Identification of tri- and tetracyclic pyrimidinediones as sirtuin inhibitors

Article abstract of DOI:10.1002/cmdc.201000030

(Chemical Equation Presented) Sirtuins are deacetylase proteins that regulate transcription, genome maintenance, longevity, and metabolism. Eight compounds were identified as a novel class of sirtuin inhibitors, with 5, 7, and 9 being the most effective as SIRT1-selective inhibitors. In U937 cells 5 induced apoptosis and displayed antiproliferative effects in Raji, DLD1, and HeLa cancer cells, suggesting potential utility in cancer treatment.

Full text of DOI:10.1002/cmdc.201000030

MED
DOI: 10.1002/cmdc.201000030
Identification of Tri- and Tetracyclic Pyrimidinediones as Sirtuin Inhibitors
Dante Rotili,[a] Domenico Tarantino,[a] Vincenzo Carafa,[b] Ester Lara,[c] Sarah Meade,[d] Giorgia Botta,[a]
Angela Nebbioso,[b] Jçrg Schemies,[e] Manfred Jung,[e] Aleksey G. Kazantsev,[d] Manel Esteller,[f] Mario F. Fraga,[c]
Lucia Altucci,[b] and Antonello Mai*[a]
Sirtuins (SIRT1–7) are a family of NAD+-dependent deacetylas-
es and/or ADP-ribosyltransferases that modify a broad range of
distinct protein substrates with various subcellular localiza-
tions.[1–3] SIRT1 and SIRT2 are mainly localized in the nucleus
(SIRT1) or the cytoplasm (SIRT2), SIRT3–5 are typically mito-
chondrial enzymes, and SIRT6 and SIRT7 are localized in the
cell nucleus and nucleolus, respectively. The requirement of
NAD+ as a cofactor for enzymatic activity, along with the abili-
ty to deacetylate both histones and a variety of non-histone
proteins, suggests that the sirtuins play a crucial role in metab-
olism and energy-dependent transcription.[4–6] SIRT1 has been
shown to deacetylate more than 30 different acetylated sub-
strates such as H4K16, H3K9, p53, forkhead box class O (FoxO),
p300, the DNA repair factor Ku70, peroxisome proliferator-acti-
vated receptor g co-activator 1a (PGC-1a), nuclear factor-kB
(NF-kB), and many others. Therefore, not surprisingly, SIRT1 is
implicated in carcinogenesis, in metabolic and cardiovascular
diseases, and in neurodegeneration.[7–10]
In addition to a protective role described for SIRT1 in DNA
damage, accumulation of mutations, and genomic instability,[11]
SIRT1 has been found to be up-regulated in human lung
cancer, prostate cancer, colon carcinoma, and leukemia.[12]
Moreover, a deleted-in-breast-cancer (DBC) protein has recently
been described as an endogenous inhibitor of SIRT1 to pro-
mote p53 acetylation and activation.[13] Cambinol (1; Figure 1),
a well-known SIRT inhibitor, is able to inactivate the transcrip-
Figure 1. Structures of known and newly designed SIRT inhibitors.
