Targeting gliomas with triazene-based hybrids: Structure-activity relationship, mechanistic study and stability

Article abstract of DOI:10.1016/j.ejmech.2019.03.048

Herein we report novel hybrid compounds based on valproic acid and DNA-alkylating triazene moieties, 1, with therapeutic potential for glioblastoma multiforme chemotherapy. We identified hybrid compounds 1d and 1e to be remarkably more potent against glioma and more efficient in decreasing invasive cell properties than temozolomide and endowed with chemical and plasma stability. In contrast to temozolomide, which undergoes hydrolysis to release an alkylating metabolite, the valproate hybrids showed a low potential to alkylate DNA. Key physicochemical properties align for optimal CNS penetration, highlighting the potential of these effective triazene based-hybrids for enhanced anticancer chemotherapy.

Full text of DOI:10.1016/j.ejmech.2019.03.048

European Journal of Medicinal Chemistry 172 (2019) 16e25
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Research paper
Targeting gliomas with triazene-based hybrids: Structure-activity
relationship, mechanistic study and stability
a
a
b
b
a
ꢀ
~
Claudia Braga , Ana R. Vaz , M. Conceiçao Oliveira , M. Matilde Marques , Rui Moreira ,
Dora Brites a, Maria J. Perry a
Research Institute for Medicines (iMed.ULisboa), Faculdade de Farmacia, Universidade de Lisboa, Av. Professor Gama Pinto, 1649-003 Lisboa, Portugal
,
*
a
ꢀ
b
ꢀ
Centro de Química Estrutural, Instituto Superior Tecnico, Universidade de Lisboa, Av. Rovisco Pais, 1049-001 Lisboa, Portugal
a r t i c l e i n f o
a b s t r a c t
Article history:
Herein we report novel hybrid compounds based on valproic acid and DNA-alkylating triazene moieties,
1, with therapeutic potential for glioblastoma multiforme chemotherapy. We identified hybrid com-
pounds 1d and 1e to be remarkably more potent against glioma and more efficient in decreasing invasive
cell properties than temozolomide and endowed with chemical and plasma stability. In contrast to
temozolomide, which undergoes hydrolysis to release an alkylating metabolite, the valproate hybrids
showed a low potential to alkylate DNA. Key physicochemical properties align for optimal CNS pene-
tration, highlighting the potential of these effective triazene based-hybrids for enhanced anticancer
chemotherapy.
Received 28 January 2019
Received in revised form
18 March 2019
Accepted 19 March 2019
Available online 22 March 2019
Keywords:
Triazenes
Histone deacetylase inhibitors
Valproic acid
© 2019 Elsevier Masson SAS. All rights reserved.
Multi-targeted glioma therapy
1. Introduction
a monomethyltriazene intermediate that spontaneously de-
composes to release the highly reactive methyldiazonium cation,
responsible for DNA alkylation [5]. However, the widespreading
nature of glioma cells disrupts TMZ efficacy, resulting in high
recurrence of glioblastoma. Proliferation and cell migration, along
with every DNA repair system activated throughout the complex
key signaling pathways of GBM, remain some of the demanding
challenges for successful therapy [6]. The anticonvulsant valproic
acid (VPA, Fig. 1) was recently identified as a short chain fatty acid
(SCFA) histone deacetylase (HDAC) inhibitor that prevents cell
proliferation in GBM brain cancer and modulates several tran-
scription factors [7]. Phase II clinical trials showed improvement in
the outcome of patients with GBM when VPA was combined with
the standard treatment (TMZ plus radiotherapy); a median overall
survival of 29.6 months was obtained with this combination,
compared to 14.6 months with the standard protocol [3,8]. Further
investigations also revealed that VPA induces chromatin loosening,
thus increasing TMZ accessibility to DNA and leading to higher
methylation levels [9].
Brain tumors have surpassed leukemia as the leading cause in
cancer deaths among children and their incidence is relatively high
in the adult population, with gliomas accounting for 82% of all
malignant brain tumors [1,2]. Glioblastoma multiforme (GBM), the
most lethal type of glioma, presents an average patient's survival
time of only 14 months. Current treatment for GBM patients in-
cludes surgical tumor resection followed by radiotherapy with
concomitant and adjuvant chemotherapy based on temozolomide
(TMZ, Fig. 1) [3]. TMZ is a triazene that has shown promising
antitumor activity not only in brain tumors, but also in a variety of
solid tumors, including malignant melanoma. TMZ can be taken
orally and, due to its small size and lipophilic features, it is able to
cross the blood-brain barrier [4]. Once in the central nervous sys-
tem, TMZ undergoes hydrolytic ring opening at physiological pH to
Abbreviations: BBB, blood-brain barrier; CDI, 1,10-carbonyldiimidazole; CNS
MPO, central nervous system multiparameter optimization; DCC, N,N0-dicyclohex-
ylcarbodiimide; ESI-HRMS, High resolution electrospray ionization mass spec-
trometry; GBM, glioblastoma multiforme; HDAC, histone deacetylase; HOBt,
hydroxybenzotriazole; MMT, monomethyltriazene; PBS, phosphate buffered saline;
SCFA, short chain fatty acid; TMZ, temozolomide; VPA, valproic acid.
* Corresponding author.
We have recently reported the activity of a triazene-VPA hybrid
compound, 1a (HYBCOM, Fig. 1), against glioma GL261 cells [10].
The loss of cell viability induced by 1a was preferentially sensed by
glioma cells and did not significantly affect normal astrocytes. In
addition to reduced cell viability, there was also a higher decrease
in cell proliferation when compared to TMZ. Remarkably, compared
E-mail address: mjprocha@ff.ulisboa.pt (M.J. Perry).
https://doi.org/10.1016/j.ejmech.2019.03.048
0223-5234/© 2019 Elsevier Masson SAS. All rights reserved.

C. Braga et al. / European Journal of Medicinal Chemistry 172 (2019) 16e25
17
compounds 1a-k in moderate to good yields.
To predict the BBB penetrance of the novel triazenes, Pfizer's
CNS multiparameter optimization (MPO) score was determined
[13]. CNS MPO scores span from 0 to 6.0 based on the optimal
ranges of several physicochemical properties, and improved BBB
penetration is usually associated to MPO values ꢀ 4. The MPO
values for 1a-k range from 4.0 to 5.5, suggesting that these com-
pounds may achieve effective drug concentrations in the brain
(Table S1).
Fig. 1. The chemical structures of valproic acid (VPA), temozolomide (TMZ), and
HYBCOM.
2.2. Biological evaluation
to TMZ, 1a enhanced the susceptibility of GL261 cells to shape
alteration, from a polar to a non-polar morphology, which may be
related to a decline in cell migratory ability. Moreover, in contrast to
TMZ, cells exposed to 1a maintained drug resistance levels close to
control conditions. To better understand the structural re-
quirements of N-acyltriazenes to target glioma cells, we herein
disclose the structure-activity relationship (SAR) and mechanistic
studies of a series of N-acyltriazenes and their anti-glioma activity.
These hybrid compounds contained differently substituted triazene
moieties and HDAC inhibitor-related short chain fatty acids,
including valproic and butyric acids, and displayed a broad range of
stereoelectronic and lipophilicity properties (Scheme 1, 1a-k).
Biological assays towards the GL261 glioma cell line revealed
analogue 1d as the most potent in the series, with selective tar-
geting of GL261 glioma cells versus SH-SY5Y neuronal cells, and
showing high induction of cell morphological alterations that led to
cell migration impairment.
