Retinoic acid conjugates as potential antitumor agents: Synthesis and biological activity of conjugates with Ara-A, Ara-C, 3(2H)-furanone, and aniline mustard moieties

Article abstract of DOI:10.1021/jm9602322

In a dual targeting approach, to explore the ability of tretinoin (all- trans-retinoic acid) to behave as a covalent carrier for cytotoxic entities, conjugates of retinoic acid with a few representative molecules, being important examples of antitumor pharmacophores (i.e., nucleoside analogues and alkylating agents), have been synthesized and tested for their cytostatic and differentiating activity. All compounds were stable to in vitro hydrolysis in human plasma and more lipophilic than the parent compounds, thus consenting enhanced uptake into the cells. Among the nucleoside analogues the Ara-C derivatives 3 and 6 and the Ara-A derivative 7 proved the most cytostatic (IC₅₀ < 0.32 μg/mL) resulting from 25- to >144-fold more active (Ara-A derivatives) or at least as equally active (Ara-C derivatives) as compared to the parent nucleosides. Compound 3, endowed with a highly lipophilic silyl moiety at the 3' and 5' positions, showed the highest differentiating activity (54% and 44% differentiated HL-60 cells at 0.2 and 0.05 μg/mL respectively). With regard to the retinoic acid conjugates of alkylating agents, compound 10 was the most cytostatic agent (IC₅₀ < 0.32 μg/mL) and the most potent differentiating agent (33-34% at 0.32 and 0.08 μg/mL). These structures may also be regarded as analogs of either retinoic acid or the cytotoxic compound.

Full text of DOI:10.1021/jm9602322

J . Med. Chem. 1997, 40, 3851-3857
3851
Notes
Retin oic Acid Con ju ga tes a s P oten tia l An titu m or Agen ts: Syn th esis a n d
Biologica l Activity of Con ju ga tes w ith Ar a -A, Ar a -C, 3(2H)-F u r a n on e, a n d
An ilin e Mu sta r d Moieties
Stefano Manfredini,*,† Daniele Simoni,‡ Roberto Ferroni,† Rita Bazzanini,† Silvia Vertuani,† Sigrid Hatse,§
J an Balzarini,§ and Erik De Clercq§
Dipartimento di Scienze Farmaceutiche, Universita` di Ferrara, via Fossato di Mortara 17-19, I-44100 Ferrara, Italy,
Istituto Farmacochimico, Universita` di Palermo, I-90133 Palermo, Italy, and Rega Institute for Medical Research,
Laboratory of Virology and Chemotherapy, Katholieke Universiteit Leuven, Minderbroedersstraat 10,
B-3000 Leuven, Belgium
Received March 25, 1996X
In a dual targeting approach, to explore the ability of tretinoin (all-trans-retinoic acid) to behave
as a covalent carrier for cytotoxic entities, conjugates of retinoic acid with a few representative
molecules, being important examples of antitumor pharmacophores (i.e., nucleoside analogues
and alkylating agents), have been synthesized and tested for their cytostatic and differentiating
activity. All compounds were stable to in vitro hydrolysis in human plasma and more lipophilic
than the parent compounds, thus consenting enhanced uptake into the cells. Among the
nucleoside analogues the Ara-C derivatives 3 and 6 and the Ara-A derivative 7 proved the
most cytostatic (IC50 < 0.32 µg/mL) resulting from 25- to >144-fold more active (Ara-A
derivatives) or at least as equally active (Ara-C derivatives) as compared to the parent
nucleosides. Compound 3, endowed with a highly lipophilic silyl moiety at the 3′ and 5′
positions, showed the highest differentiating activity (54% and 44% differentiated HL-60 cells
at 0.2 and 0.05 µg/mL respectively). With regard to the retinoic acid conjugates of alkylating
agents, compound 10 was the most cytostatic agent (IC50 < 0.32 µg/mL) and the most potent
differentiating agent (33-34% at 0.32 and 0.08 µg/mL). These structures may also be regarded
as analogs of either retinoic acid or the cytotoxic compound.
In tr od u ction
Retinoic acid itself has been reported to inhibit HIV
replication. Therefore, it has appeared of interest by
different authors to study retinoic acid covalently linked
to nucleosides, in order to enhance the uptake of the
prodrug by HIV-1-infected cells and to increase its
plasma half-life.8-10
Retinoids are a class of natural and synthetic com-
pounds structurally related to vitamin A. They have
been found to be active agents, experimentally as well
as clinically, in the prevention and therapy of tumors.1
Some retinoids have affinity for several binding proteins
(i.e., CRABP and CRBP), which were shown to be
important in their transport,and modulate cell growth
via a family of well-characterized nuclear retinoic acid
receptors (RARs, RXRs, and Z) which regulate gene
expression.2,3 This discovery led to the hypothesis that
retinoids may exert their differential biological effects,
and thus their particular chemopreventive and chemo-
therapeutic activity, via regulation of gene function by
two major pathways: (a) through direct binding of a
retinoid receptor complex to retinoid responsive ele-
ments (RAREs and RXRs) on DNA4 or (b) interacting
with other regulatory proteins.5,6 The antitumor activ-
ity of tretinoin (all-trans-retinoic acid) may be explained
to a large extent by an inhibition of cell proliferation
and induction of cellular differentiation, and as clearly
demonstrated by Bollag et al., an enhanced antitumor
effect might be achieved by its combination with low
molecular weight inducers and cytokines.7
Retinoic acid prodrugs of AZT and 3′-thia-2′,3′-
dideoxycytidine have been synthesized and tested for
their antiviral activity by Aggarwal et al. and Camplo
et al.,9,10 and it was found that these retinoic acid
derivatives enhanced the uptake of the prodrug in the
infected cells, achieving approximately 4-fold higher
intracellular concentrations than AZT itself, and in-
creased the plasma half-life of the parent nucleosides.
Surprisingly the observed antiviral activity was similar,
in both cases, to the parent nucleosides, but the cyto-
toxicity was increased by 8- and 6-fold, respectively.
However, it must be noted that these authors have not
investigated eventual concomitant effects related to the
differentiating activity of retinoic acid.
