Design, synthesis and in vitro evaluation on HRPE cells of ascorbic and 6-bromoascorbic acid conjugates with neuroactive molecules

Article abstract of DOI:10.1016/j.bmc.2004.07.043

Preliminary investigations allowed us to anticipate that conjugation of nipecotic acid with l-ascorbate (AA) gave a prodrug endowed with anticonvulsant activity in mice. In view of these results, and in order to get deepen insight into the molecular aspec

Full text of DOI:10.1016/j.bmc.2004.07.043

Bioorganic & Medicinal Chemistry 12 (2004) 5453–5463
Design, synthesis and in vitro evaluation on HRPE
cells of ascorbic and 6-bromoascorbic
acid conjugates with neuroactive molecules
a
b
a
Stefano Manfredini,a, Silvia Vertuani, Barbara Pavan, Federica Vitali,
*
Martina Scaglianti,a Fabrizio Bortolotti,a Carla Biondi,b Angelo Scatturin,a
Puttur Prasadc and Alessandro Dalpiaza
aDepartment of Pharmaceutical Sciences, via Fossato di Mortara 17–19, I-44100 Ferrara, Italy
bDepartment of Biology, General Physiology Section, via Borsari 46, I-44100 Ferrara, Italy
cDepartment of Obstetrics and Gynecology, Medical College of Georgia, Augusta, GA, USA
Received 19March 2004; revised 15 July 2004; accepted 20 July 2004
Available online 20 August 2004
Abstract—Preliminary investigations allowed us to anticipate that conjugation of nipecotic acid with L-ascorbate (AA) gave a pro-
drug endowed with anticonvulsant activity in mice. In view of these results, and in order to get deepen insight into the molecular
aspects at the base of the transport mechanism, a second generation of compounds, based on 6-bromo-6-deoxy-L-ascorbic acid
(BrAA) as the carrier molecule was designed and synthesized. Effects of the chirality of the transported drug was also investigated
on R- and S-nipecotic acid. Interaction and uptake modalities were evaluated in our in vitro model based on human retinal pigment
epithelium cells (HRPE), which expresses the membrane L-ascorbic acid (AA) SVCT2 transporters. A remarkable increase on
SVCT2 affinity was found going from AA to BrAA conjugates, that is, 11 (Ki = 1187 78lM) versus 19 (Ki = 193 14lM) and
12 (Ki = 39.8 3.2lM) versus 20, (Ki = 7.4 0.8lM). Taken together, these data are in agreement with our initial hypothesis on
the possibility to achieve better affinities by conjugation with AA analogs, and also consent to hypothesize the presence of accessory
interactions that may improve transporters recognition.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
Many different strategies have been proposed as a way
to deliver a compound from the blood to the brain,
ranging from invasive approaches to mildest methods
that involve the modification of the physicochemical
properties of drugs.4
Central nervous system (CNS) treatments are one of the
major trust in the pharmaceutical industry but, to take
advantage of many powerful new therapies, the barriers
to the brain must be reckoned with.1 In fact, the diffu-
sion of neuropharmaceutical agents from the blood into
the brain is complicated by the presence of two physio-
logical barriers: the blood–brain barrier (BBB) and the
blood–cerebrospinal fluid barrier (BCSFB). The BBB
is maintained by the endothelial tight junctions within
the brain microvasculature.2,3 The BCSFB is located
at the choroid plexuses and is formed by epithelial cells
held together at their apices by tight junctions.4
This latter approach has been largely explored in terms
of improving drug BBB penetration by increase of the
overall lipophilicity of the molecule by either reducing
the relative number of polar groups, masking them with
lipid carriers and/or attaching lipid moieties to the mol-
ecule.3 These approaches, however, were usually accom-
panied by several side processes that influence
distribution such as rapid elimination, peripheral distri-
bution and binding to plasma proteins.3,5,6 Recently, the
progress of molecular cloning and expression of genes
related to endogenous transporters, suggested the syn-
thesis of prodrugs able to interact with specific trans-
porters of native compounds.3 Several specific
transporters have been identified in boundary tissues
Keywords: Ascorbic acid conjugates; Synthesis; HRPE cells; In vitro
evaluation.
*
Corresponding author. Tel.: +39 0532 291292; fax: +39 0532
291296; e-mail: mv9@dns.unife.it
0968-0896/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.07.043

5454
S. Manfredini et al. / Bioorg. Med. Chem. 12 (2004) 5453–5463
between blood and CNS compartments. Some of them
are involved in the active supply of nutrients (i.e., glu-
cose, amino acids, vitamins) and have been used to ex-
plore prodrugs with improved brain delivery.7–10
According to this point of view, we have recently pro-
posed that drug-conjugation with L-ascorbic acid (AA)
can improve the delivery into the brain, via interaction
with the SVCT2 ascorbate transporters.10
acid, proposed for potential application in AlzheimerÕs
diseases.18–20 The biological activity was investigated
by in vitro studies involving the human SVCT2 iso-
form21 selectively expressed by human retinal pigment
epithelium (HRPE) cells.10 Finally the experimental pro-
cedures previously only communicated will be described
in detail.
SVCT1 and SVCT2 are sodium-dependent transporters
of AA,11,12 which is an essential nutrient that humans
need but are not able to synthesize.13,14 As a conse-
quence, the importance of these human transporters is
of fundamental significance to maintain the necessary
intake:15 SVCT1 is responsible for AA absorption from
intestine and of the recovery by kidney, whereas SVCT2,
expressed by neuroepithelial cells of the choroids plexus
and the retinal pigmented epithelium, allows concentra-
tion in the brain and eye. Brain, spinal cord and adrenal
glands have the highest ascorbate concentrations, this
implies a function beyond its action as a cell-specific en-
zyme co-factor.
2. Results
2.1. Chemistry
The synthetic approach has taken into account the need-
ing to select positions on L-ascorbic acid (5-(R)-[(S)-1,2-
dihydroxy-ethyl]-3,4-dihydroxy-5H-furan-2-one, AA)
and 6-bromo-6-deoxy-L-ascorbic acid (5-(R)-[2-bro-
mo-1-(S)-dihydroxy-ethyl]-3,4-dihydroxy-5H-furan-2-one,
BrAA) structure, different from those probably involved
in the interaction with the transporter. Thus, on a tenta-
tive base, and taking into account that 3,4-dehydro-
ascorbic acid does not interact with the transporters,22
we considered the intact free hydroxyl groups at posi-
tion 3 and 4 as a prerequisite for the interaction with
SVCT2. Thus, esters at 6 and 5 positions on AA and
at 5 position on BrAA have been considered. The ester
linkage was selected in view of the needing of a reversi-
ble type of bond between AA, BrAA and conjugated
drugs.
During a preliminary investigation,10 on transport capa-
bility of SVCT2 transporters, 6-deoxy-6-bromo-L-ascor-
bic acid (BrAA) displayed better affinity for the
transporters than the natural ligand AA. Moreover,
we observed that nipecotic acid, a well known antiepi-
leptic drug, that which poorly permeate the BBB,16 be-
came able of in vivo anticonvulsant effects when
conjugated to AA. These effects were paralleled by good
in vitro affinity for SVCT2 transporters. During the
same investigation we observed that two other drugs,
kynurenic and diclofenamic acids, also became able to
interact with SVCT2 transporters upon ascorbate conju-
gation.
Nipecotic and kynurenic acids needed previous protec-
tion at the amino and hydroxyl functions, respectively.
Thus nipecotic acid was protected as tert-butoxycarb-
onyl derivative (Boc) (1)23 and kynurenic acid as benzyl
derivative (5) (Scheme 1). This required prior conversion
to the kynurenic acid methyl ester 424 simply accom-
plished in MeOH under acidic catalysis rather than
anhydrous HCl as described previously.
