Derivatives of caffeic acid, a natural antioxidant, as the basis for the discovery of novel nonpeptidic neurotrophic agents

Article abstract of DOI:10.1016/j.bmc.2017.04.026

Neurodegenerative disorders, such as Parkinson's disease and Alzheimer's disease, threaten the lives of millions of people and the number of affected patients is constantly growing with the increase of the aging population. Small molecule neurotrophic age

Full text of DOI:10.1016/j.bmc.2017.04.026

Accepted Manuscript
Derivatives of caffeic acid, a natural antioxidant, as the basis for the discovery
of novel nonpeptidic neurotrophic agents
Fatemeh Moosavi, Razieh Hosseini, Hamid Rajaian, Tiago Silva, Diogo
Magalhães e Silva, Luciano Saso, Najmeh Edraki, Ramin Miri, Fernanda
Borges, Omidreza Firuzi
PII:
S0968-0896(17)30214-6
DOI:
http://dx.doi.org/10.1016/j.bmc.2017.04.026
BMC 13696
Reference:
Bioorganic & Medicinal Chemistry
To appear in:
Received Date:
Accepted Date:
7 February 2017
9 April 2017
Please cite this article as: Moosavi, F., Hosseini, R., Rajaian, H., Silva, T., Silva, D.M.e., Saso, L., Edraki, N., Miri,
R., Borges, F., Firuzi, O., Derivatives of caffeic acid, a natural antioxidant, as the basis for the discovery of novel
nonpeptidic neurotrophic agents, Bioorganic & Medicinal Chemistry (2017), doi: http://dx.doi.org/10.1016/j.bmc.
2017.04.026
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Derivatives of caffeic acid, a natural antioxidant, as the basis for the discovery of novel
nonpeptidic neurotrophic agents
Fatemeh Moosavi1,2, Razieh Hosseini1,2, Hamid Rajaian2, Tiago Silva3, Diogo Magalhães e
Silva3, Luciano Saso4, Najmeh Edraki1, Ramin Miri1, Fernanda Borges3*, Omidreza
Firuzi1*
1
Medicinal and Natural Products Chemistry Research Center, Shiraz University of Medical
Sciences, Shiraz, Iran
2 Department of Pharmacology, School of Veterinary Medicine, Shiraz University, Shiraz, Iran
3 CIQUP/Department of Chemistry and Biochemistry, Faculty of Sciences, University of Porto,
4169-007 Porto, Portugal
4
Department of Physiology and Pharmacology "Vittorio Erspamer", Sapienza University of
Rome, Italy
* Corresponding authors:
Omidreza Firuzi, MD PhD
Medicinal and Natural Products Chemistry Research Center
Shiraz University of Medical Sciences
Shiraz, Iran
Email: firuzio@sums.ac.ir
Fernanda Borges, PhD
CIQUP/Department of Chemistry and Biochemistry
Faculty of Sciences, University of Porto
4169-007 Porto, Portugal
E-mail: fborges@fc.up.pt
1

Abstract
Neurodegenerative disorders, such as Parkinson’s disease and Alzheimer’s disease, threaten the
lives of millions of people and the number of affected patients is constantly growing with the
increase of the aging population. Small molecule neurotrophic agents represent promising
therapeutics for the pharmacological management of neurodegenerative diseases. In this study, a
series of caffeic acid amide analogues with variable alkyl chain lengths, including ACAF3 (C3),
ACAF4 (C4), ACAF6 (C6), ACAF8 (C8) and ACAF12 (C12) were synthesized and their
neurotrophic activity was examined by different methods in PC12 neuronal cells. We found that
all caffeic acid amide derivatives significantly increased survival in PC12 neuronal cells in
serum-deprived conditions at 25 µM, as measured by the MTT assay. ACAF4, ACAF6 and
ACAF8 at 5 µM also significantly enhanced the effect of nerve growth factor (NGF) in inducing
neurite outgrowth, a sign of neuronal differentiation. The neurotrophic effects of amide
derivatives did not seem to be mediated by direct activation of tropomyosin receptor kinase A
(TrkA) receptor, since K252a, a potent TrkA antagonist, did not block the neuronal survival
enhancement effect. Similarly, the active compounds did not activate TrkA as measured by
immunoblotting with anti-phosphoTrkA antibody. We also examined the effect of amide
derivatives on signaling pathways involved in survival and differentiation by immunoblotting.
ACAF4 and ACAF12 induced ERK1/2 phosphorylation in PC12 cells at 5 and 25 µM, while
ACAF12 was also able to significantly increase AKT phosphorylation at 5 and 25 µM.
Molecular docking studies indicated that compared to the parental compound caffeic acid,
ACAF12 exhibited higher binding energy with phosphoinositide 3-kinase (PI3K) as a possible
molecular target. Based on Lipinski’s rule of five, all of the compounds obeyed 3 molecular
descriptors (HBD, HBA and MM) in drug-likeness test. Taken together, these findings show for
the first time that caffeic amides possess strong neurotrophic effects exerted via modulation of
ERK1/2 and AKT signaling pathways presumably by activation of PI3K and thus represent
promising agents for the discovery of neurotrophic compounds for management of
neurodegenerative diseases.
Keywords: Caffeic amides, neurotrophic agents, cell signaling, ERK1/2, AKT
Running title: Neurotrophic effects of caffeic acid amide derivatives
2