A documented approach to enhance the activity of
antitumor agents involves the targeting of the alkylat-
ing agents (i.e., simple aniline mustards) to DNA by
attaching them to DNA-affinity carriers (i.e., DNA-
intercalating ligands such as 9-aminoacridine). This
approach led to an increased intrinsic drug potency (up
to 100-fold), avoiding some of the common mechanisms
of cellular resistance to alkylating agents, and to an
* To whom correspondence should be addressed.
† Universita` di Ferrara.
‡ Universita` di Palermo.
§ Katholieke Universiteit Leuven.
X Abstract published in Advance ACS Abstracts, October 1, 1997.
S0022-2623(96)00232-4 CCC: $14.00 © 1997 American Chemical Society

3852 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 23
Notes
alteration of the pattern of formed DNA lesions and
their subsequent repair.11,12 Moreover, in the field of
site-directed drug delivery, the dual targeting approach,
which implies the use of carrier molecules that have
their own intrinsic effect, is a well-recognized concept
for drug targeting strategies.13 Significant examples
can be found in the preparation of conjugates of Ara-C
and corticosteroids to give compounds with outstanding
antitumor activity14 and in the use of naturally occur-
ring compounds (i.e., sugars, proteins, and steroids) as
covalent carriers of cytotoxic entities.15-17
These concepts, together with the well-documented
overlapping activities of retinoids, cytokines, and low
molecular weight inducers in the regulation of cell
growth,7 have led us to investigate possible conjugates
of retinoic acid with antitumor/antiviral agents in the
aim to design new potential prodrugs. Moreover, these
structures may also be regarded as analogues of either
retinoic acid or the cytotoxic compound. In particular,
we have selected a few representative molecules as
being important examples of antiviral/antitumor/dif-
ferentiating pharmacophores.
Several observations support the design and synthesis
of these conjugates: (i) there is evidence of interesting
synergistic antitumor effects by simple combination of
antiproliferative agents and retinoic acid;18-20 (ii) ret-
inoic acid could be released from its amides and esters
by cellular enzymes;9,10 (iii) the retinoic acid prodrug
moiety is endowed with differentiating properties as
well as with antiproliferative activity;1 (iv) the increased
lipophilicity of a retinoic acid conjugate may improve
both the biotransport of the prodrug through cell
membranes and the plasma half-life of the conjugated
drug.9,10
moiety as a common structural feature of compounds
inducing erythroid differentiation.
Resu lts
Ch em istr y. The conjugates 3-5, were obtained by
reaction of the corresponding nucleosides (Ara-C and
Ara-A), protected as 3′,5′-O-(1,1,3,3-tetraisopropyl-1,3-
disiloxane-1,3-diyl) derivatives (1 and 2), and retinoyl
chloride in situ prepared from retinoic acid and oxalyl
chloride in benzene (Scheme 1).9 The reaction gave both
N6- and 2′-retinoyl derivatives in the case of Ara-A (4
and 5) and solely the N4-retinoyl derivative in the case
of Ara-C (3). It is interesting to note that removal of
the silyl protecting groups, by the use of the standard
tetrabutylammonium fluoride (TBAF) procedure, worked
well in the case of 4 and 5 but resulted in the loss of
the retinoyl moiety in the case of 3. This drawback was
overcome by adopting the ammonium fluoride/metha-
nol25 (NH4F/MeOH) procedure in the presence of acetic
acid, to give the deprotected 6-8 in satisfactory yields
(43-95%). Acetic acid was effective, in the case of 3,
as a source of H+ to prevent the concomitant elimination
of the retinoic acid residue at position 4 (due to NH3
deriving from NH4F decomposition) and also the pos-
sible 2′ f 5′ acyl migration.21,26 The use of this
procedure also avoided residues of TBAF and the
difficulty of separating them from the final products.
Structural attributions were based on 1H-NMR, UV
spectroscopy, and comparison with other data obtained
by us.22 In particular, in the case of the nucleoside
derivatives (6-8), the regiochemistry of the acylation
was attributed on the basis of the 1H-NMR: a ∼1.0 ppm
downfield shift for H2′, to ∼5.3 ppm, confirmed the 2′-
O-acylation.21 Moreover, the unusual upfield shift of
about 0.2 ppm for the C20 methyl, observed in the case
of compounds 5 and 8, might be explained on the basis
of a preferred conformation in which the R position of
the ester group was in the shielding region of the purine
ring. Finally, in the case of Ara-A, the indication that
In this study we have evaluated molecular combina-
tion of retinoic acid with nucleoside analogues such as
Ara-C and Ara-A (6-8): both latter compounds are
potent antitumor and/or antiviral agents; moreover,
Ara-C and retinoic acid are also potent differentiating
agents,18-20 used in the experimental therapy of cancer.
Ara-C and Ara-A are characterized by good activity but
low distribution and high sensitivity to enzymatic
degradation in plasma, which can be both improved by
masking the amino group with lipophilic functions.13,14
Moreover, it is known that cellular enzymes hydrolyze
nucleoside prodrugs; thus, half-lives of conjugates were
evaluated in human plasma. Finally, conjugation of
retinoic acid to positions different than 5′ of the nucleo-
side (N4-retinoyl-Ara-C, N6- and 2′-retinoyl-Ara-A) might
permit phosphorylation of the conjugates by cellular
kinase.21,22
With regard to the alkylating moieties (10, 12), the
3(2H)-furanone ring is the central common structural
feature of a class of naturally occurring antitumor
agents (i.e., geiparvarin, jatrophone, and eremantholides
A, B, and C),11 and the N,N-bis(2-chloroethyl)aniline
mustard is commonly used among the clinical antitumor
drugs (i.e., melphalan).
Finally, compound 14, having two retinoyl amides
bound through a propylene bridge, may provide us with
further information on the role of the amidic bond on
prodrug behavior, in terms of activity and/or bioavail-
ability of the retinoyl amides. Moreover, Tsiftsoglou et
al.23,24 have reported on the importance of the bis-amide

Notes
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 23 3853
Sch em e 1a
a
(i) TPDS-Cl2, pyridine; (ii) retinoic acid, C2O4Cl2, benzene; (iii) NH4F/MeOH.
the amino group of 8 is free also was derived from the
UV spectra [λmax(MeOH) 258 nm].22 The same consid-
erations could be drawn the case of Ara-C (6)
[λmax(MeOH) 249 nm].27,28
For derivatives 10, 12, and 14, the starting com-
pounds were prepared as described by us (9)29 or by
literature procedure (11),30 and diamine 13 was com-
mercially available. Retinoylations were conducted as
described above to give the expected compounds in
moderate to good yields (45-90%).