According to these informations, and considering the
needs of further studies on the SVCT2 mediated trans-
port into the brain, we have undertaken the present
study in order to: (i) prepare and evaluate the corre-
sponding BrAA conjugates of the selected drugs; (ii)
prepare and evaluate the two diastereoisomers of 6-nip-
ecotyl-ascorbate, in order to explore possible effects due
to the chirality; (iii) investigate in detail the behaviour
towards the SVCT2 transporters of selected drugs and
related prodrugs obtained by conjugation with AA
and BrAA.
2.1.1. Preparation of AA conjugates. Esterification on
AA was carried out by a general procedure involving a
protected precursor, namely the 2,3-dibenzyl derivative
6,25 in the presence of N,N0-dicyclohexylcarbodiimide
(DCC) (Scheme 2). Mixtures of 6- and/or 5- and/or
5,6-di-O-ascorbates were obtained (7a–c, 8a,b, 9a–c)
with large predominance of the 6-isomers over the other
products. Structural attributions were based on 1H
NMR data, selecting the C2–H furan as diagnostic pro-
ton. In particular a deshielding effect was observed in
the case of the 5-O-ascorbates, in comparison with the
corresponding 6-O-ascorbates positional isomers. This
occurrence was confirmed also in the case of the 6-bro-
mo-5-O-ascorbates. The 6-O-ascorbates 7a, 8a, 9a were
Suitable drugs, to be used as model molecule, (Fig. 1)
were nipecotic acid, proposed as a potential anticonvul-
sant,16 kynurenic acid, proposed for the potential con-
trol of neurodegenerative disorders;17 and diclofenamic
Figure 1. Model drugs used in the study.

S. Manfredini et al. / Bioorg. Med. Chem. 12 (2004) 5453–5463
5455
Scheme 1. Reagents and conditions: (i) H2SO4 96%, anhydrous MeOH, reflux; (ii) (a) benzylbromide, K2CO3, anhydrous DMF, (b) NaOH, H2O/
MeOH (1:1).
Scheme 2. Reagents and conditions: (i) DMAP, DCC, anhydrous CH2Cl2 or CH2Cl2/DMF; (ii) TFA, anhydrous CH2Cl2; (iii) H2, C/Pd 10%,
CH3OH.
Scheme 3. Reagents and conditions: (i) DMAP, DCC, anhydrous CH2Cl2 or CH2Cl2/DMF; (ii) TFA, anhydrous CH2Cl2; (iii) H2, C/Pd 10%,
CH3OH.

5456
S. Manfredini et al. / Bioorg. Med. Chem. 12 (2004) 5453–5463
then deprotected by hydrogenolysis (11–13). In the case
of 7a, prior treatment with trifluoroacetic acid (TFA)
was required to remove the Boc-protective group (10).
The same synthetic strategy used for (R,S)-nipecotic
acid derivatives (7a), was applied for the preparation
of the corresponding (3R,2S,2R) 7a and (3S,2S,2R) 7a.
2.1.2. Preparation of BrAA conjugates. Because of BrAA
was endowed with better SVCT2 affinity as compared to
AA, BrAA derivatives of nipecotic, kynurenic and di-
clofenamic acids were prepared as second generation
compounds. To this end BrAA was prepared following
a literature procedure26 and the corresponding 3,4-di-
benzyl derivative (14) was then obtained.27 After esteri-
fication with BrAA in the presence of DCC (Scheme 3)
the 6-bromo-6-deoxy-5-O-ascorbates 15–17 were ob-
tained in satisfactory yields (60%). Next deprotection,
as for AA derivatives, gave final 19, 20 and 21.
2.2. Biology
The kinetic of the AA uptake mediated by hSVCT2 is
described in Figure 2. The rate was found to be hyper-
bolically related to AA concentration, indicating satura-
bility of the uptake process. The Michaelis–Menten
constant (Kt) and Vmax values for the transport process
referred to ascorbate alone were 36 3lM and
4.3 0.3nmol/106 cells/60min, respectively. In a previ-
ous communication we anticipated that AA conjugates
with nipecotic (AA–Nipec) and kynurenic (AA–Kynur)
acids were competitive inhibitors of the SVCT2 medi-
ated uptake of AA (Fig. 2).10
Figure 3. Kt and Vmax values of AA transport obtained in the absence
and in the presence of increasing concentrations of the BrAA
conjugates of nipecotic and kynurenic acids. The reference values for
AA alone are Kt = 36 3lM and Vmax = 4.3 0.3nmol/106 cells.
*p < 0.05, **p < 0.001 significant versus [14C]AA alone.
centration range investigated (0.5–5000lM), whereas
after conjugation with AA it became able to interact
with SVCT2 transporter. The affinity for SVCT2 of
the conjugate with nipecotic acid was not significantly
affected by chirality: as reported in Table 1, the Ki values
of AA–(R)-Nipec (1080 74lM) and AA–(S)-Nipec
(1478 83lM) appear similar to the Ki value of AA–
(R,S)-Nipec (1187 78lM) (Table 1). It is important
to note in Figure 4 that diclofenamic acid (3) was able
itself to strongly inhibit the [14C]AA uptake, showing
a Ki value (3.35 0.16lM) one order of magnitude
higher with respect to the natural ligand, AA. A similar
behaviour has been also observed for the corresponding
BrAA derivative. Moreover, the conjugation with AA
allowed to increase the diclofenamic acid affinity for
SVCT2 (AA–Diclo Ki value = 0.19 0.01 lM), whereas
The inhibition of this transport by the BrAA conjugates
of nipecotic (BrAA–Nipec) and kynurenic (BrAA–
Kinur) acids are also reported in Figure 2. As it can
be observed in Figure 3, the presence of these BrAA
conjugates inhibits the AA uptake, without influence
on the Vmax, but causing an increase of Ki value. This
behaviour indicates that these conjugates are competi-
tive inhibitors of the SVCT2 mediated uptake of AA
as found also for AA–Nipec and AA–Kynur conju-
gates.10 Figure 4 reports the inhibition of 50lM
[14C]AA uptake in the presence of varying concentra-
tions of inhibitors. As it can be observed, nipecotic acid
was not able to inhibit the [14C]AA uptake in the con-
Figure 2. Kinetics of AA uptake in HRPE and inhibition of this transport by different concentrations of BrAA conjugates of nipecotic and kynurenic
acids. These are single representative experiments performed in duplicate.

S. Manfredini et al. / Bioorg. Med. Chem. 12 (2004) 5453–5463
5457
Figure 4. Inhibition of 50lM [14C]AA uptake by AA, nipecotic and diclofenamic acids and their AA and BrAA-conjugates. These are single
representative experiments performed in duplicate.