2.2.1. Cytotoxicity and cell death in GL261 and SH-SY5Y cell lines
We first explored the cytotoxic effects of N-acyltriazenes 1a-k in
the GL261 mouse glioma cell line, at concentrations between 0.1
and 100 mM, using TMZ for comparative purposes (Figure S1). The
resulting IC50 values are presented in Table 1 and reveal that
compounds 1c-g were more potent than both TMZ and 1a in
inducing cell cytotoxicity. Remarkably, valproate derivatives 1d and
1e presented IC50 values of 12 and 7 mM, respectively, i.e., at least
10-fold more potent than the lead compound 1a. Triazene de-
rivatives with shorter acyl moieties 1i-k did not induce appreciable
cytotoxicity in the range of concentrations tested. Overall, from all
the screened compounds, we selected valproate derivates 1d
(R1 ¼ CONH2) and 1e (R1 ¼ CO2Me) at 10
mM for the subsequent
evaluation.
We previously reported that the lead compound 1a, although
affecting cell viability, did not produce significant alteration of
necrosis in GL261 cells [10]. Therefore, we decided to evaluate
whether 1d and 1e showed a similar pattern, by determining the
percentage of viable, early apoptotic and late-apoptotic/necrotic
cells by flow cytometry. As indicated in Fig. 2, hybrid compounds
1a, 1d and 1e decreased cell viability (Fig. 2A) and increased early
apoptosis (Fig. 2B). However, only 1d was able to induce late
apoptosis/necrosis in GL261 cells (Fig. 2C), in addition to being the
one causing the highest loss of cell viability among the tested
compounds (Fig. 2A).
Then, we evaluated if these compounds showed increased cell
death specificity toward glioma cells compared to other cell types,
like the differentiated SH-SY5Y neuronal cell line. Despite their
neuroblastoma cell origin, differentiated SH-SY5Y cells possess a
mature phenotype, with the presence of neuronal markers, such as
2. Results and discussion
2.1. Chemistry
Compounds 1aek were synthesized as depicted in Scheme 1
adapting a previously reported procedure [10]. Diazotization of
commercially available anilines 2, followed by reaction with
formaldehyde and methylamine, afforded the corresponding
hydroxymethyltriazenes 3 [11], which were then converted to the
respective monomethyltriazenes (MMT) 4 in the presence of a
three-fold excess of methylamine [12]. Coupling of MMTs 4 with
the appropriate carboxylic acid using CDI or DCC, combined with 1-
hydroxybenzotriazole (HOBt), afforded the desired hybrid
Scheme 1. Synthetic route for the hybrid compounds 1a-k. (i) NaNO2, conc. HCl, H2O, 0e5 ꢁC; (ii) HCHO/MeNH2, rt (iii) MeNH2, H2O, rt; (iv) R2COOH, 4a-h, coupling reagent/HOBt,
NaH, THF.

18
C. Braga et al. / European Journal of Medicinal Chemistry 172 (2019) 16e25
Table 1
IC50 values against glioma GL261 cells, half-lives (t½) in PBS buffer (10 mM, pH ¼ 7.4) and human plasma (80% v/v) for the hydrolysis to the corresponding mono-
methyltriazenes and anilines, and CNS MPO scores for 1a-k and TMZ.
Compd.
CNS MPO [13]
IC50
(
m
M)a
Half-life t1/2 (h)
pH 7.4 phosphate buffer
80% (v/v) human plasma
1a
1b
1c
1d
1e
1f
1g
1h
1i
4.2
4.0
4.2
4.6
4.1
4.0
4.0
4.0
5.4
5.5
4.8
4.9 [14]
>100b
>100b
>40
34.1 0.4
>40
>40
>40
>40
>40
>40
>40
>40
>40
47.0 3.65
11.9 5.78
6.7 1.48
58.4 6.65
80.6 8.15
>100b
>40
31.4 0.3
27.3 0.3
30.4 0.1
31.3 0.5
>40
>100b
15.2 0.7
13.9 1.1
>40
1j
1k
TMZ
>100b
>40
>40
>100b
>100b
e
e
a
Four independent experiments.
b
100
m
M was the highest concentration tested.
Fig. 2. Changes in glioma and neuronal cell viability by temozolomide (TMZ) and hybrid compounds 1a, 1d and 1e. Mouse GL261 glioma (AeC) and human SH-SY5Y neuronal (DeF)
cell lines were treated with TMZ or 1a, 1d and 1e hybrid compounds at a concentration of 10 M during 48 h. The percentages of viable cells (A,D) and cells in early- (B,E) and late-
m
apoptosis/necrosis (C,F) were determined by flow cytometry with phycoerythrin-conjugated annexin V (annexinV-PE) and 7-amino-actinomycin D (7-AAD). The three populations
were distinguished as follows: viable cells (annexin V-PE and 7-AAD negative), early apoptotic cells (annexinV-PE positive and 7-AAD negative), and cells in late stages of apoptosis
or dead cells (annexinV-PE and 7-AAD positive). All results are means SEM from three independent experiments performed in duplicate. Comparisons were conducted by one-
way ANOVA followed by multiple comparisons Bonferroni post hoc correction. **p < 0.01 vs. non-treated cells (control); $p < 0.05, $$p < 0.01 vs. cells treated with TMZ; #p < 0.05,
##p < 0.01 vs. cells treated with 1a; &p < 0.05, &&p < 0.01 vs. GL261 cells treated with the same compound.
b
III-tubulin, microtubule-associated protein-2 (MAP2), synapto-
cytoskeleton organization in GL261 cells, e.g. by promoting a switch
from a polar to a non-polar morphology and a decline in their
migratory ability properties as observed for 1a [10]. With the
exception of TMZ, all test compounds were able to increase cell
circularity, decreasing their aspect ratio (width/height) (Fig. 3),
which is compatible with a non-polar morphology of the glioma
cells [16]. Importantly, 1d and 1e were more effective than 1a in
causing such morphological alteration.
The observed morphological changes in GL261 cells by the
hybrid compound 1a were previously shown to be associated with
reduced cell migration ability [10]. As indicated in Fig. 4, the same
effect was observed for all the compounds tested in the present
study. Interestingly, 1d was also the most efficient in causing
physin, NeuN, synaptic associated protein-97 (SAP-97) and growth-
associated protein 43 (GAP43) [15]. As indicated in Fig. 2D, we
verified that hybrid compounds 1a, 1d and 1e, as well as TMZ,
reduced neuronal cell viability but with lower efficiency than in
glioma cells (Fig. 2A). Moreover, the induced levels of early
apoptosis and late necrosis on neuronal cells (Fig. 2EeF) were
roughly only one half of those produced in glioma cells (Fig. 2BeC),
revealing that the tested compounds have cell type-specific
toxicity.
2.2.2. Modulation of invasive GL261 cell properties
Next, we evaluated if the N-acyltriazenes were able to modulate

C. Braga et al. / European Journal of Medicinal Chemistry 172 (2019) 16e25
19
Fig. 3. Changes in glioma cell morphology by temozolomide (TMZ) and hybrid compounds 1a, 1d and 1e. Mouse GL261 glioma cells were treated with TMZ or the hybrid com-
pounds at a concentration of 10 M during 48 h. Representative results of one experiment are shown. Cell circularity and aspect ratio were quantified by manual tracings using
ImageJ® software. Results are expressed as mean SEM from three independent experiments. *p < 0.05, **p < 0.01 vs. non-treated cells (control); $$p < 0.01 vs. cells treated with
TMZ; #p < 0.05, ##p < 0.01 vs. cells treated with 1a. The scale bar represents 20 m.
m
m
reduction of glioma cell migration, thus showing to be highly
effective in protection against cell invasiveness. Taken together, our
data indicate 1d and 1e as the most promising compounds in
altering glioma cell morphology and decreasing cell migration
ability, properties that are implicated in glioblastoma cell invasion
and malignancy [17]. Moreover, 1d also demonstrated an elevated
necrotic potential toward glioma cells.