tional repressor BCL6 in Burkitt’s lymphoma cells through ace-
tylation.[14] More recently, tenovins have been reported to de-
[a] Dr. D. Rotili, Dr. D. Tarantino, Dr. G. Botta, Prof. A. Mai
Istituto Pasteur – Fondazione Cenci Bolognetti
Dipartimento di Chimica e Tecnologie del Farmaco
crease tumor growth in vivo as single agents through inhibi-
tion of the protein-deacetylating activities of SIRT1 and
SIRT2.[15] Finally, antisense oligonucleotides directed against
SIRT1 have been shown to induce apoptosis in A549 lung
cancer cells,[16] and SIRT1 knock-down mediated by RNA inter-
ference (siRNA) has been reported to induce growth arrest and
apoptosis in human epithelial cancer cells, whereas normal epi-
thelial cells remained unaffected.[17] This suggests a putative
role of SIRT1 as a novel target for selective killing of cancer
versus non-cancer cells. Note that salermide (2; Figure 1) has
recently been described by our research group as a SIRT inhibi-
tor endowed with cancer-specific, high apoptotic properties in
a panel of cancer cell lines.[18]
Sapienza Universitꢀ di Roma, P.le A. Moro 5, 00185 Roma (Italy)
Fax: (+39)06491491
E-mail: antonello.mai@uniroma1.it
[b] Dr. V. Carafa, Dr. A. Nebbioso, Prof. L. Altucci
Dipartimento di Patologia Generale
Seconda Universitꢀ degli Studi di Napoli
Vico L. De Crecchio 7, 80138 Napoli (Italy)
[c] Dr. E. Lara, Dr. M. F. Fraga
Department of Immunology and Oncology
National Center for Biotechnology (CNB-CSIC)
Cantoblanco, Madrid (Spain)
[d] Dr. S. Meade, Prof. A. G. Kazantsev
Harvard Medical School, Massachusetts General Hospital
CNY114 16th Street, Charlestown, MA 02129 (USA)
Pursuing our search for epigenetic-disorder-modifying
drugs[19–21] and, in particular, on sirtuin modulators,[18,22–24] we
planned to synthesize some cambinol analogues by condensa-
tion of N-phenylbarbituric acid (11) or N-phenyl-2-thiobarbitu-
ric acid (12) with 2-hydroxy-1-naphthaldehyde in dry ethanol,
by using the synthetic methodology reported by Ridi and
Aldo.[25] The expected 5-(2-hydroxynaphthalen-1-ylmethylene)-
[e] Dr. J. Schemies, Prof. M. Jung
Institute of Pharmaceutical Sciences, Albert-Ludwigs-Universitꢁt Freiburg
Albertstr. 25, 79104 Freiburg (Germany)
[f] Prof. M. Esteller
Cancer Epigenetics and Biology Program (PEBC)
Catalan Institute of Oncology (ICO), Barcelona (Spain)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201000030.
674
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 674 – 677

1-phenylbarbituric (3)[25] and -2-thiobarbituric (4) acids
(Figure 1) are also related to some thiobarbiturates described
as SIRT inhibitors by some of us,[26] and show the 2-hydroxy-1-
naphthylmethyl moiety typical of other well-known SIRT inhibi-
tors such as sirtinol,[27] salermide (2),[18] and cambinol (1)[14]
itself, and are considered to be crucial for their inhibiting activ-
ity. However, the two condensation reactions, besides the ex-
pected products 3 and 4, furnished an equimolar mixture of
two dehydration products: the 10-phenyl-7-oxa-8,10-diaza-
benzo[a]anthracene-9,11-dione (5, from barbituric acid), previ-
ously known as obtained from reaction between 11 and 2-hy-
droxy-1-naphthaldehyde at 1608C without solvent,[25] and the
10-phenyl-9-thioxo-9,10-dihydro-7-oxa-8,10-diazabenzo[a]an-
thracen-11-one (6, from thiobarbituric acid) (Figure 1;
Scheme 1).
Figure 2. hrSIRT1 and hrSIRT2 inhibitory activities of compounds 3–10
tested at 50 mm. C=control; CMB=cambinol.
the diazabenzoanthracene skeleton of 5 furnished a further
compound, 7, with high and selective SIRT1 inhibition activity
(inhibition at 50 mm: 87% against SIRT1, 0% against SIRT2).
Conversely, replacement of the tetracyclic diazabenzoanthra-
cene with the tricyclic chromenopyrimidine scaffold (5!8)
abated the anti-SIRT1 activity of the derivative, which was
promptly restored by removal of the phenyl ring at position 3
of the tricyclic ring (compound 9; inhibition at 50 mm: 99%
against SIRT1, 16% against SIRT2). In this assay, compound 10,
which has a benzoyldihydrobenzochromenone moiety, showed
moderate inhibitory activities against both SIRT1 (31% inhibi-
tion) and SIRT2 (20% inhibition). For selected derivatives (5
and 7–10) the IC50 values for hrSIRT1 and hrSIRT2 inhibition
were calculated (Table 1). Cambinol 1 and EX-527,[34] a highly
active SIRT1-selective inhibitor, were included as reference
drugs. Competition analysis performed with compound 5 and
the kinase and SIRT1/2 inhibitor Ro31-8200[35] with increasing
NAD+ concentrations (50, 250, 500, 1000, and 2000 mm) re-
vealed that in contrast to Ro31-8200,[35] compound 5 is not
competitive with respect to NAD+, thus suggesting for 5 a
binding site on SIRT1 different from the adenosine pocket (fig-
ure S1 in the Supporting Information).
