The discovery and synthesis of novel adenosine substituted 2,3-dihydro-1H-isoindol-1-ones: Potent inhibitors of poly(ADP-ribose) polymerase-1 (PARP-1)

Article abstract of DOI:10.1016/j.bmcl.2003.10.007

A series of novel 4-(N-acyl)-2,3-dihydro-1H-isoindol-1-ones have been prepared from methyl-3-nitro-2-methylbenzoate and linked through various spacers to the adenosine derivatives 11 and 12. We found that potent inhibition of poly(ADP-ribose)polymerase-1 (PARP-1) was achieved when isoindolinone was linked to adenosine by a spacer group of a specific length. Introduction of piperazine and succinyl linkers between the isoindolinone and adenosine core structures resulted in highly potent compounds 8a and 10b, which showed IC ₅₀ values of 45 and 100 nM, respectively.

Full text of DOI:10.1016/j.bmcl.2003.10.007

Bioorganic & Medicinal Chemistry Letters 14 (2004) 81–85
The discovery and synthesis of novel adenosine substituted
2,3-dihydro-1H-isoindol-1-ones: potent inhibitors of
poly(ADP-ribose) polymerase-1 (PARP-1)
Prakash G. Jagtap,* Garry J. Southan, Erkan Baloglu, Siya Ram,
Jon G. Mabley, Anita Marton, Andrew Salzman and Csaba Szabo*
Inotek Pharmaceuticals Corporation, 100 Cummings Center, Suite 419E, Beverly, MA 01915, USA
Received 12 August 2003; revised 2 October 2003; accepted 6 October 2003
Abstract—A series of novel 4-(N-acyl)-2,3-dihydro-1H-isoindol-1-ones have been prepared from methyl-3-nitro-2-methylbenzoate
and linked through various spacers to the adenosine derivatives 11 and 12. We found that potent inhibition of poly(ADP-ribose)-
polymerase-1 (PARP-1) was achieved when isoindolinone was linked to adenosine by a spacer group of a specific length. Intro-
duction of piperazine and succinyl linkers between the isoindolinone and adenosine core structures resulted in highly potent
compounds 8a and 10b, which showed IC50 values of 45 and 100 nM, respectively.
# 2003 Elsevier Ltd. All rights reserved.
1. Introduction
Various compounds that incorporate a carboxamide
group in an anti- (or cis-) configuration, within a ring
structure (i.e., lactams) were reported2 as considerably
more effective at inhibiting PARP than a prototypical
inhibitor, 3-aminobenzamide (3-AB). Of these, the sub-
stituted 3,4-dihydroisoquinolin-1(2H)-ones and iso-
quinolin-1(2H)-ones, were investigated as potential
enhancers of radiotherapeutic or chemotherapeutic
agents.2b,3 The initial impetus for the earlier work was
the concept that inhibitors of PARP would enhance the
cytotoxic effects in the treatment of cancer.4 It was
shown that high doses of the compounds were required
for these effects, suggesting that a near-complete inhibi-
tion of PARP-1 was necessary. Thus, the development
of potent PARP inhibitors for this therapeutic area
remains a valid concept.
Poly(ADP-ribose) polymerase-1 (PARP-1, EC 2.4.2.30)
also known as poly(ADP-ribose) synthetase (PARS)
and poly(ADP-ribose) transferase (PADPRT) is a
nuclear enzyme present in eukaryotes. PARP-1 is the
principal member of a family of poly(ADP-ribosyl)ating
enzymes and functions both as a DNA damage sensor
and as a signalling molecule. Upon binding to damaged
DNA, PARP-1 forms homodimers and catalyzes the
cleavage of NAD+ into nicotinamide and ADP-ribose.
The latter is used to synthesize branched nucleic acid-like
polymers with poly(ADP-ribose) units covalently
attached to nuclear acceptor proteins including histones,
transcription factors and PARP. The poly(ADP-ribosyl)-
ation of these proteins contributes to inflammatory sig-
nal transduction processes. In addition, oxidative stress-
induced over-activation of PARP consumes NAD+
and, consequently, ATP culminating in cell dysfunction
or necrosis. Activation of PARP has been implicated in
the pathogenesis of stroke, myocardial ischemia, dia-
betes, diabetes-associated cardiovascular dysfunction,
shock, traumatic central nervous system injury, arthri-
tis, colitis, allergic encephalomyelitis and various other
forms of inflammation.1 Therefore, inhibition of PARP
by pharmacological agents may prove useful for the
treatment of these diseases.