2.4. DNA alkylation assays
In our previous study, the intra- and extracellular 1a concen-
trations were measured by liquid chromatography-tandem mass
spectrometry (LC-MS/MS) and we observed that this compound
was effectively incorporated into the cells, remaining without
degradation until 72 h [10]. Nevertheless, N-Acyltriazenes are
known to undergo deacetylation to the corresponding mono-
methyltriazenes with subsequent generation of the methyl-
diazonium cation [20]. In order to assess the possibility of N-
acyltriazenes exerting their anti-glioma activity via DNA alkylation,
we incubated compound 1a with a selected single stranded oligo-
nucleotide. DNA sequences with runs of purines have been re-
ported to be more reactive towards alkylation by triazene
derivatives [21,22]. In general, the site of greatest alkylation occurs
at runs of two or three contiguous guanines, with the inner gua-
nines showing higher alkylation than isolated guanines. Supported
by these prior observations, the deoxyoligonucleotide base
sequence chosen in this study was designed as 5’ e GGG AGA e 3’.
Incubation with the single-stranded oligonucleotide was assayed at
1:3, 1:10, 1:100 and 1:250 M ratios of oligonucleotide to either
compound 1a or the corresponding monomethyltriazene 4a. No
modification was observed by high resolution mass spectrometry
when the single-stranded oligonucleotide was incubated for 7-days
2.3. HDAC inhibition assays
Since valproic acid is a known, albeit weak, HDAC inhibitor
[18], we decided to evaluate if the ability displayed by N-acyl-
triazenes to induce a non-polar morphology and decrease cell
migration in GL261 cells could be associated to HDAC inhibition.
Compounds 1a-k were studied using commercially available
HDAC fluorometric activity assay (Table S2). All compounds were
very weak HDAC inhibitors, with less than 10% inhibition of
enzyme activity at 6.25 mM. These results are in line with the weak
HDAC inhibitory activity reported for linear and branched short-
chain carboxylic acids, including valproic and butyric acids [19].
Although valproate 1b and butyrate 1i derivates were able to
reduce HDAC activity by ca. 50% at 100 mM, no obvious correlation
emerged between enzyme inhibitory potency and anti-glioma
activity.

20
C. Braga et al. / European Journal of Medicinal Chemistry 172 (2019) 16e25
acylamino acid derivatives of triazenes, where electron-attracting
substituents in the aryltriazene moiety increased the rate of hy-
drolysis, while -branching at the acyl moiety had a detrimental
a
effect on reactivity at pH 7.4 [20,23]. This discrepancy might be
ascribed to different contributions of the hydroxide and buffer-
catalyzed pathways to the hydrolysis of compounds 1 and their
N-acylamino acid counterparts. Worth of note, compound 1a also
showed to be stable when incubated with N-acetylcysteine and
glutathione (Figure S2). Remarkably, valproate derivatives of tri-
azenes, 1a-h, were stable in human plasma, showing no measur-
able decomposition over 2 days. In contrast, the less hindered n-
propyl (butyrate), 1i, and i-propyl, 1j, derivatives were hydrolyzed
to the corresponding monomethyltriazene and aniline with half-
lives of 15 and 14 h, respectively. Introducing an a-ethyl substitu-
ent in the n-propyl side chain, i.e. 1k, led to a dramatic decrease in
the reactivity towards amidases. For the less hindered triazene
derivatives 1i and 1j it was possible to monitor the hydrolysis
conversion into the corresponding products (Figure S3), suggesting
that these compounds have the potential to release the DNA-
alkylating methyldiazonium ion in the blood circulation. Indeed,
serum albumin has been reported to increase the half-lives of the
parent monomethyltriazenes [24]. The high stability of triazenes
1a-h and 1k in human plasma may also result from protein binding,
as plasma protein-bound drugs are partially protected from
hydrolysis.
Further studies must be performed to clarify the impact on the
plasma pseudo-first order hydrolysis rate constants. As previously
reported, decomposition of monomethyltriazenes is catalyzed by
divalent metal ions [25]. Blood concentrations of Cu2þ and Zn2þ are
Fig. 4. Changes in glioma cell migration ability by temozolomide (TMZ) and hybrid
compounds 1a, 1d and 1e. Mouse GL261 glioma cells were treated with TMZ or the
15 mM and 100 mM, respectively [26], and thus these ions could be
hybrid compounds at a concentration of 10 mM during 48 h. Migration was assessed
using a Boyden Chamber. Non-treated (control) and treated cells were allowed to
migrate for 6 h. Representative results of one experiment are shown. The total number
of cells per well was counted and the results represent the average of migrated cells
counted in each well. Results are expressed as mean SEM from four independent
experiments. **p < 0.01 vs. non-treated cells (control); #p < 0.05 vs. cells treated with
an important factor in determining the stability of N-acyltriazenes.
To gain further insight into the effect of metals on the reactivity of
N-acyltriazenes, we monitored by UVeVis spectroscopy the reac-
tion of compounds 1b and 1i and their monomethyltriazene
counterparts 4b and 4a, respectively, in methanolic solutions of
Cu2þ and Zn2þ (Figures S4 and S5). Compounds 1b and 1i did not
decompose to the corresponding anilines 2 in the presence of any of
the metals (Figure S4). In contrast, monomethyltriazenes 4a and 4b
were rapidly converted to the corresponding anilines in the pres-
ence of Cu2þ (Figure S5, A and B). Zinc-catalyzed decomposition
was also observed for 4a and 4b but at a slower rate when
compared with the Cu2þ-catalyzed reaction (Figure S5, C and D), in
line with the reported higher catalytic efficiency of Cu2þ on the
decomposition of monomethyltriazenes [25].
1a. The scale bar represents 20 mm.
with 1a in any of the tested concentrations. In contrast, the mon-
omethyltriazene 4a was able to modify the oligodeoxynucleotide at
a 1:250 M ratio, after 1 h incubation. The identification of two
group of peaks at m/z 946.1857 and 630.4559, assigned to a doubly-
and a triply-charged methylated oligonucleotide species, respec-
tively, indicates the formation of a monomethylated adduct (O6-
MeG). However, the data could not unequivocally establish which
guanine was methylated. The ESI(ꢂ)/HRMS mass spectra, together
with the accurate mass measurements for the identified species,
are presented in Fig. 5. These results clearly indicate that compound
1a does not alkylate the oligonucleotide, while its mono-
methyltriazene 4a counterpart was able to promote methylation in
the O6 position of guanine base(s) in vitro.
3. Conclusions
The studies herein presented led to the identification of novel
hybrid compounds with a remarkable increase in potency toward
glioma GL261 cells compared to the lead compound 1a. Cytotox-
icity assays highlighted 1d and 1e as having the highest anti-tumor
2.5. Stability evaluation
efficacy, including high potency (IC50 11.9 mM and 6.7 mM, respec-
tively) together with an effective decrease of invasive cell proper-
ties. CNS MPO scores suggest that 1d and 1e are BBB-penetrant. In
contrast to temozolomide, which undergoes hydrolysis to release
the alkylating metabolite monomethyltriazene, the valproate de-
rivatives studied 1a-h showed enhanced chemical and metabolic
stability combined with a low potential to alkylate DNA. Triazenes 1
revealed to be weak HDAC inhibitors, suggesting that this class of
enzymes is not the main target for these compounds. Overall, these
results validate valproate-triazene hybrid compounds as a new
class of anti-glioma agents with clear chemotherapeutic advan-
tages over temozolomide, and thus with potential to be used in the
chemotherapy of glioblastoma.