1. Introduction
Age related neurodegenerative disorders such as Alzheimer’s disease and Parkinson’s disease
represent heavy health burdens that result in the loss of millions of human lives every year and
1
impose huge direct and indirect socioeconomic costs . There is currently no disease-modifying
treatment available for these pathologies and finding novel therapeutic strategies is a pressing
and unmet clinical need. Nerve growth factor (NGF) is a neurotrophin that has been shown to
prevent neuronal cell damage, and also to promote neuronal regeneration in several in vitro and
in vivo models 2. Its actions are mediated through tropomyosin receptor kinase A (TrkA) and the
major neurotrophic signaling pathways: phosphoinositide 3-kinase (PI3K)/AKT, mitogen-
3
activated protein kinase (MAPK)/MEK/ERK and phospholipase C-γ1 pathways . NGF may
provide an attractive therapeutic choice for neurodegenerative disorders 4,5; however, it has poor
pharmacokinetic profile including poor stability in plasma, no oral bioavailability and poor
blood–brain barrier penetration, and may also cause adverse effects such as severe pain 4,6. These
limitations have led to the exploration of small molecules, which activate the TrkA receptor or
potentiate the neurotrophic effects of limited amounts of endogenous NGF 2.
Caffeic acid (CA) and its related compounds such as caffeic acid phenethyl ester (CAPE),
rosmarinic acid, chlorogenic acid and ferulic acid are well-known for their diverse biological
actions including antioxidant and neuroprotective effects 7,8. The neuroprotective action of these
compounds has been shown in various studies; for instance, CAPE has shown to protect
dopaminergic neurons from IFN-g/LPS-induced injury by upregulation of brain-derived
9
neurotrophic factor (BDNF) in rat midbrain slice cultures . CAPE also induced neurite
10
outgrowth in PC12 cells associated with increased expression of differentiation markers
.
Recently, it has been shown that some CA derivatives are able to potentiate neuroprotection and
11,12
induce neurogenesis via interactions with neuronal intracellular signaling pathways
.
Nevertheless, the cellular mechanisms by which these compounds induce neurotrophic effects
are not completely understood.
Following our line of research based on the development of new chemical entities derived from
naturally occurring scaffolds, we focused our efforts on the discovery of novel neurotrophic
agents based on the hydroxycinnamic acid scaffold. In this context, we synthesized a series of
caffeic amide analogues with variable chain lengths and assessed the effect of these derivatives
on the survival and induction of neurite outgrowth in PC12 cells. Moreover, the effects of these
3

compounds on TrkA receptor activation and alteration of its downstream signaling pathways,
including ERK1/2 and AKT pathways, were also examined.
4

2. Materials and methods
2.1. Reagents
Fetal bovine serum (FBS), horse serum (HS), penicillin/streptomycin, RPMI 1640, sterile
phosphate-buffered saline (PBS) and trypsin EDTA 0.25% were purchased from Biosera
(Ringmer, UK), while rat recombinant NGF-β and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl
tetrazolium bromide (MTT) were obtained from Sigma-Aldrich (St Louis, MO). Methanol was
obtained from Merck (Darmstadt, Germany). K252a was purchased from Santa Cruz
Biotechnology (Santa Cruz, CA). Anti-phospho-TrkA, anti-phospho-AKT and anti-phospho-
MAP kinase ERK1/2 (pThr202/Tyr204), anti-MAP kinase ERK1/2, as well as rabbit anti-mouse
and mouse anti-rabbit HRP-conjugated secondary antibodies were obtained from Cell Signaling
(Danvers, MA), while anti-TrkA and anti-AKT were purchased from Santa Cruz Biotechnology
(Santa Cruz, CA). Solvents were pro analysis grade and were acquired from Merck (Lisbon,
Portugal). All the other reagents used for synthesis were purchased from Sigma-Aldrich (Sintra,
Portugal). Deionized water was used in all the experiments.
2.2. Chemistry
Thin layer chromatography (TLC) was performed on pre-coated silica gel 60 F254 acquired from
Merck (Darmstadt, Germany) and spots were detected using a UV lamp at 254 nm. Reaction
progress was monitored using TLC (dichloromethane/methanol/formic acid, 9:1:0.01).
Following the extraction step, subsequent work up of the organic layers included drying over
anhydrous sodium sulphate, filtration and evaporation of solvents under reduced pressure.
Column chromatography was carried out with silica gel 60A acquired from Carlo-Erba Reactifs
(SDS, France). The crude products were purified by flash column chromatography. The fractions
containing the desired product were gathered, concentrated and the product was recrystallized.
Solvents were evaporated with a Buchi rotary evaporator.
The purity of the final products (>97% purity) was verified by high-performance liquid
chromatography (HPLC) equipped with a UV detector. Chromatograms were obtained in an
5

HPLC/DAD system, a Jasco instrument (pumps model 880-PU and solvent mixing model 880-
30, Tokyo, Japan), equipped with a commercially prepacked Nucleosil RP-18 analytical column
(250 mm x 4.6 mm, 5 µm, Macherey-Nagel, Duren, Germany), and UV detection (Jasco model
875-UV) at the maximum wavelength of 254 nm. The mobile phase consisted of a
methanol/water or acetonitrile/water (gradient mode, room temperature) at a flow rate of 1
mL/min. The chromatographic data was processed in a Compaq computer, fitted with CSW 1.7
software (DataApex, Czech Republic).
1H, C NMR and DEPT135 data were acquired at room temperature on a Brüker AMX 400
13
spectrometer operating at 400 and 100 MHz, respectively. Chemical shifts were expressed in δ
(ppm) values relative to tetramethylsilane (TMS) as internal reference and coupling constants (J)
were given in Hz. Carbon signals present in DEPT135 (Distortionless Enhancement by
Polarization Transfer) spectra were underlined. Mass spectra were obtained on a VG AutoSpec
instrument. The data were reported as m/z (% of relative intensity of the most important
fragments).
2.3. General procedure for PyBOP-assisted amidation
Caffeic acid (2.78 mmol) was dissolved in diisopropylethylamine (0.48 mL) and
dimethylformamide (4 mL). The reaction was placed on ice and PyBOP (2.78 mmol) in
dichloromethane (5 mL) was added. The mixture was stirred for 30 minutes. Then, the amine
(2.78 mmol) was added, the mixture was allowed to reach room temperature and stirred for an
additional 6 hours. The solvents were removed under reduced pressure and the residue was
dissolved in ethyl acetate (30 mL). The solution was washed with water (3 x 10 mL), HCl 1M (3
x 10 mL) and brine (10 mL). The compounds were purified by column chromatography in
dichloromethane/methanol (95:5 until 90:10) and recrystallized from dichloromethane/n-hexane.
The procedure was adapted from Gaspar et al. 2011 13.
2.3.1. (E)-3-(3,4-dihydroxyphenyl)-N-propylacrylamide (ACAF3)
1
Yield (%): 12.7. H NMR (MeOD, 400 MHz) δ (ppm) 7.40 (d, J=15.7 Hz, 1H, Hα), 7.02 (d,
J=2.0 Hz, 1H, H2), 6.92 (dd, J=8.2, 2.0 Hz, 1H, H5), 6.79 (dd, J=8.0 Hz, 1H, H6), 6.38 (d,
J=15.7 Hz, 1H, Hβ), 3.26 (t, J=7.1, 2H, CH2), 1.59 (m, 2H, CH2), 0.97 (t, J=7.4 Hz, 3H, CH3).
6

