Diketopiperazines as a tool for the study of transport across the Blood-Brain Barrier (BBB) and their potential use as BBB-shuttles

Article abstract of DOI:10.1021/ja073522o

Here we prepared and evaluated two libraries of mono-N-methylated and di-N-methylated diketopiperazines (DKPs) by parallel artificial membrane permeability assay and immobilized artificial membrane chromatography in order to obtain information on the features that govern the passage of peptidic molecules across the blood-brain barrier (BBB) by passive diffusion. On the basis of the results from these two libraries, we prepared and evaluated several DKP-baicalin and DKP-dopamine constructs. The DKPs or cyclic dipeptide scaffolds can be considered a novel family of brain delivery systems (BBB-shuttles) to transport to the brain drugs and other cargos that cannot cross the BBB unaided.

Full text of DOI:10.1021/ja073522o

Published on Web 09/01/2007
Diketopiperazines as a Tool for the Study of Transport across
the Blood-Brain Barrier (BBB) and Their Potential Use as
BBB-Shuttles
Meritxell Teixido´,† Esther Zurita,† Morteza Malakoutikhah,† Teresa Tarrago´,† and
Ernest Giralt*,†,‡
Contribution from the Institut de Recerca Biome`dica (IRB Barcelona), Parc Cient´ıfic de
Barcelona, Josep Samitier 1-5, Barcelona, Spain, and Departament de Qu´ımica Orga`nica,
UniVersitat de Barcelona, Mart´ı i Franque`s 1, Barcelona, Spain
Received May 17, 2007; E-mail: egiralt@pcb.ub.es
Abstract: Here we prepared and evaluated two libraries of mono-N-methylated and di-N-methylated
diketopiperazines (DKPs) by parallel artificial membrane permeability assay and immobilized artificial
membrane chromatography in order to obtain information on the features that govern the passage of peptidic
molecules across the blood-brain barrier (BBB) by passive diffusion. On the basis of the results from
these two libraries, we prepared and evaluated several DKP-baicalin and DKP-dopamine constructs.
The DKPs or cyclic dipeptide scaffolds can be considered a novel family of brain delivery systems (BBB-
shuttles) to transport to the brain drugs and other cargos that cannot cross the BBB unaided.
Introduction
Several severe health disorders require treatment of the brain.
These include not only neurodegenerative diseases, such as
Parkinson and Alzheimer disease, but also CNS1 diseases, such
as schizophrenia, epilepsy, and bipolar disorder. Cancer, HIV,
and even certain aspects of obesity are also considered cerebral
pharmaceutical targets. In many cases, there are promising
compounds for their treatment; however, owing to blood-brain
barrier (BBB) transport problems, >98% of these potential
agents do not reach the drug development stage.2
Figure 1. Anatomical basis of the blood-brain barrier. Endothelial cells
Drug delivery to the brain is complex, as compounds must
cross the BBB. This barrier is a natural defense mechanism
designed to keep harmful substances out of this organ. The BBB
controls the composition of brain extracellular fluid indepen-
dently of fluctuations within blood. It is also impermeable to
of the brain capillaries form tight junctions.
many environmental compounds and drugs. Nonbrain capillaries
have paracellular pathways, while brain capillaries do not have
these pathways, because of the presence of tight junctions
between endothelial cells, which form a continuous membrane
with no fenestrations. Therefore, possible transport mechanisms
are all transcellular (Figure 1).3-5
Specific transporters mediate the access of certain crucial
molecules for the brain, such as glucose, isolated amino acids,
and ions.6,7 Other compounds or drugs are dependent on
diffusion through the lipid bilayers of endothelial membranes,
which requires that these compounds exhibit a certain degree
of lipophilicity.8-10 Here we focused our attention on the passive
diffusion mechanism, as it is a nonsaturable and spontaneous
transport process.
† Parc Cient´ıfic de Barcelona.
‡ Universitat de Barcelona.
(1) Abbreviations. ACH, R-cyano-4-hydroxy-cinnamic acid; BBEC, bovine
brain endothelial cells; 2-ClTrt, 2-chlorotrityl resin; Da, dalton; DBU, 1,8-
diazabicyclo[5.4.0]-undec-7-en; CNS, central nervous system; DCM,
dichloromethane; DEA, diethylamine; DIAD, diisopropylazodicarboxylate;
DIEA, diisopropylethylamine; DIPCDI, N,N′-diisopropylcarbodiimide;
DMAP, 4-dimethylaminopyridine; DMF, N,N-dimethylformamide; ESI-
MS, electrospray ionization-mass spectrometry; Fmoc, 9-fluorenylmethoxy-
carbonyl; 1H NMR, proton nuclear magnetic resonance; HOAt, 7-aza-1-
hydroxybenzotriazole; HOBt, 1-hydroxybenzotriazole; HPLC, high-
performance liquid chromatography; HPLC-MS, high performance liquid
chromatography-mass spectrometry; MALDI, matrix-assisted laser de-
sorption ionization; MALDI-TOF, matrix-assisted laser desorption ioniza-
tion-time-of-flight; MeCN, acetonitrile; MeOH, methanol; MRI, magnetic
resonance imaging; MS, mass spectrometry; NMP, N-methylpyrrolidone;
NMR, nuclear magnetic resonance; PET, positron emission tomography;
POP, prolyl oligopeptidase; PyAOP, (7-azabenzotriazol-1-yloxy)tripyrro-
lidinophosphonium hexafluorophosphate; PyBOP, benzotriazol-1-yl-ox-
ytripyrrolidinophosphonium hexafluorophosphate; Rf, retention factor; RP-
HPLC, reverse-phase high-performance liquid chromatography; TFA,
trifluoroacetic acid; THF, tetrahydrofurane; TLC, thin layer chromatogra-
phy; tR, retention time.
(3) Reese, T. S.; Karnovsky, M. J. J. Cell Biol. 1967, 34, 207-217.
(4) Brightmann, M. W.; Reese, T. S.; Feder, N. In Capillary Permeability;
Crone, C., Lassen, N. A., Eds.; Munksgaard: Copenhagen, 1970; pp 468-
482.
(5) Kniesel, U.; Wolburg, H. Cell Mol. Neurobiol. 2000, 20, 57-76.
(6) Tsuji, A.; Tamai, I. AdV. Drug DeliV. ReV. 1999, 36, 277-290.
(7) Tamai, I.; Tsuji, A. J. Pharm. Sci. 2000, 89, 1371-1388.
(2) Pardridge, W. M. Drug DiscoVery Today 2007, 12, 54-61
9
11802
J. AM. CHEM. SOC. 2007, 129, 11802-11813
10.1021/ja073522o CCC: $37.00 © 2007 American Chemical Society

