Synthesis and biological evaluation of two glycerolipidic prodrugs of didanosine for direct lymphatic delivery against HIV

Article abstract of DOI:10.1016/j.bmcl.2007.01.062

Novel glycerolipidic prodrugs of didanosine and didanosine monophosphate designed to by-pass the hepatic first pass metabolism were synthesized and tested for their cytotoxicity and anti-HIV-1 activity. Formulation as liposomes of dipalmitoylphosphatidylcholine was elaborated. A simple quantitative HPLC-UV method was developed and validated, and ESI-MS was used for qualitative purpose. These two prodrugs exhibited promising biological activities against HIV-1 in in vitro infected cell culture.

Full text of DOI:10.1016/j.bmcl.2007.01.062

Bioorganic & Medicinal Chemistry Letters 17 (2007) 2237–2240
Synthesis and biological evaluation of two glycerolipidic prodrugs
of didanosine for direct lymphatic delivery against HIV
a
b
a
c
Muriel Lalanne, Angelo Paci, Karine Andrieux, Nathalie Dereuddre-Bosquet,
c
d
b
b
Pascal Clayette, Alain Deroussent, Micheline R e´ , Gilles Vassal,
a
e,
*
Patrick Couvreur and Didier Desma e¨ le
a
Universit e´ Paris-Sud, UMR CNRS 8612, IFR 141, F-92296 Ch aˆ tenay-Malabry, France
b
Universit e´ Paris-Sud, UPRES EA 3535, IGR, F-94805 Villejuif, France
SPI-BIO, Laboratoire de Neurovirologie, CEA, DSV/DRM, F-92265 Fontenay aux Roses, France
c
d
Service commun de spectrom e´ trie de masse, IFR54, IGR, F-94805 Villejuif, France
Universit e´ Paris-Sud, UMR CNRS 8076, IFR 141, F-92296 Ch aˆ tenay-Malabry, France
e
Received 6 December 2006; revised 18 January 2007; accepted 19 January 2007
Available online 25 January 2007
Abstract—Novel glycerolipidic prodrugs of didanosine and didanosine monophosphate designed to by-pass the hepatic first pass
metabolism were synthesized and tested for their cytotoxicity and anti-HIV-1 activity. Formulation as liposomes of dipalmitoyl-
phosphatidylcholine was elaborated. A simple quantitative HPLC-UV method was developed and validated, and ESI-MS was used
for qualitative purpose. These two prodrugs exhibited promising biological activities against HIV-1 in in vitro infected cell culture.
ꢀ
2007 Elsevier Ltd. All rights reserved.
Didanosine (ddI) (1) is a nucleosidic analog for the oral
treatment of AIDS, currently marketed as Videx , and
the fact that a variety of antiviral or antitumor nucleo-
sides conjugated with lipid moieties have been reported
previously, to the best of our knowledge no attempts
ꢁ
4
usually used in combination with other antiviral thera-
py. Didanosine is phosphorylated in vivo to active
metabolites that compete for incorporation into viral
DNA. These latter inhibit the HIV reverse transcriptase
enzyme competitively and act as a chain terminator of
DNA synthesis. Didanosine is well tolerated with chron-
ic administration, and its toxicity is uncommon and usu-
ally reversible. However, it suffers from a poor
were made to take advantage of the triglyceride metab-
olism to improve the bioavailability of ddI. As a matter
of fact, triglycerides are hydrolyzed in the small intestine
to the corresponding 2-monoglyceride and fatty acids by
pancreatic lipase and co-lipase. These metabolites are
absorbed by the enterocytes in which they are re-esteri-
fied into triglycerides and packaged into intestinal lipo-
proteins, the chylomicrons. These latter are finally
secreted into mesenteric lymph from where they reach
the systemic circulation via the thoracic lymph duct,
1
bioavailability (20–40%). This drawback is mainly due
to acidic degradation in the stomach, and to a mediocre
absorption because of its hydrophily and to the hepatic
first pass metabolism. Moreover, once in the cell, ddI
3
by-passing the liver. Therefore, we hypothesized that
ddI covalently bound to the C-2 position of a lipid
may have improved bioavailability. As triglycerides with
0
needs to be sequentially phosphorylated to the 5 -tri-
phosphate by host cell kinases to be active. The first
2
5
phosphorylation is a limiting step. Glycerolipidic pro-
C16 acid chains are predominant in nutriments, we
chose to conjugate ddI with a 1,3-dipalmitoyl lipid
drugs, by mimicking long chain triglycerides, are known
to enhance the enterocyte absorption and lymphatic
transport, therefore they by-pass the hepatic first pass
0
through a succinate linker bound to the 5 -OH group
(prodrug 2). Furthermore, in an attempt to by-pass
3
metabolism, hence enhancing bioavailability. Despite
the first cellular phosphorylation, we speculated that a
0
lipid conjugated to didanosine 5 -monophosphate
through an ether linker such as compound 3 may
have an improved activity. Comparing the intracellular
activities of the non-phosphorylated and the
mono-phosphorylated prodrugs (2 and 3) on peripheral
Keywords: Anti-HIV nucleoside; Enzyme inhibitors; Prodrugs; Lipid–
nucleoside conjugate; Bioavailability.