13C NMR (MeOD, 100 MHz): 167.9 (CONH), 147.3 (C4), 145.3 (C3), 140.7 (Cβ), 120.6 (C6),
117.1 (Cα), 115.1 (C5), 113.7 (C2), 41.0 (NHCH2), 22.3 (CH2), 10.4 (CH3). EI-MS m/z 221
(M•+, 8), 178 (40), 163 (100), 145 (18), 134 (16), 117 (10), 89 (20), 77 (11).
2.3.2. (E)-3-(3,4-dihydroxyphenyl)-N-butylacrylamide (ACAF4)
Yield (%): 50. 1H NMR (CDCl3, 400 MHz) δ (ppm) 7.54 (d, J=15.5 Hz, 1H, Hβ), 7.13 (d, J=2.0
Hz, 1H, H2), 6.97 (dd, J=8.3, 2.0 Hz, 1H, H6), 6.86 (d, J=8.2 Hz, 1H, H5), 6.20 (d, J=15.5 Hz,
1H, Hα), 5.70 (t, J=5.5 Hz, 1H, CONH), 3.37 (m, 2H, CH2), 1.54 (m, 2H, CH2), 1.38 (m, 2H,
CH2), 0.94 (t, J=6.9 Hz, 3H, CH3). 13C NMR (CDCl3, 100 MHz) δ (ppm) 167.9 (CONH), 147.3
(C4), 145.3 (C3), 140.8 (Cβ), 120.7 (C6), 117.0 (Cα), 115.1 (C5), 113.7 (C2), 38.9 (NHCH2),
31.2 (CH2), 19.7 (CH2), 12.7 (CH3). EI-MS m/z 235 (M•+, 6), 178 (35), 163 (100), 145 (15), 134
(14), 117 (12), 89 (16), 77 (10).
2.3.3. (E)-3-(3,4-dihydroxyphenyl)-N-hexylacrylamide (ACAF6)
Yield (%): 59. 1H NMR (DMSO-d6, 400 MHz) δ (ppm) 9.37 (1H, s, OH), 9.14 (1H, s, OH), 7.96
(t, J=5.5 Hz, 1H, CONH), 7.22 (d, J=15.7 Hz, 1H, Hβ), 6.92 (d, J=1.8 Hz, 1H, H2), 6.83 (dd,
J=8.1, 1.8 Hz, 1H, H6), 6.74 (d, J=8.1 Hz, 1H, H5), 6.32 (d, J=15.7 Hz, 1H, Hα), 3.14 (m, 2H,
13
CH2), 1.36 (m, 8H, (CH2)4), 0.87 (t, J=6.7 Hz, 3H, CH3). C NMR (DMSO-d6, 100 MHz) δ
(ppm) 165.4 (CONH), 147.4 (C4), 145.7 (C3), 139.0 (Cβ), 126.6 (C1), 120.5 (C6), 118.8 (Cα),
115.9 (C5), 113.9 (C2), 38.6 (NHCH2), 31.1 (CH2), 29.3 (CH2), 26.3 (CH2), 22.2 (CH2), 14.1
(CH3). EI-MS m/z (%) 263 (M•+, 5), 178 (36), 163 (100), 145 (14), 134 (14), 117 (12), 89 (18),
77 (10).
2.3.4. (E)-3-(3,4-dihydroxyphenyl)-N-octylacrylamide (ACAF8)
1
Yield (%): 45. H NMR (DMSO-d6, 400 MHz) δ (ppm) 9.19 (s, 2H, 2xOH), 7.93 (t, J=5.5 Hz,
1H, CONH), 7.22 (d, J=15.7 Hz, 1H, Hβ), 6.94 (d, J=1.8 Hz, 1H, H2), 6.83 (dd, J=8.3, 1.8 Hz,
1H, H6), 6.74 (d, J=8.1 Hz, 1H, H5), 6.32 (d, J=15.7 Hz, 1H, Hα), 3.14 (m, 2H, CH2), 1.43 (m,
13
2H, CH2), 1.26 (m, 2H, (CH2)5), 0.86 (t, J=6.7 Hz, 3H, CH3). C NMR (DMSO-d6, 100 MHz)
166.6 (CONH), 148.6 (C4), 146.9 (C3), 140.2 (Cβ), 127.8 (C1), 121.7 (C6), 120.0 (Cα), 117.1
(C5), 115.1 (C2), 39.9 (NHCH2), 32.6 (CH2), 30.6 (CH2), 30.1 (CH2), 30.0 (CH2), 27.9 (CH2),
7