Diketopiperazines as BBB-Shuttles
A R T I C L E S
Scheme 1. Synthetic Methodology for the Preparation of the First Library of DKPs
Given that there are several promising peptide-based drugs
for the treatment and diagnosis of diseases of the CNS,11,12
information on the regulation of peptide transport through the
BBB by passive diffusion is urgently required. This information
will contribute to the design of novel peptidic drugs with
improved transport properties and also to the design of peptidic
molecules that act as BBB-shuttles for drugs that cannot cross
the BBB unaided.
Diketopiperazines (DKPs) are very common in nature13,14 and
comprise a family of compounds with many potential applica-
tions of interest. Our attention was drawn to these compounds
because of an unexpected observation made during a previous
project performed in the field of evolutionary combinatorial
chemistry.15 During the synthesis and evaluation of several
control peptides designed to optimize various parameters of an
in vitro cellular assay, one of these peptides, the N-methylphe-
nylalanine trimer (Ac-N-MePhe-N-MePhe-N-MePhe-CONH2),
degraded to form the DKP N-MePhe-N-MePhe, which has the
capacity, by passive diffusion, to cross the BBB model of bovine
brain endothelial cells (BBEC assay).
the first alkaloid isolated from a lichen, Roccella fuciformis DC,
in 1877 by Stenhouse and Groves16 and later studied by Forster17
and Marcuccio,18 includes one based on the autoxidation of the
DKP N-MePhe-N-MePhe. In the 1980s Stermitz19 described a
new alkaloid, 2,5-dibenzyl-1,4-dimethylpiperazine, isolated from
Zanthoxylum aff. Arborescens L, a Mexican tree of the citrus
family. In the leaves and fruits of this tree, which is commonly
called “Yucu tism´ a” (herb to kill fish), a new family of alkaloids
was found, which included 2,5-dibenzyl-1,4-dimethyl-
piperazine. This compound is biosynthetically related to the DKP
Phe-Phe by N-methylation and reduction.
Results and Discussion
First Library of DKPs. A first library of 15 DKPs with
distinct side chains was prepared using a solid-phase methodol-
ogy (Scheme 1). For most cases, the only purification required
was a desalting step. The side chains were chosen in order to
explore the effect of adding or removing a methylene group,
adding a nitrogen or oxygen atom, and adding a larger aromatic
ring or a hydrophobic, but not aromatic, side chain. All DKPs
in this first library were characterized by HPLC, HPLC-MS,
and MALDI-TOF MS. All were obtained with purity higher
than 90%, except for 1PyrenylAla-N-MePhe, which was purified
in order to remove the salts and byproducts resulting from the
use of the Fmoc derivative 1PyrenylAla in its synthesis. In the
case of Phg-N-MePhe, two peaks were detected, corresponding
to the epimerization of CR of the phenylglycine (Phg) residue.
Phg is prone to racemize during coupling.20
The transport properties of DKP N-MePhe-N-MePhe had
never been studied, although descriptions of this compound had
been reported. The pathways used to synthesize picroroccellin,
(8) Van de Waterbeemd, H.; Camenisch, G.; Folkers, G.; Chretien, J. R.;
Raevsky, O. A. J. Drug Targeting 1998, 6, 151-165.
(9) Kelder, J.; Grootenhuis, P. D.; Bayada, D. M.; Delbressine, L. P.; Ploemen,
J. P. Pharm. Res. 1999, 16, 1514-1519.
(10) Doan, K. M. M.; Humpreys, J. E.; Webster, L. O.; Wring, S. A.; Shampire,
L. J.; Serabjit-Singh, C. J.; Adkinson, K. K.; Polli, J. W. J. Pharmacol.
Exp. Ther. 2002, 303, 1029-1037.
(11) Amer, D. A. M.; Irvine, G. B.; El-Agnaf, O. M. A. Exp. Brain Res. 2006,
173, 223-233.
(12) Kokkoni, N.; Stott, K.; Amijee, H.; Mason, J. M.; Doig, A. J. Biochemistry
2006, 45, 9906-9918.
(16) Stenhouse, J.; Groves, C. Justus Liebigs Ann. Chem. 1877, 185, 14.
(17) Forster, M. O.; Saville, W. B. J. Chem. Soc. 1922, 121, 816-827.
(18) Marcuccio, S. M.; Elix, J. A. Aust. J. Chem. 1985, 38, 1785-1796.
(19) Grina, J. A.; Ratcliff, M. R.; Stermitz, F. R. J. Org. Chem. 1982, 47, 2648-
2651.
(20) Wenschuh, H.; Beyermann, M.; Haber, H.; Seydel, J. K.; Krause, E.; Bienert,
M.; Carpino, L. A.; El-Faham, A.; Albericio, F. J. Org. Chem. 1995, 60,
405-410.
(13) Kwon, O. S.; Park, S. H.; Yun, B. S.; Pyun, Y. R.; Kim, C. J. J. Antibiot.
2001, 54, 179-181.
(14) Fu, X.; Ferreira, M. L. G.; Schmitz, F. J.; Kelly-Borges, M. J. Nat. Prod.
1998, 61, 1226-1231.
(15) Teixido, M.; Belda, I.; Zurita, E.; Llora, X.; Fabre, M.; Vilaro, S.; Albericio,
F.; Giralt, E. J. Pept. Sci. 2005, 11, 789-804.
9
J. AM. CHEM. SOC. VOL. 129, NO. 38, 2007 11803

A R T I C L E S
Teixido´ et al.
Figure 2. (a) Scheme of the PAMPA assay. (b) HPLC chromatogram from the mixture of the first library of DKPs in a linear gradient from 15 to 65%
MeCN in 15 min using a Symmetry C18 column (150 × 4.6 mm × 5 µm, 100 Å, Waters). (c) HPLC chromatograms of donor and acceptor PAMPA
compartments at time 0 and 4 h for the mixture of the first library of DKPs, in a linear gradient from 15 to 65% MeCN in 15 min using a Symmetry C18
column (150 × 4.6 mm × 5 µm, 100 Å, Waters). (d) Percentage of transport after 4 h in the PAMPA for the DKPs of the first library. *These compounds,
overlapped in the mixture, were isolated and analyzed. (e) Comparison of the evaluation of Phe-N-MePhe in the mixture and as an isolated compound.
BBB Transport Evaluation of the First Library of DKPs
by Parallel Artificial Membrane Permeability Assay
(PAMPA). Although the ideal assay to evaluate the capacity
of a compound to enter the CNS may be regarded as that which
most closely resembles the in vivo situation, other aspects must
be considered. In general terms, the method chosen to evaluate
BBB transport should be simple, automatizable with high
throughput and low cost, and require only a small amount of
test compound.
us to not be restricted to mimicking substrates of BBB trans-
porters. For this reason we used an evaluation method, PAMPA,
that predicts or evaluates only this transport mechanism.
The PAMPA assay, originally introduced by Kansy,21 uses
an artificial membrane in the form of filter-supported phospho-
lipid bilayers. In the case of mimicking the BBB, a porcine
polar lipid extract is used to coat the filter22 (Figure 2a). The
(21) Kansy, M.; Senner, F.; Gubernator, K. J. Med. Chem. 1998, 41, 1007-
1010.
Here we focused on the passive diffusion transport mecha-
nism, a nonsaturable and spontaneous process, thereby allowing
(22) Di, L.; Kerns, E. H.; Fan, K.; McConnell, O. J.; Carter, G. T. Eur. J. Med.
Chem. 2003, 38, 223-232.
9
11804 J. AM. CHEM. SOC. VOL. 129, NO. 38, 2007