Scheme 1. Reagents and conditions: a) 2-hydroxy-1-naphthaldehyde, EtOH
(anhyd), reflux, 1 h.
The four obtained compounds 3–6 were tested at 50 mm
against human recombinant (hr) SIRT1 and SIRT2 produced in
E. coli, using the substrate Fluor de Lys–SIRT1/2 (Biomol) in the
presence of NAD+, followed by the addition of the Developer
II to produce the fluorescent signal; cambinol 1 was used as a
reference drug. Unexpectedly, under these conditions, the bar-
bituric and thiobarbituric acid derivatives 3 and 4 showed very
little inhibitory activity against SIRT1 and SIRT2, whilst the tet-
racyclic compounds 5 and 6 displayed high and selective SIRT1
inhibition (85 and 56% inhibition, respectively; Figure 2).
Encouraged by these results, we designed some simpler an-
alogues of compound 5 to improve its solubility and druglike
properties, and to acquire a preliminary structure–activity rela-
tionship (SAR). In particular, we prepared the 7-oxa-8,10-diaza-
benzo[a]anthracene-9,11-dione 7 as an N-unsubstituted ana-
logue of 5, and the tricyclic 3-phenyl-2H-chromeno[2,3-d]pyri-
midine-2,4(3H)dione 8 and its 3-unsubstituted derivative 9 as
simplified analogues of 5 and 7, respectively. Compounds 7–9
were prepared by following reported procedures.[25,28,29] Finally,
a further simplification of the tetracyclic scaffold of 5 was ob-
tained by removal of the pyrimidinedione portion; thus the 2-
benzoyl-1,2-dihydrobenzo[f]chromen-3-one 10[30] (Figure 1) was
prepared, which is also structurally related to other benzochro-
mene-containing SIRT inhibitors such as splitomicin[31,32] and
HR-73.[33]
Functional tests were performed on tetracyclic and tricyclic
derivatives to ascertain their ability to inhibit SIRT enzymes in a
cellular context. Thus, p53 and a-tubulin acetylation assays
were conducted with MCF7 breast cancer cells for 6 h (p53) or
24 h (a-tubulin) using compounds 1 and 3–10 at 50 mm; the
Table 1. IC50 values of selected tetra- and tricyclic derivatives.
Compd
IC50 [mm][a]
SIRT1
SIRT2
1 (CMB)
EX-527
57.9ꢀ14.7
0.16ꢀ0.01
8.4ꢀ0.2
40.7ꢀ10.4
48.5ꢀ15.2
191.2ꢀ25.6
NI
5
7
8
9
10
4.2ꢀ0.9
125.0ꢀ24.3
5.3ꢀ2.3
NI
243.6ꢀ30.3
178.7ꢀ24.6
The new simplified analogues of 5, compounds 7–10, were
tested at 50 mm against hrSIRT1 and hrSIRT2. As depicted in
Figure 2, removal of the phenyl substituent at position 10 of
112.0ꢀ3.5
[a] NI: no inhibition.