23.5 (CH2), 15.3 (CH3). EI-MS m/z 291 (M•+, 10), 178 (34), 163 (100), 145 (13), 134 (15), 117
(13), 89 (18), 77 (8).
2.3.5. (E)-3-(3,4-dihydroxyphenyl)-N-decylacrylamide (ACAF10)
Yield (%): 55.4. 1H NMR (DMSO-d6, 400 MHz) δ (ppm) 9.20 (s, 2H, 2xOH), 7.93 (t, J=5.4 Hz,
1H, CONH), 7.22 (d, J=15.7 Hz, 1H, Hβ), 6.94 (d, J=1.6 Hz, 1H, H2), 6.83 (dd, J=8.1, 1.6 Hz,
1H, H6), 6.74 (d, J=8.3 Hz, 1H, H5), 6.32 (d, J=15.7 Hz, 1H, Hα), 1.33 (m, 16H, (CH2)8), 0.88
13
(t, J=6.9 Hz, 3H, CH3). C NMR (DMSO-d6, 100 MHz) 165.7 (CONH), 147.7 (C4), 145.9
(C3), 139.3 (Cβ), 126.9 (C1), 120.7 (C6), 119.1 (Cα), 116.2 (C5), 114.3 (C2), 39.1 (NHCH2),
31.8 (CH2), 29.7 (CH2), 29.5 (CH2), 29.4 (CH2), 29.2 (CH2), 29.1 (CH2), 26.9 (CH2), 22.6 (CH2),
14.4 (CH3). EI-MS m/z 319 (M•+, 8), 178 (40), 163 (100), 145 (14), 134 (15), 117 (11), 89 (17),
77 (10).
2.3.6. (E)-3-(3,4-dihydroxyphenyl)-N-dodecylacrylamide (ACAF12)
Yield (%): 55.4. 1H NMR (DMSO-d6, 400 MHz) δ (ppm) 9.21 (s, 2H, 2xOH), 7.93 (t, J=5.5 Hz,
1H, CONH), 7.22 (d, J=15.7 Hz, 1H, Hβ), 6.94 (d, J=1.6 Hz, 1H, H2), 6.83 (dd, J=8.1, 1.7 Hz,
1H, H6), 6.74 (d, J=8.3 Hz, 1H, H5), 6.32 (d, J=15.7 Hz, 1H, Hα), 3.14 (m, 2H, CH2), 1.30 (m,
13
20H, (CH2)10), 0.85 (t, J=6.7 Hz, 3H, CH3). C NMR (DMSO-d6, 100 MHz) 165.7 (CONH),
147.7 (C4), 145.9 (C3), 139.3 (Cβ), 126.9 (C1), 120.7 (C6), 119.1 (Cα), 116.2 (C5), 114.3 (C2),
46.3 (NHCH2), 39.1 (CH2), 31.8 (CH2), 29.7 (CH2), 29.5 (CH2), 29.4 (2xCH2), 29.2 (CH2), 29.1
(CH2), 26.9 (CH2), 22.6 (CH2), 14.5 (CH2). EI-MS m/z 347.1 (M•+ EI-MS m/z (%) 263 (M•+, 5),
178 (36), 163 (100), 145 (20), 134 (18), 117 (10), 89 (17), 77 (7).
2.4. Cell Culture
PC12 cells were a kind gift from Professor Lloyd A. Greene (Department of Pathology and Cell
Biology, Columbia University, New York, NY, USA)14. They were seeded in plates coated with
collagen obtained from rat tail and grown in RPMI 1640 medium, supplemented with 10% heat-
inactivated HS, 5% heat-inactivated FBS, 100 U/ml penicillin G, and 100 mg/ml streptomycin in
a 5% CO2 incubator at 37 ºC. Two-thirds of the growing medium was changed every 2–3 days
8

and cell passages were carried out in collagen coated dishes using trypsin 0.25% to detach cells
once a week.
2.5. Cell viability assay
PC12 cells (6 × 105 cells/mL) were seeded in 96-well plates in serum-free RPMI-1640 medium
and treated with the synthesized derivatives at the final concentrations of 1, 5 and 25 µM or NGF
at 5 or 50 ng/mL. After 48 h, cell viability was measured by the MTT reduction assay as
previously describe 15. Briefly, the media were carefully removed from the wells after
centrifugation at 500g for 10 min. Then cells were incubated with MTT solution (0.5 mg/mL
final concentration) for 90 min at 37 ◦C. Formazan crystals, which are formed due to the
reduction of MTT by dehydrogenase enzymes in viable cells, were dissolved in 200 µL DMSO
for 90 min. The absorbance was measured at 570 nm with background correction at 650 nm
using a microplate reader (Bio-Rad model). The increase in viability was calculated by dividing
the absorbance of treated cells by the absorbance of untreated control cells.
For evaluation of the role of TrkA signaling in the promotion of survival induced by caffeic acid
amides, a selective TrkA inhibitor, K252a, was co-incubated with the test compounds at 50 and
100 nM and the survival rates were compared to the cells treated with the synthesized derivatives
alone.
PC12 cells (6 × 105 cells/mL) were seeded in collagen coated 96-well plates in serum-free
medium (80 µL per well). Ten µL of NGF diluted in RPMI-1640 at the final concentrations of 5
or 50 ng/ml or the test compounds at the final concentrations of 1, 5 and 25 µM were added in
triplicate. Synthesized derivatives were first dissolved in dimethyl sulphoxide and then diluted to
the appropriate concentration in growth medium before use and added to the culture medium.
The final concentration of DMSO did not exceed 0.25%, which had no effect on the PC12 cells
(data not shown). After 48 h, cell viability was measured by the MTT reduction assay as
previously describe 15. Briefly, the media were carefully removed from the wells after
centrifugation at 500g for 10 min. Then cells were incubated with 20 µL MTT solution (0.5
mg/mL final concentration) for 90 min at 37 ◦C. Formazan crystals, which are formed due to the
reduction of MTT by dehydrogenase enzymes in viable cells, were dissolved in 200 µL DMSO
for 90 min. The absorbance was measured at 570 nm with background correction at 650 nm
9

using a microplate reader (Bio-Rad model). The increase in viability was calculated by dividing
the absorbance of treated cells by the absorbance of untreated control cells.
For evaluation of the role of TrkA signaling in promotion of survival induced by caffeic acid
amides, 50 and 100 nM K252a (a selective TrkA inhibitor), dissolved in DMSO, was added to
the culture 1 h prior to the addition of caffeic acid amides (25 µM) and NGF (5 or 50 ng/ml).
Then, MTT assay was performed as described above. The survival rates were compared to the
cells treated with the synthesized derivatives and NGF alone.
2.6. Differentiation analysis
PC12 cells (5×104 cells/mL) were cultured on collagen-coated 12-well plates in low serum RPMI
medium (2% HS and 1% FBS). After overnight growth, cells were treated with the synthesized
amides (5 and 25 µM) with or without NGF (5 and 50 ng/mL). After 3 days of incubation, the
growth medium was replaced with fresh low serum medium with the same concentrations of the
test compounds and NGF. Neurite outgrowth was quantified after 8 days of incubation, as the
proportion of neurites greater than or equal to the length of one cell body. These determinations
were made with an inverted microscope (model eclipse TS100-F). The total number of neurites,
as well as the neurites that were equal to or more than 100 µm, were counted. The number of
neurites was counted in at least five randomly selected microscopic fields (magnification power
of 20X). Each experiment was repeated at least 5 times.
2.7. Western blotting
PC12 cells (6 × 105cells/mL) were seeded in collagen-coated 35 mm dishes in normal serum
medium for 48h, then shifted to low serum medium (2% HS and 1% FBS) for 24 h prior to
exposure to test compounds (5 and 25 µM) and NGF (5 ng/mL) for the indicated time intervals.
Cells were first washed with PBS, and then scraped in ice cold lysis buffer solution (20 mM Tris
base, 150 mM NaCl, 1% Np40, 1 mM EDTA, 5% sodium deoxycholate and 0.1% SDS, pH 8.0)
containing phenylmethylsulfonyl fluoride 1 µM, Na4O7P2 10 mM, Na3VO4 2 mM, and incubated
on ice for 15 min. The cellular debris was removed by centrifugation (12,000 g for 20 min) at
10