Table 1. Inhibition constant values (Ki) of AA, BrAA, nipecotic,
kynurenic, diclofenamic acids and their conjugates with AA and
BrAA, obtained by inhibition of 50lM [14C]AA uptake on HRPE cells
or by Lineweaver–Burk analysis
HRPE cells were also proposed as a good cellular model
for in vitro studies of the endogenous activity of SVCT2
transporters. In fact, RT-PCR analysis demonstrated
that SVCT2 but not SVCT1 isoforms are expressed in
human HRPE cells. As previously anticipated,10 the ki-
netic studies for [14C]AA confirmed the existence of a
saturable transport (Fig. 1). The Kt and Vmax values ob-
tained for AA and BrAA showed that the substitution
with bromine at position 6 of AA enhances the affinity
for the SVCT2 transporter (Kt for AA = 36 3lM; Kt
for BrAA = 5.1 0.4lM) without impairing the uptake
into HRPE cells (Vmax for AA and BrAA are 4.3 0.3
and 4.0 0.3nmol/106 cells/60min, respectively).10
Compound
Ki (lM)
AA
BrAA
20.1 1.6a
2.69 0.13 a
No interactiona
1187 78a
1164 99b
1080 74a
1478 83a
192 18b
No interactiona
52.1 4.6b
7.1 1.8b
3.35 0.16a
0.19 0.01 a
21.4 1.8a
Nipecotic acid
AA–(R,S)-Nipec
AA–(R,S)-Nipec
AA–(R)-Nipec
AA–(S)-Nipec
BrAA–(R,S)-Nipec
Kynurenic acid
AA–Kynur
These interesting data suggest that modifications at
position 6 on AA, might be useful to design of a second
generation class of compounds, based on BrAA, en-
dowed with increased affinity. Thus, these new conju-
gates were synthesized as derivatives at position 5 on
BrAA of the same model drugs. The interaction with
SVCT2 transporters, concerning the parent nipecotic,
kynurenic (2) and diclofenamic (3) acids and their AA
and BrAA conjugates was then investigated. As re-
ported in Table 1, the Ki values of AA (20.1 1.6lM)
and BrAA (2.69 0.13), obtained by inhibition experi-
ments, confirm that the 6-Br-substituent on AA en-
hances the affinity for SVCT2 of about one order of
magnitude. Data reported in Table 1 confirm that nipe-
cotic and kynurenic acids are totally unable to interact
with SVCT2 transporters, whereas their AA-conjugates
show a significant inhibition potency on [14C]AA uptake
(Table 1). In view of the fact that chirality of the model
drug may also affect transporters interactions, the effects
of R- and S-nipecotic acid on the properties of the
respective AA-conjugates has been investigated. As
shown in Figure 4, no significant differences have been
found among the inhibitory properties of the AA–
(R,S)-Nipec mixture or the AA–(R)-Nipec or AA–(S)-
Nipec diasteroisomers. In fact, their Ki values are in
the same range (Table 1). These results show that the
SVCT2 transporters interacts with the ascorbic acid
conjugate of nipecotic acid irrespectively from its
chirality.
BrAA–Kynur
Diclofenamic acid
AA–Diclo
BrAA–Diclo
The conjugate with nipecotic acid has been evaluated either as a 1:1
mixture (AA–(R,S)-Nipec) or single diasteroisomers (AA–(R)-Nipec
and AA–(S)-Nipec.
a Ki values obtained by inhibition experiments.
b Ki values obtained by Lineweaver–Burk analysis.
its conjugation with BrAA caused the decrease in po-
tency (BrAA–Diclo Ki value = 21.4 1.8lM). In con-
trast with this latter finding, the Lineweaver–Burk
analysis of data, reported in Figure 2, indicate that the
BrAA conjugation of nipecotic acid increased its affinity
towards SVCT2 transporter, with respect to the
AA conjugation. The Ki values reported in Table 1
are in fact 1164 99lM for AA–(R,S)-Nipec and
192 18lM for BrAA–Nipec. A similar behaviour
was also found for kynurenic acid conjugates with Ki
values of 52.1 4.6lM for AA–Kynur and
7.1 1.8lM for BrAA–Kynur (Table 1).
3. Discussion
We have previously discovered that conjugation with
AA of model drugs, that do not normally cross BBB,
gave compounds able to reach CNS.10 We proposed,
using a model based on human HRPE cells, an interac-
tion mediated by human AA SVCT2 transporters at the
base of the observed delivery. During this study, human
On the other hand, as reported in Table 1, we have
found a remarkable increase on SVCT2 affinity going
from 11 (Ki = 1187 78lM) to 19 (Ki = 19 3 14lM).

5458
S. Manfredini et al. / Bioorg. Med. Chem. 12 (2004) 5453–5463
A similar behaviour has been found also for the kynur-
enic acid conjugates with AA (12, Ki = 39.8 3.2lM)
and BrAA (20, Ki = 7.4 0.8lM). Taken together, these
data are in agreement with our initial hypothesis on the
possibility to achieve better affinities by conjugation
with BrAA rather than AA, but also consent to suppose
the presence of accessory interactions that may improve
transporters recognition. Similarly to the AA conjugates
of nipecotic and kynurenic acids,10 the BrAA derivatives
interact toward SVCT2 as competitive inhibitors (Fig.
3). Because a competitive inhibition is indicative of an
interaction with SVCT2 in the same binding site of
AA, this implies that AA- and BrAA-conjugates may
be potentially transported by SVCT2, while the parent
compounds are certainly not. These findings are of great
significance for the design of newer derivatives endowed
with improved affinity. However, inhibition experiments
performed on HRPE cells using diclofenamic acid (3)
and its AA or BrAA conjugates 13 and 21, (Fig. 4),
showed a completely different pattern with respect to
nipecotic and kynurenic acids (2). In fact, as previously
anticipated by us,10 diclofenamic acid (3) is able itself to
interact with the SVCT2 transporters with an affinity
(Ki = 3.35 0.16lM) similar to that of BrAA. More-
over, upon conjugation with AA (13), the inhibition
potency of diclofenamic acid becomes two order of
magnitude higher (Ki = 0.19 0.01 lM) with respect to
the endogenous ligand (AA). These data, completely
unexpected, disclosed further perspectives. Therefore,
the conjugation of diclofenamic acid with BrAA (21)
was investigated, resulting in a drastic decrease in the
affinity with respect to both the parent drug (3) and
the AA-conjugate (13). In fact, the Ki value of this
BrAA-conjugate 21 (Ki = 21.4 1.8lM) appears similar
to that of the endogenous ligand (Table 1).
ica gel precoated F254 Merck plates with detection
under 254nm UV lamp and/or by spraying with a diluted
potassium permanganate solution. Nuclear magnetic
resonance (1H NMR) spectra were determined for solu-
tion in CDCl3–CD3OD–DMSO-d6 on a Bruker AC-200
spectrometer and peak positions are given in parts per
million (d) downfield from tetramethylsilane as internal
standard, whereas coupling constants (J) in Hertz. Melt-
ing points were obtained in open capillary tubes and are
uncorrected. Column chromatography was performed
with Merck 60–200mesh silica gel. Ambient temperature
was 22–25ꢁC. All drying operations were performed
over anhydrous magnesium or sodium sulfate. Micro-
analysis, unless indicated, were in agreement with calcu-
lated values within 0.4%.
5.1. Preparation of 4-hydroxyquinoline-2-carboxylic-acid
methyl ester (4)
Kynurenic acid (500mg, 2.65mmol) was suspended in
distilled MeOH (25mL) and 96% H2SO4 (30–40 drops)
was added. The clear solution was heated at reflux for
24h (TLC, CH2Cl2/MeOH, 9/1) and the solvent was
evaporated under vacuo, the residue was dissolved in
EtOAc and washed with NaHCO3 satd solution. The or-
ganic phase was dried and evaporated under vacuo. The
crude residue was purified by column chromatography
(AcOEt/hexane gradient 1/1 to 8/2). Evaporation of
appropriate fractions gave the expected compound 4.
Pale yellow solid, mp 227,29 yield 75%. 1H NMR
(DMSO-d6): d 3.96 (s, 3H, CH3), 6.66 (s, 1H, C3–H),
7.33–7.41 (m, 1H, C7–H, Ph), 7.65–7.79(m, 1H, C6–
H), 7.94 (d, 1H, J = 8.32Hz, C5–H), 8.08 (d, 1H,
J = 8.17Hz, C8–H), 11.9(br s, 1H, OH Ar). MALDI-
TOF MS: m/z 204.5Da (M+H)+ C11H9NO3 requires
203.19.