Biology. Among the retinoic acid conjugates of
nucleoside analogues, the Ara-C derivatives 3 and 6 and
the Ara-A derivative 7 proved most cytostatic. They
inhibited HL-60 cell proliferation at concentrations
below 0.32 µg/mL (Table 1), thus being approximately
144-fold more cytostatic as (compound 7) or comparably
active to (compounds 3 and 6) the reference compounds
Ara-A and Ara-C, respectively. In contrast, the Ara-A
derivatives 4 and 8 proved markedly less active but at
least 25-fold more cytostatic than the parent Ara-A. The
compounds were also evaluated on their differentiating
potential against HL-60 cell cultures. A concentration
close to their IC50 value and a 4-fold lower concentration
have been chosen to examine the effect of the test
compounds on differentiation. The Ara-C derivative 3,
in which retinoic acid was conjugated at the amino
group of the base moiety and in which the 3′ and 5′
positions of the sugar moiety were linked by highly
lipophilic silyl moieties, proved superior to the Ara-C
derivative 6, in which the 3′- and 5′-hydroxyl moieties
were not substituted. Indeed, 0.2 and 0.05 µg/mL 3
resulted in 54% and 44% differentiated HL-60 cells,
respectively, whereas 0.32 and 0.08 µg/mL 6 resulted
in 37% and 29% differentiated cells. However, the
Ara-A derivative 7 was almost as potent as 3, as a
differentiating agent, whereas its silylated derivative
was much less active in differentiating the HL-60 cell
cultures. Also, compound 8 in which retinoic acid was
Deter m in a tion of in Vitr o Hyd r olysis. To evalu-
ate the susceptibility of the compounds to spontaneous
hydrolysis, the prodrugs (6-8, 10, 12, 14) were incu-
bated in phosphate buffer (pH 7.4), free of plasma
enzymes, and in culture medium containing 20% fetal
bovine serum for up to 8 h at 37 °C, and all proved fully
stable. All compounds were also remarkably stable (t1/2
>> 360 min) toward plasma enzymes: upon incubation
in human plasma at 37 °C, they did not undergo
significant enzymatic hydrolysis for up to 6 h (data not
shown).

3854 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 23
Notes
Ta ble 1. Cytostatic and Differentiating Activity of Retinoic
Acid Conjugates against HL-60 Cells
(SAR) for the differentiating potential of the test com-
pounds. The most striking examples are the differences
in differentiating activity of the Ara-C and Ara-A
derivatives containing or lacking the lipophilic silyl
groups in the sugar moiety. For the Ara-C derivatives
3 and 6, an inverse correlation was found compared with
the Ara-A derivatives 4 and 7. Several factors may
contribute to this lack of SAR. Conjugation of the
retinoic acid moiety to different parts of the Ara-C or
Ara-A molecule may result in differential loss of the
differentiating properties of both retinoic acid and/or the
nucleosides. Since Ara-C and Ara-A presumably have
to be phosphorylated before they can act as differentiat-
ing agents on their own right, it will depend on the
intracellular stability of the test compounds as well as
their recognition by nucleoside kinases to what extent
the eventual differentiating properties will be expressed.
Also, the interaction of the retinoic acid receptor in HL-
60 cells with retinoic acid, required for optimal dif-
ferentiation, will differ from one conjugate to another
and will severely affect the differentiating potential of
retinoic acid conjugates.
differentiation
cytostatic
activity IC50,
µg/mL (µM)a
differentiation
(% NBT-
positive cells)
concentration,
µg/mL (µM)
compound
3
4
6
7
8
<0.32 (0.41)
2.08 (2.85)
<0.32 (0.6)
<0.32 (0.54)
1.96 (3.30)
<0.32 (0.68)
3.7 (7.19)
0.2 (0.26)
0.05 (0.065)
2 (2.52)
0.5 (0.63)
0.32 (0.6)
0.08 (0.15)
0.2 (0.32)
0.05 (0.08)
2 (3.36)
0.5 (0.84)
0.32 (0.68)
0.08 (0.17)
4 (7.76)
54
44
31
31
37
29
46
41
36
34
34
33
43
24
65
33
20
5
10
12
1 (1.94
4 (5.96)
1 (1.49)
14
3.71 (5.53)
ara-C
0.029 (0.12) 0.4 (1.65)
0.08 (0.33)
0.016 (0.066)
8
ara-A
20.9 (78)
50 (187.1)
10 (37.4)
12
9
Among the alkylating entities, the activity of com-
pound 10 is particularly interesting. Indeed the linkage
of the furanone 9 to retinoic acid resulted in the most
pronounced activity among the derivatives of this simple
3(2H)-furanone alkylating moiety29 which is structurally
related to the antitumor agent geiparvarin. Moreover,
it is particularly surprising to note the lack of activity
of the mustard derivative 12, since this chemical moiety
usually results in an in vitro cytotoxic and in vivo
antitumor effect. Our data that free retinoic acid is a
much more potent differentiating agent than the ret-
inoic acid conjugates are in agreement with a lower
efficiency of interaction of conjugated retinoic acid with
its receptor. However, in the case of Ara-A derivatives
significant (25- to >144-fold) increments in the cyto-
static effects were observed, but this was not the case
for Ara-C derivatives. These data are partially in
contrast with the reported, overall increased cytotoxicity
(6-8-fold) of AZT and 3′-thia-2′,3′-dideoxycytidine ret-
inoic acid conjugates which require intracellular hy-
drolysis prior to expression of their biological activity,9,10
but these authors did not report the possible effects on
the differentiating activity as compared to retinoic acid
itself.
DMSO
5.6 (71.8, 0.5%)
2.8 (35.9, 0.25%)
1.4 (17.9,0.125%)
0.03 (0.1)
0.003 (0.01)
0.0003 (0.001)
53
28
12
89
67
34
12
retinoic acid
control
a
50% inhibitory concentration. The cytotoxicity of compounds
9 and 11 has already been reported (refs 29 and 11), and they do
not express any differentiating activity.
conjugated at the 2′-hydroxyl of Ara-A proved to be a
poorer differentiating agent. In all cases the dif-
ferentiation potential was increased with regard to the
free nucleosides, but it should be mentioned that retinoic
acid itself was at least 2 to 3 orders of magnitude
superior to the conjugates to differentiate the HL-60 cell
culture.
