In vitro and in silico studies of nitrobenzamide derivatives as potential anti-neuroinflammatory agents

Article abstract of DOI:10.1080/07391102.2019.1684368

Communicated by Ramaswamy H. Sarma.

Full text of DOI:10.1080/07391102.2019.1684368

Journal of Biomolecular Structure and Dynamics
ISSN: 0739-1102 (Print) 1538-0254 (Online) Journal homepage: https://www.tandfonline.com/loi/tbsd20
In vitro and in silico studies of nitrobenzamide
derivatives as potential anti-neuroinflammatory
agents
Seda Savranoglu Kulabas, Ferah Comert Onder, Yakup Berkay Yılmaz, Adem
Ozleyen, Serdar Durdagi, Kader Sahin, Mehmet Ay & Tugba Boyunegmez
Tumer
To cite this article: Seda Savranoglu Kulabas, Ferah Comert Onder, Yakup Berkay Yılmaz, Adem
Ozleyen, Serdar Durdagi, Kader Sahin, Mehmet Ay & Tugba Boyunegmez Tumer (2019): In vitro
and in silico studies of nitrobenzamide derivatives as potential anti-neuroinflammatory agents,
Journal of Biomolecular Structure and Dynamics, DOI: 10.1080/07391102.2019.1684368
To link to this article: https://doi.org/10.1080/07391102.2019.1684368
View supplementary material
Accepted author version posted online: 23
Oct 2019.
Submit your article to this journal
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=tbsd20

LETTER TO THE EDITOR
In vitro and in silico studies of nitrobenzamide derivatives as potential anti-
neuroinflammatory agents
Seda Savranoglu Kulabas1, Ferah Comert Onder2, Yakup Berkay Yılmaz1, Adem Ozleyen1,
Serdar Durdagi3*, Kader Sahin3, Mehmet Ay2*, Tugba Boyunegmez Tumer4*
1Graduate Program of Biomolecular Sciences, Institute of Natural and Applied Sciences, Çanakkale Onsekiz
Mart University, 17020 Çanakkale, Turkey.
2Natural Products and Drug Research Laboratory, Department of Chemistry, Faculty of Science and Art,
Çanakkale Onsekiz Mart University, 17020 Çanakkale, Turkey.
3Computational Biology and Molecular Simulations Laboratory, Department of Biophysics, School of Medicine,
Bahçeşehir University, 34353 Istanbul, Turkey
4Department of Molecular Biology and Genetics, Faculty of Science and Art, Çanakkale Onsekiz Mart
University, 17020 Çanakkale, Turkey.
*E-mails: mehmetay17@gmail.com (MA, chemical synthesis), serdardurdagi@gmail.com
(SD, in silico part), tumertb@comu.edu.tr (TBT, bioactivity studies)
List of Abbreviations
Alzheimer’s diseases
Amyotrophic lateral sclerosis
Amyloid precursor protein
Antioxidant response element
Blood brain barrier
AD
ALS
APP
ARE
BBB
CNS
CO
Central nervous system
Carbon monoxide
Cyclooxygenase
COX
DMEM
DMSO
FBS
Fe2+
Dulbecco’s Modified Eagle’s Medium
Dimethyl sulfoxide
Fetal bovine serum
Free ferrous iron
Hemeoxygenease
HO-1
1

Induced fit docking
IFD
Inducible nitric oxide synthase
Kelch-like ECH-associated protein 1
Lipopolysaccharides
iNOS
KEAP1
LPS
Matthews correlation coefficient
Molecular docking and molecular dynamics
Multiple sclerosis
MCC
MD
MS
3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide
Nuclear factor-ҡB
MTT
NFҡB
NO
Nitric oxide
NO+
Nitrosonium ion
NO−
Nitroxyl radical
NADPH:quinoneoxido-reductase I
Nuclear factor-2 erythroid related factor-2
Peroxynitrite
NQOI
Nrf2
ONOO−
PARP
PBS
poly ADP-ribose polymerase
Phosphate buffered saline
Parkinson’s diseases
PD
Protein-protein interaction
Quantitative structure-activity relationships
Root mean squared error
Reactive nitrogen intermediates
Standard error of mean
PPI
QSAR
RMSE
RNOIs
SEM
SOD
SP
Superoxide dismutase
Standard precision
Tanimoto prioritization
TP
Extra precision
XP
2

INTRODUCTION
Neurodegenerative diseases are functional central nervous system (CNS) disorders that
commonly characterized by the progressive deterioration of behavioral, cognitive and motor
skills. Chronic neuroinflammation has been defined as common underlying thread for many
neurodegenerative disorders including Alzheimer’s and Parkinson’s diseases (AD and PD,
respectively), Multiple Sclerosis (MS), Amyotrophic Lateral Sclerosis (ALS), and even
psychiatric illnesses such as depression. In recent years, it has become evident that gut
microbiota can promote chronic low-grade inflammation and neuroinflammation through
mechanisms involving increased exposure of the host to harmful bacterial products
particularly to lipopolysaccharides (LPS) (Tse 2017). Systemic LPS and APP activate
macrophages and microglia cells releasing variety of cytotoxic proinflammatory mediators.
Among these effectors, particular attention has been paid to nitric oxide (NO) and its
metabolic byproducts in mediating cytotoxicity and neurotoxicity in the last decade. In
activated macrophages and microglia cells, inducible nitric oxide synthase (iNOS) produce
micromolar concentrations of NO, which then reacts with oxygen and reactive oxygen species
(ROS) including superoxide anions to form very reactive nitrogen intermediates (RNOIs),
particularly peroxynitrite ion (Tse 2017). RNOIs may result in S-nitrosylation and
nitrotyrosination of proteins that may finally lead to conformational changes (i.e.,
deformation) and modifications of their functional activities. Misfolding of Tau proteins in
AD patients and nitrotyrosination of alpha-synuclein and Tau in PD and AD cases,
respectively are two important hallmarks of neurodegeneration (Steinert, Chernova et al.
2010).
Besides protein deformation, NO and its secondary byproducts exert their cytotoxic and
neurotoxic effects through several different mechanisms. In the center of all these processes
strong and highly reactive free radical properties of these chemicals stand. Anomalous
overproduction of RNOIs may damage DNA, which could lead to over activation of poly
ADP-ribose polymerase (PARP) and cause neuronal ATP depletion and ultimately death. NO
and peroxynitrite interact with iron containing enzymes and interfere with the cellular iron
homeostasis, and irreversibly inhibit the activity of complexes I and II in mitochondrial
electron transport chain, ATP synthase, aconitase and creatine kinase, thus disrupting
mitochondrial energy production (Steinert, Chernova et al. 2010, Tse 2017). This also
presents another reason for cellular ATP depletion, which is followed by neuronal death. NO
3