4. Conclusions
5.2. Preparation of 4-benzyloxyquinoline-2-carboxylic-
acid (5)
In view of the interesting results previously communi-
cated by us, we have prepared a second generation com-
pounds, based on BrAA, that consented us to confirm
our previous hypotheses, further validating our ap-
proach. Surprisingly, diclofenamic acid behaved com-
pletely different than nipecotic and kynurenic acids,
indeed: (i) it was able itself to interact with the trans-
porters but by with a kind of pattern, different from
AA and its conjugates; (ii) conjugation to BrAA, differ-
ently from conjugation to AA, resulted detrimental for
the interaction with the transporters.
To a stirred solution, of 4 (180mg, 0.887mmol) in anhy-
drous DMF, anhydrous K2CO3 (146.5mg, 1.06mmol)
was added. Benzyl bromide (126lL, 1.06mmol) was
then added dropwise, and the resulting mixture was
heated at reflux under argon atmosphere. After 4h, sol-
vent was removed by vacuum. The residue was dissolved
in H2O/MeOH (5.25mL, 1:1) and NaOH (0.1mmol)
was added and the solution stirred at room temperature.
After 2h, the MeOH was removed and the resulting
aqueous solution was acidified (pH4–5), with 2M
HCl, and then extracted with EtOAc (3 · 30mL). The
combined organic phases were dried and the solvent re-
moved under reduced pressure. The crude residue was
purified by column chromatography (CH2Cl2/MeOH,
9/1) to provide 220mg of 5.
Moreover, in the case of R- and S-nipecotic acid deriv-
atives, the interaction and transport occurred irrespec-
tively of the chirality of the conjugated drug. These
data, are in agreement with the inability of nipecotic
acid itself to interact with transporters, being only trans-
ported after conjugation to AA or BrAA.
1
White solid, mp 143–146ꢁC. H NMR (CDCl3): d 5.44
(s, 2H, CH2–Ph), 7.29–7.78 (m, 6H, C3–H and Ph),
7.80–78.05 (m, 2H, C6–H and C7–H), 8.29(d, 1H,
J = 8.2Hz, C5–H), 8.56 (d, 1H, J = 8.3Hz, C8–H).
MALDI-TOF MS: m/z 280.2Da (M+H)+ C17H13NO3
requires 279.29.
5. Experimental
Reaction courses and product mixtures were routinely
monitored by thin-layer chromatography (TLC) on sil-

S. Manfredini et al. / Bioorg. Med. Chem. 12 (2004) 5453–5463
5459
5.3. General synthetic procedure for the preparation of the
esters 7–9, 14–16
4.88 (d, 1H, J = 2.2Hz, furan C2–H). MALDI-TOF
MS: m/z 591.1Da (M+Na)+; 607.2 (M+K)+,
C31H37NO9 requires 567.63.
To a stirred solution of the appropriate carboxylic acid
derivative (4.36mmol), DMAP (53mg, 0.436mmol) and
3,4-dibenzyloxy-5-(1,2-dihydroxy-ethyl)-5H-furan-2-one
(1.55g, 4.36mmol) in anhydrous CH2Cl2 (50mL) at 0ꢁC
DCC (1.33mg, 6.49mmol) was added and the mixture
was stirred at room temperature and under argon
atmosphere. The reaction was monitored by TLC
(EtOAc/hexane, 4/6) and when no further improvement
in the formation of the expected product was observed,
the mixture was filtered and CH2Cl2 was removed in
vacuo. The solution was then taken up in EtOAc,
washed with 0.5M HCl (20mL), brine (20mL), satd
NaHCO3 (20mL) and dried. The organic layer was
dried and removed in vacuo and the crude residue was
purified by chromatography eluting with EtOAc/hexane
(gradient 1/9 ! 4/6) to provide the expected 6-ascorbic
acid esters. The formation of the other two possible
derivatives, namely the 5-esters and/or 6,5-diesters at
the ascorbyl moiety, was also observed, in 5–25% ratio
(HPLC), as indicated below. Compound 7 was obtained
either as a mixture of two diastereoisomers starting from
racemic nipecotic acid, or as the pure R,S,R and S,S,R
diasteroisomers starting from R- and S-nipecotic acid,
respectively. These latter were obtained as reported in
literature.30
Compound 7c (3R,2S,2R and 3S,2S,2R of both posi-
tional isomers): pale yellow foam, yield 20%. H NMR
(CDCl3), d: diagnostic proton 4.83 (d, 1H, J = 2Hz, fur-
an C2–H). MALDI-TOF MS: m/z 779.9Da (M+H)+;
802.4Da (M+Na)+; 818.4 (M+K)+, C42H54N2O12 re-
quires 778.37.
1
5.5. N-Boc-piperidine-3-(R,S)-carboxylic acid [1-(S)-(3,4-
dibenzyloxy-5-oxo-2,5-dihydrofuran-2-(R)-yl)-2-bromo-
ethyl] ester (15)
1
White foam, yield 60%. H NMR (CDCl3): d 1.40–1.75
(m, 12H, C5–H, C4–H and Boc), 1.90–2.08 (m, 1H, C4–
H), 2.25–2.47 (m, 1H, C3–H), 2.60–2.95 (m, 2H, C6–H
and C2–H), 3.43–3.58 (m, 2H, hydroxy-ethyl C2–H),
3.82–4.26 (m, 2H, C6–H and C2–H), 5.01 (d, 1H,
J = 2.19Hz, furan C2–H), 5.06–5.22 (m, 4H, CH2–Ph),
5.28–5.38 (m, 1H, hydroxy-ethyl C1–H), 7.17–7.51 (m,
10H, Ph). MALDI-TOF MS: m/z 631.1Da (M+H)+;
654.1Da (M+Na)+; 670.0Da (M+K)+, C31H36BrNO8
requires 630.52.
5.6. 4-Benzyloxyquinoline-2-carboxylic acid [2-(3,4-di-
benzyloxy-5-oxo-2,5-dihydrofuran-2 (R)-yl)-2-(S)-hydr-
oxy-ethyl] ester (8a) and related positional isomer (8b)
5.4. N-Boc-piperidine-3-(R,S/R/S)-carboxylic acid
[2-(3,4-dibenzyloxy-5-oxo-2,5-dihydrofuran-2-(R)-yl)-
2-(S)-hydroxy-ethyl] ester (7a), related positional
isomer (7b) and diester (7c)
Compound 8a: white solid, mp 154–156ꢁC, yield 56%.
1H NMR (CDCl3): d 4.39–4.45 (m, 1H, hydroxy-ethyl
C2–H), 4.55–4.70 (m, 2H, hydroxy-ethyl C1–H),
4.76 (d, 1H, J = 1.84Hz, furan C2–H), 5.1–5.29(m,
6H, CH2–Ph), 7.10–7.65 (m, 17H, C3–H, C7–H and
Ph), 7.68–7.78 (m, 1H, C6–H), 8.14 (d, 1H,
J = 8.33Hz, C5–H), 8.25 (d, 1H, J = 8.49Hz, C8–H).
MALDI-TOF MS: m/z 618.4Da (M+H)+; 640.8Da
(M+Na)+; 657.2Da (M+K)+, C37H31NO8 requires
617.64.