An original study by Banasik2a compared a large num-
ber of compounds from a variety of structural classes
for their ability to inhibit PARP. The most potent
compounds contain a lactam ring as a part of their cyc-
lic systems. These findings formed the basis for many of
the polycyclic PARP inhibitors subsequently developed.
Recently,5 the growing realization that PARP may be
involved in the pathogenesis of many diseases led to the
development of numerous types of novel PARP inhibi-
tors.5a Several classes of PARP inhibitors originally
described by Banasik2a were later optimized in order to
enhance potency, improve pharmacokinetic character-
istics and increase solubility in water. Recently, Costan-
* Corresponding authors. Tel.: +1-978-232-9660x295; fax: +1-978-
232-8975; e-mail: pjagtap@inotekcorp.com; szabocsaba@aol.com
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.10.007

82
P. G. Jagtap et al. / Bioorg. Med. Chem. Lett. 14 (2004) 81–85
tino et al.5b reported a quantitative structure–activity
relationship (QSAR) analysis of several PARP inhibitors.
and coupling them with 4-substituted-isoindolinones
carrying the appropriate linkers bearing an amine or
acid group.
While much emphasis has been placed on extending the
ring systems to include tri-, tetra- and penta-cyclic sys-
tems with all the inherent problems or concerned asso-
ciated with this,6 we tried to improve the binding
characteristics of modest inhibitors by coupling them to
adenosine, a structural element of NAD, which is recog-
nized by the NAD binding site of PARP. Adenosine by
itself is only a weak inhibitor of PARP activity, but we
postulated that appropriately substituted adenosine
derivatives might show good PARP inhibitory activity.
To expand on this hypothesis, we prepared a series of
4-substituted isoindolinones and linked them through
various spacers to adenosine. We report herein, that
potent inhibition of PARP was achieved with the selec-
tion of appropriate linkers comprising cyclic or alkyl
chains.
A refluxing solution of commercially available ester 2a
and N-bromosuccinimide (NBS) in CH2Cl2 was treated
with 2,20-azobisisobutyronitrile (AIBN) to give bromo
ester 2b in excellent yield, which upon treatment with a
solution of ammonia in methanol at room temperature
furnished 4 - nitro - 2,3 - dihydro - 1H - isoindolinone 3
(Scheme 1). The nitro compound 3 was reduced to the
intermediate amino compound 4 using ammonium for-
mate and palladium on carbon (10%). Reaction of 4
with various acid chlorides in an aqueous solution of
sodium bicarbonate and ethyl acetate (Schotten Bau-
mann reaction) provided 5a–b. Compounds 9a–h were
prepared from isoindolinone 4 using a similar acylation
reaction approach, followed by hydrolysis of the corre-
sponding esters. Reaction of 5a with 4-tBOC-piperazine
in methanol at 50 ꢀC provided 6a in 59% yield. Iso-
indolinone 7 was then prepared by treating 6a with HCl
(4.0 M solution in dioxane) at 0 ꢀC to room temperature
in 98% yield. Similarly, reaction of 5a–b with various
amines in methanol provided compounds 6b–m in good
yields.
2. Results and discussion
The synthetic routes to prepare the adenosine derivatives
8a–c and 10a–h, with various combinations of the side
chains at the isoindolinone C-4 position, were designed
to use acid and base coupling reactions (Schemes 1–3).
These couplings were accomplished by introducing acid
or amine functional groups at the 50-position of 20,30-iso-
propylidene protected adenosine derivatives (11 and 12)7
The amines 6b–c and 7 were subsequently treated with
20,30-isopropylideneadenosine-50-carboxylic acid 11, as
shown in Scheme 2. This was followed by deprotection of
the adenosine-coupled product to give compounds 8a–c.
Scheme 1. Reagents and conditions: (i) NBS, AIBN, methylene chloride, reflux, 16 h, 98%; (ii) ammonia–MeOH (7 N), rt, 2 h, 81%; (iii) ammo-
nium formate, Pd/C (10%), DMF, 100 ꢀC, 30 min, 83%; (iv) ClCO(CH2)nCl, NaHCO3, EtOAc, rt, 30 min, 80–85%; (v) 4-N-tBOC-piperazine,
MeOH, 50 ꢀC, 48 h, 59%; (vi) HCl–dioxane, MeOH, 0 ꢀC to rt, 16 h, 98%; (vii) amine, MeOH, rt, 2–24 h; (viii) ClCO(CH2)nCOOMe, NaHCO3,
EtOAc, rt, 30 min; (ix) NaOH, MeOH, rt.

