'A la Carte' peptide shuttles: Tools to increase their passage across the blood-brain barrier

Article abstract of DOI:10.1002/cmdc.201300575

Noninvasive methods for efficient drug delivery to the brain is an unmet need. Molecular access to the brain is regulated by the blood-brain barrier (BBB) established by the endothelial cells of brain vessels. Passive diffusion is one of the main mechanisms that organic compounds use to travel through these endothelial cells. This passage across the BBB is determined mainly by certain physicochemical properties of the molecule such as lipophilicity, size, and the presence of hydrogen bond donors and acceptors. One emerging strategy to facilitate the passage of organic compounds across the BBB is the use of peptide shuttles.1 In using this approach the permeability in front the BBB is, clearly, determined by the combined physicochemical properties of both the cargo and the shuttle. Herein we report the synthesis of a series of variations of one of the more efficient peptide shuttles, (N-MePhe)_n. These include diverse structural features such as various backbone stereochemistries or the presence of non-natural amino acids, including halogenated residues. In several cases, we assessed the BBB permeability of both the shuttles alone and linked to a few cargos. Our results show how factors such as stereochemistry or halogen content influences the passage across the BBB and, more importantly, opens the way to a strategy of peptide shuttles 'a la carte′, in which a particular fine-tuned shuttle is used for each specific cargo. On the menu: Flexibility, chirality, and the incorporation of halogenated amino acids are, among others, tools that can be used 'a la carte' to increase the passage of peptide shuttles across the blood-brain barrier (BBB). This study paves the way for a strategy in which a particular fine-tuned shuttle can be used for the transport of a given specific drug cargo.

Full text of DOI:10.1002/cmdc.201300575

CHEMMEDCHEM
FULL PAPERS
DOI: 10.1002/cmdc.201300575
‘ꢀ la Carte’ Peptide Shuttles: Tools to Increase Their
Passage across the Blood–Brain Barrier
Morteza Malakoutikhah,[a] Bernat Guixer,[a] Pol Arranz-Gibert,[a] Meritxell Teixidꢀ,*[a] and
Ernest Giralt*[a, b]
Noninvasive methods for efficient drug delivery to the brain is
an unmet need. Molecular access to the brain is regulated by
the blood–brain barrier (BBB) established by the endothelial
cells of brain vessels. Passive diffusion is one of the main
mechanisms that organic compounds use to travel through
these endothelial cells. This passage across the BBB is deter-
mined mainly by certain physicochemical properties of the
molecule such as lipophilicity, size, and the presence of hydro-
gen bond donors and acceptors. One emerging strategy to fa-
cilitate the passage of organic compounds across the BBB is
the use of peptide shuttles.[1] In using this approach the per-
meability in front the BBB is, clearly, determined by the com-
bined physicochemical properties of both the cargo and the
shuttle. Herein we report the synthesis of a series of variations
of one of the more efficient peptide shuttles, (N-MePhe)n.
These include diverse structural features such as various back-
bone stereochemistries or the presence of non-natural amino
acids, including halogenated residues. In several cases, we as-
sessed the BBB permeability of both the shuttles alone and
linked to a few cargos. Our results show how factors such as
stereochemistry or halogen content influences the passage
across the BBB and, more importantly, opens the way to a strat-
egy of peptide shuttles ‘ꢁ la carte’, in which a particular fine-
tuned shuttle is used for each specific cargo.
Introduction
Recent years have witnessed huge research efforts channeled
into the discovery of new drugs and the development of strat-
egies for their specific delivery. It is estimated that 1.5 billion
people worldwide currently suffer from various central nervous
system (CNS) disorders, and as the elderly population increas-
es, these diseases will become a major health problem world-
wide.[2,3] Most of the drugs currently on the market for CNS dis-
orders are far from ideal because of their poor blood–brain
barrier (BBB) penetration. More than 98% of small molecules
do not cross the BBB, and the scenario is even worse for large-
molecule drugs, including peptides, recombinant proteins, and
monoclonal antibodies.[4] Transport across the BBB is one of
the main challenges in the field of brain/CNS drug targeting
using noninvasive methods.[5]
cells (which differ from those in peripheral tissues in that they
have very few endocytotic vesicles and form tight junctions
between them); however, pericytes, astrocytes, and neuronal
cells also have key functions in the BBB.[6–9]
Despite the protective function of the BBB, it does not com-
pletely isolate the brain and the CNS, as it allows basic nu-
trients such as glucose, amino acids, ions, and gases such as
O2 and CO2 to enter the brain. Influx mechanisms include the
paracellular pathway (restricted because of tight junctions) and
transcellular pathways, such as passive diffusion (normally lim-
ited to small and highly lipophilic molecules), carrier-mediated,
receptor-mediated, and absorptive-mediated transport.[7,8]
BBB shuttles or Trojan-horse moieties (molecules with the
capacity to enter the brain and carry a drug that is unable to
cross the BBB unaided) are currently one of the most promis-
ing strategies to overcome the BBB.[1,3,10,11] In the case of pas-
sive diffusion, physicochemical properties entirely determine
the capacity of molecules to cross biological membranes. Lipo-
philicity governed by the presence of polar groups and/or hy-
drogen bond donors/acceptors in the structure of a compound,
molecular weight, peptide length, and amino acid sequence
are considered the main determinants of the capacity of pep-
tides to cross the BBB.[12]
The BBB is the most important barrier protecting the CNS,
although the blood–cerebrospinal fluid (CSF) interface plays
a significant role in specific sites. The BBB is located in brain
capillaries and is formed mainly by brain capillary endothelial
[a] Dr. M. Malakoutikhah,+ B. Guixer,+ P. Arranz-Gibert, Dr. M. Teixidꢀ,
Prof. E. Giralt
Institute for Research in Biomedicine (IRB Barcelona)
Baldiri Reixac 10, Barcelona, 08028 (Spain)
E-mail: meritxell.teixido@irbbarcelona.org
ernest.giralt@irbbarcelona.org
After some years working with peptides that cross the BBB
by passive diffusion and successfully using these molecules as
BBB shuttles,[13–15] we envisaged other physicochemical features
that may be relevant for this transport and that are usually
overlooked. These properties, such as flexibility, halogenation,
and stereochemistry, can be considered tools to increase and
fine-tune the transport of compounds. In this regard, we se-
[b] Prof. E. Giralt
Department of Organic Chemistry, University of Barcelona
Martꢁ i Franquꢂs 1-11, Barcelona, 08028 (Spain)
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201300575.
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1594 – 1601 1594

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
lected a well-known family of shuttles that cross the
BBB by passive diffusion[16] and studied the effect of
factors such as halogen content and stereochemistry
on their transport capacity.