Compound 7a was obtained as a mixture of the two
diastereoisomers and/or as the single diastero-
isomer (3R,2S,2R and 3S,2S,2R): white foam, yield
1
60%. H NMR (CDCl3): d 1.30–2.00 (m, 13H, C5–H,
C4–H, Boc), 2.44–2.68 (m, 1H, C3–H), 3.2–3.8
(m, 4H, C6–H and C2–H), 3.98–4.3 (m, 3H, hydroxy-
ethyl C1–H and C2–H), 4.65 (d, 1H, J = 2Hz, furan
C2–H), 5.05–5.30 (m, 4H, CH2–Ph), 7.14–7.48 (m,
10H, Ph).
Compound 8b: pale yellow oil, yield 5%. 1H NMR
(CDCl3): diagnostic proton: d 4.90–4.95 (br, 1H, furan
C2–H). MALDI-TOF MS: m/z 879.6Da (M+H)+;
901.6Da (M+Na)+; 917.8Da (M+K)+, C54H42N2O10 re-
quires 878.92.
MALDI-TOF MS: m/z 568.8Da (M+H)+; 591.1Da
(M+Na)+; 607.2 (M+K)+, C31H37NO9 requires 567.63.
Compound 7a (3R,2S,2R) obtained from R-nipecotic
acid: pale yellow foam, yield 30%. H NMR (CDCl3)
1
diagnostic proton: 2.43–2.61 (m, 1H, C3–H), 4.64 (d,
1H, J = 1.8Hz, furan C2–H). MALDI-TOF MS: m/z
590.8Da (M+Na)+; 606.1 (M+K)+, C31H37NO9 requires
567.63.
5.7. 4-Benzyloxyquinoline-2-carboxylic acid [1-(S)-(3,4-
dibenzyloxy-5-oxo-2,5-dihydrofuran-2-(R)-yl)-2-bromo-
ethyl] ester (16)
White solid, mp 151–154ꢁC, yield 60%. 1H NMR
(CDCl3): d 3.67–3.85 (m, 2H, hydroxy-ethyl C2–H),
4.94–5.35 (m, 7H, furan C2–H and CH2–Ph), 5.54–
5.63 (m, 1H, hydroxy-ethyl C1–H), 7.05–7.85 (m, 18H,
C3–H, C6–H and C7–H, Ph), 8.22–8.35 (m, 2H, C5–H
and C8–H).
Compound 7a (3S,2S,2R) obtained from S-nipecotic
acid: pale yellow foam, yield 45%. H NMR (CDCl3)
diagnostic proton: 2.46–2.64 (m, 1H, C3–H), 4.645 (d,
1H, J = 2Hz, furan C2–H). MALDI-TOF MS: m/z
590.8Da (M+Na)+; 605.1 (M+K)+, C31H37NO9 requires
567.63.
1
MALDI-TOF MS: m/z 681.0Da (M+H)+; 704.2Da
(M+Na)+; 720.0Da (M+K)+, C37H30BrNO7 requires
680.54.
Compound 7b (3R,1S,2R and 3S,2S,2R): pale yellow
foam, yield 5%. H NMR (CDCl3), diagnostic proton:
1

5460
S. Manfredini et al. / Bioorg. Med. Chem. 12 (2004) 5453–5463
5.8. [2-(2,6-Dichlorophenylamino)-phenyl]-acetic acid [2-
(3,4-dibenzyloxy-5-oxo-2,5-dihydrofuran-2-(R)-yl)-2-(S)-
hydroxy-ethyl] ester (9a), related positional isomer (9b)
and diester (9c)
Compound 10 (3R,2S,2R) obtained from R-nipecotic
acid: pale yellow foam, yield 54%. H NMR (CDCl3)
1
diagnostic proton: 4.64 (d, 1H, J = 2Hz, furan C2–H).
MALDI-TOF MS: m/z 467.6Da (M+H)+; 489.8Da
+
(M+Na)+; 505.9(M+K) , C26H29NO7 requires 467.51.
Compound 9a: white foam, yield 65%. 1H NMR
(CDCl3): d 3.84 (s, 2H, CH2–Ar), 4.0–4.16 (m, 1H, hyd-
roxy-ethyl C2–H), 4.3, 4.4 (ABX, sist. 2H,
JAB = 11.54Hz, JAX = 6.95, JBX = 4.82Hz hydroxy-
ethyl C1–H), 4.64 (d, 1H, J = 2.07Hz, furan C2–H),
5.06–5.22 (m, 4H, CH2–Ph), 6.53 (d, 1H, J = 7.93Hz,
phenyl-acetyl H3), 6.90–7.40 (m, 16H, Ph), 6.73 (s, 1H,
NH). MALDI-TOF MS: m/z 635.4Da (M+H)+;
658.5Da (M+Na)+; 674.9Da (M+K)+, C34H29Cl2NO7
requires 634.50.
Compound 10 (3S,2S,2R) obtained from S-nipecotic
acid: pale yellow foam, yield 48%. H NMR (CDCl3)
diagnostic proton: 4.64 (d, 1H, J = 2Hz, furan C2–
H). MALDI-TOF MS: m/z 467.6Da (M+H)+;
489.8Da (M+Na)+; 505.9(M+K) +, C26H29NO7 requires
467.51.
1
5.11. Piperidine-3-(R,S)-carboxylic acid [1-(S)-(3,4-dib-
enzyloxy-5-oxo-2,5-dihydrofuran-2-(R)-yl)-2-bromo-
ethyl] ester (18)
Compound 9b: pale yellow oil, yield 15%. 1H NMR
(CDCl3), d: diagnostic proton 4.80 (d, 1H, furan C2–
H). MALDI-TOF MS: m/z 658.4Da (M+Na)+;
674.5Da (M+K)+, C34H29Cl2NO7 requires 634.50.
1
Yellow oil, yield 70%. H NMR (CDCl3): d 1.40–1.76
(m, 3H, C5–H, C4–H), 1.81–2.17 (m, 1H, C4–H),
2.60–3.20 (m, 5H, C3–H, C6–H and C2–H), 3.45–3.61
(m, 2H, hydroxy-ethyl C2–H), 5.01 (d, 1H,
J = 1.78Hz, furan C2–H), 5.05–5.25 (m, 4H, CH2–Ph),
5.28–5.38 (m, 1H, hydroxy-ethyl C1–H), 7.20–7.50 (m,
10H, Ph). MALDI-TOF MS: m/z 531.9Da (M+H)+;
553.6Da (M+Na)+; 569.7Da (M+K)+, C26H28BrNO6
requires 530.41.
Compound 9c: pale yellow oil, yield 15%. 1H NMR
(CDCl3), d: diagnostic proton 4.77 (d, 1H, J = 1.88Hz,
furan C2–H). MALDI-TOF MS: m/z 936.0Da
(M+Na)+; 952.2 (M+K)+, C48H38Cl4N2O8 requires
912.63.
5.9. [2-(2,6-Dichlorophenylamino)-phenyl]-acetic acid [1-
(S)-(3,4-dibenzyloxy-5-oxo-2,5-dihydrofuran-2-(R)-yl)-2-
bromo-ethyl] ester (17)
5.12. General procedure to remove the benzyl protecting
groups
To a solution of the appropriate esters 8a, 9a, 10, 16–18
(2.14mmol) in MeOH or EtOAc (50mL) 10% Pd/C
(200mg) was added, and the mixture was stirred 5–
24h under hydrogen atmosphere. After complete disap-
pearance of the starting material (TLC, CH2Cl2/MeOH,
9/1) the mixture was filtered on Celite pad and the sol-
vent was removed in vacuo. The crude residue was puri-
fied by column chromatography (CH2Cl2/MeOH
gradient from 9/1 to 7/3) to give the expected depro-
tected compounds 11–13, 19–21.