P. G. Jagtap et al. / Bioorg. Med. Chem. Lett. 14 (2004) 81–85
83
Scheme 2. Reagents and conditions: (i) DMF–methylene chloride, diisopropylethylamine, EDAC, rt, 24 h; then acid, water, rt, 1–2 h, 49–55%.
Scheme 3. Reagents and conditions: (i) DMF–methylene chloride, diisopropylethylamine, EDAC, rt, 24 h; then acid, water, rt, 1–2 h.
The synthetic routes for preparing compounds 10a–h
were similar except that 20,30-isopropylideneadenosine-
50-methyleneamine (12) was used for the coupling
reaction (Scheme 3). The products obtained after this
reaction were treated with TFA and water as described
above to yield 10a–h, which contain aliphatic chains
ranging from 1 to 8 methylene units between the adeno-
sine and the isoindolinone.
the peroxynitrite toxicity assay. The various sub-
stituents at the 4-position had only minor effects activ-
ity. However, significant increases in potency were
achieved when benzimidazole (6l) and N6-dimethylade-
nine (6m) were introduced at the 4-position as shown in
Tables 1 and 2, respectively. These compounds (6l and
6m) showed a 10-fold increase in activity.
When adenosine is incorporated into the 4-position of
isoindolinone using various spacers, the nature of the
linker between the isoindolinone core and the adenosine
had a large impact on efficacy. As shown in Schemes 2
and 3, adenosine derivatives 8b–c and 10a–h with var-
ious chain length [(CH2)n] (n=1–8) were prepared. The
piperazine and phenyl containing adenosine derivatives
8a and 13 (Fig. 1) represent derivative with cyclic lin-
kers. The potency was best for the succinyl and propa-
noyl linked compounds 10b and 8b (IC50=0.1 and 0.25
mM, respectively). Compound 8c, which contains a sec-
ondary amide was less active than the N-methyl amide
2,3-Dihydro-1H-isoindol-1-ones incorporate the requi-
site lactam in a five-membered ring. Polycycles contain-
ing the isoindolinone structure have been reported to be
potent PARP inhibitors.8 However, isoindolinone 1
showed only 33% inhibition at 10 mM in a cell-based
assay.9 The simple 4-substituted isoindolinones also
had poor potency (IC50 >10 mM) in this assay, but
more elaborate side chains showed improved potency.
The 4-N-substituted isoindolinones have weak to
moderate potency against the isolated enzyme and
offered only moderate protection against cell injury in

84
P. G. Jagtap et al. / Bioorg. Med. Chem. Lett. 14 (2004) 81–85
Table 1. PARP-1 Inhibition data of 4-substituted-2,3-dihydro-1H-
isoindol-1-ones
Compd
n
R
IC50, mMa
6d
6e
6f
6g
6h
6i
6j
6k
6l
7
8b
1
1
7
1
1
2
1
1
1
1
2
2
1
2
3
4
5
6
7
8
–NHMe
–NHEt
–NHMe
>30
18
30
30
>30
>30
>30
30
3
18
–NEt2
–4-N-Me-piperazine
–4-N-Me-piperazine
–Morpholine
–1,2,3,4-Tetrahydroisoquinoline
–Benzimidazole
.
–Piperazine 2HCl
NMe-50-CO-adenosine
NH-50-CO-adenosine
0.25
0.7
1.9
0.1
0.45
0.9
3
0.7
2
3
8c
Figure 1.
10a
10b
10c
10d
10e
10f
10g
10h
CO-50-NH-CH2-adenosine
CO-50-NH-CH2-adenosine
CO-50-NH-CH2-adenosine
CO-50-NH-CH2-adenosine
CO-50-NH-CH2-adenosine
CO-50-NH-CH2-adenosine
CO-50-NH-CH2-adenosine
CO-50-NH-CH2-adenosine
a IC50 values have been determined by using a commercially available
in vitro PARP-1 inhibition assay kit (Trevigen, Gaithersburg, MD,
USA) following manufacturer’s recommendations.