Table 1. Effective permeability (Pe), percentage of transport (T4h), and membrane re-
tention (R) in the PAMPA after 4 h and HPLC tR of 16 stereoisomers and control com-
pounds, propranolol and carbamazepine.[a]
Compound[b]
Pe [ꢃ106 cmsꢀ1
]
T4h [%]
R [%]
HPLC tR [min]
[c]
propranolol
carbamazepine
5.7ꢁ0.9
8.5ꢁ0.3
9.4ꢁ0.9
6.6ꢁ0.3
5.6ꢁ0.2
6.1ꢁ0.1
4ꢁ0.9
10.7ꢁ1.4
15.1ꢁ0.4
16.4ꢁ1.3
12.2ꢁ0.5
10.5ꢁ0.4
11.4ꢁ0.1
7.8ꢁ1.6
7.7ꢁ1.9
6.2ꢁ0.1
8.4ꢁ2.5
5.9ꢁ0.8
4.6ꢁ0.2
8.0ꢁ0.4
5.4ꢁ0.7
6.6ꢁ0.2
9.6ꢁ0.7
6.0ꢁ0.4
4.2ꢁ0.4
55.1ꢁ9.7
4.96
6.60
3.8
4
<1
Results and Discussion
Ac-(DDDD)-CONH2
Ac-(LLLL)-CONH2
Ac-(LLLD)-CONH2
Ac-(DDDL)-CONH2
Ac-(LDDD)-CONH2
Ac-(DLLL)-CONH2
Ac-(DLLD)-CONH2
Ac-(LDDL)-CONH2
Ac-(DDLL)-CONH2
Ac-(LLDD)-CONH2
Ac-(LLDL)-CONH2
Ac-(DDLD)-CONH2
Ac-(DLDD)-CONH2
Ac-(LDLL)-CONH2
Ac-(LDLD)-CONH2
Ac-(DLDL)-CONH2
20.7ꢁ5.0
42ꢁ0.2
Chirality
48.4ꢁ0.5
45.4ꢁ0.7
76ꢁ2.5
4.2
4.3
4.3
4.3
4.5
4.5
4.7
4.8
4.9
5
Cell membrane components such as phospholipids
and proteins are chiral. Because all these components
are commonly found enantiomerically pure in nature,
it has been postulated[17] that stereoisomers could be
discriminated in cell membranes. It was previously
shown that biomembrane models such as micelles
and vesicles can discriminate between homochiral
and heterochiral enantiomers of dipeptides.[18–21]
These studies hypothesized that in chiral aggregates
recognition occurs in a chiral environment induced in
an internal region of the aggregate by remote stereo-
genic centers. Therefore, chiral recognition cannot be
simply ascribed to noncovalent-specific interactions
between the solute and the monomers behaving as
single entities, but is instead due to the aggregate as
4ꢁ1.1
58.1ꢁ7.5
64.1ꢁ0.4
43.3ꢁ4.1
74.2ꢁ6.2
81.6ꢁ1.5
64.4ꢁ0.8
75.1ꢁ3.6
65.7ꢁ0.6
56.7ꢁ2.3
71.3ꢁ0.1
78.4ꢁ1.8
3.1ꢁ0.1
4.4ꢁ1.4
2.9ꢁ0.4
2.3ꢁ0.1
4.1ꢁ0.2
2.7ꢁ0.3
3.3ꢁ0.1
5ꢁ0.4
5.2
5.2
5.3
5.3
3ꢁ0.2
2ꢁ0.2
[a] Data are expressed as the mean ꢁSD of n=3 experiments performed in triplicate.
[b] D and L stand for D-N-MePhe and L-N-MePhe respectively. [c] Low permeability:
Pe <0.1ꢃ10ꢀ6 cm; medium permeability: 0.1ꢃ10ꢀ6 cm<Pe <1ꢃ10ꢀ6 cm; high perme-
ability: Pe >1ꢃ10ꢀ6 cm.
a
whole. Therefore, potential discrimination can
occur in regions of the aggregate/membrane that are
quite far away from the stereogenic centers. In addition, these
chiral interactions appear to confer a particular conformation
to the enantiomers and diastereomers that finally determines
the fate of the molecule when interacting with the membrane.
Furthermore, another study demonstrated that diastereomeric
peptides show distinctive lipophilicity and permeability across
the BBB.[22]
there was a direct correlation between lipophilicity and mem-
brane retention (%R; r=0.723, Figure 1c). These results indi-
cate that in this series of stereoisomeric peptides, homochiral
peptides (LLLL and DDDD) had the optimal lipophilicity. The
heterochiral peptides might show excessive lipophilicity, ham-
pering their passage across the PAMPA membrane due to ex-
cessive retention within the membrane. The observed behavior
is probably a consequence of differences in the more stable
conformations adopted by each stereoisomer in the presence
of the lipid bilayers. Our findings are consistent with previous
studies reporting that heterochiral dipeptides (LD and DL)
bond more strongly to chiral micellar phases (being trapped in
the hydrophobic micellar portion of the described model) than
homochiral enantiomers (LL and DD).[18–20] Looking now to the
presence or absence of enantiomeric discrimination along all
the stereoisomers,[19,20,25] our observations in the PAMPA mem-
brane showed that enantiomeric discrimination was remark-
able for most of the enantiomeric pairs (although it was very
low for the pairs LDDD/DLLL and DDDL/LLLD). Although there
is not a clear global trend, in general the observed enantio-
meric discrimination is higher for heterochiral enantiomers
than for homochiral peptides.
To study the effect of peptide stereochemistry on membrane
permeation, we prepared a library consisting of 16 stereoiso-
meric peptides (Table 1) and evaluated them by parallel artifi-
cial membrane permeability assay (PAMPA), which is the gold-
standard method for assessing transport across the BBB via
passive diffusion. All the peptides had four N-MePhe residues
and were designed based on our previously reported peptide,
Ac-(N-MePhe)4-CONH2, which showed high permeation in
PAMPA studies.[13,14] HPLC retention time in reversed-phase col-
umns (HPLC tR) is a good measure of the polarity of a molecule,
especially when comparing stereoisomers of the same com-
pound.[23,24] We used these retention times to rank our mole-
cules in terms of lipophilicity. We found that the homochiral
peptides (LLLL and DDDD), although less lipophilic than the
rest of the heterochiral peptides, exhibited higher permeability
and lower membrane retention (%R) in PAMPA (Table 1). Be-
tween the two homochiral versions, the all-D version showed
higher permeation than the previously reported all-L version.
One could expect that in passive diffusion, the greater the
lipophilicity, the greater the permeability across membranes.
However, for these 16 peptides, an inverse correlation was de-
tected between lipophilicity and permeability (r=0.757, Fig-
ure 1a). Furthermore, permeability correlated inversely with
membrane retention (%R; r=0.888, Figure 1b). In contrast,
ꢀ la carte peptide length
It is clear from our previous work that there is a relationship
between PAMPA permeability and the number of N-MePhe res-
idues in a BBB shuttle peptide.[13,14] In a series of N-MePhe olig-
omers (Ac-(N-MePhe)n-CONH2 n=2–10), an increase in peptide
chain length enhanced PAMPA permeability up to the peptide
with four N-MePhe residues, namely Ac-(N-MePhe)4-CONH2.