4°C, and the cell lysate were carefully transferred to micro-centrifuge tubes. The protein
concentration was measured by BCA assay kit using bovine serum albumin as a standard. The
cell lysate (25 µg) was separated on 7.5% SDS−PAGE and transferred onto the PVDF membrane
at 60 V for 3h. Nonspecific binding sites on the membranes were blocked with 4% BSA
dissolved in Tris buffer saline containing 0.1% Tween-20 (TBST) for 50 min at room
temperature. Blots were incubated with the appropriate antibodies; anti-p44/p42 MAPK
(ERK1/2, 1:1000), anti-phospho-p44/p42 MAPK (p-ERK1/2, Thr202/Tyr204, 1:1500), anti-
TrkA (1:250), anti-AKT (1:250), anti-phospho-AKT (ser473, 1:1500) and anti-phospho-TrkA
(1:1000) overnight at 4°C. After three washes with TBST, the blots were incubated with
appropriate horseradish peroxidase-conjugated rabbit anti-mouse or mouse anti-rabbit secondary
antibodies (1:1000) at room temperature for 1 h. The blots were washed 3 times with TBST, and
the proteins were detected by enhanced chemiluminescence (ECL) (GE Healthcare,
Buckinghamshire, UK) and images were obtained with G: Box chemi-XR5 GeneSys image
analyzer. The intensities of bands were calculated with Software Gene Tools SyneGene for
Windows.
2.8. Statistical analysis
All data are presented as mean S.E.M based on at least three independent experiments. All
statistical analyses were performed by ANOVA followed by appropriate post hoc test using
SPSS version 22.0 (SPSS Inc., Chicago, IL, USA). P Values of less than 0.05 were considered
significant.
2.9. Molecular docking study
Molecular docking study of CA and the promising compound of this study ACAF12 with PI3K
as conceivable molecular target was carried out using AutoDock 4.2 and AutoDock Tools 1.5.4
program. The 3D crystal structure of target protein co-crystallized with wortmannin, as cognate
ligand (PDB code: 1E7U) was downloaded from RSCB protein data bank (http://www.pdb.org).
The pre-treatment process was performed to prepare both the ligand and protein for docking
11

steps. The PDB file of protein was refined as following steps: 1) removing water molecules and
crystallographic ligand 2) polar hydrogens addition 3) calculation of collman charges 4)
generation of PDBQT file. The 3D structure of CA and ACAF12 was constructed using
Chem3D Ultra 12.0 software. Generated structures of ligands were submitted to a molecular
mechanic conformational search (MM+ force felid) and the geometry of the most stable
conformer of each ligand was energetically optimized by semi-empirical method (AM1) in
Hyperchem 7 software. In the next step, the optimized structures were transferred to ADT 1.5.4;
the Gasteiger charges were added and the degree of torsions was defined to generate PDBT
format.
The active site of protein was selected based on the ligand binding domain of wortmannin and
the grid parameter file (gpf) was created using AutoGrid. Numbers of points in x, y, z
dimensions were set to 60*60*60 with 0.375 Ǻ grid spacing and grid center was set as 23.57,
61.04, 21.72 in each dimension, respectively. Consequently, docking parameter file was defined
based on rigid macromolecule docking procedure using Lamarckian Genetic Algorithm (LGA),
with 100 runs and 2500000 energy evaluations, other parameters were left as default. The most
favorable docked pose was selected according to the minimum free energy of protein-ligand
complex and was used for further exploration of binding interaction.
2.10. Drug-likeness properties study
The theoretical study of drug-like properties of target compounds was performed in the
Molinspirations chemiinformatic on-line program (http://www.molinspiration.com). The ‘drug-
likeness’ of all synthesized compounds was assessed applying Lipinski’s ‘rule of five’.
12

3. Results
3.1. Chemistry
The structures and synthetic strategy pursued to obtain caffeic acid amide derivatives is depicted
in Scheme 1. Carboxamide derivatives were obtained by
a
benzotriazol-1-
yloxytripyrolidinophosphonium hexafluorophosphate (PyBOP) assisted amidation reaction
between the caffeic acid and the appropriate alkylamine. With the exception of ACAF3, the
compounds were obtained in moderate yields. The structural identity of all compounds was
confirmed by nuclear magnetic resonance (NMR) spectroscopy (1H NMR, 13C NMR, and
DEPT135) and electronic impact mass spectrometry (EI-MS).
Scheme 1. Synthetic strategy pursued to obtain caffeic acid amide derivatives. (a) 1. DIPEA, DMF, PyBOP,
dichloromethane, 0 °C, 30 min; 2. appropriate amine, rt, 6h.
3.2. Effects of caffeic acid amide derivatives on survival of PC12 cells in serum-
deprived conditions
Neuroprotective activity of caffeic acid amide derivatives against serum deprivation-induced cell
death was assessed by MTT assay. NGF (5, 50 ng/ml) served as a positive control in the
experiments. Cells treated with caffeic acid amides at the concentration of 25 µM, showed a
significant increase in survival compared to untreated cells. The cell viability increased to
157.6%, 201.6%, 183.1%, 157.8%, 192.6% and 230.5% for the cells treated with 25 µM of
13

ACAF3, ACAF4, ACAF6, ACAF8, ACAF10 and ACAF12, respectively. ACAF12 was the
most active compound (Figure 1).
400
∗
1µM
5µM
∗
25µM
300
∗
∗
∗
200
100
0
∗
∗
∗
∗
Figure 1: Effects of caffeic amide derivatives on survival of PC12 cells. PC12 cells were seeded in 96-well plates
in serum-free medium and were induced with synthesized caffeic acid amide derivatives at the final concentrations
of 1, 5 and 25 µM or NGF at 5 or 50 ng/ml. After 48 h, cell viability was measured by the MTT assay. The results
are plotted as the percentage cell survival in treated cells relative to untreated cells. The data are presented as the
mean S.E.M of 3-5 independent experiments. * p < 0.05.
3.3. Caffeic acid amides potentiate NGF-induced neuronal differentiation in PC12
cells
We next studied the effect of caffeic acid amides on the neurite outgrowth of PC12 cells. As it is
shown in Figure 2A-C, incubation of neuronal cells with NGF 5 and 50 ng/ml induced neurite
outgrowth. The incubation of cells with the synthesized compounds alone did not promote
neurite outgrowth (data not shown). On the other hand, the concomitant incubation of some
caffeic acid amides at the concentrations of 5 µM (ACAF4, ACAF6 and ACAF8) and 25 µM
(ACAF3, ACAF4, ACAF6 and ACAF8) with low dose NGF (5 ng/mL), significantly increased
the total number of neurites and also the number of neurites longer than 100 µM, compared to
14