and peroxynitrite together can also reduce the cellular level of antioxidants such as ascorbic
acid, plasma thiols and inhibit antioxidant enzymes such as glutathione peroxidase, NADPH
quinoneoxido-reductase (NQOI).
The brain is very sensitive organ to cellular oxidative and nitrosative stress, which are two
interconnected processes under inflammatory conditions because of its high lipid content,
oxygen consumption capacity and presence of redox active metal ions such as Cu and Fe.
Therefore, increased oxidant conditions and failures of antioxidant defense system by aberrant
NO signaling are among the prerequisites for neurodegenerative diseases. Keeping redox
homeostasis in brain is highly vital for preventing oxidative stress induced neuronal damage.
Nuclear factor-2 erythroid related factor-2 (Nrf2) is prominent signaling molecule in
cytoprotective defense system regulating inflammation to prevent neuronal death caused by
activated macrophages and microglia cells. Upon oxidative/nitrosative stress stimuli, Nrf2
dissociates from Keap1 and translocates into the nucleus where it trans-activates antioxidant
response element (ARE)-encoding genes including phase II detoxifying enzymes such as
NQO1 and Glutathione S-transferases (GSTs) and antioxidant defense enzymes like
superoxide dismutase (SOD) and catalase (Tumer, Yılmaz et al. 2018). The transcriptional
repressions of iNOS, COX-2, proinflammatory cytokines and NF-ҡB in microglia, astrocytes
and blood brain barrier (BBB) endothelial cells following Nrf2 activation have been well
established. Therefore, NO inhibitors and Nrf2 activators may serve as effective multitarget
pharmacological agents against inflammation elicited by APP and LPS.
The present study has been undertaken to investigate potent lead compounds to block
induction of iNOS and at the same time to induce the cytoprotective enzyme system under
inflammatory condition by using in vitro and in silico analyses. The benzamide functional
group is an important pharmacophore present in the structures of therapeutically important
marketed drugs such as moclobemide, remoxipride and tiapride, which are used in the
treatment of some neurological disorders. In our previous studies, a series of nitro substituted
benzamide derivatives with potent LPS induced NO inhibitory activity were reported (Tumer,
Onder et al. 2017). Continuing our interest in understanding the biological implications of the
structural characteristics in benzamide derivatives, we studied with eight nitro benzamide
derivatives (Figure 1). Moreover, in this work, binary disease quantitative structure-activity
relationships (QSAR) models, molecular docking and molecular dynamics (MD) simulations
and pharmacokinetic approaches as well as toxicity QSAR models were integrated in order to
4

investigate their therapeutic predictions, structural/dynamical properties as well as
pharmacokinetic profiles.
NO2
NO2
NO2
O
O
O
O2N
O2N
O2N
N
H
N
N
H
H
NO2
NO2
NO2
NO2
NO2
3
1
2
NO2
NO2
Cl
O
O
O
O2N
O2N
O2N
N
N
N
N
H
H
H
NO2
N
N
NO2
NO2
NO2
6
5
4
NO2
Cl
O
O
O2N
O2N
N
N
N
H
H
N
NO2
NO2
8
7
Figure 1. Structures of the synthesized compounds
MATERIALS AND METHODS
Chemicals.
Dulbecco’s Modified Eagle’s Medium (DMEM) high glucose, Dulbecco’s Modified Eagle’s
Medium F-12 Nutrient Mixture, 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide (MTT), sodium pyruvate, methanol, N, N-methylenbisacrylamide, dimethyl
sulfoxide, phosphate buffered saline (PBS) and lipopolysaccharide (LPS) from Escherichia
coli O55:B5 were purchased from Sigma-Aldrich Chemical (St Louis, Missouri, USA).
TRIzol®, Trypsin-EDTA, fetal bovine serum (FBS), horse serum,nuclear and cytoplasmic
extraction reagents NE-PER™ (Thermo Scientific, Rockford, IL, USA), PureLink® RNA
mini kit, cDNA8 synthesis kit, DNase I, TaqMan® PCR Master Mix, , TaqMan® Probe
iNOS, TaqMan® Probe β-Actin, TaqMan® Probe Nrf2, TaqMan® Probe HO-1 and
TaqMan® Probe NQO1 were purchased from Life Technologies, Thermo Fisher Scientific
5

(Waltham, MA USA). Protease/phosphatase inhibitor cocktail (100X) were obtained from
Cell Signaling Technology (Leiden, Netherlands).
Immobilon Classico Western HRP substrate (WBLUC0500) and anti-iNOS (sc-650), were
purchased from Santa Cruz Biotechnology (Bergheimer, Heidelberg, Germany), anti-Nrf2
(NBP1-32822) was purchased from Novus (Abingdon Oxon, England).
General procedure for the synthesis of aromatic amide substituted compounds (1-8).
The compounds were synthesized from various 3,5-dinitrobenzoic acid and different aniline
derivatives by using DCC and DMAP at room temperature for 12-24 hours with high and
mild yields and also, 4-dibutylamino-3,5-dinitrobenzoic acid as starting material was obtained
from 4-chloro-3,5-dinitrobenzoic acid and dibutylamine in methanol at reflux conditions for 3
h for the synthesis of compounds N-(4-Chlorophenyl)-4-(dibutylamino)-3,5-dinitrobenzamide
(7) and N-(4-Nitrophenyl)-4-(dibutylamino)-3,5-dinitrobenzamide (8) . The structures of the
studied compounds are given in Scheme 1. Detailed synthesis methods and spectral analysis
of these compounds were described in our study and its supplementary material (Güngör,
Önder et al. 2019).
Cell Culture and sample treatment.
Murine RAW macrophages (RAW264.7, ATCC®TIB-71™) cell line was kind gift of Dr. Ilya
Raskin (Rutgers University, NJ, USA). It was routinely cultured and maintained in 100 U/mL
penicillin, 100 μg/mL streptomycin, and 10% fetal bovine serum (FBS) containing high
glucose DMEM. Murine SIM-A9 microglia (ATCC® CRL-3265™) cell line was purchased
from American Type Culture Collection (ATCC). They were routinely propagated in
DMEM/F-12 supplemented with 100 U/mL penicillin, 100 μg/mL streptomycin, 10% fetal
bovine serum (FBS) and 2% horse serum. Both cells were incubated at 37 °C in humidified
air containing 5% CO2. For treatments, cells were plated at a density of 4x105 or 1.5x106
cells/mL in 24-well or 6 well plates, respectively in phenol red free media containing 10%
FBS and then incubated with or without LPS (1 μg/mL) and compounds (1-8) or with positive
control (1400W) as triplicate for 24 h. Compounds were dissolved in 10% dimethyl sulfoxide
(DMSO) and final concentration of DMSO was 0.1% in each well. All biological experiments
were repeated at least three times, and the data were represented as mean values (means ±
SD).
6