The lack of oligonucleotide methylation observed with 1a
prompted us to study the chemical stability of compounds 1a-k in
pH 7.4 buffer, at 37 ꢁC. Data depicted in Table 1 reveal that com-
pounds 1a-k hydrolyze very slowly to the corresponding mono-
methyltriazenes and anilines at physiological pH, with half-live
values higher than 25 h, i.e., the tested compounds are significantly
more stable than TMZ and its derivatives (e.g. t1/2 > 40 h for 1d
when compared to 2 h for TMZ [14] in PBS). Overall, the rate of
hydrolysis at pH 7.4 was not significantly affected by the nature of
substituents in the aryltriazene moiety or the
a-branching at the
acyl moiety. These results contrast with those reported for N-

C. Braga et al. / European Journal of Medicinal Chemistry 172 (2019) 16e25
21
Fig. 5. Comparison of the ESI(ꢂ)/HRMS profiles of 1a (HYBCOM) and 4a incubated with 5’ e GGG AGA e 30 in ammonium bicarbonate buffer (10 mM, pH 8.3) at 37 ꢁC.
4. Experimental
100e2000 m/z with an acquisition rate of 1 Hz. Data was processed
using Data Analysis 4.2 software.
4.1. General information
Compounds 3 and 4 were synthesized by previously published
methods [11,12]. 50eGGG AGAe30 was purchased from STAB VIDA
(Lisbon, Portugal) and used without further purification. The
All reagents and solvents were purchased from commercial
suppliers. All solvents were purified by standard techniques. Re-
actions were monitored by thin-layer chromatography using silica-
gel aluminum sheets (Merck Kieselgel 60 F254). Column chroma-
tography was performed using Merck silica gel 60 (230e400 mesh
ASTM). Melting points were determined using a Kofler camera
Bock-Monoscop ‘‘M’’ and are uncorrected. 1H and 13C NMR spectra
were recorded in CDCl3 solutions using a Bruker Ultra-Shield
oligonucleotide was dissolved in water and 100 mM stock solutions
were prepared and kept at ꢂ18 ꢁC until required.
4.2. General procedure for the synthesis of compounds 1
Method A: CDI-mediated amide coupling (1b-d, 1f-g and 1i-k).
To a solution of the appropriate acid (0.55 mmol) in 2 mL of THF,
was added anhydrous HOBt (81 mg, 0.6 mmol). The reaction
mixture was stirred at room temperature for 15 min. Subsequently,
CDI (97 mg, 0.6 mmol) was added in one portion, turning the so-
lution from milky white to colorless, with CO2 release. The solution
was stirred for 45 min, under nitrogen atmosphere. Triethylamine
300 MHz spectrometer; chemical shifts (d) are reported in parts
per million (ppm) relative to tetramethylsilane (TMS) as internal
standard and coupling constants (J) are expressed in hertz (Hz).
High-performance liquid chromatography was performed in an
Elite LaChrom (VWR Hitachi) system comprised of an L-2130 pump,
an L-2400 UV detector and a Rheodyne® manual sample injector,
(166 mL, 1.2 mmol) and monomethyltriazene (0.6 mmol), previously
including a 20
m
L sample loop. Chromatographic separation was
activated with NaH (60% dispersion in mineral oil, 24 mg, 0.6 mmol)
in methylene chloride (5 mL) were added and the reaction mixture
was stirred for 48 h. When complete, the reaction mixture was
quenched with 1 mL of 1 M HCl and stirred vigorously for 10 min,
and the aqueous layer was extracted with methylene chloride
(2 ꢃ 2 mL). The combined organic layers were washed sequentially
with 1 M HCl (2 mL), deionized water (2 mL), and a 1:1 aqueous
mixture of brine and saturated NaHCO3 (4 mL). The organic layer
was dried with Na2SO4 and the solvent was removed under
reduced pressure to give the crude product. All compounds were
purified by column chromatography and obtained with purity
>95%.
achieved on a Lichrospher® 100 RP-18 (5
mm) using an isocratic
solvent elution at a flow rate of 1.0 mL min-1. Water/acetonitrile
(35:65) was chosen for compounds 1a and 1i-k, while water/
acetonitrile (20:80) was preferred for compounds 1b-h. Purity of
compounds as determined by HPLC, was higher than 95%. Ac-
cording to the lmax of compounds, the UV detector was set at
300 nm. Low resolution mass spectra were recorded using a
Micromass Quattro Micro API spectrometer. High-resolution mass
spectra (HRMS) were acquired on a QqTOF Impact II (Bruker Dal-
toniks) mass spectrometer equipped with an electrospray source
operating in positive and negative mode, interfaced in line with a
Dionex UltiMate® 3000 RSLCnano (Thermo Fisher Scientific Inc.).
Method B: DCC-mediated amide coupling (1e and 1h). Similar to
Method A but using DCC as coupling agent.
Aliquots (10 mL) of the samples were analyzed by flow injection
analysis (FIA) using an isocratic gradient 50 A:50 B of 0.1% formic
acid in water (A) and 0.1% of formic acid in acetonitrile (B), at a flow
4.2.1. 1-(4-cyanophenyl)-3-methyl-3-[2-(propyl)pentanoyl]
triazene (1a)
rate of 10 m
L minꢂ1, over 15 min. The mass spectrometer parameters
were set as follows: capillary voltage: þ4.5 and ꢂ2,5 kV (positive
and negative mode, respectively); end plate offset: 500 V; nebu-
lizer: 40 psi; dry gas: 4.0 L minꢂ1; dry heather: 200 ꢁC. Prior to
analysis, the TOF analyzer was calibrated by infusion of ESI Low
Synthesis and purification were followed according to previous
reported procedures [10].
4.2.2. 1-(4-ethoxycarbonylphenyl)-3-methyl-3-[2-(propyl)
pentanoyl]triazene (1b)
Concentration Tune Mix (Agilent Technologies) at a 15 m
L min-1
flow rate. Recalibration of acquired data was performed via lock
mass internal calibration (1221.9906) located within the source.
Acquisition was performed in full scan mode over a mass range of
Method A was followed using 1-(4-ethoxycarbonylphenyl)-3-
methyltriazene and sodium valproate. The final mixture was puri-
fied by column chromatography (methylene chloride) to yield the

22
C. Braga et al. / European Journal of Medicinal Chemistry 172 (2019) 16e25
product as a brown oil (43.3%). 1H NMR (300 MHz, CDCl3)
d
8.13
yield the product as a brown oil (46.1%). 1H NMR (300 MHz, CDCl3)
7.57 (2H, d, J ¼ 8.8 Hz), 7.47 (2H, d, J ¼ 8.8 Hz), 3.65e3.74 (1H, m),
(2H, d, J ¼ 8.8 Hz), 7.63 (2H, d, J ¼ 8.8 Hz), 4.39 (2H, q, J ¼ 7.1 Hz),
3.69e3.75 (1H, m), 3.47 (3H, s), 1.70e1.83 (2H, m), 1.46e1.55 (2H,
m), 1.41 (3H, t, J ¼ 7.1 Hz), 1.24e1.33 (4H, m), 0.88 (6H, t, J ¼ 7.3 Hz);
d
3.44 (3H, s), 1.69e1.82 (2H, m), 1.45e1.57 (2H, m), 1.22e1.35 (4H,
m), 0.87 (6H, t, J ¼ 7.3 Hz); 13C NMR (75 MHz, CDCl3)
d 179.26,
13C NMR (75 MHz, CDCl3)
d
179.42, 166.19, 152.28, 130.87, 130.52,
148.00, 132.47, 123.64, 122.77, 41.88, 35.16, 27.95, 20.84, 14.28; HR-
ESI(þ)/MS: m/z calcd for C15H22BrN3O [M þ Na]þ: 362.0836; found
362.0848/364.0824.