Diketopiperazines as BBB-Shuttles
A R T I C L E S
phospholipid membrane mimics the cell membrane but has no
means for active or paracellular transport of drug molecules and
is therefore a convenient tool to evaluate the transport of
compounds by passive diffusion. Here we explored the pos-
sibility of evaluating mixtures of compounds by PAMPA. The
unstirred water layer of the standard in vitro assay does not
accurately reflect physiological conditions;23 in vivo it is
negligible, due to the small diameter of brain capillaries (∼7
µm) as well as to the efficiency of mixing due to the passage
of erythrocytes.24 Therefore, to mimic the dynamism of the in
vivo environment, we shook the plate at 100 rpm for 4 h. The
experiment was run for a time that ensures linear transport
without any back-diffusion. We recommend using these condi-
tions, which are more restrictive than those of the standard assay,
as they provide more reliable data. Last, we also evaluated
control molecules with known high brain penetration capacities,
propranolol (a well-known â-adrenergic receptor blocker) and
carbamazepine (an anticonvulsant and mood-stabilizing drug,
used primarily in the treatment of epilepsy and bipolar disorder).
increase the half-life of these compounds in the body, thereby
avoiding degradation by peptidases in the blood or associated
with cerebral microvessels.25 Several authors24,26-34 have pro-
posed that the introduction of a N-methyl group increases BBB
transport capacity. However, we wanted to check whether this
second library showed improved transport capacity or whether
they became too hydrophobic and were retained in the mem-
brane.
The N-methyl amino acids were incorporated using commer-
cially available building blocks or via N-methylation on solid-
phase35,36 using the corresponding non-N-methylated amino acid.
The yields of this second library were excellent, corroborating
the high tendency of N-methyl amino acids to form DKPs. A
comparison of the DKPs that had both amide bonds N-
methylated with those that only had one N-methylated amide
bond indicated that the presence of the second N-methyl group
in the amide bond improves BBB transport capacity (Phe,
HomoPhe) because it reduces the number of potential hydrogen
bonds that the DKP can form. However, in the case of highly
hydrophobic DKPs (2Nal, Cha, Oct, 1PyrenylAla), the addition
of a second N-methyl group can be a disadvantage, because
the corresponding DKP is more retained in the membrane. The
positive transport measured for these compounds by the PAMPA
assay as compared to the control compounds, indicated that these
compounds may cross the BBB in vivo by passive diffusion.
Moreover, comparison of the DKP scaffold (N-MePhe-N-
MePhe) to the corresponding linear peptide (Ac-N-MePhe-N-
MePhe-CONH2) highlights that the DKP is a privileged scaffold
in terms of BBB penetration potential.
BBB Transport Evaluation by Immobilized Artificial
Membrane Chromatography (IAMC). Another technique to
evaluate the passive diffusion is IAMC. This simple, fast, and
inexpensive chromatographic technique was also used to evalu-
ate our libraries (Figure 3). Developed by Pidgeon,37 IAMC
uses phospholipid molecules covalently immobilized to silica
particles at high density as the stationary phase. It has been
used to purify membrane proteins,38 to immobilize enzymes,39
and to predict transport across biological barriers.40 IAMC
exhibits a good correlation40 with in vitro cell-based assays41
and is convenient in terms of high throughput. IAMC interac-
tions include ionic, lipophilic, and hydrogen-bonding interac-
tions, which can be combined under a parameter known as
phospholipophilicity.
The DKP compounds of the first library were evaluated as a
mixture, which has the advantage of speeding up the high-
throughput of the PAMPA assay and facilitating comparison
between the compounds and establishment of rules.
A mixture of the 15 DKPs was prepared at a concentration
of 200 µM each. The mixture was analyzed by HPLC (Figure
2b) and the PAMPA assay was performed (Figure 2c,d). As a
control, the DKP Phe-N-MePhe was evaluated separately in
another well of the same PAMPA plate. The same transport
properties were found when this compound was examined alone
or in a mixture (Figure 2e). The compounds that overlapped in
the mixture were evaluated alone in another well of the same
PAMPA plate.
From the mono-N-methylated DKPs in the first library, Oct-
N-MePhe (39.4%), Cha-N-MePhe (35.1%), and 2Nal-N-MePhe
(33.7%) followed by HomoPhe-N-MePhe (26.8%) and Phe-N-
MePhe (20.8%) were the best in terms of percentage of transport
after 4 h. N-MePhe-N-MePhe and Tic-N-MePhe, which are
closely related to Phe-N-MePhe, but with an extra N-methyl
group, showed greater transport in the PAMPA assay compared
with Phe-N-MePhe. However, it was observed that increasing
flexibility (from Phg-N-MePhe to Phe-N-MePhe and then to
HomoPhe-N-MePhe) did not always lead to improved transport
of the compounds (for example, Tic-N-MePhe vs N-MePhe-N-
MePhe). It is also of interest to note the distinct transport
capacity of the two diastereomers of Phg-N-MePhe. Our results
indicate that the PAMPA assay can be used to evaluate the
transport of mixtures of compounds by passive diffusion. Using
this assay, we determined that the first library of DKPs contains
compounds with permeability values close to those of propra-
nolol (27.1%) and carbamazepine (26.4%).
(25) El-bacha, R. S.; Minn, A. Cell. Mol. Biol. 1999, 45, 15-23.
(26) Levin, L. A. J. Med. Chem. 1980, 23, 682-684.
(27) Bernards, M. D.; Hill, H. F. Anesthesiology 1992, 77, 750-756.
(28) Macheras, P. E.; Symillides, M. Y. Biopharm. Drug. Dispos. 1989, 10,
43-53.
(29) Nook, T.; Doelker, E.; Buri, P. Int. J. Pharm. 1988, 43, 119-129.
(30) Komiya, I.; Park, J. Y.; Kamani, A.; Ho, N. F. H.; Higuchi, W. I. Int. J.
Pharm. 1980, 4, 249-262.
(31) Ho, N. F. H.; Park, J. Y.; Morozowich, W.; Higuchi W. I. In Design of
Biopharmaceutical Properties through Prodrugs and Analogs; Roche, E.
B., Ed.; American Pharmaceutical Association: Washington, DC, 1977;
pp 136-227.
Synthesis and Evaluation of a Second Library of DKPs.
Not all the DKPs of the first library showed efficient transport
across the BBB PAMPA model. After choosing those with the
best capacity, we designed and prepared a second library, in
this case di-N-methylated DKPs, which were also synthesized
in solid-phase (Scheme 2) and evaluated by PAMPA (Table
1). The introduction of a second N-methyl group was done to
(32) Van Bree, J. B. M.; De Boer, A. G.; Danhof, M.; Ginsel, L. A.; Breimer,
D. D. J. Pharmacol. Exp. Ther. 1988, 247, 1233-1239.
(33) Tavakoli-Saberi, M. R.; Audus, K. L. Int. J. Pharm. 1989, 56, 135-142.
(34) Kaiser, E. T.; Kadzy, F. J. Ann. ReV. Biophys. Chem. 1987, 16, 561-581.
(35) Biron, E.; Chatterjee, J.; Kessler, H. J. Pept. Sci. 2006, 12, 213-219.
(36) Yang, L.; Chiu, K. Tetrahedron Lett. 1997, 38, 7307-7310.
(37) Pidgeon, C.; Venkataran, U. V. Anal. Biochem. 1989, 176, 36-47.
(38) Pidgeon, C.; Stevens, J.; Otto, S.; Jefcoate, C.; Marcus, C. Anal. Biochem.
1991, 194, 163-173.
(39) Chui, W. K.; Wainer, I. Anal. Biochem. 1992, 201, 237-245.
(40) Braddy, A. C.; Janaky, T.; Prokai, L. J. Chromatogr. A 2002, 966, 81-87.
(41) Pidgeon, C.; Ong, S.; Liu, H.; Qiu, X.; Pidgeon, M.; Dantzig, A. H.; Munroe,
J.; Hornback, W. J.; Kasher, J. S. J. Med. Chem. 1995, 38, 590-594.
(23) Nielsen, E.; Avdeef, A. Eur. J. Pharm. Sci. 2004, 22, 33-41.
(24) Pardridge, W. M. Peptide Drug DeliVery to the Brain; Raven Press: New
York, 1991; pp 1-357.
9
J. AM. CHEM. SOC. VOL. 129, NO. 38, 2007 11805

A R T I C L E S
Teixido´ et al.
Scheme 2. Synthetic Methodology for the Preparation of the Library of Di-N-methylated DKPs
Table 1. Percentage of Transport after 4 h, Effective Permeability
mixture by IAMC is possible using HPLC-MS; however, it is
not an easy task when one wishes to evaluate phospholipophi-
licity under physiological conditions, i.e., saline phosphate buffer
(PBS) at pH 7.4.
(Pe and log Pe), and Percentage of Retention in the Membrane in
the PAMPA Assay for the DKPs X-N-MePhe and N-MePhe-N-MeX
and Control Compounds Such as Propranolol, Carbamazepine,
and the Linear Peptide Ac-N-MePhe-N-MePhe-CONH2
The preparation and evaluation of these DKPs has helped to
establish a number of rules about the passive transport across
the BBB that will guide us in the design of a BBB-shuttle.
%
% retention
in the
transport
at 4 h
compound
Pe (cm/s)
log Pe membrane
Phe-N-MePhe
HomoPhe-N-MePhe
2Nal-N-MePhe
Cha-N-MePhe
Oct-N-MePhe
1PyrenylAla-N-MePhe
N-MePhe-N-MePhe
N-MeHomoPhe-N-MePhe
N-Me2Nal-N-MePhe
N-MeCha-N-MePhe
N-MeOct-N-MePhe
N-Me1PyrenylAla-N-MePhe
Ac-N-MePhe-N-MePhe-CONH2
propranolol
20.8 1.28 × 10-5 -4.9
26.8 1.82 × 10-5 -4.7
33.7 2.65 × 10-5 -4.6
35.1 2.87 × 10-5 -4.5
39.4 3.68 × 10-5 -4.4
5.0 0.12 × 10-5 -5.9
37.4 3.40 × 10-5 -4.5
40.2 3.86 × 10-5 -4.4
38.9 3.57 × 10-5 -4.4
34.6 2.79 × 10-5 -4.5
32.3 2.45 × 10-5 -4.6
0.5 0.02 × 10-5 -6.6
6.8 0.33 × 10-5 -5.5
27.1 2.2 × 10-5 -4.6
26.4 1.9 × 10-5 -4.7
3
0
3
4
1
76
2
0
7
17
26
83
29
22
19
carbamazepine
Most of the DKPs are in the region where a positive in vivo
CNS transport is predicted, except for 1PyrenylAla-N-MePhe
and N-Me-1PyrenylAla-N-MePhe, as these interact too strongly
with the phospholipids, which leads to retention in the mem-
brane, as observed in the PAMPA assay. The evaluation of the
Figure 3. Comparison of the values measured by PAMPA (effective
permeability, Pe) and by IAMC (capacity factor taking into account
molecular weight, K′IAM/MW4) for the DKPs and control compounds in
Table 1.
9
11806 J. AM. CHEM. SOC. VOL. 129, NO. 38, 2007