*
Corresponding author. Tel.: +33 (0)1 46 83 57 53; fax: +33 (0)1 46 83
7 52; e-mail: didier.desmaele@cep.u-psud.fr
5
0
960-894X/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.01.062

2238
M. Lalanne et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2237–2240
blood mononuclear cells (PBMC) post-infected by HIV,
we will establish if it is possible to by-pass the limiting
first-phosphorylation step by host cell kinases (Fig. 1).
the event, brief exposure of the crude product to diluted
acetic acid, followed by chromatographic purification
on silica gel column (CH Cl /MeOH, 11:2), furnished
2
2
2
1
0
phosphodiester 3 in 73% yield. This material slowly
decomposed at room temperature through cleavage of
the C-1-hypoxhanthine bond, however no substantial
decomposition was observed over several month-storage
at ꢀ30 ꢂC. Compounds 2 and 3 were characterized by
Herein, we report the syntheses of lipid conjugates of
ddI 2 and 3, the results of their qualitative and quantita-
tive analysis for quality control of the formulation, and
their antiviral activity on in vitro HIV-1 infected cells.
1
H NMR, IR, and ESI mass spectrometry. The purity
The synthesis of prodrug 2 began with the preparation
of the symmetrical 1,3-dipalmitoyl glyceride 5 easily
available in three steps from 1,3-dihydroxyacetone
evaluated by HPLC using UV detection and by ESI-
MS was estimated to be 98% for compound 2 and
1
1
90% for compound 3 (Scheme 1). The quantitative
method was validated according to the ICH recommen-
dations in order to undertake quality control of the for-
6
dimer (4) according to the procedure of McCrae.
1
2
Thus, condensation of 1,3-diglyceride 5 with glutaric
anhydride in the presence of pyridine provided acid 8
mulations. The method proved to be accurate (bias
<1.4%), with a good repeatability (Relative Standard
Deviation (%RSD) <2.9%), and intermediate precision
(%RSD <4.2%) for the determination of both com-
pounds. By using HPLC-UV dosage methods, we evi-
denced the very poor solubility of these prodrugs in
water (610.5 lg/mL in water at neutral pH) and their
good entrapment efficiency into multilamellar liposomes
(
95% yield). The final coupling of 8 with ddI (1) was
achieved using EDCI to deliver compound 2 in 61%
7
yield. The synthetic methodology, we have developed
for the synthesis of 3 involved the application of phos-
8
phoramidite chemistry. Thus, 1,3-dipalmitoyl glyceride
5
was condensed with 6-(tetrahydropyranyloxy)hexano-
ic acid (6) using DCC to furnish the triglyceride 7a in
1% yield. Deprotection of the THP protecting group
13
of dipalmitoylphosphatidylcholine (DPPC)
were prepared according to the Bangham method.
which
4
1
6
turned out to be quite troublesome. For example, aque-
ous acetic acid deprotection of 7a gave 7b in low yield
along a large amount of diglyceride 5 presumably result-
ing from acid catalyzed lactonization of 7a to caprolac-
These liposomes (prodrugs/DPPC: 1:2 (mol:mol)),
whose sizes were evaluated to 1160 ± 129 nm for both
compounds, were used in the antiviral assays.
9
tone. The use of PPTS in a refluxing mixture of
Antiviral assay. In vivo after oral administration, the
prodrugs should be modified by digestion, so they will
not be in direct contact with infected cells. The antiviral
assays were made here to check that the prodrug would
be able to deliver ddI in cellular context.