121.97, 61.30, 41.96, 35.16, 28.11, 20.87, 14.47, 14.28; HR-ESI(þ)/MS:
m/z calcd for C18H28N3O3 [M þ H]þ: 334.2125; found 334.2129.
4.2.3. 1-(4-acetylphenyl)-3-methyl-3-[2-(propyl)pentanoyl]
triazene (1c)
4.2.8. 3-Methyl-3-[2-(propyl)pentanoyl]-1-(4-tolyl)triazene (1h)
Method B was followed using 3-methyl-1-(p-tolyl)triazene and
sodium valproate. The final mixture was purified by column chro-
matography (hexane: ethyl acetate - 4:1) to yield the product as a
Method
A
was followed using 1-(4-acetylphenyl)-3-
methyltriazene and sodium valproate. The final mixture was puri-
fied by column chromatography (methylene chloride) to yield the
brown oil (yield 27.5%); 1H NMR (300 MHz, CDCl3)
d 7.51 (2H, d,
product as an orange oil (37.8%); 1H NMR (300 MHz, CDCl3)
d
8.05
J ¼ 8.3 Hz), 7.25 (2H, d, J ¼ 7.9 Hz), 3.68e3.79 (1H, m), 3.44 (3H, s),
(2H, d, J ¼ 8.8 Hz), 7.66 (2H, d, J ¼ 8.8 Hz), 3.67e3.76 (1H, m), 3.47
2.40 (3H, s) 1.71e1.83 (2H, m), 1.46e1.57 (2H, m), 1.24e1.36 (4H, m),
(3H, s), 2.63 (3H, s), 1.70e1.83 (2H, m), 1.47e1.58 (2H, m), 1.23e1.36
0.88 (6H, t, J ¼ 7.2 Hz); 13C NMR (75 MHz, CDCl3)
d 179.36, 146.91,
(4H, m), 0.88 (6H, t, J ¼ 7.3 Hz); 13C NMR (75 MHz, CDCl3)
d
197.37,
139.19, 129.93, 122.03, 41.82, 35.22, 27.70, 21.38, 20.86, 14.29; HR-
ESI(þ)/MS: m/z calcd for C16H25N3O [M þ Na]þ: 298.1890; found
298.1895.
179.38, 152.35, 136.93, 129.71, 122.21, 41.95, 35.15, 28.15, 26.85,
20.86, 14.28; HR-ESI(þ)/MS: m/z calcd for C17H26N3O2 [M þ H]þ:
304.2020; found 304.2025.
4.2.9. 3-Butyryl-1-(4-cyanophenyl)-3-methyltriazene (1i)
4.2.4. 1-(4-carbamoylphenyl)-3-methyl-3-[2-(propyl)pentanoyl]
Method
A
was followed using 1-(4-cyanophenyl)-3-
triazene (1d)
methyltriazene and butyric acid. The final mixture was purified
by column chromatography (methylene chloride) to yield the
product as a yellow solid (yield 22.9%); m.p. 55e57 ꢁC; 1H NMR
Method
A was followed using 1-(4-carbamoylphenyl)-3-
methyltriazene and sodium valproate. The final mixture was puri-
fied by column chromatography (ethyl acetate) to yield the product
as a yellow solid (56.2%); m.p. 105e107 ꢁC; 1H NMR (300 MHz,
(300 MHz, CDCl3)
d
7.74 (2H, d, J ¼ 8.8 Hz), 7.66 (2H, d, J ¼ 8.8 Hz),
3.44 (3H, s), 2.91 (2H, t, J ¼ 7.4 Hz), 1.72e1.84 (2H, m), 1.02 (3H, t,
CDCl3)
d
7.92 (2H, d, J ¼ 8.7 Hz), 7.64 (2H, d, J ¼ 8.7 Hz), 6.36 (2H, bs),
J ¼ 7.4 Hz); 13C NMR (75 MHz, CDCl3)
d 175.75, 151.97, 133.45,
3.68e3.74 (1H, m), 3.46 (3H, s), 1.69e1.81 (2H, m), 1.46e1.57 (2H,
122.78, 118.68, 112.10, 36.21, 28.19, 18.51, 14.04; HR-ESI(þ)/MS: m/z
m), 1.22e1.34 (4H, m), 0.87 (6H, t, J ¼ 7.3 Hz); 13C NMR (75 MHz,
calcd for C12H14N4O [M þ Na]þ: 253.1060; found 253.1059.
CDCl3)
d 179.41, 169.03, 151.63, 133.36, 128.70, 122.23, 41.92, 35.13,
28.08, 20.83, 14.26; HR-ESI(þ)/MS: m/z calcd for C16H25N4O2 [M þ
4.2.10. 1-(4-cyanophenyl)-3-isobutyryl-3-methyltriazene (1j)
H]þ: 305.1972; found 305.1980.
Method
A
was followed using 1-(4-cyanophenyl)-3-
methyltriazene and isobutyric acid. The final mixture was puri-
fied by column chromatography (methylene chloride) to yield the
product as an orange solid (yield 32.4%); m.p. 110e112 ꢁC; 1H NMR
4.2.5. 1-(4-methoxycarbonylphenyl)-3-methyl-3-[2-(propyl)
pentanoyl]triazene (1e)
Method B was followed using 1-(4-methoxycarbonylphenyl)-3-
methyltriazene and sodium valproate. The final mixture was puri-
fied by column chromatography (methylene chloride) to yield the
(300 MHz, CDCl3)
d
7.74 (2H, d, J ¼ 8.8 Hz), 7.67 (2H, d, J ¼ 8.8 Hz),
3.66 (1H, hept, J ¼ 6.9 Hz), 3.45 (3H, s), 1.26 (6H, d, J ¼ 6.9 Hz); 13C
NMR (75 MHz, CDCl3)
d 179.68, 152.04, 133.48, 122.78, 118.69,
product as an orange oil (33.0%); 1H NMR (300 MHz, CDCl3)
d
8.13
112.12, 32.10, 28.41, 19.41; HR-ESI(þ)/MS: m/z calcd for C12H14N4O
(2H, d, J ¼ 8.8 Hz), 7.64 (2H, J ¼ 8.8 Hz), 3.94 (3H, s), 3.69e3.75 (1H,
[M þ Na]þ: 253.1060; found 253.1063.
m), 3.47 (3H, s),1.71e1.83 (2H, m),1.47e1.58 (2H, m),1.23e1.36 (4H,
m), 0.88 (6H, t, J ¼ 7.3 Hz); 13C NMR (75 MHz, CDCl3)
d
179.37,
4.2.11. 1-(4-cyanophenyl)-3-[2(ethyl)butyryl]-3-methyltriazene
(1k)
166.63, 152.33, 130.90, 130.14, 122.00, 52.38, 41.95, 35.14, 28.09,
20.84, 14.26; HR-ESI(þ)/MS: m/z calcd for C17H26N3O3 [M þ H]þ:
320.1969; found 320.1974.