Diketopiperazines as BBB-Shuttles
A R T I C L E S
Scheme 3. Synthetic Methodology for the Preparation of DKP-Baicalin Constructs
DKPs as a BBB-Shuttle. Many active compounds cannot
cross the BBB to reach their target sites inside the CNS. A BBB-
shuttle based on a DKP or cyclic dipeptide scaffold could be
used to send or shuttle anticancer drugs, antiretroviral agents,
and other compounds that cannot cross the BBB, such as
dopamine or MRI and PET diagnostic agents, thereby avoiding
the undesired side effects of current therapy and providing a
simpler and noninvasive technique than those currently used.
The BBB-shuttle cyclic dipeptide is a nonviral agent that
could represent a new solution to bypass the problem of bio-
availability of certain types of potential CNS therapeutic drugs.
The BBB-shuttle could be useful in several therapeutic fields.
For example, the HIV virus can cross the BBB either during
primary infection or at a later stage.42-44 The brain might
therefore serve as an hidden reservoir for HIV viral replica-
tion.45,46 HIV in the brain and in the cerebrospinal fluid may
be particularly resistant to therapy, because of the failure of
antiretroviral drugs to penetrate the BBB.47 The resulting
infection leads to a number of CNS disorders, such as AIDS
dementia complex and HIV encephalopathy.48 Antiretroviral
agents that cross the BBB or BBB-shuttles able to transport
promising antivirals that cannot cross this barrier alone are
urgently required.
Baicalin, a flavonoid isolated from the plant Scutellaria
baicalensis Georgi, traditionally used in Chinese medicine, is
an example of a compound with interesting properties. It is a
non-nucleoside inhibitor of HIV-1 reverse transcriptase (NNRTI)
and an inhibitor of the prolyl oligopeptidase associated with
schizophrenia and bipolar affective disorder, but it cannot cross
the BBB.49-51
Another field of interest of a BBB-shuttle is brain tumor
cancer treatment, as many chemotherapy drugs are not suitable
for this therapy because they do not have the capacity to cross
the BBB.52 A BBB-shuttle would also be of interest in the field
of psychiatric diseases, such as schizophrenia, in epilepsy, and
in neurodegenerative disorders, such as Parkinson’s disease. As
a result of the lack of transport of dopamine across the BBB,
L-dopa, a precursor of dopamine, has been used in the treatment
of Parkinson’s disease since the 1960s. However, severe side
effects are frequently observed within a few years of L-dopa
therapy.53
To date, the methods used to administer drugs to the brain,
either for therapy or diagnosis, are sometimes invasive, such
(42) Pereira, C. F.; Nottet, H. S. L. M. Science Online: NeuroAIDS 2000, 3,
1-8.
(49) Kitamura, K.; Honda, M.; Yoshizaki, H.; Yamamoto, S.; Nakane, H.;
Fukushima, M.; Ono, K.; Tokunaga, T. AntiVir. Res. 1998, 37, 131-140.
(50) Tsai, P. L.; Tsai, T. H. Planta Med. 2004, 70, 1069-1074.
(51) Tarrago, T.; Kichik, N.; Segu´ı, J.; Giralt, E. ChemMedChem 2007, 2, 354-
359.
(52) Fortin, D.; Desjardins, A.; Benko, A.; Niyonsega, T.; Boudrias, M. Cancer
2005, 103, 2606-15.
(53) Quattrini, A.; Paggi, A.; Forastieri, L.; Del Pesce, M.; Di Bella, P. RiV.
Patol. NerV. Ment. 1979, 99, 289-97.
(43) Falangota, M. F.; Hanly, A.; Galvao-Castro, B.; Petito, C. K. J. Neuropathol.
Exp Neurol. 1995, 54, 497-503.
(44) Broadwell, R. D.; Sofroniew, M. V. Exp. Neurol. 1993, 120, 245-263.
(45) Anthonypillai, C.; Gibbs, J. E.; Thomas, S. A. Cerebrospinal Fluid Res.
2006, 3, 1.
(46) Schrager, L. K.; D’, Souza, M. P. J. Am. Med. Assoc. 1998, 280, 67-71.
(47) Glynn, S. L.; Yazdanian, M. J. Pharm. Sci. 1998, 87, 306-310.
(48) Portegies, P. J. Acquir. Immune Defic. Syndr. 1994, 7, 38-48.
9
J. AM. CHEM. SOC. VOL. 129, NO. 38, 2007 11807

A R T I C L E S
Teixido´ et al.
Scheme 4. Synthetic Methodology for the Preparation of DKP-Dopamine Constructs
as intracranial administration in the case of water-soluble
anticancer drugs. In addition, temporary alteration of BBB
integrity by osmotic disruption has been used to supplement
chemotherapy for certain patients with brain tumors.54 In other
cases, drug administration to the brain has undesired side effects
because of the high doses used to overcome the low permeability
across this barrier. Sometimes the drug can be modified to
improve its transport across the BBB,55-61 by reducing drug
size, increasing drug lipophilicity, etc. Administration of the
drug to the brain by conjugation to a biological carrier62 and
intranasal delivery63 are currently being explored to tackle BBB
permeability to drugs.
that showed the greatest capacity to cross the BBB as BBB-
shuttles. These DKPs were modified with adequate functional
groups for covalent fixation of the cargo. A DKP (BBB-shuttle)
with the capacity to link a cargo with an amino or hydroxylic
group and a DKP with the capacity to link a cargo with a
carboxylic group were prepared. The cargo was linked to the
BBB-shuttle such that, once inside the brain, it would not have
to be detached from the shuttle to be active. Nonetheless, for
future applications, the chemistry of the linkage could be
optimized to enable facile hydrolysis of this bond in the
appropriate location.64
Baicalin and dopamine were chosen as cargo. Baicalin was
linked through the carboxylic group present in its sugar moiety,
which is not needed for activity.65 Dopamine was kept with its
hydroxylic groups unprotected and was linked through its amino
group.
The DKP-cargo constructs were prepared for the best side
chains, Phe, HomoPhe, Cha, Oct, 2Nal, 1PyrenylAla, following
the solid-phase Scheme 3 (for baicalin) and Scheme 4 (for
dopamine). The DKP-baicalin and DKP-dopamine constructs
were synthesized, purified, and characterized by MALDI-TOF,
HPLC-MS, and HPLC and then evaluated by the PAMPA assay
(Tables 2 and 3). Scutellariae radix extract and isolated baicalin
The use of a BBB-shuttle approach implies that the search
for novel drugs does not become limited only to compounds
that have the capacity to cross the BBB. We selected the DKPs
(54) Kroll, R. A.; Neuwelt, E. A. Neurosurgery 1998, 42, 1083-1100.
(55) Kawai, M.; Fukuta, N.; Ito, N.; Kagami, T.; Butsugan, Y.; Maruyama, M.;
Kudo, Y. Int. J. Pept. Protein Res. 1990, 35, 452-459.
(56) Witt, K. A.; Slate, C. A.; Egleton, R. D.; Huber, J. D.; Yamamura, H. I.;
Hruby, V. J.; Davis, T. P. J. Neurochem. 2000, 75, 424-435.
(57) Weber, S. J.; Greene, D. L.; Sharma, S. D.; Yamamura, H. I.; Kramer, T.
H.; Burks, T. F.; Hruby, V. J.; Hersh, L. B.; Davis, T. P. J. Pharm. Exp.
Ther. 1991, 259, 1109-1117.
(58) Gentry, C. L.; Egleton, R. D.; Gillespie, T.; Abbruscato, T. J.; Bechowski,
H. B.; Hruby, V. J.; Davis, T. P. Peptides 1999, 20, 1229-1238.
(59) Delaet, N. G. J.; Verheyden, P.; Velkeniers, B.; Hooghe-Peters, E. L.; Bruns,
C.; Tourwe, D.; Van binst, G. Pept. Res. 1993, 6, 24-30.
(60) Wong, A.; Toth, I. Curr. Med. Chem. 2001, 8, 1123-1136.
(61) Polt, R.; Palian, M. M. Drugs Future 2001, 26, 561-576.
(62) Pardridge, W. M. Nat. ReV.: Drug DiscoVery 2002, 1, 131-139.
(63) Illum, L. J. Pharm. Pharmacol. 2004, 56, 3-17.
(64) Prokai, L.; Prokai-Tatrai, K.; Ouyang, X.; Kim, H.-S.; Wu, W.-M.;
Zharikova, A.; Bodor, N. J. Med. Chem. 1999, 42, 4563-4571.
(65) Xing, J.; Chen, X.; Sun, Y.; Luan, Y.; Zhong, D. J. Pharm. Pharmacol.
2005, 57, 743-750.
9
11808 J. AM. CHEM. SOC. VOL. 129, NO. 38, 2007

