Novel pyrimidopyrimidine derivatives for inhibition of cellular proliferation and motility induced by h-prune in breast cancer

Article abstract of DOI:10.1016/j.ejmech.2012.08.020

The human (h)-prune protein is a member of the DHH protein superfamily and it has a cAMP phosphodiesterase activity. Its overexpression in breast, colorectal and gastric cancers correlates with depth of invasion and a high degree of lymph-node metastasis. One mechanism by which h-prune stimulates cell motility and metastasis processes is through its phosphodiesterase activity, which can be suppressed by dipyridamole, a pyrimido[5,4-d]pyrimidine analogue. To obtain new and more potent agents that have high specificity towards inhibition of this h-prune activity, we followed structure-activity-relationship methodologies starting from dipyridamole and synthesised eight new pyrimido-pyrimidine derivatives. We analysed these newly generated compounds for specificity towards h-prune activities in vitro in cellular models using scintillation proximity assay for cAMP-PDE activity, cell index in cell proliferation assays and transwell methodology for two-dimensional cell migration in a top-down strategy of selection. Our findings show that two pyrimido[5,4-d]pyrimidine compounds are more effective than dipyridamole in two highly metastatic cellular models of breast cancer in vitro. Future studies will assess their therapeutic effectiveness against breast and other cancers where there is over-expression of h-prune, and in ad-hoc, proof of concept, animal models.

Full text of DOI:10.1016/j.ejmech.2012.08.020

European Journal of Medicinal Chemistry 57 (2012) 41e50
Contents lists available at SciVerse ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Novel pyrimidopyrimidine derivatives for inhibition of cellular proliferation and
motility induced by h-prune in breast cancer
,
,
c
,
,
Antonella Virgilio a 1, Daniela Spano b 1, Veronica Esposito a, Valeria Di Dato b c, Giuseppe Citarella a,
Natascia Marino c, Veronica Maffia c, Daniela De Martino b c, Pasqualino De Antonellis b c, Aldo Galeone a,
,
,
Massimo Zollo b
,c,
*
a Dipartimento di Chimica delle Sostanze Naturali, Università degli Studi di Napoli Federico II, Via Domenico Montesano 49, 80131 Naples, Italy
b Dipartimento di Biochimica e Biotecnologie Mediche, Università degli Studi di Napoli Federico II, Via Sergio Pansini 5, 80131 Naples, Italy
c CEINGE, Centro di Ingegneria Genetica e Biotecnologie Avanzate, Via Gaetano Salvatore 486, 80145 Naples, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
The human (h)-prune protein is a member of the DHH protein superfamily and it has a cAMP phos-
phodiesterase activity. Its overexpression in breast, colorectal and gastric cancers correlates with depth
of invasion and a high degree of lymph-node metastasis. One mechanism by which h-prune stimulates
cell motility and metastasis processes is through its phosphodiesterase activity, which can be suppressed
by dipyridamole, a pyrimido[5,4-d]pyrimidine analogue. To obtain new and more potent agents that have
high specificity towards inhibition of this h-prune activity, we followed structureeactivity-relationship
methodologies starting from dipyridamole and synthesised eight new pyrimidoepyrimidine deriva-
tives. We analysed these newly generated compounds for specificity towards h-prune activities in vitro in
cellular models using scintillation proximity assay for cAMP-PDE activity, cell index in cell proliferation
assays and transwell methodology for two-dimensional cell migration in a top-down strategy of selec-
tion. Our findings show that two pyrimido[5,4-d]pyrimidine compounds are more effective than
dipyridamole in two highly metastatic cellular models of breast cancer in vitro. Future studies will assess
their therapeutic effectiveness against breast and other cancers where there is over-expression of h-
prune, and in ad-hoc, proof of concept, animal models.
Received 9 February 2012
Received in revised form
29 June 2012
Accepted 13 August 2012
Available online 23 August 2012
Keywords:
h-Prune
Breast cancer
cAMP
Phosphodiesterase
Cellular motility
Ó 2012 Elsevier Masson SAS. All rights reserved.
1. Introduction
diagnosed disease present with advanced or metastatic breast
cancer (MBC), with some 40% of patients who initially present with
Breast cancer is the most common cancer among women and is
a primary cause of cancer death worldwide. Breast cancer is
a complex and heterogeneous disease as it comprises multiple
tumour entities, that are associated with distinctive histological
patterns and different biological features and clinical behaviour.
Furthermore, evidence from gene expression microarrays studies
has suggested the presence of multiple molecular subtypes of
breast cancer that might respond in different ways to different
therapies [1]. In breast cancer, metastatic spread is responsible for
virtually all of the deaths. About 6% of patients with newly
localised disease eventually progressing to MBC. MBC is a hetero-
geneous disease that has a clinical behaviour that can be difficult to
predict [2]. The choice of treatment is based on patient character-
istics (e.g., age), the presence of comorbidities, and/or the previous
treatments. In addition, physicians rely on prognostic and predic-
tive targets for therapy, the most established of which are the
expression of the oestrogen receptor (ER), progesterone receptor
(PR), and human epidermal growth factor receptor-2 (HERB-2) by
the tumour tissue. These characteristics help in determining the
most effective therapy for an individual patient with MBC [2,3].
Unfortunately, the heterogeneity of the disease and the vari-
ability in the individual presentation make the selection of widely
accepted patient-specific treatment guidelines difficult. For these
reasons, clinicians must determine which therapeutic intervention
might be most appropriate for each patient, based on individual
aspects of the patient clinical profile, and the number and sites of
metastases, vital organ involvement and previous disease-free
survival interval [3]. Thus, new therapeutic interventions have
Abbreviations: MBC, metastatic breast cancer; PDE, phosphodiesterase; DP,
dipyridamole;
CI,
cell
index;
MTS,
3-(4,5-dimethylthiazol-2-yl)-5-(3-
carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium.
* Corresponding author. CEINGE, Biotecnologie Avanzate SCARL, Via Gaetano
Salvatore 486, 80145 Napoli, Italy. Tel.: þ39 081 3737 875; fax: þ39 081 3737 711.
E-mail address: massimo.zollo@unina.it (M. Zollo).
1
These two authors contributed equally to the work.
0223-5234/$ e see front matter Ó 2012 Elsevier Masson SAS. All rights reserved.
http://dx.doi.org/10.1016/j.ejmech.2012.08.020

42
A. Virgilio et al. / European Journal of Medicinal Chemistry 57 (2012) 41e50
the need for new classes of inhibitors [4]. However, despite
significant progress in the detection and treatment of breast cancer,
MBC remains incurable. The treatment goals for MBC are therefore
focussed on prolonging patient survival and providing palliative
control of symptoms.
Metastasis is a highly complex molecular and cellular process,
during which cancer cells escape from the original tumour tissue,
invade the surrounding tissues of the primary tumour, penetrate
the lymphatic system or a blood vessel through which they travel to
distal sites, extravasate, and then colonize a second organ [5,6].
Increased invasiveness of cancer cells is a critical step in tumour
metastasis and this requires many intracellular and extracellular
changes; e.g., cancer cells loosen their adhesion to neighbouring
cells and the extracellular matrix, degrade adjacent tissues, and
show increased motility [7].
In previous studies, the enzymatic activity of the human (h)-
prune protein was shown to be involved in metastasis formation in
breast cancer [8]. The h-prune protein is a member of the DHH
(AspeHiseHis) protein superfamily, and it has a phosphodiesterase
(PDE) activity, with a preferential affinity for cAMP over cGMP [8],
and an exopolyphosphatase activity, with a preferential affinity for
short-chain inorganic polyphosphates over long-chain inorganic
polyphosphates [9]. Over-expression of h-prune in the MDA-MB-
435 breast cancer cell line promotes cell motility. Consistent with
these observations, over-expression of h-prune in breast cancer is
correlated with cancer progression and aggressiveness [10,11].
These pro-motility effects of h-prune seen in vitro are also signifi-
cantly associated with lymph-node status (N2eN3) and metastasis
formation (M1) in vivo, thus indicating that h-prune is a marker of
advanced-stage disease status in breast carcinoma [11]. Interest-
ingly, h-prune PDE activity is enhanced by interaction with nm23-
H1, and a direct correlation between increased h-prune cAMP-PDE
activity and cellular motility has been shown in the breast cancer
model [8].