methanol and ethyl acetate was more rewarding, provid-
ing the desired alcohol 7b in 80% yield. The crucial
assembly of ddI (1) with 7b was efficiently performed
by sequential treatment of 7b with 2-cyanoethyl N,N-
diisopropylchlorophoramidite (9) in the presence of
diisopropylethylamine, then 1H-tetrazole and finally
The toxicities assays were performed to ensure that there
was no increase in toxicity due to the chemical modifica-
tion and formulation (previous assays with prodrugs in
suspension evidenced cell toxicity by precipitation of pro-
drugs in culture medium). Phytohemagglutinin-P (PHA-P)-
activated peripheral blood mononuclear cells (PBMC)
treated by six concentrations (50, 17, 6, 2, 1, and 0.2 lM
for compounds 1 and 3, and 12, 4, 1.3, 0.5, 0.1, and
0.05 lM for compound 2) of each prodrug as liposomes
and, one hour later, were infected with hundred 50% tis-
sue culture infectious doses (TCID50) per 100,000 cells
8
b
ddI. After oxidative work up with aqueous iodine buf-
fered with pyridine, compound 10 was obtained in 93%
yield by purification on silica gel column (EtOAc/
MeOH/Et N, 6:1:0.5%). The final deprotection of 10
3
was performed via b-elimination using DBU. A consid-
erable amount of cleavage of the hypoxanthine–sugar
bond was observed using hydrochloric acid in the work-
up. Less acidic conditions were therefore attempted. In
1
5
O
of the HIV-1-LAI strain. This virus was amplified in vi-
1
O
6
tro on PHA-P-activated PBMC. Viral stock was titrated
using PHA-P-activated PBMC, and 50% TCID50 were
O
O
HN
HN
N
N
O
O
C15H31
C15H31
O
O
N
17
N
N
calculated using K a¨ rber’s formula. Samples were main-
N
O
O
OH
O
tained throughout the culture, and cell supernatants were
collected at day 7 post-infection and stored at ꢀ20 ꢂC.
Viral replication was measured by quantifying reverse
transcriptase (RT) activity in cell culture supernatants.
In parallel, cytotoxicity of the compounds was evaluated
in uninfected PHA-P-activated PBMC by MTT assay on
day 7. Experiments were performed in triplicate and
repeated with another blood donor. Data analyses were
O
2
O
1
ddI
HN
N
O
O
O
N
C15H31
C15H31
O
O
N
O P O
O
O
ꢁ
performed using SoftMax Pro 4.6 microcomputer soft-
OH
ware: percent of inhibition of RT activity or of cell viabil-
ity were plotted vs concentration and fitted with quadratic
curves; 50% and 90% effective doses (ED50 and ED ) and
O
3
90
Figure 1. Chemical structures of didanosine (ddI) 1 and lipid conju-
gates 2 and 3.
cytotoxic doses (CD50 and CD ) were calculated and
90
compared between ddI and ddI prodrugs (Table 1).

M. Lalanne et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2237–2240
2239
O
O
O
O
OH
OTHP
O
O
HO
HO
a,b
C15H31
C15H31
O
O
C15H31
C15H31
O
O
6
OR
OH
OH
O
O
OH
c
O
7a: R = THP
7b: R = H
4
d
5
Cl
O
i-Pr N P
2
O
O
then 1
O
g
9
O
CN
e
O
HN
N
O
O
O
O
N
O
O
C15H31
O
O
N
C15H31
C15H31
O
O P O
O
O
O
O
OH
C15H31
O
1
0
O
NC
h
8
f
2
3
Scheme 1. Synthesis of compounds 2 and 3. Reagents and conditions: (a) C15
PhH/H Cl
O, 5 ꢂC, 30 min, 70%; (c) 1.6 equiv 6, DCC, DMAP, CH
i—1.8 equiv 9, i-Pr NEt, CH
CN, 20 ꢂC, 30 min; ii—5 equiv 1H-tetrazole, 1.5 equiv ddI (1), THF, 20 ꢂC, 1 h; iii—I
0 min, 93%; (f) i—1.9 equiv DBU, THF, 1 h, 20 ꢂC; ii—AcOH, H O, 68%; (g) 2 equiv glutaric anhydride, py, DMAP, THF/CH
5%; (h) 1.2 equiv EDCI, DMAP, DMF, 50 ꢂC, 3 days, 61%.
H
31COCl, py, CHCl
3
, cat. DMAP, 3 h, 20 ꢂC, 50%; (b) NaBH
4
, THF/
2
2
2
/THF, 20 ꢂC, 48 h, 61%; (d) 1 equiv PPTS, MeOH/AcOEt, 50 ꢂC, 3 h, 80%; (e)
, THF, Py, H
O, hexane, 20 ꢂC,
Cl
, 6 h, 20 ꢂC,
2
3
2
2
3
9
2
2
2
Table 1. Anti-HIV-1-LAI effects and cytotoxicity of ddI conjugates 2
and 3 as compared to didanosine (1) on activated PBMC in vitro
of increased toxicity. Since these molecules were poorly
water soluble, adapted liposomal formulation was devel-
oped. A simple dosage method with good accuracy,
repeatability, and intermediate precision was elaborated
in order to evaluate these formulations. Further experi-
ments including absorption studies in in vitro CaCo-2
cells culture of these two new molecules in liposomes
are currently performed in comparison with ddI to
validate the here presented biomimetic strategy.