Yellow foam, yield 45%. 1H NMR (CDCl3): d 3.50–3.59
(m, 2H, hydroxy-ethyl C2–H), 3.62–3.90 (m, 2H, CH2–
Ar), 4.80–5.12 (m, 5H, furan C2–H and 4H, CH2–Ph),
5.30–5.40 (m, 1H, hydroxy-ethyl C1–H), 6.53 (d, 1H,
J = 7.92Hz, phenyl-acetyl H3), 6.82–7.59 (m, 17H, Ph
and NH). MALDI-TOF MS: m/z 698.2Da (M+H)+;
720.9Da (M+Na)+; 737.1Da (M+K)+, C34H28BrCl2-
NO6 requires 697.40.
5.10. Piperidine-3(R,S/R/S)-carboxylic acid [2-(3,4-di-
benzyloxy-5-oxo-2,5-dihydrofuran-2-(R)-yl)-2-(S)-hydr-
oxy-ethyl] ester (10)
5.13. Piperidine-3-(R,S/R/S)-carboxylic acid [2-(3,4-di-
hydroxy-5-oxo-2,5-dihydrofuran-2-(R)-yl)-2-(S)-hydroxy-
ethyl] ester (11)
To a stirred solution, of 7a (1.21g, 2.1mmol) in
anhydrous CH2Cl2 was added, in a dropwise manner,
TFA (3.8mL) and the mixture was stirred for 2h under
argon atmosphere. Most of solvent was then removed in
vacuo and the residue was added with CH2Cl2 (20mL)
and the solution washed with NaHCO3 satd
(2 · 20mL). The organic layer was dried and the solvent
removed in vacuo. The crude residue was purified by
column chromatography (CH2Cl2/MeOH 9/1) to
provide 630mg of 10.
Compound 11 was obtained as a mixture of the two
diastereoisomers and/or as the single diasteroisomer.
Mixture: pale yellow foam, yield 30%. 1H NMR
(DMSO-d6): d 1.40–1.74 (m, 3H, C5–H), 1.76–1.94 (m,
1H, C4–H), 2.57–3.20 (m, 5H, C3–H, C6–H and
C2–H), 3.78–4.35 (m, 4H, hydroxy-ethyl C1–H, C2–H
and furan C2–H), 6.17 (br s, 1H, NH). MALDI-TOF
MS: m/z 288.7Da (M+H)+, C12H17NO7 requires
287.27. Anal. (C12H17NO7) C, H, N.
Pale yellow oil, yield 63%. 1H NMR (CDCl3): d
1.63–2.10 (m, 4H, C5–H, C4–H), 2.65–3.50 (m, 5H,
C3–H, C6–H, C2–H), 4.05–4.59(m, 3H, hydroxy-ethyl
C2–H and C1–H), 4.64 (d, 1H, J = 2Hz, furan C2–H),
4.90–6.00 (m, 4H, CH2–Ph and NH), 7.05–7.5 (m,
10H, Ph). MALDI-TOF MS: m/z 469.0Da (M+H)+;
490.7 (M+Na)+; 506.6 (M+K)+, C26H29NO7 requires
467.51.
Compound 11 (3R,2S,2R) obtained from R-nipecotic
acid: pale yellow solid, mp 58–60ꢁC, yield 24%.
25
D
½aŠ þ 33 (c 0.41%, H2O); 1H NMR (DMSO-d6)
diagnostic proton: 2.30–2.38 (m, 1H, C3–H), 4.31
(d, 1H, J = 3Hz, furan C2–H). MALDI-TOF
MS: m/z 288.8Da (M+H)+; 309.8Da (M+Na)+;
325.7 (M+K)+, C12H17NO7 requires 287.27. Anal.
(C12H17NO7) C, H, N.

S. Manfredini et al. / Bioorg. Med. Chem. 12 (2004) 5453–5463
5461
Compound 11 (3S,2S,2R) obtained from S-nipecotic
acid: pale yellow solid, mp 60–62ꢁC, yield 39%.
D
nostic proton: 2.58–2.72 (m, 1H, C3–H), 4.26 (d,
2H, J = 8.02Hz, phenyl-amino H3 and H5). MALDI-
TOF MS: m/z 477.6Da (M+Na)+; 493.9Da (M+K)+,
C20H17Cl2NO7 requires 454.26. Anal. (C20H17Cl2NO7)
C, H, N.
25
1
½aŠ þ 42 (c 0.4%, H2O); H NMR (DMSO-d6) diag-
1H,
J = 2.8Hz,
furan
C2–H).
MALDI-TOF
MS: m/z 288.0Da (M+H)+, 309.9Da (M+Na)+;
325.7 (M+K)+, C12H17NO7 requires 287.27. Anal.
(C12H17NO7) C, H, N.
5.18. [2-(2,6-Dichlorophenylamino)-phenyl]-acetic acid [1-
(S)-(3,4-dihydroxy-5-oxo-2,5-dihydrofuran-2-yl)-2-bro-
mo-ethyl] ester (21)
1
5.14. Piperidine-3-(R,S)-carboxylic acid [1-(S)-(3,4-
dihydroxy-5-oxo-2,5-dihydrofuran-2-(R)-yl)-2-bromo-
ethyl] ester (19)
White foam, yield 35%. H NMR (DMSO-d6): d 3.58–
3.98 (m, 4H, CH2–Ar and hydroxy-ethyl C2–H), 5.09
(d, 1H, J = 2.16Hz, furan C2–H), 5.39–5.52 (m, 1H,
hydroxy-ethyl C1–H), 6.25 (d, 1H, J = 7.08Hz, phenyl-
acetyl H3), 6.80–6.98 (m, 2H, phenyl-acetyl H5 and
NH), 7.02–7.30 (m, 3H, phenyl-acetyl H4/H6 and phe-
nyl-amino H4), 7.52 (d, 2H, J = 7.99Hz, phenyl-amino
H3, H5), 8.72 (br s, 1H, furan C4–OH), 11.45 (br s,
1H, furan C3-OH).
Yellow solid, mp 140–143ꢁC, yield 30%. 1H NMR
(CD3OD-d6): d 1.67–2.13 (m, 4H, C5–H, C4–H), 2.8–
3.88 (m, 7H, C3–H, C6–H, C2–H and hydroxy-ethyl
C2–H), 4.64 (d, 1H, J = 2.13Hz, furan C2–H), 5.48–
5.58 (m, 1H, hydroxy-ethyl C1–H). MALDI-TOF MS:
m/z 350.7–352.5Da (M+H)+; 391.0–392.6Da (M+K)+,
C12H16BrNO6 requires 350.16. Anal. (C12H16BrNO6)
C, H, N: C calcd 41.16; found 40.91.
MALDI-TOF MS: m/z 518.2Da (M+H)+; 541.0Da
(M+Na)+; 557.0Da (M+K)+, C20H16BrCl2NO6 requires
517.15. Anal. (C20H16BrCl2NO6) C, H, N.
5.15. Hydroxyquinoline-2-carboxylic acid [2-(3,4-dihyd-
roxy-5-oxo-2,5-dihydrofuran-2-(R)-yl)-2-(S)-hydroxy-
ethyl] ester (12)
5.19. Stability
The stability in water (pH7 and 1.2) and at enzymatic
hydrolysis (porcine liver esterase, PLE) of test com-
pounds was assayed as described previously.28 Com-
pounds were stable in water from 24h to 4weeks and
showed halflives ranging from 60 to >180min
(10 > 6 > 4) in the case of PLE enzymatic hydrolysis.