The new pyrimido[5,4-d]pyrimidine derivatives have been
designed taking into account the biochemical properties of DP
(whose chemical structure is shown in Fig. 1A) and other similar
derivatives (DP analogues), properly developed as inhibitors of the
nucleoside transport [12,13]. Although the usefulness of DP as
inhibitor of the nucleoside transport has been confirmed in in vitro
experiments, its efficacy and that of some of its derivatives is
dramatically reduced in presence of the a1-acid glycoprotein (an
acute phase protein, whose plasma levels can be elevated in cancer
patients) due to the ability of DP to bind avidly to it [12,13]. Most
importantly, in a structureeactivity-relationships study involving
96 compounds [13], authors observed that the influence of the a1
-
acid glycoprotein on the activity of most of the molecules investi-
gated was less pronounced than that observed for DP, in
compounds bearing a substituted benzylamino group at the pyr-
imido[5,4-d]pyrimidine 4,8-positions. Taking into consideration
these data, we have designed 8 new compounds (Fig. 1AeB) that
can be considered derivatives of compound 1 characterized by the
same groups in 2,6-positions as DP and 4-methoxybenzylamino
groups in 4,8-positions that, in several compounds, have proven
to impair the binding to a1-acid glycoprotein.
Previous studies showed DP inhibits the ability of h-prune
purified protein to hydrolyze cAMP with an IC50 of 0.78 mM [8].
Therefore, firstly we assessed the ability of compound 1 to inhibit
the cAMP-PDE activity of h-prune purified protein and determined
its IC50 value using a scintillation proximity assay. As shown in
Supplementary Figure 1, compound 1 inhibited the cAMP-PDE
activity of h-prune purified protein with an IC50 measured value
of 0.21 mM. This value is lower if compared to the DP IC50 value, thus
indicating compound 1 is more efficient than DP in inhibiting h-
prune cAMP-PDE activity and retains a higher specificity towards h-
prune cAMP-PDE activity.
The synthesis of the 8 new compounds derived from the
compound 1 was performed by following a two steps literature
procedure starting from the tetrachloropyrimidopyrimidine [12,13]
(Fig. 1C). In this series, most of the substituents R maintain the
same number of bonds between the nitrogen atom and a hydroxyl
group, as for DP and compound 1. However, in this case, for several
substituents, the flexibility of this part of the molecule has been
decreased by introducing cyclic hydroxylamine moieties in order to
limit the conformational freedom of the eOH group, in the hypoth-
esis that it is essential for the interaction of the molecule with h-
prune. Furthermore, when chiral amines were utilized, pure enan-
tiomers were preferentially used. To evaluate the abilityof these new
molecules (Fig. 1AeB) to inhibit the h-prune protein directly in
a cellular environment, h-prune PDE activity assays were performed
on a protein extract from cells over-expressing h-prune and treated
with these compounds. For this assay, a stable clone over-expressing
h-prune cDNA was expressly created and its cellular properties were
determined. The MDA435-C100 cell line was used to generate this
stable clone that over-expressed h-prune, following a previously
described methodologyof transfection and stable clone selection [8].
This stable breast cancer clone was then assessed for expression of h-
prune by western blotting using an anti-h-prune antibody (Fig. 2A).
The parental cell line containing the empty vector plasmid was used
ascontrol. Proteinextractsfromthestable cloneline(MDA435-c100-
To identify the physiological role of h-prune, several known
selective PDE inhibitors were tested for their ability to affect h-
prune cAMP-PDE activity. Among these, dipyridamole (DP)
emerged as the most active in inhibiting the cAMP-PDE activity of
h-prune purified protein, expressed by Baculovirus Expression
system, with an IC50 of 0.78 mM [8]. To develop a therapeutic pro-
gramme aimed at identifying new, more potent and selective,
compounds that can inhibit this h-prune activity, we undertook the
preparation of new molecules based on the pyrimido[5,4-d]
pyrimidine aromatic heterocycle characteristic of DP. We per-
formed various enzymatic and biological assays to assess the effi-
cacy of the new pyrimido[5,4-d]pyrimidine compounds towards
the impairing of the h-prune PDE activity and cell migration and
proliferation, with the final goal being to identify a candidate
molecule for in vivo testing for anti-cancer properties against
metastatic breast cancer.
2. Results
2.1. Development and selection of pyrimido[5,4-d]pyrimidine
derivatives that inhibit h-prune cAMP-PDE activity
As there is no detailed information concerning the full structure
of the h-prune protein and the site of action of DP that would allow
us to use molecular modelling techniques, we decided to prepare
a series of novel eight pyrimido[5,4-d]pyrimidine derivatives and,
then, test them by suitable biological assays considering the
potential of inhibition of the two main functions of h-prune
described: (i) increased cAMP-PDE activity in breast cancer cells
due to over-expression of h-prune; (ii) enhanced cell migration due
to over-expression of h-prune [8].
h-prune) were assessed after treatment with 1
compounds, as described in the Material and Methods.
mM of these
As a useful comparison, the PDE activity of DP and compound 1,
from which the series of compounds originated, was also evaluated
using the same assay. As discussed above, this experiment was
performed using whole protein extracts of cell line MDA435 over-
expressing h-prune protein and not using the h-prune purified
protein as previously carried out in D’Angelo and collaborators [8].
For these reasons the dipyridamole induced-inhibition of h-prune

A. Virgilio et al. / European Journal of Medicinal Chemistry 57 (2012) 41e50
43
Fig. 1. Chemical structures (A and B) and synthetic strategy (C) of the pyrimido[5,4-d]pyrimidine derivatives investigated. Reagents and conditions: (i) 4-methoxybenzylamine,
K2CO3, THF (solvent), r.t.; (ii) appropriate 1ꢁ or 2ꢁ amine, 100e150 ꢁC.
cAMP-PDE activity in a cellular environment was found lower if
compared to previously data obtained on h-prune purified protein
[8]. The results point to an increased activity for all compounds 2e9
in comparison with DP and their precursor 1 (Fig. 2B).
two biological functions (cellular proliferation and cellular motility)
related to h-prune activity in tumorigenic cells to further screen the
newly generated compounds. For this reason we decided to follow
through the analyses above presented a next series of assays to
select the best compound using structureeactivity-relationship
approach. As a crucial feature of cancer cells is their high prolifer-
ation rate, we assessed the inhibition of cell proliferation by these
new compounds. The cells used as cellular model were MDA-MB-
231T, which express high level of h-prune (Fig. 3A). Based on the
cAMP-PDE activity assays results, we focused our attention on
2.2. Inhibition of cell proliferation by the newly generated pyrimido
[5,4-d]pyrimidine derivatives
To further study the efficacy of our here identified compounds
we decided to follow a “Top-Down” strategy. We have used the next
Fig. 2. Efficacy of pyrimido[5,4-d]pyrimidine derivatives in inhibiting h-prune cAMP-PDE activity. (A) Characterization of h-prune over-expressing MDA435-C100 clone. Immu-
noblotting analysis of h-prune protein expression was performed on total protein extracts from MDA435-C100 empty vector and h-prune over-expressing clones. -actin was used
b
as a control for equal loading. (B) cAMP-PDE activity assay was performed on protein extracts from MDA435-C100 h-prune over-expressing clone treated with vehicle, dipyridamole
(DP) or pyrimido[5,4-d]pyrimidine derivatives. Data are means ꢀ standard errors (SE). A representative experiment of two performed with similar results is shown.

44
A. Virgilio et al. / European Journal of Medicinal Chemistry 57 (2012) 41e50
Fig. 3. Efficacy of pyrimido[5,4-d]pyrimidine derivatives in inhibiting cell proliferation. (A) Immunoblotting analysis of h-prune protein expression was performed on total protein
extracts from MDA-MB-231T cells. -actin was used as a control for equal loading. (B) Dynamic monitoring of MDA-MB-231T cell proliferation in response to 50 M DP and other
compounds. In the graph the black line shows the time of drug addition at which the CI values were normalised. A representative experiment of two performed with similar results
is shown. Data are means ꢀ standard deviation (SD). (C) Enlargement of the dynamic monitoring of MDA-MB-231T cell proliferation in response to 50 M compound 8. The
inflection point of the curve indicates the time where 50 percent inhibition of proliferation is achieved by using 50 M compound 8.
b
m
m
m
compounds 5, 6, 8 and 9 and evaluated their ability to impair cell
proliferation in the MDA-MB-231T proliferation assay in real time
(xCelligence System).
the R substituent (Fig. 1B). Although this structural difference has
a negligible effect on the inhibition of h-prune cAMP-PDE activity
by compounds 8 and 9 (Fig. 2), the different spatial orientation of
the eOH group appears to have an important role in their inhibi-
tion of cell proliferation, with compound 8 more efficient than
compound 9. For a direct comparison of the proliferation inhibition
of these compounds, we compared the CI after 24 h of treatment.