1
3
infected by HIV-1-LAI
Compound ED50 (lM) ED90 (lM) CD50 (lM) CD90 (lM)
1
2
3
(ddI)
1.2
2.8
4.8
>12.0
>50.0
40
>12.0
>50.0
>50.0
>12.0
>50.0
28.5
Results are expressed as ED50 and ED90, concentration of drugs that
decreases the HIV replication of 50% and 90%, respectively (means of
ED from two blood donors). CD50 and CD90, concentration of drugs
to reduce the viable cell number by 50% and 90%, respectively (means
of ED from two blood donors).
Acknowledgments
The data indicated that 2 was roughly equi-active with
ddI against HIV-1. 3 was a little less active, probably
due to the ionized phosphate which could slow down
cellular absorption and might hide the by-pass of the
first phosphorylation. Another alternative hypothesis is
that the prodrug would not release the monophosphate
of ddI because aliphatic phosphate esters are known to
be difficult to be cleaved enzymatically. Further experi-
ments would be performed in order to study these
hypotheses. However, these two new compounds
showed no sign of increased cytotoxicity at tested doses
The financial support of the ‘Agence Nationale de la
Recherche’ (ANR, Grant NT05-1_43236 SYLIANNU)
is acknowledged. The authors thank MENRT for pro-
viding financial support to ML.
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.bmcl.
2007.01.062.
(
12 and 50 lM for 2 and 3, respectively). Both com-
pounds appear promising ddI prodrugs because they
keep an anti-HIV activity and would exhibit a better
bioavailability than ddI after oral administration.
References and notes
1
. Balimane, P. V.; Sinko, P. J. Adv. Drug Delivery Rev. 1999,
9, 183.
In conclusion, we have synthesized with good purity two
novel didanosine derivatives as prodrugs compounds
active against HIV-1 in infected cell culture with no sign
3
2. Tan, X.; Chu, C. K.; Boudinot, F. D. Adv. Drug Delivery
Rev. 1999, 39, 117.

2240
M. Lalanne et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2237–2240
3
. (a) Porter, C. J. H.; Charman, W. N. Adv. Drug Delivery
Rev. 1997, 25, 71; (b) Lambert, D. M. Eur. J. Pharm. Sci.
000, 11(Suppl. 2), S15, and references therein.
(s, 1H), 6.26 (m, 1H), 5.20 (q, J = 5.5 Hz, 1H), 4.50–4.20
(m, 5H), 4.12 (dd, J = 11.9, 5.4 Hz, 2H), 4.02–3.80 (m,
2H), 2.60–200 (m, 4H), 2.31 (t, J = 7.5 Hz, 6H), 1.60–1.48
2
3
1
4
. (a) For some recent examples, see: Zhang, J.-A.; Ma, L.;
Ahmad, M.; Ali, S.; Khan, A.R.; Ahmad, I. PCT Int.
Appl. 2006, WO 2006/102533; (b) Ali, S. M.; Khan, A. R.;
Ahmad, M. U.; Chen, P.; Sheikh, S.; Ahmad, I. Bioorg.
Med. Chem. Lett. 2005, 15, 2571; (c) Kucera, L.S.;
Fleming, R.A.; Ishaq, K.S.; Kucera, G.L.; Morris-Nat-
schke, S.L. U.S. Patent 0082242, 2002; (d) Hong, C. I.;
Nechaev, A.; Kirisits, A. J.; Vig, R.; West, C. R.;
Manouilov, K. K.; Chu, C. K. J. Med. Chem. 1996, 39,
(m, 6H), 1.30–1.10 (m, 52H), 0.84 (t, J = 6.5 Hz, 6H);
P
NMR (CDCl , 81 MHz) d 1.8 (broad s).
3
ꢁ
11. HPLC-UV characteristics: Nucleosil C18 column, 10 lm,
250 · 4 mm. Elution with CH OH/CH CN/0.05% aque-
3
3
ous ammonium acetate (v/v/v) pH 8 (97.5/1.5/1.5; v/v/v)
for 2 and (87.5/5/7.5; v/v/v) for 3 at 1 mL/min. Absorption
maxima = 280 nm for 2 in methanol and 254 nm for 3.