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1594 – 1601 1595

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
son’s disease. Dopamine cannot enter the brain un-
aided; however, l-DOPA crosses the BBB by means of
a large natural amino acid carrier. Once inside the
brain, l-DOPA is enzymatically converted into dopa-
mine by aromatic l-amino acid decarboxylase. How-
ever, l-DOPA administration is hindered by several
side effects caused by dopamine formation in the pe-
riphery.[26–28] l-DOPA derivatization through an amide
bond has been reported to inhibit the decarboxyla-
tion reaction in the periphery and to enhance l-
DOPA in the brain.[29–31] Furthermore, because l-DOPA
competes with other l-amino acids to gain access to
the brain, it would be desirable to deliver this drug
to this organ by a noncompetitive passive transport
mechanism.
l-DOPA is unable to cross the PAMPA membrane
unaided. In contrast, our peptide BBB shuttles trans-
ported this drug across the membrane and showed
medium to high permeability. The tetrapeptide
showed the highest PAMPA permeability (measured
as Pe or percent transport after 4 h) and lipophilicity
(measured by IAMC as kIAM) (Table 2). However, the di-
peptide with the lowest lipophilicity presented slight-
ly better PAMPA permeability than the tripeptide. In
Case 1, the longest shuttle exhibits the best transport
capacity.
Case 2, GABA: 4-aminobutyric acid
A decreased concentration of GABA, the dominant in-
hibitory neurotransmitter in the CNS, is associated
with several brain diseases. Increasing the concentra-
tion of this amino acid in the CNS may serve to treat
these brain disorders; however, GABA permeation
across the BBB is insufficient to enhance its levels in
the brain.[32–35] Some GABA derivatives show higher
permeation across the BBB than GABA alone.[36–38]
Figure 1. a) Inverse correlation between logPe of 16 stereoisomers and their logHPLC tR.
b) Inverse correlation between logPe of 16 stereoisomers and their log%R. c) Correlation
between logHPLC tR of 16 stereoisomers and their log%R.
However, when the peptide had
more than four residues, per-
meability dropped considerably.
Table 2. Effective permeability (Pe), percentage of transport (T4h) in the PAMPA, and lipophilicity (kIAM) of X-(N-
MePhe)n-CONH2 (X=l-DOPA, GABA, and NIP; n=2–4) and control compounds (propranolol, carbamazepine)
after 4 h.[a,b]
In this regard, we explored
whether H-(N-MePhe)3-CONH2
and H-(N-MePhe)2-CONH2, when
coupled to a neurodrug cargo
(l-DOPA, GABA, and NIP), show
better performance as carriers
than the original tetrapeptides.
The results are summarized in
Table 2.
Compound
Pe [ꢃ106 cmsꢀ1
]
T4h [%]
kIAM
1.9
propranolol
carbamazepine
6.6ꢁ0.4
8.5ꢁ0.3
0.6ꢁ0.03
0.4ꢁ0.03
1.1ꢁ0.1
0.2ꢁ0.02
0.3ꢁ0.01
0.4ꢁ0.06
0.9ꢁ0.1
1.5ꢁ0.2
1.4ꢁ0.1
12.2ꢁ0.6
15.1ꢁ0.4
1.2ꢁ0.05
0.9ꢁ0.07
2.4ꢁ0.2
0.5ꢁ0.04
0.6ꢁ0.03
0.8ꢁ0.1
1.9ꢁ0.3
3.1ꢁ0.3
2.8ꢁ0.2
2.1
l-DOPA-N-MePhe-N-MePhe-CONH2
l-DOPA-N-MePhe-N-MePhe-N-MePhe-CONH2
l-DOPA-N-MePhe-N-MePhe-N-MePhe-N-MePhe-CONH2
GABA-N-MePhe-N-MePhe-CONH2
GABA-N-MePhe-N-MePhe-N-MePhe-CONH2
GABA-N-MePhe-N-MePhe-N-MePhe-N-MePhe-CONH2
NIP-N-MePhe-N-MePhe-CONH2
2.4ꢁ0.1
9.7ꢁ0.4
29.6ꢁ1.3
1.7ꢁ0.05
7.2ꢁ0.1
12.5ꢁ2.2
1.6ꢁ0.01
9.6ꢁ0.01
8.1ꢁ1.3
Case 1, l-DOPA: 3,4-dihydroxy-
l-phenylalanine
NIP- N-MePhe-N-MePhe-N-MePhe-CONH2
NIP-N-MePhe-N-MePhe-N-MePhe-N-MePhe-CONH2
l-DOPA is a prodrug of dopa-
mine that has been used for the
last four decades to treat Parkin-
[a] Data are expressed as the mean ꢁSD of n=3 experiments performed in triplicate. [b] Reported by our re-
search group in 2008 and 2010.[13,14] Only results of tetrapeptides (X-N-MePhe-N-MePhe-N-MePhe-CONH2, X=
N-MePhe, Cha, 1-Nal) were already reported.
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1594 – 1601 1596

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
To test the potential of our
peptides as carriers of GABA, this
drug was attached to the pep-
tides in solid phase. Peptides
showed a small difference in
PAMPA permeability, whereas
distinctive capacity factors were
detected in IAMC (Table 2). In
Case 2, transport is less depen-
dent on length, although again,
the tetrapeptide shuttle is more
efficient.
Table 3. Effective permeability (Pe) and percentage of transport (T4h) in the PAMPA after 4 h and lipophilicity
(kIAM) of X-Y-(N-MePhe)n-CONH2 (X=l-DOPA, GABA, and NIP; Y=N-MePhe, Cha, and 2Nal; n=1–3) and control
compounds (propranolol and carbamazepine).[a,b]
Compound
Pe [ꢃ106 cmsꢀ1
]
T4h [%]
kIAM
propranolol
carbamazepine
l-DOPA-N-MePhe-N-MePhe-CONH2
l-DOPA-Cha-N-MePhe-CONH2
l-DOPA-2Nal-N-MePhe-CONH2
l-DOPA-N-MePhe-N-MePhe-N-MePhe-CONH2
l-DOPA-Cha-N-MePhe-N-MePhe-CONH2
l-DOPA-2Nal-N-MePhe-N-MePhe-CONH2
l-DOPA-N-MePhe-N-MePhe-N-MePhe-N-MePhe-CONH2
l-DOPA-Cha-N-MePhe-N-MePhe-N-MePhe-CONH2
l-DOPA-2Nal-N-MePhe-N-MePhe-N-MePhe-CONH2
GABA-N-MePhe-N-MePhe-CONH2
6.2ꢁ0.6
9.3ꢁ0.8
0.6ꢁ0.03
2.2ꢁ0.1
1.8ꢁ0.1
0.4ꢁ0.03
1.7ꢁ0.2
1ꢁ0.03
11.5ꢁ1.1
16.2ꢁ1.1
1.2ꢁ0.05
4.4ꢁ0.2
3.6ꢁ0.2
0.9ꢁ0.07
3.5ꢁ0.4
2ꢁ0.1
1.9
2.1
2.4ꢁ0.1
24.7ꢁ1
32.6ꢁ1.8
9.7ꢁ0.4
60.1ꢁ1.6
68.4ꢁ2.7
29.6ꢁ1.3
1.1ꢁ0.1
0.7ꢁ0.06
0.3ꢁ0.1
0.2ꢁ0.02
1.4ꢁ0.05
1.2ꢁ0.1
0.3ꢁ0.01
2.3ꢁ0.4
1.5ꢁ0.4
0.4ꢁ0.06
0.4ꢁ0.1
0.2ꢁ0.02
0.9ꢁ0.1
3.2ꢁ0.2
2.2ꢁ0.1
1.5ꢁ0.2
1.8ꢁ0.06
1.5ꢁ0.04
1.4ꢁ0.1
1.2ꢁ0.1
1.1ꢁ0.06
2.4ꢁ0.2