NGF treatment alone (Figure 2B and C). Caffeic acid amides at 1 µM did not potentiate
neuritogenesis induced by NGF (data not shown). Co-treatment of cells with NGF and long alkyl
chain amide derivatives (ACAF10 and ACAF12) failed to potentiate neuritogenesis induced by
NGF.
A
Control
NGF (5 ng/mL)
NGF (50ng/mL)
ACAF3 (5µM) + NGF (5 ng/mL)
ACAF3 (25µM) + NGF (5 ng/mL)
ACAF4 (5µM) + NGF (5 ng/mL)
ACAF4 (25µM) + NGF (5 ng/mL)
ACAF6 (5µM) + NGF (5 ng/mL)
ACAF6 (25µM) + NGF (5 ng/mL)
ACAF8 (5µM) + NGF (5 ng/mL)
ACAF8 (25µM) + NGF (5 ng/mL)
15

B
160
140
120
100
80
5 µM
25 µM
∗
∗
∗
∗
∗
∗
60
∗
40
20
0
C
20
15
10
5
∗
5 µM
25 µM
∗
∗
∗
∗
∗
∗
0
Figure 2: Caffeic acid amides potentiate NGF-induced neuronal differentiationin PC12 cells. PC12 cells were
seeded at a low density (5 × 104 cells/well) and exposed to thecaffeic acid amides at the concentrations of 5 and 25
µM with or without NGF (5 ng/mL) or NGF alone (5 and 50 ng/mL). After 8 days, neurite outgrowth was quantified
as the proportion of neurites greater than or equal to the length of one cell body.A: morphological changes in PC12
cells treated with indicated agents. B: The total number of neurites were counted with a phasecontrast microscope in
16

five fields. C: The neurites that were equal to or more than 100 µm were counted. Results (mean S.E.M.) are
representative of five independent experiments (n=5). * p<0.05.
3.4. Caffeic acid amides prompted neuroprotective effects in a TrkA-independent
manner
To evaluate the relation between the neuronal survival enhancement effect of caffeic acid amide
derivatives with TrkA receptor agonism, the effect of K252a, a selective TrkA inhibitor, was
examined on caffeic acid amide-promoted rescue of PC12 cells in serum-free conditions.
In these experiments, serum-free PC12 cells were treated with 25 µM of synthesized derivatives
in the presence or absence of 50 and 100 nM K252a. As expected, K252a was able to inhibit
NGF-induced rescue of PC12 cells, however, it did not significantly alter the neurotrophic effect
of the caffeic acid amides (Figure 3A). This observation indicates that TrkA signaling is not
essential for the promotion of neuronal cell survival induced by caffeic acid amide derivatives.
These results were validated by western blot analyses. Two caffeic acid amides with short and
long alkyl chain amide derivatives (ACAF4 and ACAF12, respectively) were tested for their
ability to activate TrkA receptor. To investigate whether caffeic acid amides can activate the
TrkA receptor, PC12 cells were treated with ACAF4 and ACAF12 (5 and 25 µM) and NGF 5
ng/ml as positive control for 60 min, and the cell lysates were immunoblotted for detection of
TrkA and phospho-TrkA. NGF 5 ng/mL triggered TrkA phosphorylation in PC12 cells (Figure
3B); however, none of the test compounds could cause any detectable phosphorylation in TrkA
receptor. ACAF4 and ACAF12 were also tested for their ability to potentiate NGF-induced
TrkA phosphorylation, but their concomitant incubation with NGF did not induce any alteration
in the phosphorylation level of TrkA compared to cells treated with NGF alone (Figure 3C).
A
17

350
300
250
200
150
100
50
− Κ252α
K252a 50 nM (60min)
K252a 100 nM (60min)
∗
∗
∗
∗
∗
∗
∗
∗
∗
∗ ∗
∗ ∗
∗
∗
∗
∗
∗
∗ ∗
0
18

Figure 3: Caffeic acid amides prompted neuroprotective effects in a TrkA independent manner. A: Serum-
deprived PC12 cells in 96-well plates were treated for 60 min without or with K252a (50 and 100 nM) and then were
either untreated or stimulated with caffeic acid amides (25 µM) and NGF (5 or 50 ng/ml). Subsequently, cell
viability was measured by MTT assay. Data are plotted as the percentage cell survival in treated cells relative to
untreated cells and are representative as mean S.E.M of three independent experiments (n=3). * p<0.05.B:
Cultured PC12 cells were stimulated with ACAF4 and ACAF12 alone, NGF alone, or a combination of both for 1h.
Phosphorylation of TRKA were detected by immunoblotting with antibodies to p-TRKA, TRKA. Data are
expressed as mean SEM (n=3, p<0.05).
3.5. Caffeic acid amides induce ERK1/2 activation
We next examined the effects of a short (ACAF4) and a long chain caffeic acid amide
(ACAF12) added alone or concomitant with low dose NGF to the PC12 cells on ERK1/2
signaling. As shown in Figure 4, NGF (5 ng/ml) induced, as expected, an intense
phosphorylation of ERK1/2. Also ACAF4 and ACAF12 alone significantly induced ERK1/2
phosphorylation in a dose-dependent manner. ACAF12 was more effective than ACAF4 to
induce ERK1/2 activation (Figure 4A-B). Interestingly, when the cells were treated with the test
compounds and low dose NGF, ACAF12 significantly increased NGF-induced ERK1/2
phosphorylation. Similarly, ACAF4 potentiated NGF-induced ERK1/2 activation, but to a lower
extent compared to ACAF12 (Figure 4A-B). These results suggest that the long chain alkyl
amide derivative was more effective than short chain alkyl amide derivate in activation of
ERK1/2 pathway.
19

p-ERK/ERK
Control
ACAF4 (5uM)
ACAF4 (25uM)
ACAF4 (5uM) + NGF (5 ng/mL)
ACAF4 (25uM) + NGF (5 ng/mL)
p-ERK/ERK
Control
ACAF12 (5uM)
ACAF12 (25uM)
ACAF12 (5uM) + NGF (5 ng/mL)
ACAF12 (25uM) + NGF (5 ng/mL)