Cell viability assay (MTT).
The effects of compounds on cell viability was examined at different doses (10, 20, 50, 100
µM) by using MTT assay as described previously (Tumer, Onder et al. 2017).
Nitric oxide production assay.
RAW264.7 macrophage and SIM-A9 microglia cells were seeded to 24-well plates with the
4x105 cells/mL density in phenol red free media containing 10% FBS and then cultured with
or without LPS (1 μg/mL) and various doses of compounds as triplicate for 24 hours.
Released NO concentration was assayed by using Griess reaction method as previously
described (Tumer, Yılmaz et al. 2018).
Gene expression analyses by quantitative PCR.
The gene expression levels of iNOS, Nrf2, HO-1, NQO1 and β-Actin (as a reference gene)
were investigated quantitatively by using specific TaqMan® Gene Expression Assays
(Thermo Scientific, Rockford, IL, USA). RAW264.7 macrophage and SIM-A9 microglia cells
were seeded to 24-well plates at a density of 4x105 cells/mL in phenol red free medium
supplemented 10% FBS. Cells were simultaneously treated with 1 μg/mL LPS and various
doses of compound 1 and 3 and then incubated for 24 hours. Total RNA was isolated from
cells by using Pure Link® RNA mini kit plus on-column DNase treatment according to
manufacturer's specifications (Applied Biosystems, Foster City, CA). The protocol has been
modified by using TRIzol for better quality of isolation and purification of RNA as previously
described (Tumer, Onder et al. 2017).
Cell lysate preparation and Western blot analysis.
RAW264.7 macrophage and SIM-A9 microglia cells were seeded to 6-well plates with the
1.5x106 cells/mL density in phenol red free medium containing 10% FBS and then cultured
with or without LPS (1 μg/mL) and various doses of 1 and 3 as triplicate for 6 hours. After
treatments, cells were washed with PBS and cell lysis was extracted by using RIPA buffer to
obtain total cell lysate. The cytoplasmic and nuclear fractions of cells were obtained by using
nuclear and cytoplasmic extraction reagents NE-PER™ (Thermo Scientific, Rockford, IL,
USA). The BCA protein assay kit (Thermo Scientific, Rockford, IL, USA) was performed to
7

quantify the protein content of the cell lysates which were then separated on a discontinuous
buffer system including 4% stacking gel and 8.5% SDS-PAGE. The gel was subjected to
Western blotting. Nitrocellulose membrane was incubated with iNOS, Nrf2, Lamin and β-
actin primary antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) overnight while
shaking at 4 °C. The details of the procedure were carried out as previously described in the
study of Tumer, Yılmaz et al. 2018.
Molecular docking calculations.
Before the docking, both studied ligands and proteins were prepared in physiological
conditions. The structures of studied compounds were sketched using Maestro molecular
modeling package and 3D structures were first subjected to energy minimization. Then,
LigPrep module of Maestro with OPLS3 forcefield were used to prepare the compounds. Epik
was used to generate protonation states for molecules at pH 7.4. Crystal structures of the
target proteins were retrieved from the Protein Data Bank, (PDB entries: 4NOS for iNOS and
2FLU for Keap1). The missing atoms of the proteins (i.e., hydrogen atoms) and missing (if
any) side chain atoms were added and the water molecules within the catalytic domains were
kept using protein preparation module of Maestro molecular modeling tool. The systems were
assigned at a physiological pH by PROPKA, and finally in order to relax the protein, energy
minimization was performed by the OPLS2005 forcefield method. The binding sites of the
target proteins were defined based on the co-crystallized ligands and these sites together with
the water molecules were considered as docking boxes in the simulations. The ions, small
molecules that were used to help crystallization and water molecules that were not in the
vicinity (> 5 Å) of co-crystallized ligand were removed before the docking.
Algorithms used in the docking studies included Glide’s standard precision (SP), extra
precision (XP) and Induced Fit Docking (IFD) modules in Maestro with flexible ligand
sampling. A receptor grid box was generated and certain amino acids were allowed to rotate
their side chains to add flexibility to the used target. IFD method comprises 3 consequent
stages, including (i) docking of the compounds while the receptor is rigid; (ii) refining the
complex amino acids within 5 Å of the ligand using Prime; (iii) finally redocking of the
compounds using refined binding pocket.
In Silico Pharmacokinetic Studies.
8

MetaCore/MetaDrug was used for analysis of therapeutic activity properties, pharmacokinetic
properties and toxicity effects of compounds 1 and 3. MetaCore/MetaDrug provides a
comprehensive tool to analyze compounds and their biochemical and pharmacological
behaviors in the field of drug design and development. MetaDrug uses the property of
Tanimoto Prioritization (TP) to find the similarity between analyzed compounds and
compound sets in the quantitative structure-activity relationships (QSAR) models based on
elements found in the structure. Details of the method has been previously described in our
recent paper (Kanan, Kanan et al. 2019).
Molecular Dynamics (MD) Simulations.
The complexes were hydrated in the cubic box with explicit TIP3P water molecules that have
10 Å thickness from surfaces of protein. The MD simulations were performed by Desmond.
The particle mesh Ewald method was used to calculate the long-range electrostatic interactions
and a cut-off radius of 9 Å was used for both van der Waals and Coulombic interactions. Nose-
Hoover thermostat and Martyna-Tobias−Klein protocols were used to adjust the temperature
and pressure of the systems at 310 K and 1.01325 bar, respectively. The time-step was assigned
as 2.0 fs. The default values were used for minimization and equilibration steps and finally 100-
ns production run was performed for each simulation.
Statistical analysis.
Statistical analyses were carried out by using GraphPad Prism 5 software. For the comparison
of independent samples (group means), Student’s t-test was used. Statistical significance was
considered at p<0.05. All experiments and analyses were repeated at least three times. Results
were expressed as mean ± standard error of mean (SEM).
RESULTS AND DISCUSSION
Chemistry.
In our previous study (Tumer, Onder et al. 2017), we showed the effectiveness of six different
substituted aromatic amide compounds which include nitro functional groups at the different
positions on the phenyl ring, thus, in this study, synthesized eight derivatives of aromatic
amide compounds (1-8) that contain different functional groups such as –NO2, -N(CH3)2, -
N(n-Bu)2, -CH3 and –Cl on both of their side rings, as prodrug for Ntr-based cancer therapies,
9