With regard to the retinoic acid conjugates of alkyl-
ating agents, compound 10 proved most cytostatic (IC50
< 0.32 µg/mL) and also had the strongest differentiating
potential (33-34% at 0.32 and 0.08 µg/mL). Among
compounds 12 and 14, which were equally cytostatic,
compound 14, which contained 2 units of retinoic acid,
proved slightly superior to 12 with regard to its dif-
ferentiating potential (65% and 33% at 4 and 1 µg/mL,
respectively).
In conclusion, although no clear SAR emerged from
this study, the known greater instability of the N-
nucleoside conjugates of Ara-C and Ara-A, resulting in
a partial release inside the cells of retinoic acid and the
nucleoside analogue, may be at the basis of their
observed higher activity in comparison with the other
conjugates, especially in the case of Ara-A. However,
it must be noted that the prodrug potential of our
conjugates depends on how efficiently they can be
concentrated inside the target cells, due to both the
increased plasma stability and the transport across the
cell membrane, and then reversed to the parent com-
pounds. Indeed, Ara-A conjugates showed an interest-
ing 25- to >144-fold increased cytostatic activity, as
compared to Ara-A, but with reduced differentiating
activity, as compared to retinoic acid. This occurrence
remains so far unexplained but may be also consistent
with a possible cytostatic activity expressed by the intact
conjugates.
Discu ssion
The objective of this study was to design and synthe-
size molecular combinations of retinoic acid and cyto-
toxic entities. Synergistic or additive effects might be
expected from the release of the two components of the
prodrugs. Indeed, the elevated half-life value (t1/2 >>
360 min) observed accounts for a good distribution of
the prodrugs and thus for possible candidates for in vivo
testing. However, only in the case of Ara-A derivatives
(compounds 4, 7, and 8) could a 25- to >144-fold
increment in the cytostatic activity could be observed.
From our experimental data, there is no clear-cut
correlation between cytostatic activity and differentiat-
ing potential of the test compounds. Also, there seems
to be no predictable structure-activity relationship

Notes
The low activity of the mustard 12 and the diamide
14 indicates that these molecules may act, as intact
entities, as retinoic acid analogues rather than as
carriers for the two moieties, this being partially in
contrast with the reported prominent retinoidal activity
of simple as well as complex retinamides.31,32
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 23 3855
CH2 cyclohexenyl), 1.55-1.65 (m, 2H, CH2 cyclohexenyl), 1.71
(s, 3H, CH3 cyclohexenyl), 1.98-2.09 (m, 8H, CH2 cyclohexenyl,
CH3 retinoyl, CH3 alkenyl side chain), 2.38 (s, 3H, CH3), 4.85
(d, J ) 6.3 Hz, 2H, CH2O), 5.58 (s, 1H, furanyl), 5.82 (s, 1H,
retinoyl), 6.10-6.33 (m, 4H, retinoyl), 6.60-6.70 (m, 1H,
alkenyl side chain), 6.95-7.08 (m, 1H, retinoyl). Anal.
(C30H40O4) C, H.
12: eluent EtOAc/hexane 4/6 f 1/1; brown gum; yield 90%;
UV (MeOH) λmax 348 (ꢀ 45 000), 262 (ꢀ 15 000), λmin 285 (ꢀ
11 000), 227 (ꢀ 10 000); 1H-NMR (CDCl3) δ 1.03 (s, 6H, CH3 ×
2 cyclohexenyl), 1.40-1.50 (m, 2H, CH2 cyclohexenyl), 1.55-
1.65 (m, 2H, CH2 cyclohexenyl), 1.71 (s, 3H, CH3 cyclohexenyl),
1.90-2.10 (m, 5H, CH2 cyclohexenyl, CH3), 2.39 (s, 3H, CH3),
3.53-3.62 (m, 8H, CH2 × 4), 5.82 (s, 1H, retinoyl), 6.05-6.30
(m, 4H, retinoyl), 6.55 (d, 2H, J ) 8 Hz, aryl), 6.85-7.00 (m,
1H, retinoyl), 7.43 (d, 2H, aryl), 8.08 (s, 1H, NH). Anal.
(C30H39Cl2N2O) C, H, N.
14: eluent EtOAc/hexane, 1/1; yellow solid, mp 89 °C; yield
67%; UV (MeOH) λmax 349 (ꢀ 31 700), λmin 252 (ꢀ 5000); 1H-
NMR (CDCl3) δ 1.02 (s, 12H, CH3 × 4 cyclohexenyl), 1.40-
1.50 (m, 4H, CH2 × 2 cyclohexenyl), 1.55-1.67 (m, 6H, CH2 ×
2 cyclohexenyl, CH2 diaminopropyl), 1.71 (s, 6H, CH3 × 2
cyclohexenyl), 1.90-2.10 (m, 10H, CH2 × 2 cyclohexenyl, CH3
× 2), 2.36 (s, 6H, CH3 × 2), 3.30-3.42 (m, 4H, N-CH2 × 2
diaminopropyl), 5.76 (s, 2H, retinoyl × 2), 6.07-6.28 (m, 8H,
retinoyl × 2), 6.80-6.91 (m, 4H, retinoyl × 2, NH × 2). Anal.
(C43H62N2O2) C, H, N.
Exp er im en ta l Section
Ch em istr y. Ma ter ia l a n d Meth od s. The reaction course
and product mixture were routinely monitored by thin-layer
chromatography (TLC) on silica gel-precoated F254 Merck
plates with detection under 254-nm UV lamp and/or by
spraying the plates with 10% H2SO4/MeOH and heating. After
purification (column chromatography was performed with ICN
60-200 mesh silica gel), all compounds gave analytical data
consistent with the expected structures. 1H-NMR spectra were
determined at 200 MHz for compound solutions in CDCl3 or
DMSO-d6 with a Brucker AC-200 spectrometer. Ultraviolet
spectra were recorded on a Kontron UVIKON 922 spectrom-
eter. HPLC analyses were conducted on a Waters 600E
chromatographic system, using reverse-phase Waters C18
columns (150 × 4.6 mm, 150 Å). Elemental analyses, unless
otherwise noted, were within (0.4% of the theoretical values.
P r ep a r a tion of th e P r otected Ar a -C a n d Ar a -A (1, 2).