Method
A
was followed using 1-(4-cyanophenyl)-3-
methyltriazene and 2-ethylbutyric acid. The final mixture was pu-
rified by column chromatography (methylene chloride) to yield the
product as a yellow oil (yield 38.0%); 1H NMR (300 MHz, CDCl3)
4.2.6. 1-(4-chlorophenyl)-3-methyl-3-[2-(propyl)pentanoyl]
triazene (1f)
d
7.74 (2H, d, J ¼ 8.8 Hz), 7.66 (2H, d, J ¼ 8.8 Hz), 3.45e3.54 (1H, m),
Method
A
was followed using 1-(4-chlorophenyl)-3-
3.47 (3H, s), 1.72e1.86 (2H, m), 1.55e1.69 (2H, m), 0.89 (6H, t,
methyltriazene and sodium valproate. The final mixture was puri-
fied by column chromatography (hexane: ethyl acetate - 4:1) to
yield the product as a brown oil (87.2%); 1H NMR (300 MHz, CDCl3)
J ¼ 7.4 Hz); 13C NMR (75 MHz, CDCl3)
d 178.79, 152.07, 133.49,
122.81, 118.70, 112.09, 45.65, 28.31, 25.39, 12.04. HR-ESI(þ)/MS: m/z
calcd for C14H18N4O [M þ Na]þ: 281.1373; found 281.1378.
d
7.54 (2H, J ¼ 8.9 Hz), 7.41 (2H, d, J ¼ 8.9 Hz), 3.66e3.75 (1H, m),
3.44 (3H, s), 1.69e1.82 (2H, m), 1.45e1.57 (2H, m), 1.22e1.34 (4H,
4.3. Biology
m), 0.87 (6H, t, J ¼ 7.3 Hz); 13C NMR (75 MHz, CDCl3)
d 179.29,
147.58, 134.63, 129.51, 123.35, 41.89, 35.18, 27.94, 20.86, 14.29; HR-
ESI(þ)/MS: m/z calcd for C15H22ClN3O [M þ Na]þ: 318.1344; found
318.1350/320.1324.
4.3.1. Cell cultures
The GL261 cell line was a kind gift from Dr. Geza Safrany, from
the National Research Institute for Radiobiology and Radio-
hygiene, in Hungary. This cell line is representative of
a
4.2.7. 1-(4-bromophenyl)-3-methyl-[2-(propyl)pentanoyl]triazene
(1g)
carcinogen-induced mouse syngeneic glioma model, which
carries point mutations in the K-ras and p53 genes [27], and has
features of cancer stem cells, namely elevated expression of CD133
[28]. GL261 cells were seeded in Dulbecco's Modified Eagle Me-
dium (DMEM) supplemented with 11 mM sodium bicarbonate,
Method
A
was followed using 1-(4-bromophenyl)-3-
methyltriazene and sodium valproate. The final mixture was puri-
fied by column chromatography (hexane: ethyl acetate - 4:1) to

C. Braga et al. / European Journal of Medicinal Chemistry 172 (2019) 16e25
23
38.9 mM glucose, 10% fetal bovine serum (FBS) and 1% penicillin/
streptomycin (all from Biochrom), and maintained at 37 ꢁC in a
humidified atmosphere of 5% CO2. Cells were plated at
5 ꢃ 104 cells/mL 24 h before treatments as previously described by
us [10]. Human neuroblastoma cells SH-SY5Y (SH) were a gift from
Professor Anthony Turner [29]. Cells were cultured in DMEM
supplemented with 10% FBS and 2% Antibiotic/Antimycotic (AB/
AM, Sigma-Aldrich, Sigma). The medium was changed every 2
evaluate GL261 cell morphology, cells were stained with Phalloidin
(which binds F-actin with high selectivity) conjugated with Alexa
Fluor-594 antibody (1:50, #A12381, Thermo Fisher Scientific). Cell
nuclei were stained with Hoechst 33258 dye (blue, Sigma-Aldrich)
and fluorescence was visualized using an AxioCam HR camera
adapted to an AxioScope A1® microscope (Zeiss, Germany). Merged
images of UV and fluorescence of ten random microscopic fields
were acquired per sample using Zen 2012 (blue edition, Zeiss)
software. The original magnification was 630ꢃ. For morphological
days. Culture plates were coated with Poly-
D
-Lysine (100
mg/mL)
and laminin (4
m
g/mL, Gibco) before plating the cells. For the ex-
characterization we used the tracing analysis in ImageJ® software
periments, cells were seeded at a concentration of 5 ꢃ 104 cells/mL
and maintained at 37 ꢁC in a humidified atmosphere of 5% CO2.
After 24 h proliferation, differentiation into a neuronal phenotype
was induced by adding retinoic acid (RA, Sigma-Aldrich) at a final
(1.49 version, NIH, USA) to measure cell circularity ð4p Perimeter2Þ
Area
and aspect ratio ðmajor axis
minor axisÞ. The average of at least 120 cells was
considered for in each condition.
concentration of 10
m
M in the culture medium and maintaining
4.3.6. Glioma cell migration assay
cells for 7 days. RA-containing culture medium was changed every
2 days.
GL261 glioma cells were treated with TMZ or selected hybrid
compounds at a concentration of 10 M during 48 h. After incu-
m
bation, a cell migration assay was performed in a 48-well
microchemotaxis Boyden chamber (#AP48, Neuro Probe) by us-
4.3.2. Cell treatments
Glioma cells were treated with 0.1e100
m
M of the compounds
ing
vinylpyrrolidone (PVP) surface. Each well of the lower chamber
compartment was filled with 25
L DMEM þ EGF (10%) to induce
8 mm diameter polycarbonate membranes with poly-
1b-k during 48 h. Compounds 1a and TMZ (Sigma-Aldrich) were
also used for comparative purposes, as described in our previous
work [10]. Taking into account the results of the cell viability assays,
we selected the most efficient compounds and concentrations to be
used in the subsequent analyses. SH-SY-5Y cells were also treated
with the selected compounds and concentrations only, to perform
cell death assays.
m
chemotaxis. Non-treated (control) and treated cells were allowed
to migrate for 6 h in a CO2 incubator at 37 ꢁC. After this period,
cells attached to lower side of the membrane were fixed with cold
methanol and stained with 10% Giemsa in PBS. Non-migrated
cells on the upper side of the membrane were wiped off with a
filter wiper. The rate of migration was determined by counting
cells on the lower membrane surface. At least six bright field
images per sample were acquired using an AxioCam 105 color
camera (Zeiss) adapted to an Axioskope HBO50 microscope
(Zeiss). For each experiment, at least three wells per condition
were analyzed.
4.3.3. Cell viability assay by MTS
Cell viability of the GL261 cells was assessed using
the cellular reduction of [3-(4,5-dimethylthiazol-2-yl)-5-(3-
carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium] (MTS)
by viable cells into an aqueous, soluble and coloured formazan
product, in the presence of phenazine methosulfate (PMS). After
exposure to compounds 1a-k or TMZ, cells were incubated for 1 h at
37 ꢁC with an MTS:PMS mixture (1:20, from stock solutions of 2 mg/
mL and 0.92 mg/mL, respectively) diluted 1/10 in DMEM-F12 and the
absorbance was measured at 490 nm using a PR 2100 microplate
reader (Bio-Rad Laboratories). For each experiment, the mean value
of absorbance obtained for non-treated cells (control) was consid-
ered as 100% of cell viability.
4.4. Incubation of 1a and 4a with 50e GGG AGA e 30
Compound 1a (30
with the single-stranded oligonucleotide (10
bicarbonate buffer (10 mM, pH 8.3) at 37 ꢁC. Aliquots (50
samples were withdrawn at regular time intervals and each
mixture was extracted three times with 50 L of ethyl acetate. Upon
mM, 100
m
M, 1 mM, 2.5 mM) was incubated
M) in ammonium
L) of the
m
m
m
vortexing, the mixtures were centrifuged, at 13300 rpm for 3 min
and the aqueous phase was collected for MS analysis. Experiments
were carried out in triplicate.