were examined to better understand the effect of pharmacophore groups and their iNOS
enzyme inhibition characteristics.
According to structures of the compounds, the similar group effects on the same positions
were observed within our previous and current studies. When the nitro groups exist at the -2,4
and -3,5 positions on the phenyl ring, we determined the increased activity results in in vitro
analysis and this circumstance was compared with the known iNOS inhibitors like 1400W.
Here, the compounds 3,5-dinitro-N-(2,4-dinitrophenyl)benzamide (1) and 4-methyl-3,5-
dinitro-N-(2,4-dinitrophenyl)benzamide (3) were found as highly potent iNOS inhibitors due
to their better electron withdrawing properties and mesomeric effects that suitable structural
orientations of the nitro group on positions -2,4 and -3,5 at the binding pocket of the target
structure. Similar behavior of nitro derivatives at the binding pocket of the iNOS has been
shown in our previously reported study (Tumer, Onder et al. 2017).
Furthermore, compound 3 exhibited the lowest IC50 (3.8 µM) value than compound 1 (5.8
µM) for NO production in the presence of the methyl group that shows electron donating
properties on phenyl ring.
Identification of the lead compounds: In vitro screening by NO inhibition.
Among eight nitro benzamide derivatives, two compounds have been identified as promising
lead compounds according to results obtained from NO inhibition assay in LPS induced RAW
264.7 macrophages. For each assay, a control (vehicle alone), and an induction control (1
µg/mL LPS) were used to set a baseline data. 1400W, a highly potent and selective iNOS
inhibitor, was used as positive control in both experimental and docking studies. All tested
compounds exhibited no cytotoxic effect (>90% cell viability-data not shown) on cells at 10
µM.
As shown in Figure 2, all compounds (1-8) were first screened for their effect on LPS induced
NO production in RAW 264.7 macrophage cells at 10 µM. Among eight compounds, 1 and 3
showed potent effects-more than 50% inhibition- against NO production at 10 µM with
inhibition rates of 81% and 52% respectively (p<0.05 for both). Therefore, 1 and 3 were
selected as lead compounds and the rest of the experiments were carried out to identify their
potency and mechanism of action as potential neuroprotective agents in LPS induced
macrophages and microglia cells.
10

Figure 2. Screening of compounds at 10 µM for inhibiting LPS-induced NO production in
RAW 264.7 macrophage cells. The figure shows the mean results of the triplicate experiment
for each treatment group. 1400 W: Positive control. C: Control. Each bar represents the mean
±SEM, *p<0.02, ***p<0.001.
The effects of 1 and 3 on cell viability.
We performed MTT assay in both RAW 264.7 macrophage and SIM-A9 microglia cell lines
to determine the concentrations at which 1 and 3 reveal no cytotoxic effects. As seen in Figure
3, treatment with 3 did not affect the cell viabilities in both cell lines up to highest tested
concentration (100 µM). However, macrophage and microglia cell lines demonstrated slightly
different sensitivity towards 1. While in macrophages for all tested doses cell viabilities were
more than 80%, in SIM-A9 at 100 µM the cell viability decreased slightly to 77.5 %.
Figure 3. Effects of 1 and 3 on cell viability in (A) RAW 264.7 macrophage and (B) SIM-A9
microglia cell lines. The figure shows the mean results of the triplicate experiment for each
treatment group. C: Control, **p<0.05, ***p<0.001.
11

Effects of lead compounds on NO/iNOS signaling in LPS induced macrophage and
microglia cells.
Considerable evidence has shown that iNOS-derived NO secreted form microglia and
macrophage cells is one the major sources of cytotoxic mediators in brain. The reactions of
NO with molecular oxygen, transition metals and superoxide result in the formation potent
oxidizing agents such as nitrosonium ion (NO+), nitroxyl radical (NO−), peroxynitrite
(ONOO−), and many other secondary metabolites generated from these precursors with
profound cytotoxic effects. The neurotoxic effect of NO and its byproducts may involve
several mechanisms such as protein dysfunction through S-nitrosylation or tyrosine nitration,
promoting apoptotic and necrotic cell death, direct DNA damage and/ or inhibiting DNA
repair mechanisms (Steinert, Chernova et al. 2010, Tse 2017). Elevated levels of nitrate and
nitrite were detected in the cerebrospinal fluid of patients with MS and the cerebrospinal fluid
concentration of these cytotoxic compounds was found to be correlated with the breakdown of
the BBB (Thiel and Audus 2001). NO and peroxynitrite have been implicated in the
pathogenesis of MS, AD and PD where NO was found to be neurotoxic to oligodendrocytes
and neurons, and inhibited the mitochondrial respiratory chain (Thiel and Audus 2001).
Therefore, iNOS has emerged as a functionally important source of oxidative/nitrosative brain
injury. For this reason, pharmacologic inhibition of this enzyme is a rational approach in the
prevention as well as treatment of various neurodegenerative diseases. Under pathological
conditions, a known iNOS inhibitor, 1400W, stabilizes blood pressure but has a direct adverse
effect on evoked potential amplitudes that does not appear in physiological conditions.
Because of this effect, the interpretation of the effects on neurovascular coupling is limited;
however, there was no clear positive effect. Thus, there is still some room for novel potent
inhibitors of iNOS that have high binding affinity and specificity as well as with desired
pharmacokinetic profiles. Therefore, targeted approaches to intervene NO/iNOS signaling
pathway should be considered. In the present study, when both RAW 264.7 macrophages and
SIM-A9 microglia cells were simultaneously treated with different concentrations of these
two molecules (0.5-12 µM) and LPS (1 µg/mL), as seen in Figure 4, NO production was
significantly inhibited in a dose-dependent manner.
The IC50 values of these two molecules were investigated for NO production and were
resulted in 5.8 µM for 1 and 3.8 µM for 3, which are comparable to IC50 of 1400W (1.84 µM)
12