Compound 2: peak at 13.6 min corresponding to
Mw = 900.6 (m/z 901.6, [M+H]
+
and m/z 923.6,
+
1771; (e) van Wijk, G. M. T.; Hostetler, K. Y.; van den
[M+Na] , positive ESI mass spectrum). Compound 3:
main peak at 9.8 min, corresponding to Mw = 980.6 (m/z
Bosch, H. J. Lipid Res. 1992, 33, 1211; (f) Steim, J. M.;
Camaioni Neto, C.; Sarin, P. S.; Sun, D. K.; Sehgal, R. K.;
Turcotte, J. G. Biochem. Biophys. Res. Commun. 1990,
ꢀ
979.6, [MꢀH] , negative ESI mass spectrum) with a minor
impurity at 9.1 min (underprotected phosphotriester 10
(m/z 1032.6, [MꢀH] )).
ꢀ
171, 451.
5. Tso, P. Adv. Lipid Res. 1985, 21, 143.
6. Bentley, P. H.; McCrae, W. J. Org. Chem. 1970, 35, 2082.
7. Characterization data for compounds 2: colorless oil,
12. ICH (Q2A) and (Q2B): (CPMP/ICH/381/95), (CPMP/
ICH/281/95), 1994. Calibration curves = linear response
2
(10–100 lg/mL) (r P 0.995). Limits of detection/quanti-
2
D
0
ꢀ1
½
aꢁ ¼ ꢀ7 (c 0.7, CHCl ); IR (neat, cm ) m 3600–2500,
fication = 2.6/7.8 lg/mL for 2 and 1.2/3.7 lg/mL for 3. For
each sample, n = 6.
3
1
1
1
617, 1588, 1544; H NMR (CDCl , 1% D O, 300 MHz) d
3 2
3.0 (br s, 1H), 8.15 (s, 1H), 8.00 (s, 1H), 6.21 (t,
13. Li, H.; Song, J. H.; Park, J. S.; Han, K. Int. J. Pharm.
2003, 258, 11.
J = 4.6 Hz, 1H), 5.17 (quint, J = 5.0 Hz, 1H), 4.40–4.15
m, 5H), 4.07 (dd, J = 12.0, 5.8 Hz, 2H), 2.55–2.45 (m,
H), 2.36 (t, J = 7.3 Hz, 2H), 2.33 (t, J = 7.4 Hz, 2H), 2.23
t, J = 7.6 Hz, 4H), 2.18–1.95 (m, 2H), 1.86 (quint,
J = 7.4 Hz, 2H), 1.60–1.48 (m, 4H), 1.30–1.10 (m, 48H),
.80 (t, J = 6.9 Hz, 6H).
(
2
(
14. Bangham, A. D.; Hill, H. W.; Miller, N. G. A. Methods
Membr. Biol. 1974, 1, 1, Briefly, prodrugs and DPPC were
dissolved in organic solvents. Then, solvents were evapo-
rated under vacuum. The lipidic film was then rehydrated
with phosphate buffer and filterated through a 0.8 lm filter
0
ꢁ
8
. (a) Caruthers, M. H. Science 1985, 230, 281; (b) Sinha, N.
D.; Biernat, J.; K o¨ ster, H. Tetrahedron Lett. 1983, 24,
(Sartorius , France), three times.
15. Barre-Sinoussi, F.; Chermann, J. C.; Rey, F.; Nugeyre, M.
T.; Chamaret, S.; Gruest, J.; Dauguet, C.; Axler-Blin, C.;
Vezinet-Brun, F.; Rouzioux, C.; Rozenbaum, W.; Mon-
tagnier, L. Science 1983, 220, 868.
16. Roisin, A.; Robin, J. P.; Dereuddre-Bosquet, N.; Vitte, A.
L.; Dormont, D.; Clayette, P.; Jalinot, P. J. Biol. Chem.
2004, 279, 9208.
5843.
9
. Miyashita, N.; Yoshikoshi, A.; Grieco, P. A. J. Org.
Chem. 1977, 42, 3772.
0. Characterization data for compounds 3: colorless oil,
1
2
0
ꢀ1
½
aꢁ ¼ ꢀ13 (c 1.1, CHCl ); IR (neat, cm ) m 3600–2500,
D
955, 2916, 2850, 1736, 1693, 1588, 1468, 1219, 1095, 1063;
H NMR (CDCl
3
2
1
3
, 1% D
2
O, 300 MHz) d 8.50 (s, 1H), 8.07
17. K a¨ rber, G. Arch. Exp. Path. Pharmak. 1931, 162, 480.