As shown in Table 1, compound 8 impaired cell proliferation more
effectively than the other compounds. The time-course analysis of
Previous studies showed that 20
mg/ml DP (corresponding to
a final of 40 M concentration) inhibits cell growth [14]. Addi-
m
tionally in our recently published data we found that the concen-
tration of DP able to significantly impair cell proliferation, in breast
cancer cell lines, corresponds to 50
assay DP and the pyrimido[5,4-d]pyrimidine derivatives, including
compound 1, at the concentration of 50 M. As shown in Fig. 3B,
mM [15]. Therefore we decide to
m
cell proliferation in presence of 50 mM compound 8 showed that
MDA-MB-231T cells treated with DP and the other pyrimido[5,4-d]
pyrimidine derivatives did not show any significant changes in the
cell index (CI), as a measure of cell proliferation, during the first 5 h
of their addition. After this time, there was a significant lower
increase in the CI values of treated cells compared to vehicle
(DMSO)-treated cells. These data show that compound 1 blocked
cell proliferation more effectively than DP at the same molar
concentration (Fig. 3B). Inhibition of cell proliferation by
compound 5 was comparable to that of DP during the first 27 h of
treatment, although at longer times of treatment compound 5
showed greater inhibition than DP (Fig. 3B). For compounds 6 and
9, inhibition of cell proliferation remained generally similar to that
of DP throughout the assay period. In contrast, compound 8 was
much more efficient than DP in inhibiting cell proliferation
throughout the remaining assay period (Fig. 3B). Comparison of the
effects on cell proliferation of the most efficient compounds with
inhibition of the h-prune cAMP-PDE activity, showed that the
inhibition for compound 8 in the cell proliferation assay increases
relative to DP from 9 h to around 45 h of incubation, as there is no
further increase in CI seen with compound 8. In contrast, in
comparison with DP, the smaller anti-proliferative effects of
compound 9 are lost over the next 30 h of treatment (Fig. 3B). A
comparison of the chemical structure of compounds 8 and 9 shows
that they differ only in the spatial orientation of the eOH group in
the 50% inhibition of proliferation, measured as CI value, is ach-
ieved 7 h and 20 min after drug addition (which was performed at
1 h and 27 min from cell proliferation experiment start) (Fig. 3C
and Table S1). Furthermore, only the compounds 1, 8 and 9 showed
a statistically significant decrease in CI values at 24 h treatment
compared to DP (Table 1). In vitro proliferation assays in presence
of two different concentrations of compounds 8 and 9 were also
performed to evaluate the doseeresponse effects on inhibiting cell
proliferation. As showed in the upper and bottom panels of Fig. 4A,
the increase of compounds concentration caused an increase of
their proliferation inhibitory effects. We compared the CI curves of
Table 1
Cell index values from the in real time cell proliferation assays after 24 h of
treatment.
Drugs
CI 24 h after
SE
p value of
drugs addition
compound vs DP
DMSO
DP
1
5
6
1.777
1.472
1.185
1.407
1.452
0.770
1.239
0.000
0.008
0.028
0.030
0.025
0.016
0.035
e
e
0.0006
0.1336
0.5496
<0.0001
0.0026
8
9

A. Virgilio et al. / European Journal of Medicinal Chemistry 57 (2012) 41e50
45
Fig. 4. Doseeresponse curves of compounds 8 and 9. (A) The graphs show the dynamic monitoring of MDA-MB-231T cell proliferation in response to the DMSO control (0) and
50 M and 100 M of compounds 8 (upper panel) and 9 (bottom panel). Black marker line is the time of compound addition to which CI values were normalised. Data are
m
m
means ꢀ SD. For each panel, a representative experiment of two independent in vitro proliferation assays performed with similar results is shown. Middle panel is an enlargement of
the dynamic monitoring assays shown in upper and bottom panels within the time window of 6e10 h. (B) Representative end-point cell proliferation assay (of two experiments
performed with similar results) with MDA-MB-231T cells treated with DMSO (control) or compound 8 (as indicated). Data are means ꢀ SE.

46
A. Virgilio et al. / European Journal of Medicinal Chemistry 57 (2012) 41e50
compounds 8 and 9, at 50 and 100
mM concentrations and nor-
malised at time of drug addition, within the time window of 6e
10 h (see middle panel of Fig. 4A). The corresponding single
curves within this time window are shown in Supplementary
Figure 2. This comparison showed that the inhibition of cell
proliferation induced by compound 9 at 100
mM concentration was
comparable to that obtained by compound 8 at 50
mM concentra-
tion (Fig. 4A middle panel). These data overall are indicating that:
(i) the compound 8 is more potent than compound 9 on impairing
cell proliferation and (ii) the compound 9 needs higher concen-
trations to achieve the same effect on cellular proliferation in
comparison of compound 8. For this reason the higher potency of
compound 8 may explain its efficacy on inhibiting cell proliferation
over compound 9. Furthermore, as shown in Table 2, evaluating
over the time cell proliferation inhibition (at the time of 15, 20, 30
and 40 h), we found that the 100
compound 9 was not able to inhibit the cell proliferation in
comparison of compound 8 used instead at 50 M. It should be
mM concentration used for
m
noted that while compound 8, during the time of in real time
proliferation assay, is able to impair the cellular proliferation
(measured as CI values) (see data at 15 h 0.73e0.52; at 20 h 0.74e
0.48; at 30 h 0.77e0.37; at 40 h 0.68e0.22), compound 9 is per-
forming with less efficacy, thus showing a potential high grade of
degradation within the cell, being the CI values increased within
the time of treatment (see data in Table 2). Indeed a sign of
potential degradation into the cell was observed (see data at 15 h:
0.94e0.76; at 20 h: 1.04e0.82; at 30 h: 1.42e1.01; at 40 h: 1.70e
1.09). Overall these data show lower potency and efficacy of
compound 9 compared to compound 8, and allow us to conclude
that this latest is more effective and potent than compound 9 on
inhibiting cell proliferation. Then to further confirm this dose-
dependent inhibition of cell proliferation induced by the
compound 8, we performed end-point cell proliferation assays
(MTS assays) in a range of compound 8 concentrations. As previ-
ously already shown, this assay confirmed that the compound 8
inhibited the cellular proliferation of MDA-MB-231T breast cancer
cells in a dose-dependent inhibition (Fig. 4B).
Fig. 5. Efficacy of pyrimido[5,4-d]pyrimidine derivatives in inhibiting cell motility. (A)
Motility assays were performed on MDA-MB-231T cells treated with vehicle (DMSO),
DP or the other compounds at a concentration of 8
m
M. Data are means ꢀ SE. A
representative experiment of two independent cell motility assays performed with
similar results is shown. (B) M-prune endogenous expression in 4T1-Luc cells.
Immunoblotting analysis of m-prune protein expression performed on total protein
extracts from 4T1-Luc cells.
assays were performed on 4T1-Luc cells treated with vehicle (DMSO), compounds 8
and 9 at a concentration of 8
b-actin was used as a control for equal loading. (C) Motility
m
M. Data are means ꢀ SE. A representative experiment of
two independent cell motility assays performed with similar results is shown.
2.3. Inhibition of cell motility by the newly generated pyrimido[5,4-d]
pyrimidine derivatives
motility was compound 9. In this case, the different spatial orien-
tation of the eOH group in compounds 8 and 9 (see Fig. 1B)
appeared to have an important role against cell motility in this
MDA-MB-231T breast cancer model (see Fig. 5A, in the “cell 2D
migration assay”).