Figure 4: Caffeic acid amides induce the activation of ERK1/2. PC12 cells were stimulated with 5 ng/ml NGF
alone and ACAF4 and ACAF12 (5 and 25 µM) alone, and with a combination of both for 2h. A: Phosphorylated
form and total ERK were detected by immunoblotting. B: The band density of the phosphorylated ERK1/2 was
normalized to the band density of the total ERK1/2 present in the same lane and then each value of the ratio p-
ERK/ERK were divided by p-ERK/ERK of NGF (5 ng/ml). Each value of caffeic acid amides(5 and 25 µM) alone
was expressed relative to control. The results are shown as mean SEM (n=4, p < 0.05).
3.6. ACAF12 induces the phosphorylation of AKT
We also tested whether or not the neurotrophic effects of ACAF4 and ACAF12 (5 and 25 µM)
are linked to AKT signaling pathway. PC12 cells were treated with ACAF4 and ACAF12 (5 and
25 µM) either alone or concomitant with low dose NGF for 2h (Figure 5). ACAF12 at 5 and 25
µM significantly increased AKT phosphorylation. On the other hand, ACAF4 did not increase
AKT phosphorylation. Furthermore, co-incubation of caffeic acid amide derivatives with NGF
did not increase the AKT phosphorylation to a degree stronger than that induced by NGF alone
(Figure5A-B).
21

B
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
∗
∗
Figure 5: ACAF12 induces the activation of AKT. PC12 cells were incubated with NGF (5 ng/ml) alone and
ACAF4 and ACAF12 (5 and 25 µM) alone, and with a combination of both for 2h. A: Phosphorylated form and
total AKT were detected by western blot analysis.B: Activation of AKT was determined by normalization of the
22

band density from the phosphorylated form (p) with that of the total form and then each value of the ratio p-
AKT/AKT were divided by p-AKT/AKT of NGF (5 ng/ml). Values obtained with caffeic acid amides(5 and 25 µM)
alone were compared with control. Data are shown as mean SEM (n=4, p < 0.05).
3.7. In silico docking simulation of compounds CA and ACAF12 with PI3K
Several studies have suggested that PI3K and its downstream pathways are involved in the
regulation of neuronal survival and neurogenesis both in vitro and in vivo 16-18. PI3K can activate
downstream kinases such as AKT, PKC and ERK family members 19. Therefore, molecular
docking study was carried out to investigate the binding and mode of interaction of CA and its
promising amide derivative (ACAF12) with PI3K as a possible molecular target responsible for
neurotrophic activity of synthesized compounds 20. To this end, flexible ligand/rigid
macromolecule docking was performed using crystal structure of PI3K (PDB ID: 1E7U) with its
cognate ligand, wortmannin. Docking procedure was further optimized and validated via re-
docking of the co-crystallized conformation of cognate ligand (wortmannin) into the PI3K
binding site. Consequently, RMSD (root mean square deviation) of the best-docked
conformation of the native ligand from the experimental one was used for validation of the
docking study. The best-docked and experimental conformation of wortmannin correlated quite
well with an RMSD of 1.73Ǻ and binding free energy of −8.96 Kcal/mol (Table 1). Validated
docking protocol was used for molecular docking study of test compounds under investigation.
The results are presented in Table 1. The binding profile of CA and ACAF12 docked with PI3K
(PDB code: 1E7U) is depicted in Figure 6. ACAF12 and CA demonstrated different binding
orientations in PI3K active site. In the case of ACAF12, both of phenolic hydroxyl groups were
involved in H-bond interaction with Asp841 and Asp964 of PI3K active site. In return,
orientation of phenolic part of CA resulted to only one H-bond interaction between meta-
phenolic hydroxyl moiety and Glu880. In addition, carbonyl group of dodecyl amide derivative
(ACAF12) was involved in H-bond network interaction with Ile-881 and Val882 of N-terminal
lobe; while carbonyl of acidic derivative (CA) interacted with Lys833 of binding site. It is
evident from Figure 6 that additional dodecylamide pendant of ACAF12 got directed toward C-
terminal lobe covered by hydrophobic residues such as Met953, Phe961 and Ile963 and produced
hydrophobic interactions with this hydrophobic cavity. Such particular induced binding pose of
23

ACAF12 into the PI3K active site could justify its superior binding free energy (∆Gb=-6.70
Kcal/mol) and observed neuroprotective effect against serum deprivation-induced cell death
compared to its acidic counterpart CAF (∆Gb=-5.05 Kcal/mol) and sustenance PI3K as its
probable molecular target underlying the neurotrophic effect observed by amide derivatives of
HCAs.
24

Table 1. Binding interactions and results of molecular docking study of caffeic acid (CA), caffeic acid amide
analogues (ACAF) and wortmannin.
Compound
CA
Ki (µm)
Amino acid
∆Gb
(Kcal/mol)
Atom of the
ligand
3-OH
C=O
3-OH
4-OH
C=O
3-OH
4-OH
C=O
3-OH
4-OH
C=O
4-OH
C=O
3-OH
4-OH
3-OH
4-OH
C=O
C=O
---
Glu 880
Lys 833
Asp841
Asp964
Val882
Asp841
Asp964
Val882
Asp841
Asp964
Val882
Asp964
Val882
Asp841
Asp964
Asp841
Asp964
Ile-881
Val882
---
-5.05
-7.29
199.83
4.57
3.59
3.06
ACAF3
ACAF4
ACAF6
-7.43
-7.52
-6.41
-6.14
19.87
31.78
ACAF8
ACAF10
-6.70
-8.96
12.28
ACAF12
0.28
Wortmannin*
25