Scheme 4. Reagents and conditions: (i) succinic anhydride, chloro-
form, reflux, 2.5 h; (ii) DMF–CH2Cl2, diisopropylethylamine, EDAC,
12, rt, 24 h; then TFA, water, rt, 1–2 h.
Table 2. PARP-1 Inhibition data of isoindolinones 4, 6m and 8a
Compd
R
IC50, mMa
Table 3. PARP-1 Inhibition data of compounds 1 and 13–18
Compd
% PARP-1 inhibition9 at 10 mM
IC50, mMa
4
30
1
13
33
51
>30
10
14
15
16a
17a
17b
68
30
n.d. at 10 mM
>300
n.d.
75
28
52
5
6m
3
>300
a See Table 1. n.d., not determined.
3-Aminobenzamide (16a) is a commercially available
PARP inhibitor that has an IC50 in the range of 30 mM
in the isolated PARP-1 enzyme assay.10 To see the effect
of an adenosine linked 3-aminobenzamide analogue on
PARP-1 inhibition, a succinyl linker was coupled to
3-AB through the 3-position to adenosine as shown in
Scheme 4. In the cell based assay,9 3-AB (16a) showed
28% inhibition of PARP-1 at 10 mM, whereas adenosine
linked 3-AB derivative 17a showed 52% inhibition at 10
mM (Table 3). In the same assays, the 50-piperazine
amide 15, adenosine, inosine and the biphenyl derivative
17b had only weak effects on PARP activity.
EB-47 (8a)
0.045
a See Table 1.
derivative 8b. Compounds 10e–h, showed deleterious
effects on activity. Compound 13, with a styrene group
incorporated into the linker, had reduced potency.
However, the piperazine analogue EB-47 (8a) showed
100% inhibition at 200 nM and exhibited the lowest
IC50 (0.045 mM). The inosine-linked compound 14 (Fig.
1) was 10-fold less potent than the adenosine analogue
10f.
The difference in potency between the weakly active
compounds 1 and 6d–m, and the potent compounds 10f
(98% inhibition at 1 mM) and its inosine analogue (14,
Y=OH, n=6; 68% inhibition at 10 mM), suggests that
the adenosine is recognized by a pocket within the active

P. G. Jagtap et al. / Bioorg. Med. Chem. Lett. 14 (2004) 81–85
85
site, as may be expected to exist for the binding of
NAD+. Rolli and colleagues11 postulated that two
hydrogen-bonding interactions between the adenosine
of NAD+ and the NAD+ binding site of PARP in
their recent report. These are via the primary amino
group and the adjacent ring nitrogen (N1) of adenosine.
This is consistent with the weak PARP activity of ade-
nosine and the weak activity of the other nucleosides
hypoxanthine and inosine, which lack the 6-amino
functionality.12
Damage and Stress Signaling to Cell Death; de Murcia Shall,
S., Ed.; Oxford University Press: New York, 2000; p 177.
5. (a) Southan, G. J.; Szabo, C. Curr. Med. Chem. 2003, 10,
321. (b) Costantino, G.; Macchiarulo, A.; Camaioni, E.;
Pellicciari, R. J. Med. Chem. 2001, 44, 3786. (c) Zhang, J.
In PARP as a Therapeutic Target; Zhang, J., Ed.; CRC:
Boca Raton, FL, 2002; p 239.
6. (a) An inherent problem for the plainer heteroaromatic
compounds is poor solubility in water and other routinely
used organic solvents. Another concern for plainer com-
pounds is the potential for mutagenic and/or carcinogenic
activity as seen for 2-acetamidocarbazole, 2-acylamino-
fluorine, 5-fluoroquinoline and dibenz[a,j] acridine. War-
shawsky, D.; talaska, G.; Xue, W.; Schneider, J. Crit. Rev.
Toxicol. 1996, 26, 213. (b) Heflich, R. H.; Neft, R. E.
Mutat. Res. 1994, 318, 73.
3. Conclusion
A series of novel adenosine-linked analogues of iso-
indolinone have been synthesized and evaluated in an in
vitro PARP-1 inhibition assays. The cyclic derivative of
3-aminobenzamide 4, and many smaller 4-substituted
isoindolinones (6d–k) demonstrated weak PARP-1
inhibition. Introducing benzimidazole and N6-dimethyl-
adenine to the isoindolinone core resulted in a 10-fold
increase in potency. However, adenosine linked via suc-
cinyl and propanyl spacers (10b and 8b) yielded highly
potent compounds. In addition, the piperazine linked
adenosine analogue 8a exhibited the greatest potency
with an IC50 value of 45 nM, which is 650-fold more
potent than the parent isoindolinone core.13 The current
work illustrates the synergism of having two discreet
binding functions on the same molecule.