Diketopiperazines as BBB-Shuttles
A R T I C L E S
Figure 4. HPLC chromatograms of donor and acceptor PAMPA compartments at 4 h for the S. radix extract and the isolated baicalin, in a linear gradient
from 0 to 100% MeCN in 15 min using a Symmetry C18 column (150 × 4.6 mm × 5 µm, 100 Å, Waters).
Table 2. Percentage of Transport after 4 h and Effective
Permeability (Pe and log Pe) Found in the PAMPA for
N-MeOct and N-Me2Nal have been shown to be the most
DKP-Baicalin Constructs and Baicalin (Cargo)
the artificial model BBB membrane. The constructs based on
promising ones for the transport of dopamine and baicalin.
However, the optimal BBB-shuttle may vary with the cargo.
%
transport
In both cases, the linkage between the BBB-shuttle and the
compound
at 4 h
Pe (cm/ s)
log Pe
cargo stayed intact during diffusion through the PAMPA model.
Moreover, the POP inhibition activity of the DKP Phe(p-NH-
CO-baicalin)-N-MePhe was checked (48% at 100 µM) and was
found to be as good as that of free baicalin (51% at 100 µM).
We also established methodology for the preparation of DKP-
shuttles for other cargo, such as anticancer drugs, nanoparticles,
and MRI and PET diagnostic agents (See the Supporting
Information).
a
a
a
baicalin
-
-
-
DKP Phe(p-NH-CO-baicalin)-N-MePhe
DKP Phe(p-NH-CO-baicalin)-N-
MeHomoPhe
DKP Phe(p-NH-CO-baicalin)-N-MeCha
DKP Phe(p-NH-CO-baicalin)-N-Me2Nal
DKP Phe(p-NH-CO-baicalin)-N-MeOct
DKP Phe(p-NH-CO-baicalin)-N-Me-
1PyrenylAla
0.03
1.33 × 10-8 -7.7
a
a
a
-
-
-
0.76
7.22
1.04
0.01
3.65 × 10-7 -6.4
3.70 × 10-6 -5.4
4.97 × 10-7 -6.3
0.19 × 10-8 -8.7
a These compounds were not detected in the acceptor well of the PAMPA
after 4 h.
Experimental Section
Materials and Methods. Protected amino acids, handles, and resins
were supplied by Luxembourg Industries (Tel-Aviv, Israel), Neosystem
(Strasbourg, France), Calbiochem-Novabiochem AG (Lau¨felfingen,
Switzerland), Bachem AG (Bubendorf, Switzerland), or Iris Biotech
(Marktredwitz, Germany). PyBOP was supplied by Calbiochem-
Novabiochem AG. DIEA, DEA, DIPCDI, DMAP, KCN, KHSO4,
ninhydrin, and â-mercaptoethanol were obtained from Fluka Chemika
(Buchs, Switzerland). HOBt was purchased from Albatros Chem. Inc.
(Montreal, Canada). PyAOP was supplied by Applied Biosystems and
HOAt from GL Biochem Shanghai Ltd (Shanghai, China). Solvents
for peptide synthesis and RP-HPLC were obtained from Scharlau or
SDS (Barcelona, Spain). Trifluoroacetic acid was supplied by Kali-
were also evaluated by the PAMPA assay (Figure 4). S. radix
extract and isolated baicalin did not cross the PAMPA assay,
but the BBB-shuttle constructs based on the DKP did carry
baicalin through this assay (Table 2). The transport of the com-
pounds was measured by HPLC of the acceptor compartments
and confirmed by HPLC-MS and/or MALDI-TOF spectrom-
etry, which also showed that there was no loss in the integrity
of the DKP-cargo constructs during the PAMPA assay.
Although our DKP-cargo constructs enabled only limited
transport of dopamine and baicalin, they nevertheless greatly
increased the transport of these challenging compounds across
9
J. AM. CHEM. SOC. VOL. 129, NO. 38, 2007 11809

A R T I C L E S
Teixido´ et al.
Table 3. Percentage of Transport after 4 h and Effective Permeability (Pe and log Pe) in the PAMPA for DKP-Dopamine Constructs and
Also Dopamine (Cargo) and L-Dopa (Current Drug Used in the Market)
%
compound
transport at 4 h
Pe (cm/ s)
log Pe
a
a
a
a
dopamine
L-dopa
-
-
-
-
-
a
a
-
DKP Phe(p-NH-CH2-CO-NH-dopamine)-N-MePhe
DKP Phe(p-NH-CH2-CO-NH-dopamine)-N-MeHomoPhe
DKP Phe(p-NH-CH2-CO-NH-dopamine)-N-MeCha
DKP Phe(p-NH-CH2-CO-NH-dopamine)-N-Me2Nal
DKP Phe(p-NH-CH2-CO-NH-dopamine)-N-MeOct
DKP Phe(p-NH-CH2-CO-NH-dopamine)-N-Me-1PyrenylAla
0.08
0.02
0.05
0.24
0.67
0.04
3.61 × 10-8
0.76 × 10-8
2.28 × 10-8
1.15 × 10-7
3.21 × 10-7
1.90 × 10-8
-7.4
-8.1
-7.6
-6.9
-6.5
-7.7
a These compounds were not be detected in the acceptor well of the PAMPA after 4 h.
Chemie (Bad Wimpfen, Germany). Other chemicals used were
purchased from Aldrich (Milwaukee, WI) and were of the highest purity
commercially available. All commercial reagents and solvents were
used as received, with the exception of DCM and DMF. DMF was
bubbled with nitrogen to remove volatile contaminants and stored over
activated 4 Å molecular sieves (Merck, Darmstad, Germany). DCM
was passed through a short column of Al2O3 (in the case of DCM used
for peptide synthesis). PAMPA plates and PAMPA system solution
were from pION (Woburn, MA USA). Porcine polar brain lipid extract
(PBLEP) was purchased from Avantis Polar Lipids (Alabaster, AL).
IAM column (10 × 4.6 mm, 12 µm, 300 Å, IAM.PC.DD2 column)
was from Regis Technologies Inc. (Morton Grove, IL). Mass spectra
were recorded on a MALDI Voyager DE RP time-of-flight (TOF)
spectrometer (PE Biosystems, Foster City, CA), using an ACH matrix.
HPLC chromatograms were recorded on a Waters model Alliance 2695
with a photodiode array detector 996 from Waters (Waters, Milford,
MA) using a Symmetry C18 column (150 × 4.6 mm × 5 µm, 100 Å,
Waters), H2O (0.045% TFA) and MeCN (0.036% TFA) as solvents at
1 mL/min, and Millenium version 4.0 software. HPLC-MS [Waters
model Alliance 2796, quaternary pump, UV/vis dual λ absorbance
detector Waters 2487, ESI-MS model Micromass ZQ and Masslynx
version 4.0 software (Waters)] was done using a Symmetry 300 C18
column (150 × 3.9 mm × 5 µm, 300 Å, Waters) with H2O (0.1%
formic acid) and MeCN (0.07% formic acid) as solvents at 1 mL/min.
The products were purified in a Waters 600 with a dual λ absorbance
detector (Waters 2487, Waters) and a Symmetry C18 column (100 ×
30 mm × 5 µm, 100 Å, Waters) with H2O (0.1% TFA) and MeCN
(0.05% TFA) as solvents at 10 mL/min. 1H NMR spectra were obtained
at 25 °C with a Varian Mercury 400 NMR spectrometer using CD3OD
as a solvent.
50:0.1, v/v). 1H NMR (CD3OD, 400 MHz, δ ppm): 4.98 (s, 2H), 7.96
(d, J ) 21, 1H), 8.26 (dd, J ) 4, J ) 20, 1H), 1.56-1.24 (m, 12H),
8.56 (d, J ) 4, 1H).
General Protocols for Solid-Phase Synthesis. Syntheses were
performed on a 100 µmol scale/each, in all cases L-amino acids were
used. Solid-phase peptide elongation and other solid-phase manipula-
tions were done manually in polypropylene syringes, each fitted with
a polyethylene porous disk. Solvents and soluble reagents were removed
by suction. Washings between synthetic steps were done with DMF (5
× 0.5 min) and DCM (5 × 0.5 min) using 5 mL of solvent/g of resin
each time. During couplings the mixture was allowed to react with
intermittent manual stirring.
Identification Tests. The Kaiser colorimetric assay was used for
the detection of solid-phase-bound primary amines,66 while the De
Clercq test was used for secondary amines bound to solid-phase.67
Initial Conditioning of Resin. The p-MBHA resin was conditioned
by washing with DCM (5 × 30 s) followed by a 40% TFA solution in
DCM (1 × 30 s and 2 × 5 min). This acid treatment was followed by
a neutralization step with DIEA 5% in DCM (3 × 2 min) and finally
the resin was washed with DCM (5 × 30 s).
Coupling of Nbb Handle to p-MBHA Resin. Nbb handle (97 mg,
5 equiv) in DCM (3 mL/g of resin) and DIPCDI (77 µL, 5 equiv) were
sequentially added to the p-MBHA resin (164 mg). The mixture was
left to stir overnight. The solvent was then removed by suction, the
resin was thoroughly washed, and the reaction was monitored with the
Kaiser test.
Boc Group Removal. The Boc group was removed by treating the
resin with 40% (v/v) TFA in DCM (3-4 mL/g of resin, 2 × 10 min).
In solution, the Boc group was removed using 50% TFA in DCM for
1 h. DCM was then evaporated under reduced pressure.
Fmoc Group Removal. The Fmoc group was removed by treating
the resin with 20% piperidine in DMF (3-4 mL/g of resin, 2 × 1 min
and 1 × 20 min). To remove the Fmoc group in solution, the compound
was dissolved in DMF (100 mg/mL) and the solution was kept at
0 °C; DEA was then added to reach 5%. The reaction was left to reach
room temperature and then left to stand at room temperature for 3 h
under stirring. To remove the base and the remaining Fmoc byproduct,
the compound was purified.
Coupling Methods: Method 1, Coupling of the First Amino Acid
onto the Nbb Handle. Amino acid derivative (4 equiv), DIPCDI (62
µL, 4 equiv), and DMAP (5 mg, 0.4 equiv) in DCM (1-3 mL/g of
resin) were sequentially added to the resin. The mixture was allowed
to react with intermittent manual stirring for 30 min. The solvent was
then removed by suction, the resin was thoroughly washed, and the
coupling was repeated once.
ZGP-AMC was obtained from Bachem (Bubendorf, Switzerland).
POP was obtained by expression in Escherichia coli and subsequent
affinity purification using an His tail fusion as reported previously.51
Fluorescence was measured using a Bio-Tek FL600 fluorescence plate
reader (Bio-Tek Instruments, Winooski, VT).
Synthesis of Nbb Handle. 4-Bromomethyl-3-nitrobenzoic acid (4.5
g, 17.3 mmols) was placed in a round-bottom flask provided with reflux
system, and an aqueous NaHCO3 saturated solution was added (135
mL). The reaction was stirred and allowed to stand at 90-95 °C. The
progress of the reaction was monitored by TLC (chloroform/MeOH/
AcOEt, 100:50:0.1, v/v) or by HPLC of an acidified aliquot in a gradient
from 0 to 100% of MeCN in 15 min in a Symmetry C18 column (150
× 4.6 mm × 5 µm, 100 Å, Waters). Retention time for the final product
was 7.1 min; retention time for the initial product was 10.4 min. The
reaction was completed in 30 min. The solution obtained was
hot-filtered, the reaction was quenched with HCl (12N) to pH 1-2,
the crude product was extracted with AcOEt (3 × 100 mL), and the
organic fractions were combined, washed with a saturated NaCl aqueous
solution, and dried over anhydrous MgSO4. The AcOEt was evaporated
in vacuo. The product was obtained as a pale yellow solid. Yield: 80%,
2.73 g. HPLC tR: 7.12 min (linear gradient 0-100% MeCN in 15 min).
HPLC-MS: 196 Da. TLC: Rf ) 0.3 (chloroform/MeOH/AcOEt, 100:
Method 2, Coupling of the Second Amino Acid. Protected amino
acid (5 equiv) and PyAOP (260 mg, 5 equiv) in DMF (1-3 mL/g of
(66) Kaiser, E.; Colescott, R. L.; Bossiger, C. D.; Cook, P. I. Anal. Biochem.
1970, 34, 595-598.
(67) Madder, A.; Farcy, N.; Hosten, N. G. C.; De Muynck, H.; De Clercq, P.
J.; Barry, J.; Davis, A. P. Eur. J. Org. Chem. 1999, 2787-2791. Note:
The p-nitrophenyl ester of disperse red 1 is prepared from the commercially
available disperse 1.
9
11810 J. AM. CHEM. SOC. VOL. 129, NO. 38, 2007