3′,5′-O-TPS-Ara-C and -Ara-A were obtained by standard
procedure. 1: syrup; yield 63%.33 2: syrup; yield 67%.34
Gen er a l P r oced u r e for th e P r ep a r a tion of Retin oic
Acid Con ju ga tes (3-5, 10, 12, 14). To a solution of retinoic
acid (0.106 g, 0.36 mmol) in dry benzene (10 mL) was added
oxalyl chloride (46 mL, 0.54 mmol), and the reaction mixture
was stirred, at ambient temperature and protected from light,
for 2 h under positive argon pressure. The deep-yellow
solution was then evaporated under high vacuum (1 × 10-3
bar), and the residue was redissolved in dry benzene (5 mL)
and slowly added, at 0 °C and under argon positive pressure,
to a solution of the appropriate compound 3-5, 9, 11 (0.36
mmol), and 13 (0.18 mmol), in dry benzene containing DMAP
(catalytic amount) and TEA (1.5 mL, 1.8 mmol). After 1 h
(TLC analysis) the reaction was usually complete. The reac-
tion mixture was then diluted with benzene (20 mL) and
evaporated to dryness to give a residual oil which was purified
by silica gel column chromatography.
3: eluent CH2Cl2/MeOH, 9.5/0.5; yellow foam; yield 40%;
1H-NMR (DMSO-d6) δ 0.92-1.08 (m, 34H, iPr × 4, CH3 × 2
cyclohexenyl), 1.41-1.50 (m, 2H, CH2 cyclohexenyl), 1.55-1.63
(m, 2H, CH2 cyclohexenyl), 1.72 (s, 3H, CH3), 1.90-2.10 (m,
5H, CH2 cyclohexenyl and CH3); 2.36 (s, 3H, CH3); 3.78-4.20
(m, 4H, H3′,4′,5′,5′′), 4.30-4.50 (m, 1H, H2′), 5.61 (pseudo-t,
1H, OH2′), 5.70 (d, J ) 7.5 Hz, 1H, H-5), 5.74 (br s, 2H, H1′,
retinoyl), 6.12-6.43 (m, 4H, retinoyl), 6.95-7.18 (m, 2H, NH,
retinoyl), 7.50 (d, 1H, H-6), 8.45 (br s, 1H, NH).
4: eluent EtOAc/hexane, 3/7 f 4/7; yellow foam; yield 45%;
1H-NMR (CDCl3) δ 0.95-1.10 (m, 34H, iPr × 4, CH3 × 2
cyclohexenyl), 1.45-1.50 (m, 2H, CH2 cyclohexenyl), 1.61-1.69
(m, 2H, CH2 cyclohexenyl), 1.72 (s, 3H, CH3), 2.0-2.10 (m, 5H,
CH2 cyclohexenyl, CH3), 2.36 (s, 3H, CH3), 3.60-3.65 (m, 1H,
H5′), 4.0-4.10 (m, 2H, H4′, H5′′), 4.32-4.39 (dd, 1H, J ) 8.5,
3 Hz, H3′), 4.55-4.60 (m, 1H, H2′), 5.55 (br s, 1H, OH2′), 5.94
(d, 1H, J ) 5 Hz, H1′), 6.09-6.37 (m, 6H, retinoyl, NH), 6.50
(s, 1H, H2), 7.0-7.14 (m, 1H, retinoyl), 8.14 (s, 1H, H8).
5: eluent EtOAc/hexane, 3/7 f 4/7; yellow foam; yield 25%;
1H-NMR (CDCl3) δ 1.01-1.25 (m, 34H, iPr × 4, CH3 × 2
cyclohexenyl), 1.40-1.50 (m, 2H, CH2 cyclohexenyl), 1.55-1.65
(m, 2H, CH2 cyclohexenyl), 1.70 (s, 3H, CH3), 1.90-2.05 (m,
5H, CH2 cyclohexenyl, CH3), 2.19 (s, 3H, CH3), 3.92-4.0 (m,
1H, H4′), 4.10-4.20 (m, 1H, H5′), 4.20-4.35 (m, 1H, H5′′),
4.95-5.10 (m, 1H, H3′), 5.50-5.60 (m, 1H, H2′), 5.70 (br s,
2H, NH2), 6.14-6.35 (m, 5H, retinoyl), 6.49 (d, 1H, J ) 5.5
Hz, H1′), 6.85-7.0 (m, 1H, retinoyl), 7.98 (s, 1H, H2), 8.28 (s,
1H, H8).
Gen er a l P r oced u r e for Dep r otection of th e Nu cleo-
sid e Der iva tives 3-5. To a solution of the protected 3-5
(0.26 mmol) in dry MeOH (10 mL) under positive argon
pressure was added acetic acid (0.039 g, 0.65 mmol) followed
by NH4F (0.115 mg, 3 mmol). The reaction was monitored on
TLC (CH2Cl2/MeOH, 9/1) and heated at reflux conditions if
necessary. After evaporation to dryness the residue was
purified by silica gel column chromatography.
6: eluent CH2Cl2/MeOH, 9.5/0.5; orange solid, mp 111 °C;
yield 43%; UV (MeOH) λmax 370 (ꢀ 32 000), 249 (ꢀ 14 000), 298
1
(ꢀ 8000), λmin 270 (ꢀ 4000), 230 (ꢀ 6000); H-NMR (DMSO-d6)
δ 1.02 (s, 6H, CH3 × 2 cyclohexenyl), 1.40-1.50 (m, 2H, CH2
cyclohexenyl), 1.54-1.65 (m, 2H, CH2 cyclohexenyl), 1.71 (s,
3H, CH3), 1.95-2.05 (m, 5H, CH2 cyclohexenyl, CH3), 2.34 (s,
3H, CH3), 3.45-3.60 (m, 2H, H5′,5′′), 3.65-3.75 (m, 1H, H4′),
3.80-4.0 (m, 2H, H3′,2′), 5.05 (br s, 1H, OH5′), 5.40-5.45 (m,
2H, OH3′,2′), 6.05-6.40 (m, 6H, H1′, retinoyl), 6.90-7.03 (m,
1H, retinoyl), 7.20 (d, 1H, J ) 7.5 Hz, H-5), 8.05 (d, 1H, H-6),
10.45 (br s, 1H, NH). Anal. (C29H40N3O6) C, H, N.