1.4ꢁ0.1
0.7ꢁ0.2
0.5ꢁ0.04
2.8ꢁ0.1
2.5ꢁ0.4
0.6ꢁ0.03
4.6ꢁ1.4
3ꢁ0.7
>128[c]
102ꢁ6.2
Case 3, NIP: nipecotic acid
1.7ꢁ0.05
GABA-Cha-N-MePhe-CONH2
GABA-2Nal-N-MePhe-CONH2
10.6ꢁ0.1
13.7ꢁ0.03
7.2ꢁ0.1
The level of GABA in the brain
can also be increased by using
either GABA inhibitors, which
prevent this unusual amino acid
from being used by neurons and
glial cells, or GABA receptor ago-
nists. Nipecotic acid fulfills these
two requirements; however, its
inability to gain access to the
brain limits its application for
the treatment of diseases affect-
ing this organ.[34,39,40]
GABA-N-MePhe-N-MePhe-N-MePhe-CONH2
GABA-Cha-N-MePhe-N-MePhe-CONH2
GABA-2Nal-N-MePhe-N-MePhe-CONH2
GABA-N-MePhe-N-MePhe-N-MePhe-N-MePhe-CONH2
GABA-Cha-N-MePhe-N-MePhe-N-MePhe-CONH2
GABA-2Nal-N-MePhe-N-MePhe-N-MePhe-CONH2
NIP-N-MePhe-N-MePhe-CONH2
13ꢁ0.04
22.9ꢁ0.4
12.5ꢁ2.2
28.2ꢁ6.7
30.2ꢁ4.4
1.6ꢁ0.01
10.1ꢁ0.1
14.1ꢁ0.02
9.6ꢁ0.01
20.4ꢁ0.7
33.9ꢁ0.3
8.1ꢁ1.3
0.8ꢁ0.1
0.8ꢁ0.3
0.4ꢁ0.04
1.9ꢁ0.3
6.3ꢁ0.3
4.4ꢁ0.3
3.1ꢁ0.3
3.7ꢁ0.07
3.1ꢁ0.08
2.8ꢁ0.2
2.5ꢁ0.2
2.4ꢁ0.1
NIP-Cha-N-MePhe-CONH2
NIP-2Nal-N-MePhe-CONH2
NIP-N-MePhe-N-MePhe-N-MePhe-CONH2
NIP-Cha-N-MePhe-N-MePhe-CONH2
NIP-2Nal-N-MePhe-N-MePhe-CONH2
NIP-N-MePhe- N-MePhe-N-MePhe-N-MePhe-CONH2
NIP-Cha-N-MePhe-N-MePhe-N-MePhe-CONH2
NIP-2Nal-N-MePhe-N-MePhe-N-MePhe-CONH2
>128[b]
30.4ꢁ1.5
As third cargo, we conjugated
NIP to the peptides and assessed
the transport of the resulting
constructs in PAMPA and IAMC
(Table 2). Here the tripeptide
showed higher PAMPA permea-
[a] Data are expressed as the mean ꢁSD of n=3 experiments performed in triplicate. [b] Reported by our re-
search group in 2008 and 2010.[13,14] Only results of tetrapeptides (X-N-MePhe-N-MePhe-N-MePhe-CONH2, X=
N-MePhe, Cha, 1-Nal) were already reported. [c] Retention time >60 min in the IAMC HPLC column.
bility than the dipeptide and the tetrapeptide. In summary, we
found here a first example of the ‘ꢄ la carte’ concept; that is,
there is an optimal length for the (N-MePhe)n shuttle, but for
each cargo this length can be different (n=4 for l-DOPA and
GABA; n=3 for NIP).
PAMPA permeability compared with tetrapeptides. The tripep-
tide Cha-(N-MePhe)2-CONH2 showed the best shuttling per-
formance for GABA. In the case of nipecotic acid, we observed
that the N-MePhe tetrapeptide coupled to NIP exhibited better
PAMPA permeability than tetrapeptides with Cha and 2Nal. Di-
peptides enhanced NIP transport considerably for this cargo,
especially in the case of Cha-N-MePhe-CONH2, which showed
a greater capacity to transfer this neurodrug than the other
peptides.
Incorporation of non-natural amino acids
Chemical diversity of peptides can be widely expanded by in-
corporation of non-natural amino acids carrying a variety of
special side chains. To explore the use of non-natural amino
acids as tools to modulate BBB transport we first explored the
use of two non-halogenated Phe analogues, cyclohexylalanine
(Cha) and 2-naphtylalanine (2Nal), as N-terminal residues in our
peptide BBB shuttles.
Finally, we decided to explore the effect of halogenation as
a way to modulate peptide shuttle lipophilicity and conse-
quently their capacity to cross the BBB.[41–45] For this purpose
we studied the effect of chlorination of the phenylalanine resi-
due on the dipeptide (H-N-MePhe-N-MePhe-CONH2) and tri-
peptide (H-N-MePhe-N-MePhe-N-MePhe-CONH2) neurodrug
shuttle conjugates. On the one hand, we synthesized (p-Cl)-N-
MePhe analogues of those constructs. N-Methylation of resi-
dues substituted on their aromatic rings such as Tyr, 3-pyridyla-
lanine (3-Pal), and (p-Cl)Phe is not trivial.[46] However, we used
a protocol (see Supporting Information) that allows N-methyla-
tion of (p-Cl)Phe in solid phase with great purity and without
the formation of byproducts. On the other hand, we prepared
the corresponding non-N-methylated chlorinated compounds,
namely H-(p-Cl)Phe-N-MePhe-CONH2 and H-(p-Cl)Phe-(p-Cl)Phe-
Hence the three neurodrug cargos (l-DOPA, GABA, and NIP)
were attached to Cha- and 2Nal-containing peptides of varia-
ble length, and the behavior in PAMPA and IAMC assays was
studied (Table 3). For l-DOPA, the dipeptides probably have
the optimum lipophilicity required to penetrate the PAMPA
phospholipids, with dipeptide Cha-N-MePhe-CONH2 displaying
the greatest capacity to carry this drug through the PAMPA
membrane. For GABA, Cha and 2Nal tripeptides showed the
best performance as shuttles and greatly improved their
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1594 – 1601 1597

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Table 4. Effective permeability (Pe), percentage of transport (T4h), membrane retention (R) in the PAMPA after 4 h, and lipophilicity (kIAM) of X-(Y)n-N-MePhe-
CONH2 (X=l-DOPA, GABA, and NIP; Y=N-MePhe, (p-Cl)N-MePhe and [(p-Cl)Phe]; n=1–2) and control compounds (propranolol, carbamazepine).[a]
Compound
Pe [ꢃ106 cmsꢀ1
]
T4h [%]
R [%]
kIAM
propranolol
carbamazepine
6.2ꢁ0.6
9.3ꢁ0.8
0.6ꢁ0.03
1.9ꢁ0.1
1.6ꢁ0.01
0.4ꢁ0.03
1.5ꢁ0.1
0.4ꢁ0.03
0.2ꢁ0.02
0.7ꢁ0.1
1.8ꢁ0.3
0.3ꢁ0.01
1.7ꢁ0.03
1.6ꢁ0.1
0.9ꢁ0.1
1.5ꢁ0.03
1.7ꢁ0.1
1.5ꢁ0.2
3.1ꢁ0.2
2ꢁ0.3
11.5ꢁ1.1
16.2ꢁ1.1
1.2ꢁ0.05
3.9ꢁ0.2
3.3ꢁ0.02
0.9ꢁ0.07
3ꢁ0.2
33.3ꢁ4
1.9
2.1
<1
l-DOPA-N-MePhe-N-MePhe-CONH2
l-DOPA-(p-Cl)N-MePhe-N-MePhe-CONH2