It should be noted here that the cell motility process is affected
by cytoskeleton rearrangements due to assembly of F-actin and
its polymerization and depolarization in focal adhesions. This
function is regulated by h-prune in a context of the interactions
Previous reports have shown that h-prune over-expression in
the MDA-MB-435 breast cancer cell line promotes cell motility [8]
and is correlated with breast cancer progression and aggressive-
ness [10,11]. To investigate the ability of these new pyrimido[5,4-d]
pyrimidine derivatives to inhibit cell motility, we performed in vitro
cell motility assay on the MDA-MB-231T metastatic breast cancer
cells [16]. We analysed inhibition of cell motility by the four best
compounds in the PDE-activity assay, namely 5, 6, 8, 9 and 1 and DP
and regulation of additional protein partners (e.g. GSK-3b) and by
the specific cAMP-PDE activity of h-prune [17]. However, these
are not the only factors that can influence cell motility. This
would thus explain the different relative behaviours of
compounds 5 and 6 in those PDE assays (Fig. 2B) and cell motility
assays (Fig. 5A).
for comparison at a concentration of 8 mM. As shown in Fig. 5A, the
inhibition of cellular motility of the compound 1 was slightly
greater with respect to DP. However, among the new derivatives
generated from compound 1, the best one for the inhibition of cell
These data indicate that compound 9 can impair cell motility in
this two-dimensional (2D) cell migration assay (MDA-MB-231T
Table 2
Cell index values from the in real time cell proliferation assays in presence of
compounds 8 and 9 at the indicated time points.
breast cancer model) with
a greater efficacy compared to
compound 8, which was similar to DP to impair cell motility. To
further validate this, we evaluated the ability of these compounds
to reduce cell motility in an additional model of breast cancer in
mouse, the 4T1-Luc cell metastatic model [18], which express
a sustained high level of the murine prune homologue (Fig. 5B). As
shown in Fig. 5C, both compounds 8 and 9 at a concentration of
Time-interval (h) 50
m
M
100
m
M
50
m
M
100 mM
compound 8 compound 8 compound 9 compound 9
CI SE CI SE CI SE CI SE
15
20
30
40
0.73 0.016 0.52 0.024 0.94 0.013 0.76 0.006
0.74 0.011 0.48 0.024 1.04 0.022 0.82 0.010
0.77 0.021 0.37 0.020 1.42 0.042 1.01 0.021
0.68 0.029 0.22 0.008 1.70 0.050 1.09 0.029
8
mM inhibited this 4T1-Luc two-dimensional (2D) migration ability
compared to DMSO.

A. Virgilio et al. / European Journal of Medicinal Chemistry 57 (2012) 41e50
47
3. Discussion
nucleosides or bases constitutes an intrinsic resistance mechanism
to anti-metabolite inhibitors of de novo purine or pyrimidine
biosynthesis [26,27]. The pyrimidoepyrimidine cardiovascular
agent DP inhibits equilibrative nucleoside transporters [28e30].
Several studies have shown that DP significantly increases the
cytotoxic and anti-tumour activity of a variety of inhibitors of de
novo nucleotide synthesis by blocking the uptake of nucleosides for
salvage [31e39]. Moreover, by blocking nucleoside efflux, DP can
also exacerbate antifolate-induced nucleotide pool imbalances and
increase the cytotoxicities of these inhibitors. These anti-tumour
activities and all the other properties of these pyrimido[5,4-d]
pyrimidines derivatives will be investigated in our future studies.
Recently, some novel pyrimido[5,4-d]pyrimidines were effi-
ciently synthesized and evaluated for antibacterial activity against
Mycobacterium tuberculosis strain H37Rv [40]. Additionally, a patent
deposited in USA (US 2003229051, 2003) shows pyrimido[5,4-d]
pyrimidines derivatives as irreversible inhibitors of tyrosine
kinases and their potential use in cancer and in the treatment of
inflammatory diseases, such atherosclerosis, restenosis, endome-
triosis and psoriasis. These data make our newly discovered anti-h-
prune derivatives very attractive for testing of their anti-
inflammatory properties. Future studies will address these topics.
Our structureeactivity-relationship approach here has been
valuable for the identification of new molecules that can impair h-
prune cAMP-PDE activity and the increased cellular motility in
these highly metastatic models of breast cancer. Through the
combination of these assays, we have identified two compounds
that show increases on their anti-cancer properties, as measured
through inhibition of cellular proliferation and motility of these
highly metastatic tumours cell lines. Future studies will be directed
to the characterization of the site of interaction between h-prune
and these newly generated inhibitors. We envision the use of
nuclear magnetic resonance (NMR) studies for the identification of
the small-molecule binding site and for the optimization of the
inhibitor properties [19,20] and thus following studies on RTKs
c-Met inhibitors as reviewed by Accornero et al., 2010 [21].
Thus, future NMR studies performed on the pyrimido[5,4-d]
pyrimidine compounds generated in the present study should
provide insights into their interaction with the h-prune binding
pocket, and thus will allow the design and the synthesis of further
analogues to provide more potent compounds around those
scaffolds. In particular, the identification of the h-prune region that
interacts with these new pyrimido[5,4-d]pyrimidine compounds
will provide further insights into the molecular mechanisms
underlying their ability to impair h-prune function in cancer.
It is known that h-prune regulates cell motility by two different
means of action: through its cAMP-PDE activity and through its
interactions with various protein partners, which include nm23-H1
4. Conclusion
In this study we identified new molecules that can impair
h-prune cAMP-PDE, cellular proliferation and increased cellular
motility in breast cancer metastatic models. Both the newly iden-
tified compounds 8 and 9 are useful targets for future studies that
will mainly address their properties as anti-cancer (anti-metastatic)
and potentially being new anti-inflammatory candidate molecules
as being h-prune involved in those several mechanism functionally
related.
[8,22], GSK-3
b
[17], paxillin [17], vinculin [17], and ASAP1 [23]. H-
, paxillin, vinculin and ASAP1 are all localised within
prune, GSK-3
b
focal adhesions. In particular, a role for h-prune and GSK-3b in the
regulation of the disassembly of focal adhesions, to promote cell
migration, has been demonstrated [17]. However, if all of these
proteins form a unique molecular complex or whether the h-prune
interaction with its protein partners is mutually exclusive has not
yet been determined. The h-prune region involved in the interac-
tion with its protein partners was previously determined only for
5. Experimental protocols
5.1. Synthetic procedures and materials and methods
nm23-H1 and GSK-3
b [24]; studies to resolve the region of
proteineprotein binding site(s) of h-prune with its other protein
partners are ongoing in our laboratory. Moreover, the role of each
protein interaction for the promotion of cell motility has not yet
been determined. The ability of DP to impair h-prune-induced cell
motility has been ascribed to its properties as an inhibitor of the
cAMP-PDE activity of h-prune [8], whereas DP did not affect the
subcellular localisation of h-prune, indicating that the PDE activity
is not required for the localisation of h-prune to focal adhesions
[17]. In the future, we will determine which function of h-prune is
abrogated by these newly generated pyrimido[5,4-d]pyrimidine
compounds, and if these compounds can impair protein function
through the binding of h-prune with its other protein partners in
the focal adhesions compartments.
Questions are raised at this stage about the anti-proliferative
properties of those newly identified compounds. We have already
shown an anti-proliferative effect of DP [8], as have Curtin and
colleagues [25]. At this time, we cannot be certain whether the
molecular mechanism underlying those biological effects of DP is
the same as these newly identified small molecules. In particular,
these molecules can mimic metabolites (purine and pyrimidines)
and might have effects on DNA synthesis, acting against de novo
purineepyrimidine biosynthesis and potentially being antagonists
of nucleotide transport in DNA synthesis. Curtin NJ and colleagues
[25] showed that pyrimido[5,4-d]pyrimidines act as anti-
metabolites, although they were designed as anti-cancer agents,
and they still represent a major class with wide therapeutic
applications today. Salvage of extracellular purine or pyrimidine
The general synthetic strategy used to prepare the investigated
compounds (1e9) was similar to that proposed by Curtin et al. [13]
in which the preparation of several substituted pyrimido[5,4-d]
pyrimidines was described (Fig. 1B). General procedure for the first
reaction: the 4-methoxybenzylamine (1.1 ml, 8 mmol) was added
to a stirred solution of 2,4,6,8-tetrachloropyrimido[5,4-d]pyrimi-
dine (a) (490 mg, 1.80 mmol) in dry THF (12 mL) containing K2CO3
(1.8 g, 13.5 mmol). The reaction mixture was stirred (1 h) at room
temperature under dry conditions until TLC analysis confirmed the
absence of starting materials. H2O (20e30 mL) was added, the
reaction mixture was stirred for 20 min, and the precipitated solid
was collected by filtration. Chromatography on silica afforded the
intermediate compound 2,6-dichloro-4,8-dibenzylaminopyrimido
[5,4-d]pyrimidine (b) (yields 98%). For the second reaction, inter-
mediate b (0.1e1 mmol) was dissolved in an excess (1e5 mL) of the
required amine, and the mixture was stirred at 100e150 ꢁC, until
TLC analysis confirmed the absence of starting materials (typically
24e48 h). After the mixture was cooled, the crude product was
precipitated by addition of H2O (30e50 mL) and collected by
filtration and/or extraction with EtOAc. The product was purified by
chromatography on silica (yields 97%).