7: eluent CH2Cl2/MeOH, 9/1; syrup; yield 66%; UV (MeOH)
λmax 372 (ꢀ 29 000), 216 (ꢀ 21 500), λmin 294 (ꢀ 7700), λshoulder
240 (ꢀ 10 000); 1H-NMR (DMSO-d6) δ 1.03 (s, 6H, CH3 × 2
cyclohexenyl), 1.40-1.50 (m, 2H, CH2 cyclohexenyl), 1.54-1.62
(m, 2H, CH2 cyclohexenyl), 1.69 (s, 3H, CH3), 1.95-2.02 (m,
5H, CH2 cyclohexenyl, CH3), 2.33 (s, 3H, CH3), 3.60-3.69 (m,
2H, H5′,5′′), 3.74-3.85 (m, 1H, H4′), 4.10-4.28 (m, 2H, H3′,2′),
5.10-5.18 (br, 1H, OH5′), 5.70-5.75 (br, 2H, OH2′,3′), 6.01-
6.43 (m, 6H, H1′, retinoyl × 5), 6.75 (s, 1H, H2), 6.96-7.10
(m, 1H, retinoyl), 8.09 (s, 1H, H8), 10.05 (br s, 1H, NH). Anal.
(C30H39N5O5) C, H, N.
8: eluent CH2Cl2/MeOH, 9.5/0.5; syrup; yield 95%; UV
(MeOH) λmax 366 (ꢀ 27 000), 258 (ꢀ 15 500), λmin 290 (ꢀ 8000),
229 (ꢀ 8500); 1H-NMR (DMSO-d6) δ 1.03 (s, 6H, CH3 × 2
cyclohexenyl), 1.40-1.50 (m, 2H, CH2 cyclohexenyl), 1.54-1.62
(m, 2H, CH2 cyclohexenyl), 1.72 (s, 3H, CH3), 1.85-2.05 (m,
5H, CH2 cyclohexenyl, CH3), 2.17 (s, 3H, CH3), 3.60-3.80 (m,
2H, H5′,5′′), 3.83-3.88 (m, 1H, H4′), 4.48 (q, 1H, J ) 5.8, 5.9
Hz, H3′), 5.15 (br, 1H, OH5′), 5.30 (φt, 1H, J ) 5.8, 5.7 Hz,
H2′), 5.54 (s, 1H, retinoyl), 5.86 (d, 1H, J ) 5.2 Hz, OH3′),
6.09-6.30 (m, 4H, retinoyl), 6.46 (d, 1H, J ) 5.7 Hz, H1′),
6.93-7.01 (m, 1H, retinoyl), 7.26 (br s, 2H, NH2), 8.09 (s, 1H,
H2), 8.20 (s, 1H, H8). Anal. (C30H39N5O5) C, H, N.
Hyd r olysis of th e Com p ou n d s. To 100 µL of phosphate
buffer (0.2 M, pH 7.4), human plasma, or culture medium
containing 20% fetal bovine serum was added 10 µL of a
solution of one of the compounds (10 mg/mL in DMSO), and
the mixture was incubated at 37 °C in a water bath. At
various time intervals (0-8 h), 20 µL of the samples was
withdrawn and introduced in a quartz cuvette containing 1
10: eluent EtOAc/hexane, 1/9 f 2/8; oil; yield 45%; UV
(MeOH) λmax 360 (ꢀ 20 000), 297 (ꢀ 17 000), λmin 325 (ꢀ 15 000),
263 (ꢀ 11 000); 1H-NMR (CDCl3) δ 1.03 (s, 6H, CH3 × 2
cyclohexenyl), 1.40 (s, 6H, CH3 × 2 furan), 1.41-1.49 (m, 2H,

3856 J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 23
Notes
(4) Fanjul, A.; Dawson, M. I.; Hobbs, P. D.; J ong, L.; Cameron, J .
F.; Harlev, E.; Graupner, G.; Lu, X. P.; Pfahl, M. A new class of
retinoids with selective inhibition of AP-1 inhibits proliferation.
Nature 1994, 372, 107-111.
mL of water. Optical density values were measured at the
following wavelengths: 6 at 362, 351, and 274 nm; 7 at 376,
351, and 251 nm; 8 at 366, 351, and 259 nm; 10 at 360, 351
and 261 nm. In the case of 12 and 14 (having λmax overlapping
with that of the retinoic acid) the hydrolyses were monitored
by HPLC: 20 µL of samples was withdrawn, diluted with
methanol (800 µL), and poured at -21 °C. The supernatants
were filtered and analyzed by HPLC using an acetonitrile-
water solvent system.
(5) Mori, H.; Tanaka, T. Chemopreventive retinoids. Drugs Future
1993, 18, 737-742.
(6) Dawson, M. I.; J ong, L.; Hobbs, P. D.; Cameron, J . F.; Chao, W.;
Pfahl, M.; Lee, M.; Shroot, B.; Pfahl, M. Conformational effects
on retinoid receptor selectivity. 2. Effects of retinoid bridging
group on retinoid × receptor activity and selectivity. J . Med.
Chem. 1995, 38, 3368-3383.
Biology. In h ibition of Tu m or Cell Gr ow th . HL-60 cells
were obtained from the American Type Culture Collection
(ATCC, Rockville, MD). All assays were performed in 96-well
microtiter plates. To each 200-µL well were added 7.5 × 104
HL-60 cells (100 µL) and a given amount of the test compound
(100 µL). The cells were then allowed to proliferate for 72 h
at 37 °C in a humidified CO2-controlled atmosphere. The
growth of the cells was linear during this incubation period.
At the end of the incubation period, the cells were counted in
a Coulter counter (Coulter Electronics Ltd., Harpenden, Herts,
England). The IC50 (50% inhibitory concentration) was defined
as the concentration of compound that reduced the number of
tumor cells by 50%.
(7) Bollag, W.; Majewski; J ablonska, S. Cancer combination che-
motherapy with retinoids: experimental rationale. Leukemia
1994, 8, 1453-1457 and references cited therein.
(8) Poli, G.; Kinter, A. L.; J ustement, J . S.; Bressler, P.; Kehrl, J .
H.; Fauci, A.S. Retinoic acid mimics transforming growth factor
in the regulation of human immunodeficiency virus expression
in monocytic cells. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 2689-
2693.