Diketopiperazines as BBB-Shuttles
A R T I C L E S
resin) were sequentially added to the resin followed with DIEA (255
µL, 15 equiv). The mixture was left to react with intermittent manual
stirring for 1 h. The solvent was removed by filtration, the resin was
washed as indicated above, and the coupling was repeated two more
times. The extent of the coupling was checked by the De Clercq test.
Amino Acid N-Alkylation. The N-methylation of the amino acid
derivatives was performed using the methods described by Biron et
al.35 and Yang et al.36 This process can be divided in to three steps:
(A) protection and activation with O-NBS, (B) Mitsunobu reaction,
and (C) O-NBS removal.
(A) Protection and Activation with O-NBS. To perform the
protection, O-NBS (88 mg, 4 equiv) and collidine (170 µL, 10 equiv)
in NMP were added to the resin. The reaction was left with intermittent
manual stirring for 1 h and this step was repeated once and checked
by the Kaiser test.
(B) Mitsunobu Reaction. The Mitsunobu reagents, triphenylphos-
phine (133 mg, 5 equiv), and MeOH (24 µL, 10 equiv) in dry THF
were added to the resin and left for 1 min; afterward, without filtering,
DIAD (98 µL, 5 equiv) was added in dry THF and left for another 10
min.
(C) O-NBS Removal. To proceed to O-NBS removal, â-mercap-
toethanol (70 µL, 10 equiv) and DBU (75 µL, 5 equiv) in NMP were
added to the resin, and the mixture was left to react for 5 min under an
argon atmosphere. This process was repeated once.
Phe-N-MePhe, 9.7 min; HomoPhe-N-MePhe, 10.3 min; Cha-N-MePhe,
11.1 min; 2Nal-N-MePhe, 11.3 min; Oct-N-MePhe, 11.4 min; 1Pyre-
nylAla-N-MePhe, 13.5 min; PyrAla-N-MePhe, 5.9 min; Gly-N-MePhe,
6.7 min; Pro-N-MePhe, 7.8 min; Phg-N-MePhe, 8.8 and 9.4 min;
N-MePhe-N-MePhe, 10.6 min; Pro(Phe)-N-MePhe, 10.8 min; Tic-N-
MePhe, 10.4 min; Tyr-N-MePhe, 8.3 min; N-MeTrp restricted-N-
MePhe, 10.6 min. Mass spectrometry (MALDI-TOF): Phe-N-MePhe,
309.0 Da; HomoPhe-N-MePhe, 325.2 Da; Cha-N-MePhe, 315.2 Da;
2Nal-N-MePhe, 359.7 Da; Oct-N-MePhe, 303.2 Da; 1PyrenylAla-N-
MePhe, 433.2 Da; PyrAla-N-MePhe, 310.1 Da; Gly-N-MePhe, 219.6
Da; Pro-N-MePhe, 258.5 Da; Phg-N-MePhe, 295.6 Da; N-MePhe-N-
MePhe, 323.2 Da; Pro(Phe)-N-MePhe, 335.1 Da; Tic-N-MePhe, 321.7
Da; Tyr-N-MePhe, 325.7 Da; N-MeTrp restricted-N-MePhe, 358.5 Da.
Yield: Phe-N-MePhe, 5.2%; HomoPhe-N-MePhe, 24%; Cha-N-MePhe,
5.1%; 2Nal-N-MePhe, 4.5%; Oct-N-MePhe, 9.9%; 1PyrenylAla-N-
MePhe, 4.2% (after RP-HPLC purification), PyrAla-N-MePhe, 7.8%;
Gly-N-MePhe, 8.3%; Pro-N-MePhe, 15%; Phg-N-MePhe, 33%; N-
MePhe-N-MePhe, 33%; Pro(Phe)-N-MePhe, 19%; Tic-N-MePhe, 5.6%;
Tyr-N-MePhe, 41%; N-MeTrp restricted-N-MePhe, 7.6%.
Characterization of the Second Library of DKPs. Reverse phase
HPLC: linear gradient from 0 to 100% MeCN in 15 min using a
Symmetry C18 column (150 × 4.6 mm × 5 µm, 100 Å, Waters). tR:
N-MePhe-N-MePhe, 10.4 min; N-MePhe-N-MeHomoPhe, 11.2 min;
N-MePhe-N-Me2Nal, 11.9 min; N-MePhe-N-MeOct, 12.4 min; N-
MePhe-N-MeCha, 12.5 min; N-MePhe-N-Me-1PyrenylAla, 14.0 min.
Mass spectrometry (MALDI-TOF): N-MePhe-N-MePhe, 323.2 Da;
N-MePhe-N-MeHomoPhe, 337.1 Da; N-MePhe-N-Me2Nal, 373.2 Da;
N-MePhe-N-MeOct, 317.2 Da; N-MePhe-N-MeCha, 329.2 Da; N-Me-
Phe-N-Me-1PyrenylAla, 447.2 Da. Yield: N-MePhe-N-MePhe, 50%;
N-MePhe-N-MeHomoPhe, 75%; N-MePhe-N-Me2Nal, 61%; N-MePhe-
N-MeOct, 90%; N-MePhe-N-MeCha, 95%; N-MePhe-N-Me-1Pyreny-
lAla, 60%.
Introduction of the Anchoring Moiety by Reductive Amination.
The anchoring moiety was introduced by adding glyoxylic acid (37
mg, 5 equiv) in DMF with 1% AcOH to the resin and leaving it for 30
min. The solvent and reagents were removed by filtration, the resin
was washed, and then NaBH3CN (19 mg, 3 equiv) in DMF with 1%
AcOH was added to perform the reduction for 1 h.
Coupling of the Cargo. Depending on the nature of the cargo, it
was linked to the BBB-shuttle through several types of chemical bonds.
For Cargo with a Carboxylic Group (e.g., Baicalin). The
corresponding BBB-shuttle provided with a NH2 group was used and
was reacted with cargo-COOH (224 mg of baicalin, 5 equiv), using
PyBOP (260 mg, 5 equiv) and HOAt (196 mg, 15 equiv) as coupling
reagents and DIEA (256 µL, 15 equiv) as a base in DMF for 2 h.
For Cargo with an Amine or Alcohol Group (e.g., Dopamine).
This reaction was done in two steps. First, the carboxylic group was
activated with DIPCDI (78 µL, 5 equiv) and HOAt (68 mg, 5 equiv)
in DMF for 30 min, after which the solvent was removed by suction
and the resin was washed. The coupling was performed using NH2-
cargo (98 mg dopamine, 5 equiv) in DMF. The mixture was left to
react with intermittent manual stirring overnight. The solvent was
removed by filtration and the resin was washed.
Characterization of the DKP Phe(p-NH-CO-cargo)-N-MePhe
and Its Analogous DKP Phe(p-NH-CO-cargo)-N-MeX, Where X
) HomoPhe, Cha, 2Nal, 1PyrenylAla, Oct and Cargo ) Baicalin.
Reverse phase HPLC: linear gradient from 0 to 100% MeCN in 15
min using a Symmetry C18 column (150 × 4.6 mm × 5 µm, 100 Å,
Waters). tR: Phe(p-NH-CO-baicalin)-N-MePhe, 9.3 min; Phe(p-NH-
CO-baicalin)-N-MeHomoPhe, 8.9 min; Phe(p-NH-CO-baicalin)-N-
MeCha, 8.6 min; Phe(p-NH-CO-baicalin)-N-Me2Nal, 9.5 min; Phe(p-
NH-CO-baicalin)-N-Me-1PyrenylAla, 9.5 min; Phe(p-NH-CO-
baicalin)-N-MeOct, 10.5 min (confirmed by HPLC-MS). Mass
spectrometry (MALDI-TOF): Phe(p-NH-CO-baicalin)-N-MePhe, 752.2
Da; Phe(p-NH-CO-baicalin)-N-MeHomoPhe, 766.1 Da; Phe(p-NH-
CO-baicalin)-N-MeCha, 758.1 Da; Phe(p-NH-CO-baicalin)-N-Me2Nal,
801.4 Da; Phe(p-NH-CO-baicalin)-N-Me-1PyrenylAla, 876.3 Da; Phe-
(p-NH-CO-baicalin)-N-MeOct, 745.4 Da. Total yields after synthesis
and RP-HPLC purification: Phe(p-NH-CO-baicalin)-N-MePhe, 1.2%;
Phe(p-NH-CO-baicalin)-N-MeHomoPhe, 1.7%; Phe(p-NH-CO-ba-
icalin)-N-MeCha, 2.4%; Phe(p-NH-CO-baicalin)-N-Me2Nal, 2.5%;
Phe(p-NH-CO-baicalin)-N-Me-1PyrenylAla, 1.4%; Phe(p-NH-CO-
baicalin)-N-MeOct, 1.1%.
Neutralization-Cyclization and Cleavage All in One Step. Only
by treating the resin with 10% DIEA in DCM (3 × 10 min) was
neutralization-cyclization and cleavage accomplished. The filtrates
were collected, DCM was evaporated under N2, and the residue was
solved in H2O:MeCN (1:1) and lyophilized.
Product Workup and RP-HPLC Purification. The mono and di-
N-methylated DKPs of the two libraries which did not need to be
purified by RP-HPLC were directly desalted using a DOWEX MR-3
mixed bed resin overnight. The remaining compounds were purified
by reverse-phase HPLC using a Symmetry C18 column (100 × 30 mm
× 5 µm, 100 Å, Waters), at 10 mL/min flow with the following
solvents: A, H2O with 0.1% TFA; B, MeCN with 0.05% TFA. The
DKP-dopamine and DKP-baicalin constructs, as well as the DKP
Phe(p-NH2)-N-MePhe, N-MePhe-N-MePhe(p-NH2), and Phe(p-NH-
CH2-COOH)-N-MePhe, were directly purified by RP-HPLC.
Product Characterization. The identity of the compounds synthe-
sized was confirmed using MALDI-TOF mass spectrometry or HPLC-
MS. Purity was checked by reverse phase HPLC.
Characterization of the DKP Phe(p-NH-CH2-CO-NH-cargo)-
N-MePhe and Its Analogous DKP Phe(p-NH-CH2-CO-NH-
cargo)-N-MeX, Where X ) HomoPhe, Cha, 2Nal, 1PyrenylAla, Oct
and Cargo ) Dopamine. Reverse phase HPLC: linear gradient from
15 to 65% MeCN in 15 min using a Symmetry C18 column (150 × 4.6
mm × 5 µm, 100 Å, Waters). tR: Phe(p-NH-CH2-CO-NH-
dopamine)-N-MePhe, 7.8 min; Phe(p-NH-CH2-CO-NH-dopamine)-
N-MeHomoPhe, 7.8 min; Phe(p-NH-CH2-CO-NH-dopamine)-N-
MeCha, 9.4 min; Phe(p-NH-CH2-CO-NH-dopamine)-N-Me2Nal, 9.6
min; Phe(p-NH-CH2-CO-NH-dopamine)-N-Me-1PyrenylAla, 11.7
min; Phe(p-NH-CH2-CO-NH-dopamine)-N-MeOct, 10.6 min (con-
firmed by HPLC-MS). Mass spectrometry (HPLC-MS): Phe(p-NH-
CH2-CO-NH-dopamine)-N-MePhe, 516.5 Da; Phe(p-NH-CH2-
CO-NH-dopamine)-N-MeHomoPhe, 530.3 Da; Phe(p-NH-CH2-
Characterization of the First Library of DKPs. Reverse phase
HPLC: linear gradient from 0 to 100% MeCN in 15 min using a
Symmetry C18 column (150 × 4.6 mm × 5 µm, 100 Å, Waters). tR:
9
J. AM. CHEM. SOC. VOL. 129, NO. 38, 2007 11811