*The best-docked and experimental conformation of wortmannin correlated quite well with an RMSD of 1.73Ǻ.
Figure 6. Comparative binding mode of CA (∆Gb=-5.05 Kcal/mol) and ACAF12 (∆Gb=-6.70Kcal/mol) (backbones
were colored in green and orange, respectively) in PI3K active site. H-bond interactions were demonstrated as light
blue lines.
3.8. Drug-likeness
In order to achieve insight about the structural features that cause the observed experimental
results, we evaluated the molecular properties important for drug accessibility to its molecular
target. Based on the Lipinski’s rule of five, molecular properties of the active compounds must
pass at least 3 of the 4 rules in order to consider them as suitable drug candidates: H-bond donors
(HBD) ≤ 5, H-bond acceptors (HDA) ≤ 10, molecular mass (MM) ≤ 500, and the calculated logP
21
(clogP) ≤ 5 . Calculated molecular descriptors are presented in Table 2. The results indicated
that values of three molecular descriptors (HBD, HBA and MM) were within the acceptable
limits of Lipinski’s rule for CA and its alkyl amide derivatives.
26

Table 2. Assessment of the molecular descriptors of drug-likeness applying Lipinski’s rule of five.
Compound cLogP≤5
NH/OH≤5
N/O≤10
MW≤500
221.26
235.28
263.34
291.39
319.44
347.50
180.16
1.68
2.24
3.25
4.26
5.27
6.28
0.94
4
4
4
4
4
4
4
3
3
3
3
3
3
3
ACAF3
ACAF4
ACAF6
ACAF8
ACAF10
ACAF12
CAF
cLogP, Calculated decimal logarithm of octanol/water partition coefficient. NH/OH, Number of H-bond donors.
N/O, Number of H-bond acceptors. MW, Molecular weight.
27

4. Discussion
In this study, we synthesized a series of caffeic acid amide analogues with variable alkyl chain
lengths and tested their neurotrophic action in PC12 neuronal cells. We found that these
compounds are capable of increasing cell survival in serum-deprived conditions and also
enhancing the differentiation induced by NGF. We also found that the neurotrophic effects of
caffeic acid amides are not mediated by direct agonistic action on TrkA receptors, but rather
through the activation of ERK1/2 and AKT signaling pathways.
CA is a phenolic compound widely distributed in some species of the plant kingdom and
22
abundantly present in the human diet . Previous studies have revealed neuroprotective effects
of CA and its derivatives against oxidative stress and ischemic injury 23,24. For example, it has
been shown that caffeic acid ester fraction from the plant Erigeron breviscapus, increases
25
neuronal cell viability in rat primary cultured microglia . Furthermore, CAPE has been shown
26
to block 6-hydroxydopamine-inducedneurotoxicity in cerebellar granule neurons and also to
be effective in induction of differentiation in PC12 cells 10.
A previous study of our group recently showed that caffeic acid esters possess neurotrophic
20
activities . As an improved stability is generally considered advantageous in vivo, in order to
increase the stability of caffeic acid esters we synthetized caffeic acid amide derivatives with
variable alkyl chain lengths (C3-C12). The amide analogues have higher hydrolytic energies of
activation and are able to avoid enzymatic hydrolysis by esterases 27.
We first examined the effects of caffeic acid amides in protection of neuronal cells against
serum deprivation-induced cell death. It was found that pretreatment with caffeic acid amides at
the concentration of 25 µM increased neruonal survival compared to untreated cells. Another CA
derivative, CAPE, has been previously demonstrated to have neuroprotective effects against
28
H2O2-induced cytotoxicity in PC12 cells and also against IFN-γ/LPS-induced injury in nigral
dopaminergic neurons 9. Moreover, several other polyphenolic compounds have been reported to
enhance neuronal survival in serum deprived neuronal cells. For instance, Lin and colleagues
have shown that in PC12 cells, luteolin, a flavonoid compound, decreases serum withdrawal-
induced cytotoxicity 29. Other examples of polyphenolic neuroprotective compounds that
stimulate proliferation of PC12 cells include epigallocatechin-3-gallate 30, 5-hydroxy-
3,6,7,8,3',4'-hexamethoxyflavone and nobiletin 31.
28

This study also revealed that co-incubation of some of the synthesized caffeic acid amides
(ACAF3, ACAF4, ACAF6 and ACAF8) with low dose NGF (5 ng/ml), increased the total
number of neurites and also those longer than 100 µm compared to NGF alone. Although long
alkyl chain caffeic acid amides (ACAF10 and ACAF12 amides) were active compounds in
enhancement of neuronal survival, they did not show any significant effect on neurite outgrowth.
Since the incubation time in the neurite assay was significantly longer than survival assay (8 vs.
2 days) longer chain amides may exert toxic effects in a longer incubation time. Correlation of
the length of the alkyl chain with toxicity has been previously shown 32.
Different studies have shown that age-related disorders such as Alzheimer’s disease are
33
associated with NGF deficiency . Hence, the NGF-potentiating action of caffeic acid amide
derivatives could be considered of therapeutic value in neurodegenerative diseases related to
NGF insufficiency. Several studies have identified a variety of polyphenols that potentiate
neurotrophins’ (NGF and BDNF) effect in neurite outgrowth induction. For instance, green tea
polyphenols potentiate the neuritogenic effect of NGF and BDNF in PC12 cells 34,35. Genistein, a
polyphenol compound belonging to the class of flavonoids has also the potential to enhance
36
NGF-induced neurite outgrowth in PC12 cells . Another compound, isorhamnetin, a flavonol
isolated mainly from G. biloba plant, has also been shown to enhance the neuronal
differentiation induced by NGF in cultured PC12 cells 37.
We also demonstrated that K252a, an inhibitor of TrkA receptor, did not significantly alter the
effect of caffeic acid amide derivatives on cell survival. Additionally, western blot experiments
showed that ACAF4 and ACAF12 (5 and 25 µM) did not induce TrkA phosphorylation. Further,
none of the test compounds increased NGF-induced TrkA phosphorylation. Previous studies
have shown that inhibition of TrkA receptors block most of the effects prompted by NGF in
PC12 cells, including rescue from cell death induced by serum withdrawal 38. Our findings
suggest that the effects of caffeic acid amides in rescuing cells from serum deprivation-induced
death and promoting NGF-induced neuronal differentiation are regulated via pathways that are
independent of direct activation of TrkA receptor, and probably act through downstream
pathways.
Other studies have also shown that neurotrophic action of phenolic compounds such as 5-
31 39
hydroxy-3, 6, 7, 8, 3', 4'-hexamethoxyflavone and nobiletin are not associated with TrkA
activation in PC12 cells. There are indeed only few compounds that directly activate Trk
29