1H and 13C NMR spectra were acquired on a Varian Unity
Inova700 spectrometer and the residual solvent signal or TMS were
used as an internal standard. Spectra were recorded in DMSO-d6 as
solvent. NH signals appeared as broad singlets exchangeable with
D2O. MS (ESIþ) was recorded with an API-2000 triple quadrupole
mass spectrometer equipped with
a turbo ion-spray source

48
A. Virgilio et al. / European Journal of Medicinal Chemistry 57 (2012) 41e50
(Applied Biosystem; Thornhill, ON, Canada). NMR and MS experi-
ments were performed at “Centro di Servizi Interdipartimentale di
Analisi Strumentale”, Università degli Studi di Napoli Federico II.
TLC was performed on silica gel 60 plates (Merck, precoated), with
CHCl3/MeOH as a mobile phase and was visualized with UV light at
254 and 365 nm. Chromatography was conducted on silica gel
(Kieselgel 60, 70e230 mesh). All reagents were purchased from
Aldrich Chemical Co. and were used without further purification.
NHCH2CHOHCHOH), 3.72 (6H, 2ꢂ ArOCH3), 4.49 (4H, 2ꢂ CH2Ar),
4.57 (2H, 2ꢂ CH2OH), 4.72 (2H, 2ꢂ CHOH), 5.91 (2H, 2ꢂ
NHCH2CHOHCH2OH), 6.87 (4H, 4ꢂ AreH), 7.29 (4H, 4ꢂ AreH), 7.51
(2H, 2ꢂ NHCH2Ar); 13C NMR
2 ꢂ 44.4, 2 ꢂ 44.9, 2 ꢂ 55.1, 2 ꢂ 63.9,
d
2 ꢂ 72.7, 4 ꢂ 113.7, 4 ꢂ 129.2, 2 ꢂ 131.9, 2 ꢂ 134.0, 2 ꢂ 155.0,
2 ꢂ 155.4, 2 ꢂ 158.3; MS ESI (m/z): C28H36N8O6 581 (M þ H)þ.
5.1.6. (3S,30S)-1,10-(4,8-bis(4-Methoxybenzylamino)pyrimido[5,4-
d]pyrimidine-2,6-diyl)dipyrrolidin-3-ol (8) and (3R,30R)-1,10-(4,8-
bis(4-methoxybenzylamino)pyrimido[5,4-d]pyrimidine-2,6-diyl)
dipyrrolidin-3-ol (9)
5.1.1. 1,10-(4,8-bis(4-Methoxybenzylamino)pyrimido[5,4-d]
pyrimidine-2,6-diyl)dipiperidin-4-ol (2)
1H NMR
CHCH2NCH2CH),
d
1.30 (4H, 2ꢂ CHCH2NCH2CH), 1.74 (4H, 2ꢂ
1H NMR
d
1.86 (2H, 2ꢂ NCH2CHCH(OH)), 1.98 (2H, 2ꢂ
3.13 (4H, CHNCH), 3.67 (2H,
2ꢂ
NCH2CHCH(OH)), 3.55 (4H, 2ꢂ NCH2CH2), 3.57 (4H, 2ꢂ
NCH2CH(OH)), 3.71 (6H, 2ꢂ ArOCH3), 4.35 (2H, 2ꢂ NCH2CH(OH)),
4.61 (4H, 2ꢂ CH2Ar), 4.99-5.10 (2H, 2ꢂ OH), 6.88 (4H, 4ꢂ AreH),
CH2N(CH2)3CHOH), 3.71 (6H, 2ꢂ ArOCH3), 4.34 (4H, 2ꢂ CHNCH),
4.56 (4H, 2ꢂ CH2Ar), 4.66 (2H, 2ꢂ OH), 6.86 (4H, 4ꢂ AreH), 7.30
(4H, 4ꢂ AreH); 13C NMR
d
4 ꢂ 36.1, 4 ꢂ 44.4, 2 ꢂ 44.5, 2 ꢂ 55.1,
7.34 (4H, 4ꢂ AreH); 13C NMR
d
2 ꢂ 44.4, 2 ꢂ 44.5, 2 ꢂ 54.8,
2 ꢂ 67.6, 4 ꢂ 113.7, 4 ꢂ 129.2, 2 ꢂ 131.9, 2 ꢂ 134.0, 2 ꢂ 155.0,
2 ꢂ 55.1, 2 ꢂ 69.4, 2 ꢂ 79.3, 4 ꢂ 113.7, 4 ꢂ 129.2, 2 ꢂ 131.9, 2 ꢂ 134.0,
2 ꢂ 155.0, 2 ꢂ 157.7, 2 ꢂ 158.3; MS ESI (m/z): C30H36N8O4 573
(M þ H)þ.
2 ꢂ 157.7, 2 ꢂ 158.3; MS ESI (m/z): C32H40N8O4 601 (M þ H)þ.
5.1.2. 1,10-(4,8-bis(4-Methoxybenzylamino)pyrimido[5,4-d]
pyrimidine-2,6-diyl)dipiperidin-3-ol (diastereomeric mixture) (3)
5.2. Cell culture
1H NMR
d
1.45 (2H, 2ꢂ NCH2CHCH2CH(OH)), 1.52 (2H, 2ꢂ
NCH2CH(OH)CH), 1.55 (2H, 2ꢂ NCH2CHCH2CH(OH)), 1.77 (2H, 2ꢂ
NCH2CH(OH)CH), 3.50 (4H, 2ꢂ NCH2CH2), 3.57 (4H, 2ꢂ
NCH2CHOH), 3.72 (6H, 2ꢂ ArOCH3), 4.35 (2H, 2ꢂ NCH2CH(OH)CH2),
4.61 (4H, 2ꢂ CH2Ar), 4.81 (2H, 2ꢂ OH), 6.88 (4H, 4ꢂ AreH), 7.34
MDA-435-C100 empty vector and h-prune over-expressing cells
were obtained as previously described [8]. MDA-435-C100 stable
clones, mouse mammary gland tumour cell line 4T1-Luc and
human mammary gland tumour cell line MDA-MB-231T were
grown in high glucose Dulbecco’s modified Eagle’s medium
(DMEM; Invitrogen) supplemented with 10% (v/v) foetal bovine
(4H, 4ꢂ AreH), 6.81 (2H, 2ꢂ NH); 13C NMR
d
2 ꢂ 23.9, 2 ꢂ 33.1,
2 ꢂ 44.4, 2 ꢂ 46.1, 2 ꢂ 53.5, 2 ꢂ 55.1, 2 ꢂ 66.3, 4 ꢂ 113.7, 4 ꢂ 129.2,
2 ꢂ 131.9, 2 ꢂ 134.0, 2 ꢂ 155.0, 2 ꢂ 157.7, 2 ꢂ 158.3; MS ESI (m/z):
serum (Invitrogen), 2 mM L-Glutamine (Invitrogen), and 1% (v/v)
C
32H40N8O4 601 (M þ H)þ.
antibiotics (10,000 U/ml penicillin, and 10 mg/ml streptomycin
(Invitrogen)). The cells were grown at 37 ꢁC in a humidified
atmosphere of 95% air and 5% CO2 (v/v).
5.1.3. (1,10-(4,8-bis(4-Methoxybenzylamino)pyrimido[5,4-d]
pyrimidine-2,6-diyl)bis(piperidine-3,1-diyl))dimethanol
(diastereomeric mixture) (4)
5.3. Immunoblotting
1H NMR
d
1.23 (2H, 2ꢂ NCH2CH(CH2OH)CHCH2), 1.39 (2H, 2ꢂ
NCH2CHCH2), 1.56 (2H, 2ꢂ NCH2CHCH2), 1.67 (2H, 2ꢂ
NCH2CH(CH2OH)CHCH2), 1.76 (2H, 2ꢂ NCH2CH(CH2OH)), 2.67 (2H,
2ꢂ NCH(CH2)2), 2.89 (2H, 2ꢂ NCH(CH2)2), 3.34 (4H, 2ꢂ CH2OH), 3.34
(4H, 2ꢂ NHCH2CH(CH2OH)), 3.71 (6H, 2ꢂ ArOCH3), 4.58 (4H, 2ꢂ
CH2Ar), 4.66 (2H, 2ꢂ OH), 6.86 (4H, 4ꢂ AreH), 7.34 (4H, 4ꢂ AreH);
Total protein extracts from cells were prepared and used for
immunoblotting as previously described [8] using the A59 rabbit
anti-prune antibody, and detected by a horseradish peroxidasee
conjugated anti-rabbit (1:3000) antibody (Amersham). A mouse
anti-b-actin antibody (1:5.000; SigmaeAldrich) was used as the
loading control.