l-DOPA-(p-Cl)Phe-N-MePhe-CONH2
l-DOPA-N-MePhe-N-MePhe-N-MePhe-CONH2
l-DOPA-(p-Cl)N-MePhe-(p-Cl)N-MePhe-N-MePhe-CONH2
l-DOPA-(p-Cl)Phe-(p-Cl)Phe-N-MePhe-CONH2
GABA-N-MePhe-N-MePhe-CONH2
GABA-(p-Cl)N-MePhe-N-MePhe-CONH2
GABA-(p-Cl)Phe-N-MePhe-CONH2
GABA-N-MePhe-N-MePhe-N-MePhe-CONH2
GABA-(p-Cl)N-MePhe-(p-Cl)N-MePhe-N-MePhe-CONH2
GABA-(p-Cl)Phe-(p-Cl)Phe-N-MePhe-CONH2
NIP-N-MePhe-N-MePhe-CONH2
<1
2.4ꢁ0.1
7.2ꢁ0.1
16.1ꢁ0.4
9.7ꢁ0.4
67ꢁ1
5.1ꢁ0.4
12.7ꢁ1.1
29.8ꢁ3.4
25.5ꢁ5.2
86.8ꢁ2.4
0.8ꢁ0.07
0.5ꢁ0.04
1.5ꢁ0.2
3.7ꢁ0.6
0.6ꢁ0.03
3.5ꢁ0.05
3.2ꢁ0.3
1.9ꢁ0.3
3.1ꢁ0.05
3.5ꢁ0.5
3.1ꢁ0.3
6.2ꢁ0.3
4.1ꢁ0.6
>128[b]
<1
1.7ꢁ0.05
6.2ꢁ0.1
6.5ꢁ2.4
15.3ꢁ3
7.1ꢁ0.1
<1
7.2ꢁ0.1
20.3ꢁ3.4
35.8ꢁ0.2
53.5ꢁ1.1
96.3ꢁ3.3
1.6ꢁ0.01
5.7ꢁ0.1
<1
<1
7.6ꢁ2.1
<1
12.1ꢁ0.9
28.5ꢁ0.6
NIP-(p-Cl)N-MePhe-N-MePhe-CONH2
NIP-(p-Cl)Phe-N-MePhe-CONH2
NIP-N-MePhe-N-MePhe-N-MePhe-CONH2
NIP-(p-Cl)N-MePhe-(p-Cl)N-MePhe-N-MePhe-CONH2
NIP-(p-Cl)Phe-(p-Cl)Phe-N-MePhe-CONH2
7.2ꢁ0.1
9.6ꢁ0.01
56.2ꢁ1.4
103.7ꢁ1.4
[a] Data are expressed as the mean ꢁSD of n=3 experiments performed in triplicate. [b] Retention time >60 min in the IAMC HPLC column.
N-MePhe-CONH2, in order to compare the respective influence
of N-methylation and chlorination on lipophilicity and PAMPA
permeability.
ed-N-methylated peptide was the most permeating analogue,
while it was less lipophilic than the chlorinated peptide alone.
For all three cargos, the peptide with chlorinated Phe showed
greater PAMPA permeability than that with N-methylated Phe.
For the tripeptide, with all cargos chlorinated peptide with
higher lipophilicity showed less PAMPA permeability than
chlorinated-N-methylated peptide. This greater lipophilicity
may cause retention of the peptide in the membrane, as ob-
served in PAMPA experiments (Table 4). Plotting lipophilicity
versus PAMPA permeability provided correlations of r=0.998,
r=0.933, and r=0.836 for NIP, GABA, and l-DOPA dipeptides,
respectively. However, for the tripeptides, a good correlation
between lipophilicity and PAMPA permeability was not found
except for GABA (r=0.971). These findings suggest that in
order to obtain the greatest permeability, either chlorination or
chlorination-N-methylation should be applied, ‘ꢁ la carte’ de-
pending on the peptide sequence, as tools to fine-tune shuttle
transport capacity.
As expected, chlorination increased the lipophilicity of all
the analogues tested with all cargos relative to the parent
compounds (Table 4). Data also suggest differences in the lipo-
philicity of chlorinated peptides [H-(p-Cl)Phe-N-MePhe-CONH2
or H-(p-Cl)Phe-(p-Cl)Phe-N-MePhe-CONH2] and chlorinated-N-
methylated peptides [H-(p-Cl)N-MePhe-N-MePhe-CONH2 or
H-(p-Cl)N-MePhe-(p-Cl)N-MePhe-N-MePhe-CONH2], with the
(p-Cl)Phe analogue, unexpectedly, being more lipophilic than
the (p-Cl)N-MePhe one. In addition, we observed that the pep-
tide with chlorinated Phe was more lipophilic than that with
N-methylated Phe. Moreover, the addition of two chlorine
atoms (Table 4) gave rise to a greater difference between the
lipophilicity of chlorinated-N-methylated peptide and N-methy-
lated
peptide
[H-(p-Cl)N-MePhe-(p-Cl)N-MePhe-N-MePhe-
CONH2 vs. H-N-MePhe-N-MePhe-N-MePhe-CONH2], beyond
that of the addition of a single chlorine [H-(p-Cl)N-MePhe-N-
MePhe-CONH2 vs. H-N-MePhe-N-MePhe-CONH2].
Conclusions
We found that chlorination significantly increased the effi-
ciency of peptide passage across the PAMPA phospholipids.
For the dipeptides, when GABA and NIP were used as cargos,
chlorinated peptide exhibited greater transport than the N-me-
thylated and chlorinated-N-methylated peptides. These results
indicate that while N-methylation of Phe enhanced PAMPA per-
meability relative to Phe, N-methylation of (p-Cl)Phe decreased
peptide transport through the PAMPA membrane, probably
due to the impact of N-methylation on the peptide conforma-
tion in a way that decreases the interaction between peptide
and membrane and consequently decreases both lipophilicity
and permeability.[47] However, in the case of l-DOPA, chlorinat-
We have introduced the concept of ‘ꢁ la carte’ BBB peptide
shuttles, which could prompt those in the field to abandon
the search for a universal shuttle and, instead, to evolve an op-
timal shuttle for each cargo by fine-tuning structural features
such as peptide length and halogen content. This concept
could likely find application in other areas of drug delivery re-
lated to passage through other biological barriers, such as oral
or transdermal drug delivery. We also observed that BBB per-
meability is not severely hampered (and can even be en-
hanced) when modifying the configuration of backbone ste-
reogenic centers. This observation opens the way to the future
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1594 – 1601 1598

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Fmoc group removal: The Fmoc group was removed by treating
the resin with 20% piperidine in DMF (3–4 mL(g resin)ꢀ1, 2ꢃ1 min
and 1ꢃ10 min). To remove the Fmoc group from Fmoc-N-MePhe-
OH and Fmoc-Pro-OH, an additional treatment with DBU, toluene,
piperidine, DMF (5:5:20:70; 1ꢃ5 min) was performed.
use of d-amino acid containing BBB shuttles in order to im-
prove their ‘in vivo’ stability in the presence of proteases.