A R T I C L E S
Teixido´ et al.
CO-NH-dopamine)-N-MeCha, 522.5 Da; Phe(p-NH-CH2-CO-NH-
dopamine)-N-Me2Nal, 566.2 Da; Phe(p-NH-CH2-CO-NH-dopamine)-
N-Me-1PyrenylAla, 640.4 Da; Phe(p-NH-CH2-CO-NH-dopamine)-
N-MeOct, 510.4 Da. Total yields after synthesis and RP-HPLC
purification: Phe(p-NH-CH2-CO-NH-dopamine)-N-MePhe, 10.8%;
Phe(p-NH-CH2-CO-NH-dopamine)-N-MeHomoPhe, 9.1%; Phe(p-
NH-CH2-CO-NH-dopamine)-N-MeCha, 7.2%; Phe(p-NH-CH2-
CO-NH-dopamine)-N-Me2Nal, 10.9%; Phe(p-NH-CH2-CO-NH-
dopamine)-N-Me-1PyrenylAla, 9.6%; Phe(p-NH-CH2-CO-NH-
dopamine)-N-MeOct, 8.2%.
also confirmed by MALDI-TOF spectrometry and/or HPLC-MS of
aliquots of the acceptor wells.
The phospholipid mixture used was a porcine polar brain lipid
extract. Composition: 12.6% phosphatidylcholine (PC), 33.1% phos-
phatidylethanolamine (PE), 18.5% phosphatidylserine (PS), 4.1%
phosphatidylinositol (PI), 0.8% phosphatidic acid, and 30.9% of other
compounds. The percentage of transport after 4 h was calculated as
was the effective permeability, using eq 1
2CA(t)
-218.3
Pe )
log 1 -
× 10-6 cm/s
(1)
[
]
Synthesis and Characterization of the Control Linear Dipeptide
Ac-N-MePhe-N-MePhe-CONH2. The linear dipeptide was synthesized
on a 50 µmol scale using Sieber resin. For the coupling of the first
amino acid, Fmoc-N-MePhe (81 mg, 4 equiv), PyBOP (104 mg, 4
equiv) and HOAt (79 mg, 12 equiv) in DMF (1-3 mL/g of resin), and
DIEA (103 µL, 12 equiv) were sequentially added to the resin. The
mixture was left to react with intermittent manual stirring for 90 min.
The solvent was removed by filtration, the resin was washed, and the
coupling was repeated once. The extent of the coupling was checked
by the Kaiser test.
t
CD(t0)
where t is time (h); CA(t) is the compound concentration at the acceptor
well at time t, and CD(t0) is the compound concentration at the donor
well at 0 h.
Immobilized Artificial Membrane Chromatography. Retention
times were determined using an IAMC column with phosphatidylcho-
line (PC), the major phospholipid in cell membranes, which was
covalently immobilized (10 × 4.6 mm, 12 µm, 300 Å, IAM.PC.DD2
column, Regis Technologies Inc.).
The compounds were detected by UV absorption at 220 nm. The
chromatograms were obtained using an HPLC working isocratically
with a mobile phase containing 10 mM phosphate buffer, 2.7 mM KCl,
and 137 mM NaCl at pH 7.4 and 20% (v/v) MeCN. The retention times
(tR) were transformed into capacity factors (K′IAM) following eq 2
For the second amino acid, Fmoc-Phe (78 mg, 4 equiv), PyBOP
(104 mg, 4 equiv) and HOAt (79 mg, 12 equiv) in DMF (1-3 mL/g
of resin), and DIEA (103 µL, 12 eq) were sequentially added to the
resin. The mixture was left to react with intermittent manual stirring
for 90 min. The solvent was removed by filtration, the resin was washed,
and the coupling was repeated two more times. The extent of the
coupling was checked by the De Clercq test.
K′IAM ) (tR - t0)/(t0 - tconnectors
)
(2)
N-methylation of the second amino acid was performed on the solid-
phase. The N-terminal acetylation was done using Ac2O (236 µL, 50
equiv) and DIEA (428 µL, 50 equiv) in DCM (1-3 mL/g of resin).
The mixture was left to react with intermittent manual stirring for 20
min. The solvent was removed by filtration, and the resin was washed.
The cleavage was performed using with 2% (v/v) TFA in DCM (10
mL/g of resin, 4 × 3 min). The filtrates were collected, and the DCM
was evaporated under N2. The residue was dissolved in H2O:MeCN
(1:1) and then lyophilized. Reverse phase HPLC: linear gradient from
0 to 100% MeCN in 15 min using a Symmetry C18 column (150 × 4.6
mm × 5 µm, 100 Å, Waters). tR: Ac-N-MePhe-N-MePhe-CONH2, 8.2
min (confirmed by HPLC-MS). Mass spectrometry (MALDI-TOF):
Ac-N-MePhe-N-MePhe-CONH2, 381.2 Da. Total yield after synthesis
and RP-HPLC purification: Ac-N-MePhe-N-MePhe-CONH2, 41%.
where tR is the compound retention time (min); t0 is the citric acid
retention time (min), and tconnectors is the citric acid retention time without
column (min).
POP Activity Assay. POP activity was determined as follows: For
each reaction, activity buffer (131 µL, 100 mM of Na/K phosphate
buffer, pH 8.0) was preincubated for 15 min at 37 °C with POP (7
nM) and the corresponding inhibitor solution (3 µL). A 5.3 mM stock
solution of inhibitor was prepared in H2O:MeCN (1:1, v/v). After
preincubation, ZGP-AMC (10 µL, 3 mM in 40% of 1,4-dioxane) was
added, and the reaction was incubated for 1 h at 37 °C. The reaction
was stopped with sodium acetate (150 µL, 1M, pH 4), at which point
the formation of AMC was measured fluorimetrically. The excitation
and emission wavelengths were 360/40 and 485/20 nm, respectively.
A control with the same amount of MeCN was performed.
Parallel Artificial Membrane Permeability Assay. The PAMPA
assay was used to determine the capacity of compounds to cross the
BBB by passive diffusion.68,69 The effective permeability of the
compounds was measured in triplicate at an initial concentration of
200 µM. The buffer solution was prepared from a concentrated one,
commercialized by pION, following the manufacturer’s instructions.
pH was adjusted to 7.4 using a 0.5M NaOH solution. The compound
of interest was dissolved to the desired concentration (200 µM). The
PAMPA sandwich was separated and the donor well was filled with
200 µL of the compound solution to be studied. The acceptor plate
was placed into the donor plate, ensuring that the underside of the
membrane was in contact with buffer. Then 4 µL of the mixture of
phospholipids (20 mg/mL) in dodecane was added to the filter of each
well and 200 µL of buffer solution was added to the each acceptor
well. The plate was covered and incubated at room temperature in a
saturated humidity atmosphere for 4 h under orbital agitation at 100
rpm. After the 4 h, 150 µL/well from the donor plate and 150 µL/well
from the acceptor plate were transferred to HPLC vials, and 100 µL/
each sample was injected in a HPLC reverse phase Symmetry C18
column (150 × 4.6 mm × 5 µm, 100 Å, Waters). The transport was
Conclusions
In this paper, we have introduced, for the first time, the
concept of using DKPs as shuttles to the brain for the delivery
of drugs that cross the BBB with difficulty. This new approach
is illustrated with the cases of dopamine and baicalin, and for
both compounds, our data using the PAMPA assay demonstrates
that the new prodrugs are able to cross a model of the BBB
membrane. Our results are only preliminary and cell-based in
vitro assays and in vivo studies will follow as well as the
application of the DKPs as BBB-shuttles for other potential CNS
drugs and diagnostic agents.
From the synthetic point of view, the all-in-solid-phase
methods described are highly convenient, as they allow the
effective synthesis of two libraries of mono- and di-N-methy-
lated DKPs, as well as the DKP-dopamine and DKP-baicalin
constructs. A comparison of the transport properties of all the
DKPs synthesized highlights the importance of the capacity to
form hydrogen bonds for the optimization of permeability across
the BBB. It is important to take into account that the simple
rule “the smaller the number of potential hydrogen bonds that
(68) Balimane, P. V.; Pace, E.; Chong, S.; Zhu, M.; Jemal, M.; Van Pelt, C. K.
J. Pharm. Biomed. Anal. 2005, 39, 8-16.
(69) Rezai, T.; Yu, B.; Millhauser, G. L.; Jacobson, M. P.; Lokey, R. S. J. Am.
Chem. Soc. 2006, 128, 2510-2511.
9
11812 J. AM. CHEM. SOC. VOL. 129, NO. 38, 2007