a selective and irreversible inhibitor of iNOS. In LPS-induced SIM-A9 microglia cells, same
compounds gave slightly higher IC50 values, which were found to be 6.1 µM for 1, 10.8 µM
for 3 and 2.2 µM for 1400W. Microglia cells are gaining remarkable interest in recent years
due their role in different neurodegenerative and neurological disorders (Tse 2017). SIM-A9
cell line has been recently isolated glial culture from murine cerebral cortices (Nagamoto-
Combs, Kulas et al. 2014). It is the first spontaneously immortalized microglial cell clone
demonstrating major characteristics of cultured primary microglia and serves as a suitable
model for the research studies of neuroinflammation when challenged with LPS or amyloid β.
Since the establishment of this cell line as in vitro model is very recent, the number of
research studies is also very limited, besides there is no study investigating any specific iNOS
inhibitor on this cell line. Recently, NO inhibitory capacities of L-carnitine and vateriferol, a
new resveratrol trimer, on SIM-A9 cell line have been reported (Gill, Raman et al. 2018,
Samaradivakara, Samarasekera et al. 2018). Accordingly, L-carnitine inhibited LPS induced
NO release at relatively high concentrations (Gill, Raman et al. 2018). When the L-carnitine
concentration exceeds 5 mM, the rate of LPS induced NO inhibition could reach to 50%.
Also, treatment of LPS-induced SIM-A9 cells with vateriferol at 14.7 and 73.6 µM resulted in
a dose dependent significant inhibition, however, at both doses, the rate of inhibition could
not reach to 50% (Samaradivakara, Samarasekera et al. 2018). Therefore, the inhibition
potencies of 1 and 3 on NO/iNOS signaling at quite low micro molar ranges (3.8-10.8 µM) in
both macrophages and microglia cells enable these two molecules promising as lead
neuroprotective agents.
13

Figure 4. (A) Effect of 1 and 3 on NO production in LPS induced RAW 264.7 macrophage
cells. (B) Effect of 1 and 3 on NO production in LPS induced SIM-A9 microglia cells. The
cells were treated with LPS (except for negative control-C) and various doses of the 1 and 3.
The figures show the mean results of the triplicate experiment. C: Control. Each bar
represents the mean ±SEM, **p<0.05 ***p<0.001.
14

iNOS is not expressed in normal brain but expression can be induced in microglia cells and
macrophages only under pathological situation. The mechanism of induction involves
transcription of mRNA and novel protein synthesis. In the pathogenesis of various
neurodegenerative disorders such as MS, AD and PD, both iNOS protein and mRNA
expressions were found to be markedly elevated in the active lesions of the patients’ brain
(Thiel and Audus 2001, Steinert, Chernova et al. 2010).
In the present study, treatment of RAW 264.7 macrophages and SIM-A9 microglia cells with
1 (5-20 µM) resulted in substantial down regulations in both mRNA and protein levels of
iNOS (Figure 5). In macrophages, compound 1 at 10 and 20 µM suppressed LPS-induced
iNOS mRNA level 56% and 69%, respectively as compared to LPS-induced control cells
(Figure 5A). Similar to mRNA, iNOS protein expression also decreased 52% and 60% by 1 at
10 and 20 µM, respectively. Compound 3 (10 and 20 µM) in macrophages also suppressed
LPS-stimulated iNOS expression significantly at both mRNA (72% and 62%, respectively)
and protein levels (66% and 60%, respectively). Although, compound 3 resulted in significant
suppression in both mRNA and protein levels of LPS induced microglia cells, the pattern of
downregulation was not dose dependent and also it was not as potent as this seen in
macrophages (Figure 5B). However, compound 1 at 10 and 20 µM suppressed the LPS-
induced iNOS expression at mRNA level 60% and 61%, respectively. Protein levels were also
suppressed by 5, 10 and 20 µM treatment of compound 1 dose dependently by 30%, 53% and
58%, respectively.
Therefore, potent inhibition and downregulation capacities of both compounds, particularly
compound 3 on NO/iNOS signaling at three level of expression (activity, mRNA and protein
levels) support potential anti-neuroinflammatory and neuroprotective effects of these
compounds on degenerative disorders of CNS.
15

Figure 5. (A) Effect of 1 and 3 on iNOS mRNA and protein expression in LPS induced RAW
264.7 macrophage cells. (B) Effect of 1 and 3 on iNOS mRNA and protein expression in LPS
induced SIM-A9 microglia cells. The cells were treated with LPS (except for negative
control-C) and 5, 10 and 20 µM doses of the 1 and 3. The figures show the mean results of the
triplicate experiment. C: Control. Each bar represents the mean ±SEM, **p<0.05,
***p<0.001.
In Silico Investigation of Studied Compounds at the Binding Pocket of iNOS.
In order to investigate binding interactions of studied molecules, all eight nitro derivatives
have been docked to the binding pocket of iNOS using Glide/SP, Glide/XP and Induced Fit
(IFD) docking algorithms. Same docking protocol has been applied to the well-known iNOS
inhibitor 1400W and Ronopterin as positive control compounds. Docking scores confirmed
the potent molecules 1 and 3 (Table S1 of the Supporting Information). Docking scores also
fits well with the experimental findings (i.e., compounds 1 and 3 shows promising binding
interactions with the target enzyme but their IC50 values are slightly bigger than 1400W).
When the top-docking poses of compounds 1 and 3 were superimposed, it can be seen that
they have similar binding interactions with the target residues (Figure 6). Binding cavity for
both compounds were formed with following residues: Glu377, Gln263, Ala282, Arg388,
Tyr373, Trp372, Gly371, Arg266, Arg381, Val352, Pro350, Phe369 and Trp463. Compound
1400W has also similar binding cavity with the 1 and 3. Salt bridge interaction and hydrogen
bond were constructed with Glu377 for compound 1. Corresponding interactions were formed
16

with Arg266, Gln263 and Glu377 for compound 3 (Figure 6). Glu377 also forms a salt bridge
interaction with 1400W.
17

Figure 6. 2D and 3D ligand interaction diagrams of compounds 1 and 3 at the binding pocket
of iNOS using top-docking poses (Glide/XP). Results were compared with positive control
1400W and Ronopterin.
Top-docking poses of compounds 1, 3, 1400W and ronopterin were used as initial structure
and all-atom classical molecular dynamics (MD) simulations were performed in order to
investigate time-dependent dynamics of both ligands and target atoms. Fluctuations of ligand
at the binding pocket can be investigated by root mean square fluctuation (RMSF). Figure S1
at the supplementary information shows RMSF of iNOS based on C atoms. Figure
represents which regions of protein have higher mobilities compared to other parts. Especially
regions of residue numbers 20-50, 70-80, 190-195 and 320-340 show higher mobility. Most
mobile regions of ligand can be also monitored throughout the MD simulations. Ligand
RMSF may give insights on how ligand fragments interact with the binding pocket residues
and their entropic role in the binding event. So, translational (based on protein) and rotational
(based on ligand itself) RMSF of ligand has been computed (considering only heavy atoms).
Figure 7 shows rotational and translational RMSF of the compound 3. It can be seen that
especially atom numbers 15-19 and 20-22 at the ligand show higher translational motion
during the simulations. The region of atom numbers 24-28 has higher rotational motion.
18