7. (a) Epp, J. B.; Widlanski, T. S. J. Org. Chem. 1999, 64,
293. (b) Lee, J.; Kang, S. U.; Kang, M. K.; Chun, M. W.;
Jo, Y. J.; Kwak, J. H.; Kim, S. Bioorg. Med. Chem. Lett.
1999, 9, 1365.
8. (a) Ator, M. A.; Bihovsky, R.; Chatterjee, S.; Dunn, D.;
Hudkins, R. I. US Patent application publication,
0028815 A1, 2002. (b) Li, J.-H.; Zhang, J.; Jackson, P. F.;
Maclin, K. M. US Patent 6,306,889 B1, 2001.
9. Cell protection assay: Raw murine macrophages were
treated with test compounds for 15 min prior to the addi-
tion of peroxynitrite (750 mM) for a further 15 min. For the
measurement of PARP activity, the media was removed
and replaced with 0.5 mL HEPES (pH 7.5) containing
3
0.01% digitonin and H-NAD (0.5 mCi mLꢁ1, final con-
centration of NAD+ in buffer is 20 nM/L) for 20 min. The
cells were then scraped from the wells and placed in
Eppendorf tubes containing 200 mL of 50% (w/v) ice-cold
trichloroacetic acid (TCA). The tubes were then placed at
4 ꢀC. After 4 h, the tubes were centrifuged at 1800g for 10
min and the supernatant removed. The pellets were washed
twice with 500 mL ice-cold 5% TCA. The pellets were solu-
bilized in 250 mL NaOH (0.1M) containing 2% SDS over-
night at 37ꢀC and the PARP activity was then determined
by measuring the radioactivity incorporated using a Wallac
scintillation counter. The solubilized protein (250 mL) was
mixed with 5 mL of scintillant (ScintiSafe Plus, Fisher Sci-
entific) before being counted for 10 min. The compounds
described as inhibitors of PARP-1 are capable of permeating
whole cells, a property necessary for therapeutic efficacy.
10. (a) Shall, S. J. Biochem.(Tokyo) 1975, 77, 2. (b) Purnell,
M. R.; Whish, W. J. Biochem. J. 1980, 185, 775.
Acknowledgements
This work was supported by grants from the National
Institutes of Health (R44DK054099 and R44GM058986).
References and notes
1. For recent reviews on the PARP-1 inhibitors see: (a)
Zhang, J., Ed. PARP as a Therapeutic Target; CRC: Boca
Raton, FL, 2002. (b) Szabo, C., Ed. Cell Death: The Role
of PARP; CRC: Boca Raton, FL, 2000. (c) Szabo, C.;
Virag, L. Pharmacol. Rev. 2002, 54, 375.
2. (a) Banasik, M.; Komura, H.; Shimiyama, M.; Ueda, K.
J. Biol. Chem. 1992, 267, 1569. (b) Suto, M. J.; Turner,
W. R.; Arundel-Suto, C. M.; Werbel, L. M.; Sebolt-Leo-
polt, J. S. Anticancer Drug Res. 1991, 7, 107.
3. (a) Suto, M. J.; Turner, W. R.; Werbel, L. M. US Patent
5,177,075, 1993. (b) Showalter, H.D.H. US Patent
5,391,554, 1995.
4. Curtin, N. J.; Golding, B. T.; Griffin, R. J.; Newel, D. R.;
Roberts, M. J.; Srinivasen, S.; White, A. In From DNA
11. Rolli, V.; Ruf, A.; Augustin, A.; Schulz, G.; Menissier-de;
Murcia, J.; de Murcia, G. In From DNA Damage and
Stress Signaling to Cell Death; de Murcia, G., Shall, S.,
Eds.; Oxford University Press: New York, 2000; p 35.
12. Virag, L.; Szabo, C. FASEB J. 2001, 15, 99.
13. Compound 8a (EB-47) was used for in vivo experimenta-
tion. This piperazine linked adenosine derivative exerts
cytoprotective effects in oxidatively damaged cells, and
shows protection in in vivo models of reperfusion injury
and inflammation. The in vivo experimental results will be
published elsewhere.