Diketopiperazines as BBB-Shuttles
A R T I C L E S
can be formed, the better the transport” is inappropriate, as it
can lead to compounds that are too hydrophobic and conse-
quently retained in the lipid membrane. In addition, from the
methodological point of view, we have demonstrated the
usefulness of performing PAMPA assays for the evaluation of
mixtures. In our format, both the throughput and the reliability
of the method are improved.
Our BBB-shuttle cyclic dipeptide may provide a new solution
to bypass the problem of bioavailability of certain types of
potential CNS therapeutic drugs. This approach could be useful
to transport promising antiretroviral agents and chemotherapy
drugs across the BBB and into the brain, thereby reducing the
severe side effects caused by the high doses currently used to
overcome low permeability. Our BBB-shuttle approach implies
that the search for novel drugs does not become limited only to
compounds that have the capacity to cross the BBB.
Acknowledgment. This work was supported by Ministerio
Educacio´n y Ciencia-FEDER (BIO 2005-00295 and CON-
SOLIDER) and the Generalitat de Catalunya [Xarxa de Refer-
e`ncia en Biotecnologia and Grups Consolidats (2005SGR-
00663)].
Supporting Information Available: Methodology for the
preparation of DKP-shuttles for other cargo, including charac-
terization data. This material is available free of charge via the
Internet at http://pubs.acs.org.
JA073522O
9
J. AM. CHEM. SOC. VOL. 129, NO. 38, 2007 11813