Experimental Section
Materials and methods
Coupling methods
Protected amino acids and resins were supplied by Luxembourg
Industries (Tel-Aviv, Israel), Neosystem (Strasbourg, France), Calbio-
chem-Novabiochem AG (Laꢅfelfingen, Switzerland), Bachem AG
(Bubendorf, Switzerland), or Iris Biotech (Marktredwitz, Germany).
PyBOP was supplied by Calbiochem-Novabiochem AG. DIEA, ninhy-
drin, and b-mercaptoethanol were obtained from Fluka Chemika
(Buchs, Switzerland). HOAt was purchased 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 purchased from KaliChemie (Bad Wimpfen,
Germany). Other chemicals used were purchased from Aldrich (Mil-
waukee, WI, USA) and were of the highest purity commercially
available. 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, USA). IAMC
column (10ꢃ4.6 mm, 12 mm, 300 ꢆ, IAM.PC.DD2 column) was from
Regis Technologies Inc. (Morton Grove, IL, USA). Mass spectra were
recorded on a MALDI Voyager DE RP time-of-flight (TOF) spectrom-
eter (PE Biosystems, Foster City, CA, USA) using an ACH matrix.
High-resolution mass spectra were recorded on an LTQ-FT Ultra
(Thermo Scientific). HPLC chromatograms were recorded on
a Waters model Alliance 2695 with photodiode array detector 996
from Waters (Waters, Milford, MA, USA) using a Symmetry C18
column (150ꢃ4.6 mmꢃ5 mm, 100 ꢆ, Waters), solvents: H2O
(0.045% TFA) and MeCN (0.036% TFA), flow: 1 mLminꢀ1 and Millen-
ium software version 4.0. HPLC–MS [Waters model Alliance 2796,
quaternary pump, UV/Vis dual absorbance detector Waters 2487,
ESI-MS model Micromass ZQ and Masslynx version 4.0 software
(Waters)] was done using a Symmetry 300 C18 column (150ꢃ
3.9 mmꢃ5 mm, 300 ꢆ, Waters), solvents: H2O (0.1% formic acid)
and MeCN (0.07% formic acid), flow: 1 mLminꢀ1. The products
were purified in a Waters 600 with dual absorbance detector
(Waters 2487), and a Symmetry C18 column (100ꢃ30 mmꢃ5 mm,
100 ꢆ, Waters), solvents H2O (0.1% TFA) and MeCN (0.05% TFA),
Method 1, coupling of the first amino acid onto the Sieber
resin: N-Protected N-methylated phenylalanine (4 equiv, 160.5 mg),
PyBOP (4 equiv, 208 mg), and HOAt (12 equiv, 163.3 mg) were
added sequentially to the resin in DMF (3 mL) followed by DIEA
(12 equiv, 204 mL). The mixture was allowed to react with intermit-
tent manual stirring for 1.5 h. The solvent was removed by filtra-
tion; the resin was washed with DMF (5ꢃ30 s) and CH2Cl2 (5ꢃ
30 s). The extent of coupling was monitored by the Kaiser colori-
metric assay.
Method 2, coupling of second amino acid and the following
amino acid onto the Sieber resin: The procedure was the same as
for the first, except that N-protected phenylalanine was used. The
coupling was repeated two more times, and the extent of coupling
was checked by the De Clercq test.
Amino acid N-alkylation
The N-methylation of phenylalanine was performed by using the
method described by Miller and Scanlan.[51] This process can be di-
vided into three steps: A) protection and activation with o-nitro-
benzenesulfonyl chloride (o-NBS), B) deprotonation and methyla-
tion, and C) o-NBS removal.
A) Protection and activation with o-NBS: To perform the protec-
tion, o-NBS (3 equiv, 67 mg) and collidine (5 equiv, 66 mL) in CH2Cl2
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) Deprotonation and methylation: Methyl p-nitrobenzensulfo-
nate (4 equiv, 86.9 mg) and MTBD (3 equiv, 43 mL) in DMF were
added to the resin and left for 30 min, and this step was repeated
once.
flow: 10 mLminꢀ1
.
C) o-NBS removal: To remove o-NBS, b-mercaptoethanol
(10 equiv, 70 mL) and DBU (5 equiv, 75 mL) in DMF were added to
the resin, and the mixture was left to react for 10 min under a nitro-
gen atmosphere. This process was repeated once for 40 min.
General protocols for solid-phase synthesis
Syntheses were performed on a 100-mmol scale each; in all cases l-
amino acids were used. Solid-phase peptide elongation and other
solid-phase manipulations were done manually in polypropylene
syringes, each fitted with a polyethylene porous disk. Solvents and
soluble reagents were removed by suction. Washings between syn-
thetic steps were done with DMF (5ꢃ30 s) and CH2Cl2 (5ꢃ30 s)
using 5 mL solvent per gram of resin each time. During couplings
the mixture was allowed to react with intermittent manual stirring.
Levodopa coupling
Amine-protected levodopa (4 equiv, 168 mg), PyBOP (4 equiv,
218 mg), and HOAt (12 equiv, 163.3 mg) were sequentially added
to the resin in DMF (3 mL) followed by DIEA (12 equiv, 204 mL). The
mixture was allowed to react with intermittent manual stirring for
1.5 h. The solvent was removed by filtration, and the resin was
washed with DMF (5ꢃ30 s) and CH2Cl2 (5ꢃ30 s). The coupling was
repeated two more times in the case of coupling on N-MePhe. The
extent of coupling was checked by the De Clercq test or the Kaiser
colorimetric assay.
Identification tests: The Kaiser colorimetric assay[48] was used for
the detection of solid-phase bound primary amines, while the De
Clercq test[49] was used for secondary amines bound to the solid
phase.
Initial conditioning of resin: The Sieber resin[50] was conditioned
by washing with CH2Cl2 (5ꢃ30 s) and DMF (5ꢃ30 s) followed by
a 20% piperidine solution in DMF (2ꢃ1 min and 1ꢃ10 min) to
remove the Fmoc group. Finally, the resin was washed with DMF
(5ꢃ30 s).
GABA/Nip coupling
The procedure was the same as for levodopa, except that Fmoc-
GABA-OH or Fmoc-Nip-OH was used.
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1594 – 1601 1599

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Cleavage of the peptides
CAðtÞ
ð2Þ
T% ¼
ꢃ 100
CDðt0Þ
Final amide peptides were cleaved from the resin using 2% TFA in
CH2Cl2 (6ꢃ3 min).
in which t is time (h), CA(t) is the compound concentration in the
acceptor well at time t, and CD(t0) is the compound concentration
in the donor well at 0 h. The membrane retention (%R) was calcu-
lated from the difference between the total starting amount and
the amounts in donor and acceptor compartments at the end of
the experiment (t=4 h).
Product workup and RP-HPLC purification
After cleavage of the peptides, the solvent was evaporated by N2.
The residue was dissolved in H2O/MeCN (1:1) and then lyophilized.