Figure 7. Translational and rotational RMSF of compound 3 at the binding pocket of the
target.
Torsional profile of the compound 3 has been also investigated throughout the simulation
(Figure S2). The figure summarizes the conformational changes of every rotatable bond at the
compound 3. Each torsional angle at the rotatable bond is accompanied by a radial and bar
plots. Radial plots show the torsional angle changes during the simulations (the beginning of
the simulation is in the center of the radial plot and the time evolution is plotted radially
outwards).
19

Protein-compound 3 interactions have been monitored throughout the MD simulations. Figure
8 represents the crucial residues for the ligand contacts. It can be seen that especially Cys200,
Gln263, Arg266, , Val352, Asn354, Arg381 and Arg388 have higher interaction fractions
with the ligand at the binding pocket. Similar analysis has been carried out for the 1400W and
ronopterin it is found that especially Pro350, Val352, Trp372 and Glu377 have crucial roles in
binding for 1400W (Figure S3) and Cys200, Gln263, Pro350, Gly371, Tyr373, Glu377,
Asp382 for ronopterin (Figure S4).
Figure 8. Protein-compound 3 contacts throughout the MD simulations.
MM/GBSA continuum solvation approach is well-defined method for estimating the free
energy of binding of small ligands to a specific biological target (Shityakov, Sohajda et al.
2014, Shityakov, Roewer et al. 2017). Derived trajectory frames throughout the MD
simulations were used in MM/GBSA calculations. MM/GBSA calculations were performed for
compounds 1, 3 as well as positive controls 1400W and Ronopterin. (Table S1) Results of
selected compounds gave similar binding energies with positive controls.
Our computational investigation results showed that selected compounds (1 and 3) have similar
interaction energy profiles with the two selected positive control compounds in this study.
20

The effects of 1 and 3 on Nrf2 mRNA Levels and Its Transactivated Target Gene
Expression.
Oxidative stress is accepted as crucial player in the pathogenesis of various chronic
inflammatory diseases including neurodegenerative disorders. Nrf2 activators attenuate
multiple pathogenic processes in which disruption of redox homeostasis and chronic
inflammation coexist including various neurodegenerative disorders (Tumer, Yılmaz et al.
2018).
In the present work, we studied the effects of compounds 1 and 3 on the mRNA expression of
Nrf2 and its downstream enzymes under inflammatory conditions in both macrophages and
microglia cells. As shown by Figure 9, 1 at 10 and 20 µM increased the mRNA expression of
Nrf2 in a dose dependent manner both in macrophages (1.4-1.7 fold, respectively) and
microglia cells (2.0-2.4 fold, respectively) as compared to control group. Treatment with 3,
resulted in very similar effects at same doses in macrophage (1.8 and 2.0 fold upregulation,
respectively) and microglia cells (2.5 and 2.9 fold upregulation, respectively).
Figure 9. Effect of 1 and 3 on Nrf2 mRNA expression in LPS induced RAW 264.7
macrophage cells and SIM-A9 microglia cells. The cells were treated with LPS and 5, 10 and
20 µM doses of the 1 and 3. The figures show the mean results of the triplicate experiment. C:
Control. Each bar represents the mean ±SEM, **p<0.05, ***p<0.001.
Both compounds 1 and 3 also resulted in a marked upregulation in the expression of NQO1
enzyme, a prototypic Nrf2 target in both LPS induced inflammation model of macrophage
and microglia cells. In the present study, as presented in Figure 10, compound 1 at 10 and 20
µM upregulated mRNA level of NQO1 3.2 and 10.1-fold in macrophages and 8.8 and 17.6-
fold, respectively in microglia cells. Compound 3 at same doses resulted in more potent
21

increases which were 5.3 and 10.6-fold in macrophages and 14.8 and 26.3-fold, respectively
in microglia cells.
Figure 10. Effect of 1 and 3 on NQO1 mRNA expression in LPS induced RAW 264.7
macrophage cells and SIM-A9 microglia cells. The cells were treated with LPS (except for
negative control-C) and 5,10 and 20 µM doses of the 1 and 3. The figures show the mean
results of the triplicate experiment. C: Control. Each bar represents the mean ±SEM,
**p<0.05, ***p<0.001.
In terms of hemeoxygenease (HO-1), compounds 1 and 3 did not result in similar positive
effects. 1 µg/mL LPS treatment caused a slight upregulation (around 4 fold) at the mRNA
level of HO-1 in both macrophage and microglia cells as reported in previous research studies
where LPS above 100ng/mL concentration significantly induces expression of this enzyme
(Figure 11) (Ye, Wang et al. 2014). On the other hand, treatment with both compounds 1 and
3 resulted in repression of the mRNA expression to a similar level, which was observed in
non-treated control cells. HO-1 catalyzes the degradation of heme groups to carbon
monoxide (CO), free ferrous iron (Fe2+) and biliverdin. By the action of biliverdin reductase,
the latter product is metabolized to bilirubin which can remove hydroxyl radicals, superoxide
anions, and singlet oxygen. Bilirubin also prevents protein and lipid peroxidation as well as
exhibits strong antioxidant, anti-apoptotic and anti-inflammatory functions. Therefore, HO-1
indirectly exerts cytoprotection through the biological activities of its downstream products.
Different studies on microglial and macrophage cells demonstrated that upregulation of HO-1
22