(9) Camplo, M.; Faury, P.; Charvet, A. S.; Graciet, J . C.; Chermann,
J . C.; Kraus, J . C. Synthesis and comparative anti-HIV activities
of new acetylated 2′,3′-dideoxy-3′-thiacytidine analogues. Eur.
J . Med. Chem. 1994, 29, 357-362.
(10) Aggarwal, K. S.; Gogu, R. S.; Rangan, S. R. S.; Agrawal, K. C.
Synthesis and biological evaluation of prodrugs of Zidovudine.
J . Med. Chem. 1990, 33, 1505-1510.
(11) Gourdie, T. A.; Valu, K. K.; Gravatt, G. L.; Boritzki, T. J .;
Baguley, B. C.; Wakelin, L. P. G.; Wilson, W. R.; Woodgate, P.
D.; Denny, W. A. DNA-Directed alkylating agents. 1. Structure-
activity relationships for acridine-linked mustards: conse-
quences of varying the reactivity of the mustard. J . Med. Chem.
1990, 33, 3014-3019 and references cited therein.
(12) Manfredini, S.; Baraldi, P. G.; Bazzanini, R.; Guarneri, M.;
Simoni, D.; Balzarini, J .; De Clercq, E. Geiparvarin analogues.
4. Synthesis and cytostatic activity of geiparvarin analogues
bearing a carbamate moiety or a furocoumarine fragment on the
alkenyl side chain. J . Med. Chem. 1994, 37, 2401-2405.
(13) Meijer, D. K. R.; J ansen, R. W.; Molema, G. Drug targeting
systems for antiviral agents: options and limitations. Antiviral
Res. 1992, 18, 215-258.
(14) Wipf, P.; Wenjie, L. Prodrugs of Ara-C. Drugs Future 1994, 19,
49-54.
(15) Lundberg, B. Cytotoxic activity of two new lipophilic steroid
nitrogen carbamates incorporated into low-density lipoprotein.
Anti Cancer Drug Des. 1994, 9, 471-476.
(16) Robarge, M. J .; Repa, J . J .; Hanson, K. K.; Seth, S.; Clagett-
Dame, M.; Abou-Issa, H.; Curley, R. W., J r. N-linked analogs of
retinoid O-glucuranides: potential cancer chemopreventive/
chemotherapeutic agents. Bioorg. Med. Chem. Lett. 1994, 4,
2117-2122.
(17) Van Lier, J . E.; Kan, G.; Autenrieth, D.; Nigam, V. N. Steroid-
nucleosides possible novel agents for cancer chemotherapy.
Nature 1977, 267, 522-524.
(18) Ponzoni, M.; Lanciotti, M.; Montaldo, P. G.; Cornaglia-Ferraris,
P. Gamma interferon, retinoic acid and cytosine arabinoside
induce neuroblastoma differentiation mechanism. Cell. Mol.
Neurobiol. 1991, 11, 397-413.
(19) Hassan, H. T.; Rees, J . K. H. Triple combination of retinoic acid,
low concentration of cytarabine and dimethylformamide induces
differentiation of human acute myeloid leukemic blasts. Che-
motherapy 1986, 10, 631-636.
(20) Chomienne, C.; Balitrand, N.; Degos, L.; Abita, J . P. 1-â-D-
Arabinofuranosyl cytosine and retinoic acid in combination
accelerates and increaseas monocyte differentiation of myeloid
leukemic cells. Leuk. Res. 1986, 10, 631-636.
(21) (a) Backer, C. D.; Haskell, T. H.; Putt, S. R.; Sloan, B. J . Prodrugs
of 9-â-D-arabinofuranosyladenine 2. Synthesis and evaluation
of a number of 2′,3′- and 3′,5′-di-O-acyl-derivatives. J . Med.
Chem. 1979, 22, 273-279. (b) Backer, D. C.; Kumar, S. D.;
Waites, W. J .; Arnett, G.; Shannon, W. M.; Higichi, W. I.;
Lambert, W. Synthesis and evaluation of a series of 2′-O-acyl
derivatives of 9-â-D-arabinofuranosyladenine as antiherpes
agents. J . Med. Chem. 1984, 27, 270-274.
Differ en tia tion of HL-60 Cell Cu ltu r es. The differen-
tiating activity of the test compounds was examined in HL-60
cell cultures and performed as previously described.35 Briefly,
106 HL-60 cells (exposed to the test compounds for 6 days)
were suspended in 1 mL of RPMI-1640 culture medium
containing 20% fetal bovine serum and 2 mM L-glutamine.
Then 1 mL of a freshly prepared nitro blue tetrazolium (NBT)
solution [2 mg/mL, containing phorbol 12-myristate 13-acetate
(PMA) at 200 ng/mL in phosphate-buffered saline, pH 7.2] was
added to the cell suspension, and the mixture was further
incubated for 40 min in a shaking water bath at 37 °C. After
this incubation period, the mixture was cooled on ice during 5
min to stop the reaction. Then the cells were centrifuged for
10 min at 1000 rpm at 4 °C, and the supernatant was carefully
removed. The cell pellet was resuspended in 0.5 mL of RPMI-
1640 medium, kept on ice, and protected from light until the
percentage of cells containing dark blue formazan precipitates
was determined in a hemocytometer by light microscopy. For
each sample, at least a total of 200 cells were scored, and the
percentage of formazan-positive cells was calculated. Cell
cultures that were not exposed to the test compounds were
included in this assay procedure and found to contain not more
than 12% NBT-positive cells.
Ack n ow led gm en t. This work was supported by
grants from the Belgian Nationaal Fonds voor Ge-
neeskundig Wetenschappelijk Onderzoek (3.3010.91),
the Belgian Fonds voor Geneeskundig Wetenschappelijk
Onderzoek (3.0180.95), the Belgian Econcerteerde Onder-
zoeksacties (95/5), and the Consiglio Nazionale delle
Ricerche (Progetto Finalizzato Chimica Fine 2,
93.02895.PS72 and 96.01104). We thank Lizette van
Berckelaer for excellent technical assistance.
Refer en ces
(1) (a) Hong, W. K.; Itri, M. I. Retinoids and human cancer. In The
Retinoids, Biology, Chemistry and Medicine, 2nd ed.; Sporn, M.