13C NMR
d
2 ꢂ 25.6, 2 ꢂ 27.7, 2 ꢂ 39.4, 2 ꢂ 44.4, 2 ꢂ 46.6, 2 ꢂ 49.7,
2 ꢂ 55.1, 2 ꢂ 65.4, 4 ꢂ113.7, 4 ꢂ129.2, 2 ꢂ131.9, 2 ꢂ134.0, 2 ꢂ155.0,
2 ꢂ 157.7, 2 ꢂ 158.3; MS ESI (m/z): C34H44N8O4 629 (M þ H)þ.
5.4. Protein expression and purification in Baculovirus
5.1.4. (1R,10R,2R,20R)-2,20-(4,8-bis(4-Methoxybenzylamino)
pyrimido[5,4-d]pyrimidine-2,6-diyl)bis(azanediyl)dicyclohexanol
Protein expression was performed using Baculovirus expression
system as previously described [8].
(5)
1H NMR
d
1.09 (2H, 2ꢂ (NH)CHCH(OH)CH), 1.13 (2H, 2ꢂ (NH)
5.5. Characterization of the pyrimido[5,4-d]pyrimidine derivatives
action on h-prune PDE activity
CHCH2CH), 1.19 (2H, 2ꢂ (NH)CHCH(CH2)3), 1.20 (2H, 2ꢂ (NH)
CH(CH2)CH), 1.61 (2H, 2ꢂ (NH)CHCH(CH2)3), 1.62 (2H, 2ꢂ (NH)
CH(CH2)CH), 1.66 (2H, 2ꢂ (NH)CHCH2CH), 1.78 (2H, 2ꢂ (NH)
CHCH(OH)CH), 2.72 (2H, 2ꢂ (NH)CH), 3.38 (2H, 2ꢂ (NH)
CHCH(OH)), 3.72 (6H, 2ꢂ ArOCH3), 4.0 (2H, 2ꢂ (NH)CH), 4.62 (4H,
2ꢂ CH2Ar), 4.81 (2H, 2ꢂ OH), 6.86 (4H, 4ꢂ AreH), 7.30 (4H, 4ꢂ Are
For the determination of the ability of compound 1 to inhibit the
h-prune purified protein cAMP-PDE activity scintillation proximity
assays were performed as previously described [8] using the
Phosphodiesterase [3H]cAMP SPA enzyme assay (Amersham),
according to the manufacturer’s protocol. In particular, 400 ng of
purified enzymes were incubated for 15 min at 30 ꢁC in a reaction
mix containing 50 mM TriseHCl pH 7.5, 8.3 mM MgCl2, 1.7 mM
H), 7.40 (2H, 2ꢂ NHCH2Ar); 13C NMR
d
2 ꢂ 24.8, 2 ꢂ 25.0, 2 ꢂ 33.8,
2 ꢂ 34.4, 2 ꢂ 44.4, 2 ꢂ 55.1, 2 ꢂ 56.9, 2 ꢂ 75.6, 4 ꢂ 113.7, 4 ꢂ 129.2,
2 ꢂ 131.9, 2 ꢂ 134.0, 2 ꢂ 155.0, 2 ꢂ 157.7, 2 ꢂ 158.3; MS ESI (m/z):
C
34H44N8O4 629 (M þ H)þ.
EGTA, 0.005
tions of compound 1 (0.0001
0.2 M, 0.4 M, 0.8 M, 1.6 M, 3.2
tions were terminated by adding 50
mCi/m
l [3H]cAMP in presence of increasing concentra-
mM, 0.0001
m
M, 0.001 M, 0.01
m
mM,
5.1.5. (2R,20R)-3,30-(4,8-bis(4-Methoxybenzylamino)pyrimido[5,4-d]
pyrimidine-2,6-diyl)bis(azanediyl)dipropane-1,2-diol (6) and
(2S,20S)-3,30-(4,8-bis(4-methoxybenzylamino)pyrimido[5,4-d]
pyrimidine-2,6-diyl)bis(azanediyl)dipropane-1,2-diol (7)
m
m
m
m
mM, 6.4
m
M, 10 mM). The reac-
m
l of Yttrium SPA PDE beads
and mixed well to redistribute the settled beads and allowed to
stand at room temperature for 20 min. The reactions were counted
on a scintillation counter. All reactions, including buffer-only
blanks, were conducted in triplicate. Enzyme activities were
1H NMR
d
3.20 (2H, 2ꢂ NHCH), 3.35 (2H, 2ꢂ NHCH2CHOH), 3.35
(2H, 2ꢂ NHCH2CHOHCHOH), 3.42 (2H, 2ꢂ NHCH), 3.62 (2H, 2ꢂ

A. Virgilio et al. / European Journal of Medicinal Chemistry 57 (2012) 41e50
49
calculated according to the amount of radio-labelled product
5.7. xCelligence system
detected, according to the manufacturer’s protocol. The IC50 was
determined by plotting log compound 1 concentration on X-axis
and % cAMP-PDE inhibition on Y-axis. The percentage inhibition
curves were plotted using four determinations. The IC50 value was
calculated by using Sigmoidal doseeresponse (Prism Release 5.04;
GraphPad Software Inc, USA). A total of two independent sets of
experiments were performed.
Cellular extracts were prepared to measure h-prune PDE
activity. Briefly the MDA435-C100 cells over-expressing h-prune
were plated onto 6-well dishes with 2 ml growth medium/well the
day before treatment. After 24 h, the medium was replaced with
fresh medium containing the vehicle (DMSO or PEG/PBS) or the
The Real-Time Cell Analysis (RTCA) Instrument (Roche) uses an
xCelligence system which is based on the RT-CES (Real-Time Cell
Electronic Sensor) system, that allows label-free dynamic moni-
toring of cell proliferation and viability in real time [41,42]. Briefly,
this system uses an electronic readout of impedance to non-
invasively quantify adherent cell proliferation and viability in
real-time. The cells are seeded into standard microtiter plates that
contain microelectronic sensor arrays. The interactions of the cells
with the electronic biosensors generate a cell-electrode impedance
response that not only indicates the cell viability, but also correlates
with the number of cells seeded in the well. Cell-sensor impedance
is expressed as an arbitrary unit called the Cell Index (CI) [41,42].
The cell culture conditions on the sensor device are the same as
those in the tissue disks described above. For real-time calculations
of the CI during the in vitro proliferation assay, the background
impedance of each sensor well of the E-plate 16 (ACEA Biosciences,
indicated compound at 1
mM concentration, and the cells were
incubated over night at 37 ꢁC 5% CO2. Treated and untreated cells
were washed three times with ice-cold phosphate-buffered saline
and lysed with lysis buffer (50 mM TriseHCl, pH 7.4, 1% Nonidet
P40, 0.25% sodium deoxycholate, 150 mM NaCl, 1 mM EDTA, 1 mM
phenylmethylsulfonyl fluoride, 1 mM Na3VO4, 1 mM NaF, 1 mM
Inc) was measured in the absence of cells. To do this, 40 ml cell
benzamidine, 1
mg/ml aprotinin, and 1
m
g/ml leupeptin) at 4 ꢁC. The
culture media was added into each well for the baseline
measurements, followed by addition of the cells to the sensor wells
(20,000 MDA-MB-231T cells/well). Four replicates for each exper-
imental point were used, so four single readings of CI from four
different wells for each experimental point were obtained. After
cell seeding, the sensor device was left at room temperature for
30 min to reduce the variability resulting from edge effects [43],
then it was put into the RTCA Instrument inside a 37 ꢁC, 5% CO2
incubator and the CI of each well was monitored automatically with
the xCelligence system software. To ensure consistent initial cell
conditions for the proliferation assays, each independent
compound or DMSO (placebo) were added once after the CI
reached a range of 1.0e1.2, indicating 50%e60% cell confluence.
After addition of the compounds or DMSO, the plate with the cells
was mounted back into the RTCA Instrument inside a 37 ꢁC, 5% CO2
incubator. The CI was automatically and continuously monitored by
xCelligence system software every 5 min for up to 54e56 h.