has a strong anti-inflammatory effect. This has been evidenced both in cell cultures and
animal models treated with different pharmacological agents as well as natural compounds
able to activate Nrf2/HO-1 (Tumer, Yılmaz et al. 2018). However, opposite evidences are also
available such as decreasing neuroprotective effects of epigallocatechin-3-gallate and
glatiramer acetate by upregulation of HO-1 in experimental autoimmune encephalomyelitis
model (Janssen, Fiebiger et al. 2015). Moreover, it is also well established that HO-1 is
overexpressed in the brains of patients with AD, particularly in the hippocampus and cerebral
cortex region, as well as in nigralastroglia and in dopaminergic neuronal Lewy bodies of PD
patients. This up-regulation has been linked to the late phase of neurodegeneration and has
been proposed as biomarker of AD and PD (Nitti, Piras et al. 2018).
Figure 11. Effect of 1 and 3 on HO-1 mRNA expression in LPS induced RAW 264.7
macrophage cells and SIM-A9 microglia cells. The cells were treated with LPS (except for
negative control-C) and 5,10 and 20 µM doses of the 1 and 3. The figures show the mean
results of the triplicate experiment. C: Control. Each bar represents the mean ±SEM,
**p<0.05, ***p<0.001.
Although the role played by HO-1 in oxidative stress and neuroinflammation is highly
complex and is not completely understood due to contradictory results obtained, it has been
recently postulated that upregulation of HO-1 in glial compartment of the brain may promote
bioenergetics failure by affecting iron metabolism and mitochondrial activity (Nitti, Piras et
al. 2018). In fact, it is proposed that the dysregulation of heme degradation pathway through
the alteration of iron metabolism and the mitochondrial impairment may lead to
neurodegeneration not only when it occurs in glial cells, but also in neurons (Nitti, Piras et al.
2018). When the intensity and duration of HO-1 up-regulation increase beyond a level this
23

may increase tissue damage mainly through heme-derived iron, which favors oxidative stress
in mitochondria and other compartments (Nitti, Piras et al. 2018). In this aspect, the
smoothing effect of compound 1 and 3 on the expression of HO-1 may be the advantages of
these compounds in case of long term uses if they can be developed as potential
neuroprotective lead compounds. However, more studies are warranted in this aspect.
The effect of compound 1 and 3 on nuclear accumulation of Nrf2 protein.
Although both 1 and 3 resulted in substantial elevation in the mRNA expression of Nrf2,
nuclear accumulation of this protein is essential for its function as transcription factor for
activating Nrf2-ARE pathway. Some activators might promote the release of Nrf2 from its
inhibitor Keap1 resulting in increased level of Nrf2 in nucleus. In the present work, as shown
by Figure 12, 1 at 5, 10 and 20 µM increased the accumulation of Nrf2 protein to nuclear
region in a dose dependent manner 1.7, 1.9 and 5.1-fold, respectively as compared control
group. Treatment with 3, resulted in a more prominent effect, 2.1, 3.1 and 7.1-fold increases
at 5, 10 and 20 µM doses, respectively when compared to control group. Protein results
together with those obtained from mRNA expression studies may indicate that 1 and 3
regulate Nrf2 expression at both transcriptional and posttranscriptional level. It has been
evidenced that therapeutic compounds such as chemopreventive agents disorder Keap1-Nrf2
PPI by direct binding to Keap1 pocket to where the 16-mer peptide of Neh2 domain of Nrf2
binds. To investigate whether compound 1 and 3 activate Nrf2 by disrupting PPI, we
performed in silico analyses.
24

Figure 12. Effect of 1 and 3 on Nrf2 protein expression in LPS induced RAW 264.7
macrophage cells. The cells were treated with LPS and 5, 10 and 20 µM doses of the 1 and 3.
The figures show the mean results of the triplicate experiment. C: Control. Each bar
represents the mean ±SEM, **p<0.05, ***p<0.001.
In Silico Investigation of Studied Compounds at the Binding Pocket of Keap1.
In order to understand the molecular mechanisms of studied molecules at the interactions of
Keap1, these molecules were docked to the Keap1 target. Results showed that compounds 1
and 3 have higher docking scores at the binding pocket compared to other nitro derivatives
(Table S2 of the Supporting Information). 2D and 3D ligand interactions diagrams of
compounds 1 and 3 which shows top-docking poses has been represented at the Figure 13.
These interactions were also compared with a well-known Keap1 Nrf2 PPI inhibitor LH601A.
Although compounds 1 and 3 have common binding interactions, their bioactive
conformations at the Keap1 was slightly different. While crucial salt bridge and hydrogen
bonding interactions were constructed with Arg380, Ser363, Asn414, Arg415 and Tyr525
with compound 1, corresponding interactions were Arg380, Asn382, Tyr334 and Tyr572 in
compound 3. Positive control molecule was constructed similar interactions (i.e., Arg380,
Asn382 and Tyr525).
Current work represented herein has shown that both compounds 1 and 3 might be promising
hits and represent highly promising starting point for the development of potent
neuroprotective agents against neurodegenerative diseases. In this aspect, further studies
including in silico target based structural optimization of these two compounds and also
validation of their potency by in vitro and in vivo models will be crucial. However, it is well
known that numerous lead compounds have failed in animal studies and clinical trials due to
issues related with ADME and toxicity. For this reason, in the present study we evaluated
both compounds in terms of pharmacokinetic and toxicity parameters by using in silico
techniques.
Pharmacokinetic and Toxicity Predictions of Compounds 1 and 3
Identification of novel compounds against neurodegenerative diseases are hot topics of both
academy and industry, however, these compounds face various challenges such as specificity
and potential toxicities as they may interfere with other pathways and cellular processes
25

regulated by iNOS or Keap1. For this reason, we have decided to evaluate chemical
properties, possible metabolites, potential therapeutic values and toxicities of selected potent
ligands. This allowed us to get a thorough understanding and prediction of their biological
activity when compared to hundreds of other drugs using MetaCore from Clarivate Analytics.
Toxicity properties of 1 and 3 were studied with 26-different toxicity QSAR models. (Table
S3 of the Supporting Information) Compound 3 did not show significant high toxicity profiles
except the AMES model. Compound 3 showed no toxicity predictions in 18 toxicity QSAR
models and it showed slightly larger values in 7 toxicity QSAR models than the threshold
values. Compound 1 showed no toxicity predictions in 13 toxicity QSAR models and in
slightly larger toxicity predictions in 5 QSAR models. It shows high risks for following
toxicity models: AMES, Anemia, Carcinogenecity (both in mouse female and rat male
models), hepatotoxicity, liver necrosis, and in liver weight gain models. (Table S3 of the
Supporting Information) When ADME and protein binding QSAR models were investigated
with both compounds 1 and 3, it is found that both compounds can permeate in BBB
penetration model. (Table S4 of the Supporting Information) In the Table, the data is
expressed as log values of the ratio of the metabolite concentrations in brain and plasma.
Cutoff is -0.3 and larger values indicate that the metabolite is more likely to enter the brain.
Both compounds have desired lipophilicity values. Although both compounds have higher
affinity to human serum albumin, their retention time (i.e., the model is based on retention
times of compounds assayed by HPLC using an immobilized HSA column) is low.
26