4.3.4. Determination of cell death by necrosis and apoptosis
GL261 glioma and SH-SY5Y-neuron cell lines were treated with
TMZ or selected hybrid compounds at a concentration of 10 mM
during 48 h. A phycoerythrin-conjugated annexin V (V-PE) and 7-
aminoactinomycin-D (7-AAD) mixture (Guava Nexin® Reagent,
Millipore, MA, USA) was used to determine the percentage of
viable, early-apoptotic and late-apoptotic/necrotic cells by flow
cytometry. After incubation, adherent cells were collected by
trypsinization and added to the cells present in the incubation
media. After centrifugation, the cell pellet was resuspended in PBS
containing 1% bovine serum albumin (BSA), stained with Guava
Nexin® Reagent according to manufacturer's instructions, and
analyzed on a Guava easyCyte 5HT flow cytometer (Guava Nexin®
Software module, Millipore). Two readings were performed for
each sample.
4.5. Hydrolysis in phosphate buffered saline
Isotonic phosphate buffer (10 mM, pH 7.4, 4 mL) and ACN (1 mL)
were added into a Falcon tube placed in a water bath thermostated
at 37 ꢁC. After temperature stabilization, 150
mL of stock solution of
1a-k in ACN (10 mM) were added. Several aliquots were taken at
regular times, injected and analyzed by HPLC. These experiments
were carried out in duplicate.
4.6. Hydrolysis in human plasma
Human plasma from 13 healthy donors was collected in sodium
heparinate, divided into 2 mL fractions and stored at ꢂ18 ꢁC until
required. Before usage, frozen samples were thawed gradually to
avoid denaturation of plasma proteins. Each 2 mL aliquot of human
plasma was diluted to 80% (v/v) by the addition of 0.5 mL of PBS
(10 mM, pH 7.4) and the mixture was placed in a water bath ther-
4.3.5. Glioma cell morphological assay
GL261 glioma cells were treated with TMZ or selected hybrid
compounds at a concentration of 10 mM during 48 h. After incu-
bation, cells in coverslips were fixed with freshly prepared 4% (w/v)
paraformaldehyde in PBS, permeabilized with 0.1% Triton X-100
Solution and incubated with blocking solution (10% FBS in Tween-
20-Tris-buffered saline) during 30 min at room temperature. To
mostated at 37 ꢁC. After temperature stabilization, 75
stock solution in ACN (10 mM) were added. Aliquots of 100
taken at regular times and transferred to eppendorfs containing
m
L of a 1a-k
L were
m

24
C. Braga et al. / European Journal of Medicinal Chemistry 172 (2019) 16e25
preclinical rationale and published clinical results, Expert Opin. Investig.
200 mL ACN to precipitate the plasma proteins. After vortexing,
Drugs 21 (2012) 1391e1415, https://doi.org/10.1517/13543784.2012.694425.
[8] A.V. Krauze, S.D. Myrehaug, M.G. Chang, D.J. Holdford, S. Smith, J. Shih,
P.J. Tofilon, H.A. Fine, K. Camphausen, A phase 2 study of concurrent radiation
therapy, temozolomide, and the histone deacetylase inhibitor valproic acid for
patients with glioblastoma, Int. J. Radiat. Oncol. Biol. Phys. 92 (2015) 986e992,
https://doi.org/10.1016/j.ijrobp.2015.04.038.
followed by centrifugation at 12000 rpm for 10 min, the superna-
tants were collected and analyzed by HPLC. These experiments
were carried out in duplicate for 1a-h and 1k and in triplicate for
1i-j.
[9] K.A. Van Nifterik, J. Van den Berg, B.J. Slotman, M.V.M. Lafleur, P. Sminia,
L.J.A. Stalpers, Valproic acid sensitizes human glioma cells for temozolomide
4.7. HDAC inhibition assay
and g-radiation, J. Neuro Oncol. 107 (2012) 61e67, https://doi.org/10.1007/
s11060-011-0725-z.
HDAC inhibitory activity of N-acyltriazenes 1a-k was deter-
mined using the Fluor-de-Lys HDAC activity assay kit (Enzoli-
fesciences, BML-AK500-0001). The assay was performed according
to the manufacturer's instructions. The source of HDAC enzymatic
activity is a HeLa nuclear extract, provided with the kit. Compounds
1a-k were resuspended in DMSO (20 mM) and assayed in duplicate
[10] R. Pinheiro, C. Braga, G. Santos, M.R. Bronze, M.J. Perry, R. Moreira, D. Brites,
~
A.S. Falcao, Targeting gliomas: can a new alkylating hybrid compound make a
difference? ACS Chem. Neurosci. 8 (2017) 50e59, https://doi.org/10.1021/
acschemneuro.6b00169.
[11] T.P. Ahern, H. Fong, K. Vaughan, Open-chain nitrogen compounds. Part II.
Preparation, characterization, and degradation of 1(3)-Aryl-3(1)-methyl-
triazenes; the effect of substituents on the reaction of diazonium salts with
methylamine, Can. J. Chem. 55 (1977) 1701e1709, https://doi.org/10.1139/
v77-240.
at 100, 25 and 6.25 mM. Trichostatin A and GNX-4784 [30] were
used as positive controls. The reaction was starting by adding HDAC
fluorometric substrate [Boc-Lys(Ac)-AMC], which contains acety-
lated lysine, to the inhibitors and HeLa nuclear extract mixture.
Deacetylation sensitized the substrate, and treatment with the
lysine developer produced the fluorophore, and fluorescence was
measured after 15 min of incubation at excitation 360 nm and
emission 460 nm.
[12] S.C. Cheng, J. Iley, Synthesis of 1-aryl-3-methyltriazenes by base-promoted
decomposition of 1-aryl-3-hydroxymethyl-3-methyltriazenes, J. Chem. Res.
(1983) 320e321.
[13] T.T. Wager, X. Hou, P.R. Verhoest, A. Villalobos, Moving beyond rules: the
development of a central nervous system multiparameter optimization (CNS
MPO) approach to enable alignment of drug like properties, ACS Chem.
Neurosci. 1 (2010) 435e449, https://doi.org/10.1021/cn100008c.
[14] R.L. Svec, L. Furiassi, C.G. Skibinski, T.M. Fan, G.J. Riggins, P.J. Hergenrother,
Tunable stability of imidazotetrazines leads to a potent compound for glio-
blastoma, ACS Chem. Biol. 13 (2018) 3206e3216, https://doi.org/10.1021/
acschembio.8b00864.
[15] J. Kovalevich, D. Langford, Considerations for the use of SH-SY5Y neuroblas-
toma cells in neurobiology, Methods Mol. Biol. 1078 (2013) 9e21, https://
doi.org/10.1007/978-1-62703-640-5_2.
[16] S.P. Carey, C.M. Kraning-Rush, R.M. Williams, C.A. Reinhart-King, Biophysical
control of invasive tumor cell behavior by extracellular matrix micro-
architecture, Biomaterials 33 (2012) 4157e4165, https://doi.org/10.1016/
j.biomaterials.2012.02.029.
[17] I. Koh, J. Cha, J. Park, J. Choi, S.-G. Kang, P. Kim, The mode and dynamics of
glioblastoma cell invasion into a decellularized tissue-derived extracellular
matrix-based three-dimensional tumor model, Sci. Rep. 8 (2018), https://
doi.org/10.1038/s41598-018-22681-3.
[18] M.-S.I. Abaza, A.-M. Bahman, R.J. Al-Attiyah, Valproic acid, an anti-epileptic
drug and a histone deacetylase inhibitor, in combination with proteasome
inhibitors exerts antiproliferative, pro-apoptotic and chemosensitizing ef-
fects in human colorectal cancer cells: underlying molecular mechanisms,
Int. J. Mol. Med. 34 (2014) 513e532, https://doi.org/10.3892/ijmm.2014.
1795.