B., Roberts, A. B., Goodman D. S., Eds.; Raven Press: New York,
1994; pp 597-630. (b) Livrea M. A., Vidali, G., Eds. Retinoids
from Basic Science to Clinical Applications; Birkhauser: Basel,
1994. (c) Nagpal, S.; Chandraratna, R.A.S. Retinoids as anti-
cancer agents. Curr. Pharmacol. Des. 1996, 2, 295-316.
(2) (a) Mangelsdorf, D. J .; Umesons, K.; Evans, R. M. The retinoid
receptors. In The Retinoids, Biology, Chemistry and Medicine,
2nd ed.; Sporn, M. B., Roberts, A. B., Goodman D. S., Eds.; Raven
Press: New York, 1994; pp 319-350. (b) Sani, B. P.; Shealy, Y.
F.; Hill, D. L. N-(4-Hydroxyphenyl)retinamide: interactions with
retinoid-binding proteins/receptors. Carcinogenesis 1995, 16,
2531-2434.
(22) (a) Manfredini, S.; Baraldi, P. G.; Bazzanini, R.; Bortolotti, F.;
Simoni, D.; Ashida, N.; Machida, H. Enzymatic synthesis of 2′-
O-acyl prodrugs of 1-(â-D-arabinofuranosyl)-5(E)-(2-bromovinyl)-
uracil (Sorivudine, BV-araU) and of 2′-O-acyl-araU, -araC and
-araA. Manuscript submitted. (b) Baraldi, P. G.; Bazzanini, R.;
Manfredini, S.; Simoni, D.; Robins, M. J . Facile access to 2′-O-
acyl prodrugs of 1-(â-D-arabinofuranosyl)-5(E)-(2-bromovinyl)-
uracil (BVAraU) via regioselective esterase-catalyzed hydrolysis
of 2′,3′,5′-triesters. Tetrahedron Lett. 1993, 34, 3177-3180.
(23) Tsiftsoglou, A. S.; Pappas, I. S.; Niopas, I. Structural, cellular
and pharmacological implications of neoplastic cell differentia-
tion induced by ureido derivatives of pyridine (UDPs). In
Chemistry and Properties of Biomolecular Systems; Rizzarelli,
E., Theophanides, T., Eds.; Kluwer Academic Publishers: Dor-
drecht, 1991; pp 189-208.
(3) Greiner, E. F.; Kirfel, J .; Greschik, H.; Do¨rflinger, U.; Becker,
P.; Mercep, A.; Schu¨le, R. Functional analysis of retinoid Z
receptor beta, a brain-specific nuclear orphan receptor. Proc.
Natl. Acad. Sci. U.S.A. 1996, 93, 10105-10110.

Notes
J ournal of Medicinal Chemistry, 1997, Vol. 40, No. 23 3857
(24) Tsiftsoglou, A. S.; Bhargava, K. K.; Rittman, L. S.; Sartorelli,
A. C. Distribution of the inducer of differentiation bis-acetyl-
diaminopentane in murine erythroleukemia cells. J . Cell. Phys-
iol. 1981, 106, 419-424.
(25) Zhang, W.; Robins, M. J . Removal of silyl protecting groups from
hydroxyl functions with ammonium fluoride in methanol. Tet-
rahedron Lett. 1992, 33, 1177-1180.
(26) Ryu, E. K.; Ross, R. J .; Matsishita, T.; MacCoss, M.; Hong, C. I.;
West, C.R. Phospholipid-nucleoside conjugates. 3. Syntheses and
preliminary biological evaluation of 1-1-â-Darabinofuranosyl-
cytosine 5′-monophosphate-L-1,2-dipalmitin and selected 1-â-
Darabinofuranosylcytosine 5′-diphosphate-L-1,2-diacylglycerols.
J . Med. Chem. 1982, 25, 1322-1329.
(27) Montgomery, J . A.; Thomas, H. J . Acyl derivatives of 1-â-D-
arabinofuranosylcytosine. J . Med. Chem. 1971, 15, 116-118.
(28) Beranek, J .; Drasar, P. Selective O-acylation. Preparation of
(31) Shealy, Y. F.; Frye, J . L.; O’Dell, C. A.; Thorpe, M. C.; Kirk, M.
C.; Cobrun W. C.,J r.; Sporn, M.B. Synthesis and properties of
some 13-cis and all-trans-retinamides. J . Pharm. Sci. 1984, 73,
745-751.
(32) Shealy, Y. F.; Frye, J . L.; Schiff, L. J . N-(Retinoyl)amino acids.
Synthesis and chemopreventive activity in vitro. J . Med. Chem.
1988, 31, 190-196.
(33) Robins, M.J .; Wilson, J .S.; Sawyer, L.; J ames, M.N.G. Nucleic
acid related compounds. 41. Restricted furanoseconformations
of 3′,5′-O-(1,1,3,3-tetraisopropyldisilox-1,3-diyl)nucleosides pro-
vide a convenient evaluation of anomeric configuration. Can. J .
Chem. 1983, 61, 1911-1920.
(34) J iang, C.; Suhadolnik, R.J .; Baker, D.C. Oxygen-18-labeled
adenosine and 9-(beta-D-arabinofuranosyl)adenine (ARA-A): syn-
thesis, mass spectrometry, and studies of oxygen-18-induced
carbon-13 NMR chemical shifts. Nucleosides Nucleotides 1988,
7, 271-294.
(35) Balzarini, J .; Vertsuyf, S.; Goebels, J .; Sobis, H.; Vandeputte,
M.; De Clercq, E. The human immunodeficiency virus (HIV)
inhibitor 9-(2-phosphonylmethoxyethyl)adenine (PMEA) is a
strong inducer of differentiation of several tumor cell lines. Int.
J . Cancer 1995, 61, 130-137.
1-(2,3,5-tri-O-acetyl-â-D-arabinopentanosyl)cytosine
hydro-
chloride. Collect. Czech. Chem. Commun. 1977, 42, 366-369.
(29) Baraldi, P. G.; Guarneri, M.; Manfredini, S.; Simoni, D.; Balzari-
ni, J .; De Clercq, E. Synthesis and cytostatic activity of geipar-
varin analogues. J . Med. Chem. 1989, 32, 284-288.
(30) Everett, J . L.; Ross, W. C. J . Aryl-2-halogenoalkylamines. Part
II. J . Chem. Soc. 1949, 1972-1983.
J M9602322