Figure 13. 2D and 3D ligand interaction diagrams of compounds 1 and 3 at the binding
pocket of Keap1 using top-docking poses (Glide/SP). Results were compared with positive
control molecule LH601A.
Conclusions.
27

In conclusion, in the current work we showed that the compounds 3,5-dinitro-N-(2,4-
dinitrophenyl)benzamide (1) and 4-methyl-3,5-dinitro-N-(2,4-dinitrophenyl)benzamide (3)
display marked inhibitory potential on NO/İNOS signaling as well as potent activator
properties on Nrf2 and associated cytoprotective enzymes in LPS-stimulated macrophages
and microglial cells. It is evident from our current and previous work that when the nitro
groups exist at the -2,4 and -3,5 positions on the phenyl ring, the activity of the substituted
aromatic amide compounds increasing due to their better electron withdrawing properties and
mesomeric effects at the binding pocket of the target structures. Molecular docking, MD
simulations and ADMET analysis also presented that both compound 1 and 3 can be
classified as druggable molecule in drug design. Therefore, nitrobenzamide scaffolds may
represent a promising template and starting point, which may lead to novel small molecules to
manage neuroinflammatory diseases.
AUTHOR INFORMATION
Author contributions
T.B.T. designed the project, supervised all contributions, wrote and finalized the draft
manuscript. F.C.O. and M.A. synthesized the compounds wrote the draft of the relative
section. S.S.K., Y.B.Y. and A.O. performed bioactivity studies, assisted data analysis, and
construction of figures/ tables of the related parts. S.D. and K.S. conceived and performed the
molecular modeling/ADMET studies and wrote the draft of the relative sections. All authors
read and approved the final manuscript.
Conflict of interest
We have no conflict of interest whether of a financial or other nature.
Acknowledgements
Synthesis part of this study was partially supported by a research grant from the Scientific and
Technological Research Council of Turkey-TUBITAK (Grant No. 110T754).
28

References
Gill, E. L., S. Raman, R. A. Yost, T. J. Garrett and V. Vedam-Mai (2018). "l-Carnitine
Inhibits Lipopolysaccharide-Induced Nitric Oxide Production of SIM-A9 Microglia Cells."
ACS chemical neuroscience 9(5): 901-905.
Güngör, T., F. C. Önder, E. Tokay, Ü. G. Gülhan, N. Hacıoğlu, T. T. Tok, A. Çelik, F.
Köçkar and M. Ay (2019). "Prodrugs for nitroreductase based cancer therapy-2 novel
amide/Ntr combinations targeting PC3 cancer cells." European Journal of Medicinal
Chemistry.
Janssen, A., S. Fiebiger, H. Bros, L. Hertwig, S. Romero-Suarez, I. Hamann, C.
Chanvillard, J. Bellmann-Strobl, F. Paul and J. M. Millward (2015). "Treatment of chronic
experimental autoimmune encephalomyelitis with epigallocatechin-3-gallate and glatiramer
acetate alters expression of heme-oxygenase-1." PLoS One 10(6): e0130251.
Kanan, T., D. Kanan, I. Erol, S. Yazdi, M. Stein and S. Durdagi (2019). "Targeting the
NF-κB/IκBα complex via fragment-based E-Pharmacophore virtual screening and binary
QSAR models." Journal of Molecular Graphics and Modelling 86: 264-277.
Nagamoto-Combs, K., J. Kulas and C. K. Combs (2014). "A novel cell line from
spontaneously immortalized murine microglia." Journal of neuroscience methods 233: 187-
198.
Nitti, M., S. Piras, L. Brondolo, U. Marinari, M. Pronzato and A. Furfaro (2018).
"Heme Oxygenase 1 in the Nervous System: Does It Favor Neuronal Cell Survival or Induce
Neurodegeneration?" International journal of molecular sciences 19(8): 2260.
Samaradivakara, S. P., R. Samarasekera, S. M. Handunnetti, O. J. Weerasena, A. A. Al-
Hamashi, J. T. Slama, W. R. Taylor, Q. Alhadidi, Z. A. Shah and L. Perera (2018). "A
Bioactive Resveratrol Trimer from the Stem Bark of the Sri Lankan Endemic Plant Vateria
copallifera." Journal of natural products 81(8): 1693-1700.
Shityakov, S., N. Roewer, C. Förster and J.-A. Broscheit (2017). "In silico investigation
of propofol binding sites in human serum albumin using explicit and implicit solvation
models." Computational biology and chemistry 70: 191-197.
Shityakov, S., T. Sohajda, I. Puskás, N. Roewer, C. Förster and J.-A. Broscheit (2014).
"Ionization states, cellular toxicity and molecular modeling studies of midazolam complexed
with trimethyl-β-cyclodextrin." Molecules 19(10): 16861-16876.
Steinert, J. R., T. Chernova and I. D. Forsythe (2010). "Nitric oxide signaling in brain
function, dysfunction, and dementia." The Neuroscientist 16(4): 435-452.
Thiel, V. E. and K. L. Audus (2001). "Nitric oxide and blood–brain barrier integrity."
Antioxidants and Redox Signaling 3(2): 273-278.
Tse, J. K. (2017). "Gut microbiota, nitric oxide, and microglia as prerequisites for
neurodegenerative disorders." ACS chemical neuroscience 8(7): 1438-1447.
Tumer, T. B., F. C. Onder, H. Ipek, T. Gungor, S. Savranoglu, T. T. Tok, A. Celik and
M. Ay (2017). "Biological evaluation and molecular docking studies of nitro benzamide
derivatives with respect to in vitro anti-inflammatory activity." International
immunopharmacology 43: 129-139.
Tumer, T. B., B. Yılmaz, A. Ozleyen, B. Kurt, T. T. Tok, K. M. Taskin and S. S.
Kulabas (2018). "GR24, a synthetic analog of Strigolactones, alleviates inflammation and
promotes Nrf2 cytoprotective response: In vitro and in silico evidences." Computational
biology and chemistry 76: 179-190.
Ye, M., Q. Wang, W. Zhang, Z. Li, Y. Wang and R. Hu (2014). "Oroxylin A exerts
anti-inflammatory activity on lipopolysaccharide-induced mouse macrophage via Nrf2/ARE
activation." Biochemistry and Cell Biology 92(5): 337-348.
29