ChemMedChem 2010, 5, 674 – 677
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
675

MED
acetylation levels were determined by Western blot analyses
with specific antibodies. In the acetyl-p53 assay, pretreatment
of MCF7 cells with etoposide was made as reported. In the
acetyl-a-tubulin assay, a sample of SAHA (5 mm) was added as
a positive control. p53 acetylation at Lys382 was taken as a
readout of SIRT1 inhibition, whereas a-tubulin acetylation was
taken as a readout of SIRT2 inhibition. Total ERKs (extracellular
signal-regulated kinases) were used for equal loading. As de-
picted in Figure 3A, in the acetyl-p53 assay compound 5 and,
Figure 3. Determination of A) p53 and B) a-tubulin acetylation levels in
MCF7 cells treated with 1 and 3–10 at 50 mm (6 h for p53; 24 h for a-tubu-
lin). Total ERKs were used for equal loading. SAHA (5 mm) was added in the
acetyl-a-tubulin test as a positive control. ETO=etoposide; SAHA=suberoy-
lanilide hydroxamic acid.
Figure 4. A) Cell-cycle effects and B) apoptosis induction exerted by com-
pounds 3–10 (50 mm, 50 h) in the human U937 leukemia cell line.
to a lesser extent, compounds 7 and 9 gave an appreciable
signal increase relative to control, in agreement with their high
SIRT1 inhibition; the other compounds were less or not effec-
tive in this test. Regarding acetylation of a-tubulin (Figure 3B),
a faint signal was observed with cambinol 1 and compound
10, whereas the other tested compounds did not elicit any
acetyl-a-tubulin signal under these conditions.
tested cell lines starting from a concentration of 15 mm. The
most pronounced effect was observed in the DLD1 colon
cancer cell line; after 72 h incubation at 15 mm, only 49% of
the cells were viable, compared with 536% of the control in
normal proliferation.
In conclusion, we report herein the identification of a novel
class of SIRT inhibitors (compounds 3–10) bearing either a tet-
racyclic or a tricyclic scaffold. Among them, compound 5, its
desphenyl derivative 7, and the tricyclic counterpart of 7, com-
pound 9, showed high and selective SIRT1 inhibition activity.
Western blot analyses were performed in MCF7 cells to deter-
mine the effects of 1 and 3–10 (tested at 50 mm) on the acety-
lation of p53 and a-tubulin, two known respective substrates
for SIRT1 and SIRT2 activity. In the p53 assay, compounds 5, 7,
and 9 increased the levels of acetyl-p53, confirming their SIRT1
inhibition. In the a-tubulin assay, only cambinol 1 and com-
pound 10 gave weak induction of the acetylated protein. Addi-
tionally, compounds 5, 7, and 9 affect the cell cycle of human
U937 leukemia cells, and 5 showed 28.4% of apoptosis when
tested at 50 mm in U937 leukemia cells. Moreover, 5 displayed
high antiproliferative effects in three diverse cancer cell lines
(Raji, DLD1, and HeLa), thus suggesting this compound as a
promising candidate for further optimization toward a drug for
cancer treatment.
The tetra- and tricyclic compounds 3–10 were tested at
50 mm for 50 h in human U937 leukemia cells, along with cam-
binol 1 and EX-527 as reference drugs at the same concentra-
tion, to determine their effects on cell cycle, apoptosis induc-
tion (measured as caspase 3–7 activation), and granulocytic dif-
ferentiation (measured as CD11c expression). In U937 cells,
none of the tested compounds showed significant cytotoxicity
up to 50 mm. Compounds 5, 6, and 10 elicited cell-cycle arrest
at the G1 phase, and FACS analysis for apoptosis indicated that
the majority of compounds tested were able to induce pro-
grammed cell death, with compound 5 being the most effi-
cient (28.4% of apoptosis, Figure 4). In the CD11c test to evalu-
ate the effects of compounds 3–10 on granulocytic differentia-
tion, none of the compounds, including reference compounds
1 and EX-527 (figure S2 in the Supporting Information),
showed induction of differentiation.
Owing to its apoptotic property, compound 5 was also
tested at increasing concentrations for 24, 48, and 72 h on
three cancer cell lines: Raji lymphoblastoid cells derived from
Burkitt’s lymphoma, DLD1 colon carcinoma cells, and HeLa cer-
vical carcinoma cells. Cell proliferation was quantified by 3-
Acknowledgements
(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide
This work was supported by grants from AIRC, FIRB, RETI FIRB,
European Union (Cancerdip; ATLAS), Fondazione Roma, and
R.J.G. and Carmen’s Foundations.