1
White solid, mp 138ꢁC, yield 40%. H NMR (DMSO-
d6): d 4.12–4.28 (m, 1H, hydroxy-ethyl C1–H), 4.40–
4.52 (m, 2H, hydroxy-ethyl C2–H), 4.87 (d, 1H,
J = 1.8Hz, furan C2–H), 6.78 (s, 1H, C3–H), 7.30–7.47
(m, 1H, C7–H), 7.65–7.80 (m, 1H, C6–H), 7.95 (d, 1H,
J = 8.4Hz, C5–H), 8.08 (d, 1H, J = 8.8Hz, C8–H),
8.46 (br s, 1H, furan C4–OH), 11.24 (br s, 1H, furan
C3–OH), 12.06 (br s, 1H, OH Ar). MALDI-TOF MS:
m/z 348.5Da (M+H)+; 370.8Da (M+Na)+; 386.8Da
5.20. HRPE cell culture
HRPE cell line was routinely grown in 1:1 mixture of
DulbeccoÕs modified EagleÕs and HamÕs F12 media,
supplemented with 10% FBS, 50lmg/mL streptomycin
and 50IU/mL penicillin (Gibco Laboratories, Invitro-
gen Italia, Milan, Italy) at 37ꢁC in 5% CO2. Cells
employed for uptake measurements10 were seeded in
24-well tissue culture plates and grown to confluence
(2–3days).
(M+K)+,
(C16H13NO8) C, H, N: H calcd 3.77; found 3.75.
C16H13NO8
requires
347.28.
Anal.
5.16. Hydroxyquinoline-2-carboxylic acid [1-(S)-(3,4-
dihydroxy-5-oxo-2,5-dihydrofuran-2-(R)-yl)-2-bromo-
ethyl] ester (20)
White solid, mp 203–206ꢁC, yield 74%. 1H NMR
(DMSO-d6): d 3.81–4.05 (m, 2H, hydroxy-ethyl C2–
H), 5.15 (d, 1H, J = 2.13Hz, furan C2–H), 5.63–5.71
(m, 1H, hydroxy-ethyl C1–H), 6.56 (s, 1H, C3–H),
7.28–7.44 (m, 1H, C7–H), 7.65–7.80 (m, 1H, C6–H),
7.93 (d, 1H, J = 8.26Hz, C5–H), 8.08 (d, 1H,
J = 7.7Hz, C8–H), 12.12 (br s, 1H, OHAr). MALDI-
TOF MS: m/z 411.9Da (M+H)+; 432.8Da (M+Na)+;
449.3Da (M+K)+, C16H12BrNO7 requires 410.17. Anal.
(C16H12BrNO7) C, H, N.
5.21. SVCT2 transporter interactions
Transport assays were performed following the method
described by Rajan et al.21 Briefly, the uptake buffer was
prepared fresh each time, the composition was: 25mM
Hepes/Tris (pH7.5), 140mM NaCl, 5.4mM KCl,
1.8mM CaCl2, 0.8mM MgSO4, 5mM glucose. DTT
(1mM) (Sigma, St. Louis, MO, USA) was also added
to the uptake buffer to prevent the oxidation of AA.
At this concentration, DTT had no effect on the
transport process. Incubation time for the transport
measurements was 60min (within the time course of
linear uptake, Fig. 2) at 37ꢁC, after this time the uptake
buffer containing the radioactive substrate was aspirated
off and cells were washed with 2 · 2mL of ice-cold
uptake buffer. Cells were then solubilized in 250lL of
0.2M NaOH solution containing 0.5% CHAPS
(Sigma), transferred to vials and radioactivity associated
with the cells was evaluated by liquid scintillation
spectrometry.
5.17. [2-(2,6-Dichlorophenylamino)-phenyl]-acetic acid [2-
(3,4-dihydroxy-5-oxo-2,5-dihydrofuran-2-(R)-yl)-2-(S)-
hydroxy-ethyl] ester (13)
White solid, mp 152–154ꢁC, yield 73%. 1H NMR
(DMSO-d6): d 3.84 (s, 2H, CH2–Ar), 3.93–4.14 (m,
3H, hydroxy-ethyl C1–H and C2–H), 4.59(d, 1H,
J = 2.41Hz, furan C2–H), 6.23 (d, 1H, J = 7.93Hz, phe-
nyl-acetyl H3), 6.79–6.90 (m, 1H, J = 6.98Hz, phenyl-
acetyl H3), 7.01–7.30 (m, 4H, Ph and NH), 7.53 (d,

5462
S. Manfredini et al. / Bioorg. Med. Chem. 12 (2004) 5453–5463
The kinetics of [14C]AA (6mCi/mmol; NEN Life Sci-
ence, Boston, MA, USA) uptake, mediated by SVCT2,
was analyzed using concentration of AA ranging from
2.5 to 600lM. The concentration of [14C]AA ranged
from 2.5 to 50lM and was kept constant at 50lM. Data
were analyzed by nonlinear regression of Michaelis–
Menten equation.
4. Scherrmann, J. M. Drug delivery to brain via the blood–
brain barrier. Vasc. Pharmacol. 2002, 38, 349–354.
5. Hansch, C.; Bjorkroth, J. P.; Leo, A. Hydrophobicity and
central nervous system agents: on the principle of minimal
hydrophobicity in drug design. J. Pharm. Sci. 1987, 76,
663–687.
6. Testa, B.; Carrupt, P. A.; Gaillard, P.; Billois, F.; Weber,
P. Lipophilicity in molecular modeling. Pharm. Res. 1996,
13, 335–343.
Inhibition of AA transport was determined by adding
the indicated concentrations of unlabelled compounds
to plated cells along with either [14C]AA at fixed concen-
tration 50lM, or [14C]AA ranging from 2.5 to 100lM.
The unlabelled inhibitor concentrations displacing 50%
of [14C]AA (IC50 values) were obtained by computer
analysis of displacement curves. Inhibitory binding con-
stants (Ki values) were derived from the IC50 values
according to the Cheng and Prusoff equation
Ki = IC50/(1+[C*]/Kt*), where [C*] is the concentration
of the [14C]AA and Kt* its Michaelis–Menten constant.29
Ki values were also obtained by Lineweaver–Burk anal-
ysis for competitive inhibitors, using the equation
Ki = [inhibitor]/{(KM,obs/KM)ꢀ1.0}, where KM,obs is the
apparent Michaelis–Menten constant obtained in the
presence of a fixed concentration of inhibitor. All calcu-
lations were performed using the computer program
Graph Pad Prism (GRAPHPAD).
7. Tamai, I.; Tsuji, A. Transporter-mediated permeation of
drugs across the blood–brain barrier. J. Pharm. Sci. 2000,
89, 1371–1388.
8. Ohnishi, T.; Maruyama, T.; Higashi, S.; Awazu, S. Blood
brain barrier transport of L-tyrosine conjugates: a model
study for the brain targeting using large neutral aminoacid
transport system. J. Drug Target 2000, 8, 395–401.
9. Bonina, F. P.; Arenare, L.; Ippolito, R.; Boatto, G.;
Battaglia, G.; Bruno, V.; de Caprariis, P. Synthesis
pharmacokinetics and anticonvulsant activity of 7-chlo-
rokynurenic acid prodrugs. Int. J. Pharm. 2000, 202,
79–88.
10. Manfredini, S.; Pavan, B.; Vertuani, S.; Scaglianti, M.;
Compagnone, D.; Biondi, C.; Scatturin, A.; Tanganelli, S.;
Ferraro, L.; Prasad, P.; Dalpiaz, A. Design, synthesis and
activity of ascorbic acid prodrugs of nipecotic, kynurenic
and diclofenamic acids, liable to increase neurotropic
activity. J. Med. Chem. 2002, 45, 559–562.