The peptides were purified by reversed-phase (RP)-HPLC using
a symmetry C18 column (100ꢃ30 mmꢃ5 mm, 100 ꢆ, Waters), at
a 10 mLminꢀ1 flow rate with the following solvents: A, H2O with
0.1% TFA; B, MeCN with 0.05% TFA.
Immobilized artificial membrane chromatography (IAMC)
Retention times were determined using an IAMC column with
phosphatidylcholine (PC), the major phospholipid in cell mem-
branes, which was covalently immobilized (10ꢃ4.6 mm, 12 mm,
300 ꢆ, IAM.PC.DD2 column, Regis Technologies Inc.). The com-
pounds were detected by UV absorption at 220 nm. The chromato-
grams were obtained by 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 (kIAM) following Equa-
tion (3):
Product characterization
The identity of the compounds synthesized was confirmed by
MALDI-TOF MS, HPLC–MS and HRMS. Purity was checked by RP-
HPLC using a symmetry C18 column. Products showed purity of
ꢂ95 (see Supporting Information).
Parallel artificial membrane permeability assay (PAMPA)
The PAMPA[52] was used to determine the capacity of compounds
to cross the BBB by passive diffusion. The effective permeability of
the compounds was measured at an initial concentration of
200 mm. The buffer solution was prepared from a concentrated
one, commercialized by pION, and following the manufacturer’s in-
structions. The pH was adjusted to 7.4 using a 0.5m solution of
NaOH. The compound of interest was dissolved in buffer solution
or water (in the case of cargos alone, evaluated by HPLC–MS) and
1-propanol (20%, co-solvent) to the desired concentration
(200 mm). The PAMPA sandwich was separated, and the donor well
was filled with 200 mL of the compound solution of interest. The
acceptor plate was placed into the donor plate, ensuring that the
underside of the membrane was in contact with buffer. The mix-
ture of phospholipids (4 mL, 20 mgmLꢀ1) in dodecane was added
to the filter of each well, and buffer solution (200 mL) was added to
each acceptor well. The plate was covered and incubated at room
temperature under an atmosphere of saturated humidity for 4 h
with orbital agitation at 100 rpm. After the 4 h, 150 mL per well
from the donor plate and 150 mL per well from the acceptor plate
were transferred to HPLC vials, and 100 mL from each sample were
injected onto an HPLC RP Symmetry C18 column (150ꢃ4.6 mmꢃ
5 mm, 100 ꢆ, Waters). Transport was also confirmed by MALDI-TOF
mass spectrometry to verify that the compound had kept its integ-
rity. For PAMPA evaluated by HPLC–MS the same procedure was
used, except that buffer solution was replaced by water and after
the 4 h, 100 mL per well from the donor plate and 100 mL per well
from the acceptor plate were transferred to HPLC–MS vials, and
10 mL from each donor and 20 mL from each acceptor were inject-
ed onto an HPLC–MS apparatus.
kIAM ¼ ðtRꢀt0Þ=t0
ð3Þ
in which tR is the compound retention time (min), and t0 is the
citric acid retention time (min), indicating the column dead time.
HPLC–MS experiment for quantification of GABA and NIP
The HPLC–MS analysis was performed using a Waters 2795 Separa-
tion Module, operating in positive ion mode with a Waters 2487
Dual l Absorbance detector. The HPLC column used was an RP
Symmetry column (4.6ꢃ150 mm, 5 mm, C18). The samples were
eluted at 1 mLminꢀ1 using different linear gradients of solvents A
(H2O + 0.1% formic acid (v/v)) and B (MeCN + 0.07% formic acid
(v/v)). The detection was performed at 220 nm. External standards
at concentration of 0.2, 0.4, 1, 2, 4, 10, 20, 40, 80, 100, 120, 160,
and 200 mm were prepared by diluting the 1000 mm stock solution
with 20% 1-propanol in water. Prior to the experiments, the com-
pounds were analyzed by HPLC–MS using scan mode to determine
the retention time of each. For the experiments, single-ion moni-
toring (SIM) was used instead of a scan mode.
Acknowledgements
This work was supported by MCI-FEDER (Bio2008-00799) and the
Generalitat de Catalunya (XRB and 2009SGR-1005).
Keywords: blood–brain barrier · drug delivery · membranes ·
peptides
The phospholipid mixture used was a porcine polar brain lipid ex-
tract. Composition: phosphatidylcholine (PC) 12.6%, phosphatidyl-
ethanolamine (PE) 33.1%, phosphatidylserine (PS) 18.5%, phos-
phatidylinositol (PI) 4.1%, phosphatidic acid 0.8%, and 30.9%
other compounds. The effective permeability after 4 h was calculat-
ed using Equation (1); the percentage of transport after was calcu-
lated using Equation (2):
[1] M. Malakoutikhah, M. Teixido, E. Giralt, Angew. Chem. Int. Ed. 2011, 50,
7998–8014; Angew. Chem. 2011, 123, 8148–8165.
[2] W. M. Pardridge, Brain Drug Targeting: The Future of Brain Drug Develop-
ment, Cambridge University Press, Cambridge, 2001.
[3] L. Malavolta, F. R. Cabral, Neuropeptides 2011, 45, 309–316.
[4] W. M. Pardridge, Drug Discovery Today 2007, 12, 54–61.
[5] W. M. Pardridge, Pharm. Sci. Technol. Today 1999, 2, 49–59.
[6] L. L. Rubin, J. M. Staddon, Annu. Rev. Neurosci. 1999, 22, 11–28.
ꢀ
ꢁ
ꢀ218:3
2 ꢃ CAðtÞ
CDðt0Þ
Pe ¼
ꢃ Log 1ꢀ
ꢃ 10ꢀ6cm sꢀ1
ð1Þ
t
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1594 – 1601 1600

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
[7] A. G. De Boer, P. J. Gaillard, Annu. Rev. Pharmacol. Toxicol. 2007, 47, 323–
355.
[31] G. Cannazza, A. Di Stefano, B. Mosciatti, D. Braghiroli, M. Baraldi, F.
Pinnen, P. Sozio, C. Benatti, C. Parenti, J. Pharm. Biomed. Anal. 2005, 36,
1079–1084.
[32] H. Sasaki, Y. Mori, J. Nakamura, J. Shibasaki, J. Med. Chem. 1991, 34,
628–633.
[33] V. Carelli, F. Liberatore, L. Scipione, G. Giorgioni, A. Di Stefano, Bioorg.
Med. Chem. Lett. 2003, 13, 3765–3769.
[34] A. Kakee, H. Takanaga, T. Teraskui, M. Naito, T. Tsuruo, Y. Sugiyama, J.
Neurochem. 2001, 79, 110–118.
[8] Y. Chen, L. Liu, Adv. Drug Delivery Rev. 2012, 64, 640–665.
[9] “Drug Delivery Strategies: BBB Shuttles”, R. Prades, M. Teixidꢀ, E. Giralt
in Nanostructured Biomaterials for Overcoming Biological Barriers, (Eds.:
M. J. Alonso, N. S. Csaba), RSC Drug Discovery Series No. 22, 2012.
[10] W. M. Pardridge, Curr. Opin. Pharmacol. 2006, 6, 494–500.
[11] R. Prades, S. Guerrero, E. Araya, C. Molina, E. Salas, E. Zurita, J. Selva, G.