[19] K.M. Gilbert, A. DeLoose, J.L. Valentine, E.K. Fifer, Structureeactivity rela-
tionship between carboxylic acids and T cell cycle blockade, Life Sci. 78 (2006)
2159e2165, https://doi.org/10.1016/j.lfs.2005.09.047.
[20] M. de J. Perry, E. Carvalho, E. Rosa, J. Iley, Towards an efficient prodrug of the
alkylating metabolite monomethyltriazene: synthesis and stability of N-acy-
lamino acid derivatives of triazenes, Eur. J. Med. Chem. 44 (2009) 1049e1056,
https://doi.org/10.1016/J.EJMECH.2008.06.022.
Acknowledgements
~
^
This work was supported by Fundaçao para a Ciencia e Tecno-
logia (FCT, Portugal) through grants UID/00100/2013 to Centro de
Química Estrutural and UID/04138/2013 to iMed.ULisboa. Joint
funding from FCT and the COMPETE Program through grant
SAICTPAC/0019/2015 is gratefully acknowledged. CB also thanks
FCT for a PhD fellowship under the MedChemTrain program (PD/
BD/128239/2016). We also acknowledge the financial support from
FCT and Portugal 2020 to the National Mass Spectrometry Network
(RNEM LISBOA-01-0145-FEDER-402-022125 IST).
We kindly thank S. Vultaggio and C. Mercurio from IFOM - FIRC
Institute of Molecular Oncology Foundation (Milano, Italy) for the
HDAC activity measurements.
Appendix A. Supplementary data
Supplementary data related to this article can be found at
https://doi.org/10.1016/j.ejmech.2019.03.048.
[21] M.B. Kroeger Smith, L.A. Taneyhill, C.J. Michejda, R.H.J. Smith, Base sequence
selectivity in the alkylation of DNA by 1,3-Dialkyl-3-acyltriazenes, Chem. Res.
Toxicol. 9 (1996) 341e348, https://doi.org/10.1021/tx9500742.
[22] J.A. Hartley, W.B. Mattes, K. Vaughan, N.W. Gibson, DNA sequence specificity
of guanine N7-alkylations for a series of structurally related triazenes, Carci-
nogenesis 9 (1988) 669e674.
References
[1] R.L. Siegel, K.D. Miller, A. Jemal, Cancer statistics, CA Cancer J. Clin. 66 (2016)
7e30, https://doi.org/10.3322/caac.21332.
[23] A.S. Monteiro, J. Almeida, G. Cabral, P. Severino, P.A. Videira, A. Sousa,
R. Nunes, J.D. Pereira, A.P. Francisco, M.J. Perry, E. Mendes, Synthesis and
evaluation of N-acylamino acids derivatives of triazenes. Activation by
tyrosinase in human melanoma cell lines, Eur. J. Med. Chem. 70 (2013) 1e9,
https://doi.org/10.1016/j.ejmech.2013.09.040.
[24] F. Delben, S. Paoletti, G. Manzini, C. Nisi, Influence of serum albumins on
decomposition rates of para-substituted 1-phenyl-3-methyltriazenes and 5-
(3-methyl-1-triazeno)imidazole-4-carboxamide in near physiological condi-
tions, J. Pharm. Sci. 70 (1981) 892e897.
[25] J. Iley, R. Moreira, E. Rosa, Triazene drug metabolites. Part 10. Metal-ion cat-
alysed decomposition of monoalkyltriazenes in ethanol solutions, J. Chem.
Soc. Perkin Trans. 2 (1991) 81e86, https://doi.org/10.1039/p29910000081.
[26] Biology Data Book, second ed., Federation of American Societies for Experi-
mental Biology, Bethesda, Maryland, 1974.
[2] Q.T. Ostrom, H. Gittleman, P.M. de Blank, J.L. Finlay, J.G. Gurney, R. McKean-
Cowdin, D.S. Stearns, J.E. Wolff, M. Liu, Y. Wolinsky, C. Kruchko, J.S. Barnholtz-
Sloan, American brain tumor association adolescent and young adult primary
brain and central nervous system tumors diagnosed in the United States in
2008-2012, Neuro Oncol. 18 (2016) i1ei50, https://doi.org/10.1093/neuonc/
nov297.
[3] R. Stupp, W.P. Mason, M.J. van den Bent, M. Weller, B. Fisher, M.J.B. Taphoorn,
K. Belanger, A.A. Brandes, C. Marosi, U. Bogdahn, J. Curschmann, R.C. Janzer,
S.K. Ludwin, T. Gorlia, A. Allgeier, D. Lacombe, J.G. Cairncross, E. Eisenhauer,
R.O. Mirimanoff, European organisation for Research and treatment of cancer
brain tumor and radiotherapy groups, national cancer Institute of Canada
clinical trials group, radiotherapy plus concomitant and adjuvant temozolo-
mide for glioblastoma, N. Engl. J. Med. 352 (2005) 987e996, https://doi.org/
10.1056/NEJMoa043330.
[4] S.S. Agarwala, J.M. Kirkwood, Temozolomide, a novel alkylating agent with
activity in the central nervous system, may improve the treatment of
advanced metastatic melanoma, Oncol. 5 (2000) 144e151.
[5] C.L. Moody, R.T. Wheelhouse, The medicinal chemistry of imidazotetrazine
prodrugs, Pharmaceuticals (Basel) 7 (2014) 797e838, https://doi.org/10.3390/
ph7070797.
ꢀ
ꢀ
[27] T. Szatmari, K. Lumniczky, S. Desaknai, S. Trajcevski, E.J. Hídvegi, H. Hamada,
ꢀ
ꢀ
G. Safrany, Detailed characterization of the mouse glioma 261 tumor model
for experimental glioblastoma therapy, Cancer Sci. 97 (2006) 546e553,
https://doi.org/10.1111/j.1349-7006.2006.00208.x.
[28] A. Wu, S. Oh, S.M. Wiesner, K. Ericson, L. Chen, W.A. Hall, P.E. Champoux,
þ
W.C. Low, J.R. Ohlfest, Persistence of CD133
cells in human and mouse
glioma cell lines: detailed characterization of GL261 glioma cells with cancer
stem cell-like properties, Stem Cell. Dev. 17 (2008) 173e184, https://doi.org/
10.1089/scd.2007.0133.
[6] J. Zhang, M.F.G. Stevens, T.D. Bradshaw, Temozolomide: mechanisms of action,
repair and resistance, Curr. Mol. Pharmacol. 5 (2012) 102e114.
[7] S. Berendsen, M. Broekman, T. Seute, T. Snijders, C. van Es, F. de Vos, L. Regli,
P. Robe, Valproic acid for the treatment of malignant gliomas: review of the
[29] N.D. Belyaev, K.A.B. Kellett, C. Beckett, N.Z. Makova, T.J. Revett, N.N. Nalivaeva,

C. Braga et al. / European Journal of Medicinal Chemistry 172 (2019) 16e25
25
N.M. Hooper, A.J. Turner, The transcriptionally active amyloid precursor
protein (APP) intracellular domain is preferentially produced from the 695
isoform of APP in a b-Secretase-dependent pathway, J. Biol. Chem. 285 (2010)
R. Dal Zuffo, M.C. Fulco, M.G. Rozio, A. Mai, G. Dondio, S. Minucci, C. Mercurio,
Discovery, synthesis, and pharmacological evaluation of spiropiperidine
hydroxamic acid based derivatives as structurally novel histone deacetylase
(HDAC) inhibitors, J. Med. Chem. 54 (2011) 3051e3064, https://doi.org/
10.1021/jm200146u.
41443e41454, https://doi.org/10.1074/jbc.M110.141390.
[30] M. Varasi, F. Thaler, A. Abate, C. Bigogno, R. Boggio, G. Carenzi, T. Cataudella,