in the inclu-
sion complex. Therefore, it is reasonable to propose that
the concentration of the inclusion complex present in the
solution is equal to initial concentrations of lorazepam
(0.1 mM). Furthermore, the hydroxypropylcyclodextrins
used are a mixture of variously substituted cyclodextrins.
For instance, HP-a-CD is a mixture of mono-, di-, three-,

6082
R. Kobetic´ et al. / Tetrahedron Letters 42 (2001) 6077–6082
1
23. Chromatographic study was performed with Merck TLC
plates with C18-silica gel (5×20 cm) with a fluorescent
indicator. Methanol was used as the solvent with slow
increases of water (gradient chromatography). A similar
solvent was used for reverse phase column chromatogra-
phy. We were not able to detect nor isolate any UV-
absorbing compound from the solution of L in aqueous
HP-a-CD except lorazepam. Therefore, we believe that
formation of acetal A under these conditions is unlikely.
24. In the first experiment lorazepam (3.2 mg; 0.01 mol) and
hydroxypropyl-a-cyclodextrin (140 mg; ꢀ0.1 mmol) in
50 mL were stirred and sonicated at room temperature
overnight. The aqueous solution was extracted with chlo-
roform, and the chloroform solution was dried over
anhydrous sodium sulfate and evaporated. The TLC
analysis in chloroform shows the formation of one domi-
nant product with Rf 0.26. Our 1H NMR is consistent
with the monochloro derivative of I for which experimen-
tal data are available (Pharmazie 1988, 43, 362). In the
second experiment the same aqueous solution with the
same composition of lorazepam and hydroxypropyl-a-
cyclodextrin was refluxed for 45 minutes. In the begin-
ning white crystals of lorazepam were on the surface of
the solution and the walls of the round-bottomed flask.
After several minutes the solution became clear. The
cooled reaction mixture was extracted with chloroform
(3×15 mL), the chloroform solution dried over anhydrous
sodium sulfate and evaporated to dryness. Considering
both TLC chromatography in chloroform and H NMR
spectra there are no traces of lorazepam in the residue.
There are three products detected by the TLC. The major
product has Rf values of 0.22, while the lorepazam value
should be 0.04 (thin-layer chromatography was per-
formed using plastic-based 0.25 mm thick silica gel 60
F-254 plates with chloroform as solvent). In the 1H NMR
(CDCl3) spectra signals that corresponds to lorazepam
such as two doublets at 5.048 (J=9 Hz) and 4.541 (J=9
Hz) for the CHOH unit are not present and aromatic
signals are shifted to higher ppm values (ꢀ0.6 ppm). This
is all in agreement with the assumption that product P is
formed in the course of the reaction.
25. Due to the low solubility of b-cyclodextrin in water the
ES-MS of aqueous 100 mM b-CD was not possible to
record. On the other hand, results from the 10 mM
aqueous b-CD ES-MS studies were not reproducible.
26. Adding glycerol to 100 mmol aqueous a- and g-cyclodex-
trin causes the precipitation of the cyclodextrin. This is
not the case with freshly prepared aqueous HP-a-CD,
HP-b-CD, and HP-g-CD. Nevertheless, the instrument
injection needles become quickly clogged and precipita-
tion is formed in the injection chamber which makes
spray inconsistent and obtained results hard to repro-
duce. Therefore, our studies with glycerol as an additive
were carried out with 10 mM aqueous hydroxypropylcy-
clodextrins where the formation of the lorazepam
cyclodextrin inclusion complexes are not as strong.
.