11. Tsukaguchi, H.; Tokui, T.; Mackenzie, B.; Berger, U. V.;
Chen, X. Z.; Wang, Y.; Brubaker, R. F.; Hediger, M. A. A
family of mammalian Na+-dependent L-ascorbic acid
transporters. Nature 1999, 399, 70–75.
12. Friedman, P. A.; Zeidel, M. L. Victory at C. Nat. Med.
1999, 5, 620–621.
Statistical analysis was performed by ANOVA followed
by DunnettÕs t-test. Difference was considered statisti-
cally significant at P values less than 0.05.
13. Rice, M. E. Ascorbate regulation and its neuroprotective
role in the brain. Trends Neurosci. 2000, 23, 209–216.
14. Nishikimi, M.; Fukuyama, R.; Minoshima, S.; Shimizu,
M.; Yaghi, K. Cloning and chromosomal mapping of the
human non functional gene for L-gulono-c-lactone oxi-
dase, the enzyme for L-ascorbic acid biosynthesis missing
in man. J. Biol. Chem. 1994, 269, 23215–23222.
15. Sotiriou, S.; Gispert, S.; Cheng, J.; Wang, Y.; Chen, A.;
Hoogstraten-Miller, S.; Miller, G. F.; Kwon, O.; Levine,
M.; Guttentag, S. H.; Nussbaum, R. L.; Sotiriou, S.;
Gispert, S.; Cheng, J.; Wang, Y.; Chen, A.; Hoogstraten-
Acknowledgements
This work was supported by a research grant from the
University of Ferrara and by National Institutes of
Health Grant (HD37150).
Appendix A. Elemental analyses
Compound
Formula
Anal. Calcd (found)
H
C
N
11
11 (3R,2R,1S)
11 (3S,2R,1S)
C12H17NO7
C12H17NO7
5.96 (5.94)
5.96 (5.95)
5.96 (5.94)
3.77 (3.75)
3.77 (3.77)
4.61 (4.60)
2.95 (2.94)
3.12 (3.12)
50.17 (50.19)
50.17 (50.20)
50.17 (50.17)
55.34 (55.30)
55.88 (55.91)
41.16 (40.91)
46.85 (46.88)
46.46 (46.32)
4.88 (4.87)
4.88 (4.88)
4.88 (4.87)
4.03 (4.02)
3.08 (3.09)
4.00 (4.01)
3.41 (3.40)
2.71 (2.72)
C
12H17NO7
C16H13NO8
20H17Cl2NO7
12
13
19
20
21
C
C12H16BrNO6
C16H12BrNO7
C20H16BrCl2NO6
Miller, S.; Miller, G. F.; Kwon, O.; Levine, M.; Guttentag,
S. H.; Nussbaum, R. L. Ascorbic-acid transporter Slc23a1
is essential for vitamin C transport into the brain and for
perinatal survival. Nat. Med. 2002, 8, 514–517.
16. Lambert, D. M. Rational and application of lipids as
prodrug carriers. Eur. J. Pharm. Sci. 2000, 130,
1016S–1022S.
17. Hull, M.; Lieb, K.; Fiebich, B. L. Antiinflammatory drugs:
a hope for AlzheimerÕs disease? Expert Opin. Invest. Drugs
2000, 9, 671–683.
References and notes
1. Filmore, D. Breeching the blood–brain barrier. Mod. Drug
Discov. 2002, 6, 22–27.
2. De Vries, H. E.; Kuiper, J.; De Boer, A. G.; van Berkel, T.
J.; Breimer, D. D. The blood brain barrier in neuroin-
flammatory diseases. Pharm. Rev. 1997, 49, 143–155.
3. Padridge, W. M. Targeting neurotherapeutic agents
through the blood–brain barrier. Arch. Neurol. 2002, 59,
35–40.

S. Manfredini et al. / Bioorg. Med. Chem. 12 (2004) 5453–5463
5463
18. Halliday, G.; Robinson, S. R.; Shepherd, C.; Krill, J.
AlzheimerÕs disease and inflammation: a review of cellular
and therapeutic mechanisms. Clin. Exp. Pharm. Physiol.
2000, 27, 1–8.
25. Kato, K.; Terao, S.; Shimamoto, N.; Hirata, M. Studies
on scavengers of active oxygen species. 1. Synthesis and
biological activity of 2-O-alkylascorbic acids. J. Med.
Chem. 1988, 31, 793–798.
19. Hodgkins, P. S.; Schwarz, R. Metabolic control of
kynurenic acid formation in the rat brain. Dev. Neurosci.
1998, 20, 408–416.
20. Scharf, S.; Mander, A.; Ugoni, A.; Vajda, F.; Christoph-
idis, N. A double-blind, placebo-controlled trial of diclof-
enac/misoprostol in AlzheimerÕs disease. Neurology 1999,
53, 197–201.
21. Liang, W. J.; Johnson, D.; Jarvis, S. M. Vitamin C
transport systems of mammalian cells. Mol. Membr. Biol.
2001, 18, 87–95.
22. Rumsey, S. C.; Welch, R. W.; Garraffo, H. M.; Ge, P.; Lu,
S. F.; Crossman, A. T.; Kirk, K. L.; Levine, M. Specificity
of ascorbate analogs for ascorbate transport. Synthesis
and detection of [(125)I]6-deoxy-6-iodo-L-ascorbic acid
and characterization of its ascorbate-specific transport
properties. J. Biol. Chem. 1999, 274, 23215–23222.
23. Kim, Y. C.; Zhao, L. X.; Kim, T. H.; Je, S. M.; Kim, E.
K.; Choi, H.; Chae, W. G.; Park, M.; Choi, J.; Jahng, Y.;
Lee, E. S. Design and synthesis of anticonvulsive agents as
gamma-vinylGABA-based potential dual acting prodrugs
and their biological activities. Bioorg. Med. Chem. Lett.
2000, 10, 609–614.
26. Glenn, A. C. 5,6-Anhydro-L-ascorbic acid. A reactive
intermediate for the formation of 6-substituted deriva-
tives of L-ascorbic acid. Carbohydr. Res. 1984, 134,
321–326.
27. Raic-Malic, S.; Svedruzic, D.; Gazivoda, T. Synthesis and
antitumor activities of novel pyrimidine derivates of 2,3-
O,O-dibenzyl-6-deoxy-L-ascorbic acid and 4,5-didehydro-
5,6-dideoxy-L-ascorbic acid. J. Med. Chem. 2000, 43,
4806–4811.
28. Manfredini, S.; Baraldi, P. G.; Bazzanini, R.; Bortolotti,
F.; Vertuani, S.; Ashida, N.; Machida, H. Enzymatic
synthesis of 20-O-acyl-prodrugs of 1-(b-D-arabinofurano-
syl)-5-(E)-(2-bromovinyl)uracil (sorivudine, BV-araU) and
of 20-O -acyl-araU, -araC and -araA. Antiv. Chem.
Chemother. 1998, 25, 9–16.
29. Cheng, Y. C.; Prusoff, W. H. Relationships between
the inhibition constant (ki) and the concentration of
inhibitor which cause 50 per cent inhibition (IC50) of an
enzymatic reaction. Biochem. Pharmacol. 1973, 22,
3099–3108.
30. Akkerman, A. M.; De Jongh, D. K.; Veldstra, H.
Synthetic oxytocics. I. 3-(Piperidyl-(N)-methyl-) indoles
and related compounds. Rec. Trav. Chim. Pays-Bas 1951,
70, 899–916.
24. Spath, E. Constitution of kynurenic acid. Monatsh 1921,
42, 89–95.