(MTT) assay as previously described.[18] As shown in Figure 5,
compound 5 was able to significantly decrease, in both a
dose- and time-dependent fashion, the proliferation of the
676
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 674 – 677

[10] A. Mai, D. Rotili, S. Valente, A. G. Kazantsev, Curr. Pharm. Des. 2009, 15,
3940–3957.
[11] L. R. Saunders, E. Verdin, Oncogene 2007, 26, 5489–5504.
[12] M. F. Fraga, R. Agrelo, M. Esteller, Ann. NY Acad. Sci. 2007, 1100, 60–74.
[13] W. Zhao, J. P. Kruse, Y. Tang, S. Y. Jung, J. Qin, W. Gu, Nature 2008, 451,
587–590.
[14] B. Heltweg, T. Gatbonton, A. D. Schuler, J. Posakony, H. Li, S. Goehle, R.
Kollipara, R. A. Depinho, Y. Gu, J. A. Simon, A. Bedalov, Cancer Res. 2006,
66, 4368–4377.
[15] S. Lain, J. J. Hollick, J. Campbell, O. D. Staples, M. Higgins, M. Aoubala,
A. McCarthy, V. Appleyard, K. E. Murray, L. Baker, A. Thompson, J. Math-
ers, S. J. Holland, M. J. Stark, G. Pass, J. Woods, D. P. Lane, N. J. West-
wood, Cancer Cell 2008, 13, 454–463.
[16] Y. Sun, D. Sun, F. Li, L. Tian, C. Li, L. Li, R. Lin, S. Wang, Lung Cancer
2007, 58, 21–29.
[17] J. Ford, M. Jiang, J. Milner, Cancer Res. 2005, 65, 10457–10463.
[18] E. Lara, A. Mai, V. Calvanese, L. Altucci, P. Lopez-Nieva, M. L. Martinez-
Chantar, M. Varela-Rey, D. Rotili, A. Nebbioso, S. Ropero, G. Montoya, J.
Oyarzabal, S. Velasco, M. Serrano, M. Witt, A. Villar-Garea, A. Imhof, J. M.
Mato, M. Esteller, M. F. Fraga, Oncogene 2009, 28, 781–791.
[19] A. Mai, S. Massa, R. Pezzi, S. Simeoni, D. Rotili, A. Nebbioso, A. Scogna-
miglio, L. Altucci, P. Loidl, G. Brosch, J. Med. Chem. 2005, 48, 3344–
3353.
[20] A. Mai, D. Rotili, D. Tarantino, P. Ornaghi, F. Tosi, C. Vicidomini, G. Sbar-
della, A. Nebbioso, M. Miceli, L. Altucci, P. Filetici, J. Med. Chem. 2006,
49, 6897–6907.
[21] A. Mai, D. Cheng, M. T. Bedford, S. Valente, A. Nebbioso, A. Perrone, G.
Brosch, G. Sbardella, F. De Bellis, M. Miceli, L. Altucci, J. Med. Chem.
2008, 51, 2279–2290.
[22] A. Mai, S. Massa, S. Lavu, R. Pezzi, S. Simeoni, R. Ragno, F. R. Mariotti, F.
Chiani, G. Camilloni, D. A. Sinclair, J. Med. Chem. 2005, 48, 7789–7795.
[23] A. Mai, S. Valente, S. Meade, V. Carafa, M. Tardugno, A. Nebbioso, A. Gal-
mozzi, N. Mitro, E. De Fabiani, L. Altucci, A. Kazantsev, J. Med. Chem.
2009, 52, 5496–5504.
[24] D. A. Sinclair, A. Mai, (Sirtuin-inhibiting sirtinol analogues for cancer
treatment) WO2007084162, 20060406, 2007.
[25] M. Ridi, G. Aldo, Gazzetta Chimica Italiana 1952, 82, 13–22.
[26] U. Uciechowska, J. Schemies, R. C. Neugebauer, E. M. Huda, M. L.