Furthermore in vitro proliferation assays in real-time were per-
lysates were incubated in ice for 30 min and then centrifuged at
14,000 g for 10 min. The protein concentrations of the resulting
supernatants were determined using the Bradford protein assay
(Bio-Rad Laboratories, Hercules, CA). Equal amounts of whole
protein extract were used for the cAMP-PDE assays, using the
Phosphodiesterase [3H]cAMP SPA enzyme assay (Amersham),
according to the manufacturer’s protocol. Briefly, total protein
extracts were diluted as required and assessed in a 100
mix containing 50 mM TriseHCl pH 7.5, 8.3 mM MgCl2, 1.7 mM
EGTA, 0.005 Ci/
l [3H]cAMP. Reactions were incubated for 15 min
at 30 ꢁC, then terminated by adding 50
l of Yttrium SPA PDE beads
ml reaction
m
m
m
and mixed well to redistribute the settled beads and allowed to
stand at room temperature for 20 min. The reactions were counted
on a scintillation counter. All reactions, including buffer-only
blanks, were conducted in triplicate. Enzyme activities were
calculated according to the amount of radio-labelled product
detected, according to the manufacturer’s protocol. A total of two
independent sets of experiments were performed.
formed in presence of 50 and 100 mM compounds 8 and 9 to valuate
doseeresponse effects in inhibiting cell proliferation. A total of two
5.6. 2D cell migration assays
independent experiments was performed.
100,000 MDA-MB-231Tcells, resuspended in serum free medium
5.8. End-point proliferation assay (MTS assays)
(200
at a concentration of 8
chamber of an 8 m pore size 24 well-plates Transwell insert
(Corning Incorporated, Costar, Corning, NY, USA). Complete medium
(600 l) containing 0.5% FBS, as chemoattractant, and DMSO
(control) or each specific compound at a concentration of 8 M was
m
l)inthepresenceofDMSO(control)oreachspecificcompound
mM, were plated in triplicate into the upper
Proliferation assays were performed using the CellTiter 96
Aqueous One Solution Cell Proliferation Assay (Promega). 104 MDA-
MB-231T cells/well were plated into 96-well plates and treated
with DMSO (control) or DMSO containing compound 8 (at the
indicated concentrations) for 24 h. Each experimental point was
m
m
m
added to the lower chamber. The 4T1-luc migration assay was per-
formed as described with a few changes: 25,000 cells were seeded in
the upper chamber; the low chamber contained complete medium
supplemented with 10% FBS. MDA-MB-231T cells were lived to
migrate for 3 h at 37 ꢁC 5% CO2, while the 4T1-Luc cells were lived to
migrate for 4 h at 37 ꢁC 5% CO2. After migration cells were fixed with
2.5% glutaraldehyde and stained in Gill’s Hematoxilin No.1.
Each of the triplicate chamber was divided in nine fields. A
picture of each field was taken from optical microscopy and
migrating cells were counted by using Cell Counter ImageJ program.
As control of normal migration, the experimental point “Not
Treated” (only the DMEM growth medium) was settled down. Each
compound was then compared to the Placebo (DMSO 0.001%)
experimental point, as a control of their dissolving solution, and
to the DP experimental point to identify the pyrimido[5,4-d]
pyrimidine derivative with the best migration inhibitory activity. A
total of two independent experiments was performed.
assessed in quadruplicate. For each experimental point, 20 ml (3-
(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-
sulfophenyl)-2H-tetrazolium) (MTS) solution (1.90 mg/ml) was
added to each well, and the cells were incubated for 2 h at 37 ꢁC.
Absorbance was then measured at 490 nm using a microtiter plate
reader (VICTOR3 1420 Multilabel Counter, Perkin Elmer). For each
experimental point, the means of absorbance and their standard
errors (SE) were calculated. A total of two independent sets of
experiments were performed.
5.9. Statistical analysis
The statistical comparison of the data from the in vitro treat-
ments was performed using Students’ t test. Statistical significance
was established at p ꢃ 0.05.
Disclosure: The authors declare that they have no competing
interests as defined by European Journal of Medicinal Chemistry, or

50
A. Virgilio et al. / European Journal of Medicinal Chemistry 57 (2012) 41e50
[16] M. Deckers, M. van Dinther, J. Buijs, I. Que, C. Löwik, G. van der Pluijm, P. ten
Dijke, The tumor suppressor Smad4 is required for transforming growth
factor beta-induced epithelial to mesenchymal transition and bone metastasis
of breast cancer cells, Cancer Res. 66 (2006) 2202e2209.
other interests that might be perceived to influence the results and
discussion reported in this paper.
[17] T. Kobayashi, S. Hino, N. Oue, T. Asahara, M. Zollo, W. Yasui, A. Kikuchi,
Glycogen synthase kinase 3 and h-prune regulate cell migration by modu-
lating focal adhesions, Mol. Cell. Biol. 26 (2006) 898e911.
[18] K. Tao, M. Fang, J. Alroy, G.G. Sahagian, Imagable 4T1 model for the study of
late stage breast cancer, BMC Cancer 8 (2008) 228.
Acknowledgements
We would like to thank N.J. Curtin and R.J. Griffin at AICR-UK
(New Castle, UK) for valuable discussions about strategies to
develop new anti-h-prune compounds.
[19] T. Diercks, M. Coles, H. Kessler, Applications of NMR in drug discovery, Curr.
Opin. Chem. Biol. 5 (2001) 285e291.
This work was supported by the following grants: FP6-EET
pipeline LSH-CT-2006-037260 (MZ), FP7-Tumic HEALTH-F2-2008-
201662 (MZ), Associazione per la Ricerca sul Cancro AIRC (MZ),
Associazione Italiana contro la lotta al Neuroblastoma (MZ), PRIN-
MIUR PRIN (E5AZ5F) 2008 (MZ). Legge n.5, Regione Campania
2007-8 (MZ). Valeria Di Dato and Daniela Spano were supported by
Fondazione San Paolo University of Naples, Federico II, DBBM.
Veronica Maffia, Natascia Marino, Pasquale De Antonellis were
supported by EET-pipeline LSH-CT-2006-037260 and Tumic
HEALTH-F2-2008-201662). Daniela De Martino is supported by
Associazione Italiana contro la lotta al Neuroblastoma (MZ).
[20] H.O. Villar, J. Yan, M.R. Hansen, Using NMR for ligand discovery and optimi-
zation, Curr. Opin. Chem. Biol. 8 (2004) 387e391.
[21] P. Accornero, L.M. Pavone, M. Baratta, The scatter factor signaling pathways as
therapeutic associated target in cancer treatment, Curr. Med. Chem. 17 (25)
(2010) 2699e2712, Review.
[22] L. Garzia, A. D’Angelo, A. Amoresano, S.K. Knauer, C. Cirulli, C. Campanella,
R.H. Stauber, C. Steegborn, A. Iolascon, M. Zollo, Phosphorylation of nm23-H1
by CKI induces its complex formation with h-prune and promotes cell
motility, Oncogene 27 (2008) 1853e1864.
[23] T. Müller, U. Stein, A. Poletti, L. Garzia, M. Rothley, D. Plaumann, W. Thiele,
M. Bauer, A. Galasso, P. Schlag, M. Pankratz, M. Zollo, J.P. Sleeman, ASAP1
promotes tumor cell motility and invasiveness, stimulates metastasis forma-
tion in vivo, and correlates with poor survival in colorectal cancer patients,
Oncogene 29 (2010) 2393e2403.
[24] N. Marino, M. Zollo, Understanding h-prune biology in the fight against
cancer, Clin. Exp. Metastasis 24 (2007) 637e645.
Appendix A. Supplementary data
[25] H.D. Thomas, K. Saravanan, L.Z. Wang, M.J. Lin, J.S. Northen, H. Barlow,
M. Barton, D.R. Newell, R.J. Griffin, B.T. Golding, N.J. Curtin, Preclinical evalu-
ation of a novel pyrimidopyrimidine for the prevention of nucleoside and
nucleobase reversal of antifolate cytotoxicity, Mol. Cancer Ther. 8 (2009)
1828e1837.
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.ejmech.2012.08.020.
[26] J.A. Belt, N.M. Marina, D.A. Phelps, C.R. Crawford, Nucleoside transport in
normal and neoplastic cells, Adv. Enzym. Regul. 33 (1993) 235e252.
[27] J. Walling, From methotrexate to pemetrexed and beyond. A review of the
pharmacodynamic and clinical properties of antifolates, Invest. New Drugs 24
(2006) 37e77.