Egea, C. Lopez-Iglesias, M. Teixidꢀ, M. J. Kogan, E. Giralt, Biomaterials
2012, 33, 7194–7205.
[35] Y. Zhang, G. Q. Liu, Brain Res. 1998, 808, 1–7.
[12] K. A. Witt, T. J. Gillespie, J. D. Huber, R. D. Egleton, T. P. Davis, Peptides
2001, 22, 2329–2343; R. D. Egleton, T. P. Davis, Peptides 2001, 22,
2329–2343.
[13] M. Malakoutikhah, M. Teixido, E. Giralt, J. Med. Chem. 2008, 51, 4881–
4889.
[36] V. E. Shashoua, J. N. Jacob, R. Ridge, A. Campbell, R. J. Baldessarini, J.
Med. Chem. 1984, 27, 659–664.
[37] H. H. Frey, W. Lçscher, Neuropharmacology 1980, 19, 217–220.
[38] L. Galzigna, L. Garbin, M. Bianchi, A.; Marzotto, Arch. Int. Pharmacodyn.
Ther. 1978, 235, 73–85.
[14] M. Malakoutikhah, R. Prades, M. Teixido, E. Giralt, J. Med. Chem. 2010,
53, 2354–2363.
[15] M. Teixidꢀ, E. Zurita, M. Malakoutikhah, T. Tarrago, E. Giralt, J. Am. Chem.
Soc. 2007, 129, 11802–11813.
[16] E. G. Chikhale, N.-Y. Ng, P. S. Burton, R. T. Borchardt, Pharm. Res. 1994,
11, 412–419.
[17] E. M. Arnett, J. M. Gold, J. Am. Chem. Soc. 1982, 104, 636–639.
[18] O. Cruciani, S. Borocci, R. Lamanna, G. Mancini, A. L. Segre, Tetrahedron:
Asymmetry 2006, 17, 2731–2737.
[19] C. Bombelli, S. Borocci, O. Cruciani, G. Mancini, D. Monti, A. L. Segre, A.
Sorrenti, M. Venanzi, Tetrahedron: Asymmetry 2008, 19, 124–130.
[20] C. Bombelli, S. Borocci, F. Lupi, G. Mancini, L. Mannina, A. L. Segre, S.
Viel, J. Am. Chem. Soc. 2004, 126, 13354–13362.
[39] R. Barrett-Jolley, Br. J. Pharmacol. 2001, 133, 673–678.
[40] E. F. Ali, W. E. Bondinell, P. A. Dandridge, J. S. Frazee, E. Garvey, G. R.
Girard, C. Kaiser, T. W. Ku, J. J. Lafferty, G. I. Moonsammy, H. J. Oh, J. A.
Rush, P. E. Setler, O. D. Stringer, J. W. Venslavsky, B. W. Volpe, L. M.
Yunger, C. L. Zirkle, J. Med. Chem. 1985, 28, 653–660.
[41] H. Fischer, M. Kansy, A. Avdeef, Eur. J. Pharm. Sci. 2007, 31, 32–42.
[42] C. L. Gentry, R. D. Egleton, T. Gillespie, T. J. Abbruscato, H. B. Bechowski,
V. J. Hruby, T. P. Davis, Peptides 1999, 20, 1229–1238.
[43] S. J. Weber, D. L. Greene, S. D. Sharma, H. I. Yamamura, T. H. Kramer, T. F.
Burks, V. J. Hruby, L. B. Hersh, T. P. Davis, J. Pharmacol. Exp. Ther. 1991,
259, 1109–1117.
[44] H. M. Liu, X. F. Liu, J. L. Yao, C. L. Wang, Y. Yu, R. Wang, J. Pharmacol. Exp.
Ther. 2006, 319, 308–316.
[21] S. C. D. N. Lopes, A. Fedorov, M. A. R. B. Castanho, ChemMedChem 2006,
1, 723–728.
[22] K. A. Wit, C. A. Slate, R. D. Egleton, J. D. Huber, H. I. Yamamura, V. J.
Hruby, T. P. Davis, J. Neurochem. 2000, 75, 424–435.
[23] C. J. Bowerman, W. Liyanange, A. J. Federation, B. L. Nilsson, Biomacro-
molecules 2011, 12, 2735–2745.
[24] B. M. Liederer, T. Fuchs, D. Vander Velde, T. J. Siahaan, R. T. Borchardt, J.
Med. Chem. 2006, 49, 1261–1270.
[25] A. Sorrenti, M. Diociaiuti, V. Corvaglia, P. Chistolini, G. Mancini, Tetrahe-
dron: Asymmetry 2009, 20, 2737–2741.
[45] T. J. Abbruscato, S. A. Williams, A. Misicka, A. W. Lipkowski, V. J. Hruby,
T. P. Davis, J. Pharmacol. Exp. Ther. 1996, 276, 1049–1057.
[46] M. Teixidꢀ, F. Albericio, E. Giralt, J. Pept. Res. 2005, 65, 153–166.
[47] E. Biron, J. Chatterjee, O. Ovadia, D. Langenegger, J. Brueggen, D.
Hoyer, H. A. Schmid, R. Jelinek, C. Gilon, A. Hoffman, H. Kessler, Angew.
Chem. Int. Ed. 2008, 47, 2595–2599; D. Langenegger, J. Brueggen, D.
Hoyer, H. A. Schmid, R. Jelinek, C. Gilon, A. Hoffman, H. Kessler, Angew.
Chem. Int. Ed. 2008, 47, 2595–2599; Angew. Chem. 2008, 120, 2633–
2637.
[48] E. Kaiser, R. L. Colescott, C. D. Bossinger, P. I. Cook, Anal. Biochem. 1970,
34, 595–598.
[49] A. Madder, N. Farcy, N. G. C. Hosten, H. De Muynck, P. J. De Clercq, J.
Barry, A. P. Davis, Eur. J. Org. Chem. 1999, 2787–2791.
[26] F. H. McDowell, J. E. Lee, West. J. Med. 1970, 113, 44–46.
[27] R. A. Hawkins, A. Mokashi, I. A. Simpson, Exp. Neurol. 2005, 195, 267–
271.
[28] A. Di Stefano, P. Sozio, L. S. Cerasa, Molecules 2008, 13, 46–68.
[29] F. Pinnen, I. Cacciatore, C. Cornacchia, P. Sozio, A. Iannitelli, M. Costa, L.
Pecci, C. Nasuti, F. Cantalamessa, A. Di Stefano, J. Med. Chem. 2007, 50,
2506–2515.
[50] P. Sieber, Tetrahedron Lett. 1987, 28, 2107–2110.
[51] S. C. Miller, T. S. Scanlan, J. Am. Chem. Soc. 1997, 119, 2301–2302.
[52] M. Kansy, F. Senner, K. Gubernator, J. Med. Chem. 1998, 41, 1007–1010.
[30] A. Di Stefano, M. Carafa, P. Sozioa, F. Pinnena, D. Braghirolia, G. Orlan-
doa, G. Cannazzab, M. Ricciutellic, C. Marianeccid, E. Santucci, J. Con-
trolled Release 2004, 99, 293–300.
Received: December 31, 2013
Published online on March 24, 2014
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 1594 – 1601 1601