Schmitt, R. Meier, E. Verdin, M. Jung, W. Sippl, ChemMedChem 2008, 3,
1965–1976.
[27] C. M. Grozinger, E. D. Chao, H. E. Blackwell, D. Moazed, S. L. Schreiber, J.
Biol. Chem. 2001, 276, 38837–38843.
[28] F. Yoneda, R. Hirayama, M. Yamashita, J. Heterocycl. Chem. 1982, 19,
301–304.
[29] X. Chen, K. Tanaka, F. Yoneda, Chem. Pharm. Bull. 1990, 38, 307–311.
[30] F. Medda, R. J. Russell, M. Higgins, A. R. McCarthy, J. Campbell, A. M.
Slawin, D. P. Lane, S. Lain, N. J. Westwood, J. Med. Chem. 2009, 52,
2673–2682.
[31] A. Bedalov, T. Gatbonton, W. P. Irvine, D. E. Gottschling, J. A. Simon, Proc.
Natl. Acad. Sci. USA 2001, 98, 15113–15118.
^
Figure 5. Dose-dependent antiproliferative activity of compound 5 at 0 ( ),
~
&
5 ( ), 15 ( ), 25 (ꢁ), and 50 mm ( ) toward a panel of cancer cell lines: A) Raji
lymphoblastoid cells derived from Burkitt’s lymphoma, B) DLD1 colon carci-
noma cells, and C) HeLa cervical carcinoma cells.
*
[32] R. C. Neugebauer, U. Uchiechowska, R. Meier, H. Hruby, V. Valkov, E.
Verdin, W. Sippl, M. Jung, J. Med. Chem. 2008, 51, 1203–1213.
[33] S. Pagans, A. Pedal, B. J. North, K. Kaehlcke, B. L. Marshall, A. Dorr, C.
Hetzer-Egger, P. Henklein, R. Frye, M. W. McBurney, H. Hruby, M. Jung, E.
Verdin, M. Ott, PLoS Biol. 2005, 3, e41.
Keywords: antiproliferation
diazabenzoanthracene · inhibitors · sirtuins
·
apoptosis
·
[34] A. D. Napper, J. Hixon, T. McDonagh, K. Keavey, J. F. Pons, J. Barker, W. T.
Yau, P. Amouzegh, A. Flegg, E. Hamelin, R. J. Thomas, M. Kates, S. Jones,
M. A. Navia, J. O. Saunders, P. S. DiStefano, R. Curtis, J. Med. Chem. 2005,
48, 8045–8054.
[35] J. Trapp, A. Jochum, R. Meier, L. Saunders, B. Marshall, C. Kunick, E.
Verdin, P. Goekjian, W. Sippl, M. Jung, J. Med. Chem. 2006, 49, 7307–
7316.
[1] J. M. Denu, Curr. Opin. Chem. Biol. 2005, 9, 431–440.
[2] S. Michan, D. Sinclair, Biochem. J. 2007, 404, 1–13.
[3] D. M. Taylor, M. M. Maxwell, R. Luthi-Carter, A. G. Kazantsev, Cell. Mol.
Life Sci. 2008, 65, 4000–4018.
[4] J. C. Milne, J. M. Denu, Curr Opin Chem. Biol. 2008, 12, 11–17.
[5] S. Lavu, O. Boss, P. J. Elliott, P. D. Lambert, Nat. Rev. Drug Discovery 2008,
7, 841–853.
[6] S. Pillarisetti, Recent Pat. Cardiovasc. Drug Discovery 2008, 3, 156–164.
[7] B. Zschoernig, U. Mahlknecht, Biochem. Biophys. Res. Commun. 2008,
376, 251–255.
Received: January 22, 2010
Revised: February 18, 2010
Published online on April 13, 2010
[8] B. L. Tang, C. E. Chua, Mol. Aspects Med. 2008, 29, 187–200.
[9] T. F. Outeiro, O. Marques, A. Kazantsev, Biochim. Biophys. Acta Mol. Basis
Dis. 2008, 1782, 363–369.
ChemMedChem 2010, 5, 674 – 677
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
677