[28] S.A. Baldwin, J.R. Mackey, C.E. Cass, J.D. Young, Nucleoside transporters:
molecular biology and implications for therapeutic development, Mol. Med.
Today 5 (1999) 216e224.
[29] A.M. Wright, W.P. Gati, A.R. Paterson, Enhancement of retention and cyto-
toxicity of 2-chlorodeoxyadenosine in cultured human leukemic lympho-
blasts by nitrobenzylthioinosine, an inhibitor of equilibrative nucleoside
transport, Leukemia 14 (2000) 52e60.
[30] W. Schaper, Dipyridamole, an underestimated vascular protective drug, Car-
diovasc. Drugs Ther. 19 (2005) 357e363.
[31] J.L. Grem, Biochemical modulation of fluorouracil by dipyridamole: preclinical
and clinical experience, Semin. Oncol. 19 (1992) 56e65.
[32] S.B. Howell, D. Hom, R. Sanga, J.S. Vick, I.S. Abramson, Comparison of the
synergistic potentiation of etoposide, doxorubicin, and vinblastine cytotox-
icity by dipyridamole, Cancer Res. 49 (1989) 3178e3183.
[33] N. Ramu, A. Ramu, Circumvention of adriamycin resistance by dipyridamole
analogues: a structureeactivity relationship study, Int. J. Cancer 43 (1989)
487e491.
[34] N.L. Lehman, P.V. Danenberg, Modulation of RTX cytotoxicity by thymidine
and dipyridamole in vitro: implications for chemotherapy, Cancer Chemother.
Pharmacol. 45 (2000) 142e148.
References
[1] B. Weigelt, J.S. Reis-Filho, Histological and molecular types of breast cancer: is
there a unifying taxonomy? Nat. Rev. Clin. Oncol. 6 (2009) 718e730.
[2] E. Andreopoulou, G.N. Hortobagyi, Prognostic factors in metastatic breast
cancer: successes and challenges toward individualized therapy, J. Clin. Oncol.
26 (2008) 3660e3662.
[3] C. Bernard-Marty, F. Cardoso, M.J. Piccart, Facts and controversies in systemic
treatment of metastatic breast cancer, Oncologist 9 (2004) 617e632.
[4] J. O’Shaughnessy, Extending survival with chemotherapy in metastatic breast
cancer, Oncologist 10 (2005) 20e29.
[5] D. Spano, C. Heck, P. De Antonellis, G. Christofori, M. Zollo, Molecular networks
that regulate cancer metastasis, Semin Cancer Biol. 22 (2012) 234e249.
[6] D. Spano, M. Zollo, Tumor microenvironment: a main actor in the metastasis
process, Clin Exp Metastasis 4 (2012) 381e395.
[7] P.S. Steeg, Tumor metastasis: mechanistic insights and clinical challenges, Nat.
Med. 12 (2006) 895e904.
[8] A. D’Angelo, L. Garzia, A. André, P. Carotenuto, V. Aglio, O. Guardiola, G. Arrigoni,
A. Cossu, G. Palmieri, L. Aravind, M. Zollo, Prune cAMP phosphodiesterase binds
nm23-H1 and promotes cancer metastasis, Cancer Cell. 5 (2004) 137e149.
[9] M. Tammenkoski, K. Koivula, E. Cusanelli, M. Zollo, C. Steegborn, A.A. Baykov,
R. Lahti, Human metastasis regulator protein H-prune is a short-chain exo-
polyphosphatase, Biochemistry 47 (2008) 9707e9713.
[10] A. Forus, A. D0Angelo, J. Henriksen, G. Merla, G.M. Maelandsmo, V.A. Florenes,
S. Olivieri, B. Bjerkehagen, L.A. Meza-Zepeda, F. del Vecchio Blanco, C. Muller,
F. Sanvito, J. Kononen, J.M. Nesland, O. Fodstad, A. Reymond, O.P. Kallioniemi,
G. Arrigoni, A. Ballabio, O. Myklebost, M. Zollo, Amplification and over-
expression of PRUNE in human sarcomas and breast carcinomas-a possible
mechanism for altering the nm23-H1 activity, Oncogene 20 (2001) 6881e6890.
[11] M. Zollo, A. Andre, A. Cossu, M.C. Sini, A. D0Angelo, N. Marino, M. Budroni,
F. Tanda, G. Arrigoni, G. Palmieri, Overexpression of h-prune in breast cancer is
correlated with advanced disease status, Clin. Cancer Res. 11 (2005) 199e205.
[12] H.C. Barlow, K.J. Bowman, N.J. Curtin, A.H. Calvert, B.T. Golding, B. Huang,
P.J. Loughlin, D.R. Newell, P.G. Smith, R.J. Griffin, Resistance-modifying agents.
Part 7: 2,6-disubstituted-4,8-dibenzylaminopyrimido[5,4-d]pyrimidines that
inhibit nucleoside transport in the presence of alpha1-acid glycoprotein
(AGP), Bioorg. Med. Chem. Lett. 10 (2000) 585e589.
[13] N.J. Curtin, H.C. Barlow, K.J. Bowman, A.H. Calvert, R. Davison, B.T. Golding,
B. Huang, P.J. Loughlin, D.R. Newell, P.G. Smith, R.J. Griffin, Resistance-modi-
fying agents. 11. (1) Pyrimido[5,4-d]pyrimidine modulators of antitumor drug
activity. Synthesis and structure-activity relationships for nucleoside trans-
port inhibition and binding to alpha1-acid glycoprotein, J. Med. Chem. 47
(2004) 4905e4922.
[35] J.A. Nelson, S. Drake, Potentiation of methotrexate toxicity by dipyridamole,
Cancer Res. 44 (1984) 2493e2496.
[36] P.B. Desai, R. Sridhar, Potentiation of cytotoxicity of mitoxantrone toward
CHO-K1 cells in vitro by dipyridamole, Pharm. Res. 9 (1992) 178e181.
[37] C.R. Boyer, P.L. Karjian, G.M. Wahl, M. Pegram, S.T. Neuteboom, Nucleoside
transport inhibitors, dipyridamole and p-nitrobenzylthioinosine, selectively
potentiate the antitumor activity of NB1011, Anticancer Drugs 13 (2002)
29e36.
[38] M. Rodrigues, F. Barbosa Jr., J.R. Perussi, Dipyridamole increases the cytotox-
icity of cisplatin in human larynx cancer cells in vitro, Braz. J. Med. Biol. Res.
37 (2004) 591e599.
[39] N.J. Curtin, K.J. Bowman, R.N. Turner, B. Huang, P.J. Loughlin, A.H. Calvert,
B.T. Golding, R.J. Griffin, D.R. Newell, Potentiation of the cytotoxicity of thy-
midylate synthase (TS) inhibitors by dipyridamole analogues with reduced
alpha1-acid glycoprotein binding, Br. J. Cancer 80 (1999) 1738e1746.
[40] A.H. Bacelar, M.A. Carvalho, M.F. Proença, Synthesis and in vitro evaluation of
substituted pyrimido[5,4-d]pyrimidines as a novel class of anti-Mycobacte-
rium tuberculosis agents, Eur. J. Med. Chem. 45 (2010) 3234e3239.
[41] J.Z. Xing, L. Zhu, J.A. Jackson, S. Gabos, X.J. Sun, X.B. Wang, X. Xu, Dynamic
monitoring of cytotoxicity on microelectronic sensors, Chem. Res. Toxicol. 18
(2005) 154e161.
[14] N. Van Larebeke, C. Dragonetti, M. Mareel, Effect of dipyridamole on invasion
of five types of malignant cells in organ culture, Clin. Exp. Metastasis 7 (1989)
645e657.
[15] D. Spano, J.C. Marshall, N. Marino, D. De Martino, A. Romano, M. N. Scop-
pettuolo, A. M. Bello, V. Di Dato, L. Navas, G. De Vita, C. Medaglia, P.S. Steeg
and M. Zollo. Dipyridamole prevents triple-negative breast-cancer progres-
sion. Clin. Exp. Metastasis. (in press).
[42] J.Z. Xing, L. Zhu, S. Gabos, L. Xie, Microelectronic cell sensor assay for detection
of cytotoxicity and prediction of acute toxicity, Toxicol. in Vitro 20 (2006)
995e1004.
[43] B.K. Lundholt, K.M. Scudder, L. Pagliaro, A simple technique for reducing edge
effect in cell-based assays, J. Biomol. Screen. 8 (2003) 566e570.