Diaryl hydrazones as multifunctional inhibitors of amyloid self-assembly

Article abstract of DOI:10.1021/bi3012059

The design and application of an effective, new class of multifunctional small molecule inhibitors of amyloid self-assembly are described. Several compounds based on the diaryl hydrazone scaffold were designed. Forty-four substituted derivatives of this core structure were synthesized using a variety of benzaldehydes and phenylhydrazines and characterized. The inhibitor candidates were evaluated in multiple assays, including the inhibition of amyloid β (Aβ) fibrillogenesis and oligomer formation and the reverse processes, the disassembly of preformed fibrils and oligomers. Because the structure of the hydrazone-based inhibitors mimics the redox features of the antioxidant resveratrol, the radical scavenging effect of the compounds was evaluated by colorimetric assays against 2,2-diphenyl-1-picrylhydrazyl and superoxide radicals. The hydrazone scaffold was active in all of the different assays. The structure-activity relationship revealed that the substituents on the aromatic rings had a considerable effect on the overall activity of the compounds. The inhibitors showed strong activity in fibrillogenesis inhibition and disassembly, and even greater potency in the inhibition of oligomer formation and oligomer disassembly. Supporting the quantitative fluorometric and colorimetric assays, size exclusion chromatographic studies indicated that the best compounds practically eliminated or substantially inhibited the formation of soluble, aggregated Aβ species, as well. Atomic force microscopy was also applied to monitor the morphology of Aβ deposits. The compounds also possessed the predicted antioxidant properties; approximately 30% of the synthesized compounds showed a radical scavenging effect equal to or better than that of resveratrol or ascorbic acid.

Full text of DOI:10.1021/bi3012059

Article
pubs.acs.org/biochemistry
Diaryl Hydrazones as Multifunctional Inhibitors of Amyloid Self-
Assembly
Bela Torok,† Abha Sood,† Seema Bag,† Rekha Tulsan,† Sanjukta Ghosh,† Dmitry Borkin,†
́
̈
̈
Arleen R. Kennedy,† Michelle Melanson,† Richard Madden,† Weihong Zhou,† Harry LeVine, III,‡
and Marianna Torok*,†
̈
̈
†Department of Chemistry, University of Massachusetts Boston, 100 Morrissey Boulevard, Boston, Massachusetts 02125-3393,
United States
‡Department of Cellular and Molecular Biochemistry, Center on Aging, Center for Structural Biology, University of Kentucky, 800
South Limestone Street, Lexington, Kentucky 40536-0230, United States
S
* Supporting Information
ABSTRACT: The design and application of an effective, new class
of multifunctional small molecule inhibitors of amyloid self-
assembly are described. Several compounds based on the diaryl
hydrazone scaffold were designed. Forty-four substituted deriva-
tives of this core structure were synthesized using a variety of
benzaldehydes and phenylhydrazines and characterized. The
inhibitor candidates were evaluated in multiple assays, including
the inhibition of amyloid β (Aβ) fibrillogenesis and oligomer
formation and the reverse processes, the disassembly of preformed
fibrils and oligomers. Because the structure of the hydrazone-based
inhibitors mimics the redox features of the antioxidant resveratrol,
the radical scavenging effect of the compounds was evaluated by
colorimetric assays against 2,2-diphenyl-1-picrylhydrazyl and superoxide radicals. The hydrazone scaffold was active in all of the
different assays. The structure−activity relationship revealed that the substituents on the aromatic rings had a considerable effect
on the overall activity of the compounds. The inhibitors showed strong activity in fibrillogenesis inhibition and disassembly, and
even greater potency in the inhibition of oligomer formation and oligomer disassembly. Supporting the quantitative fluorometric
and colorimetric assays, size exclusion chromatographic studies indicated that the best compounds practically eliminated or
substantially inhibited the formation of soluble, aggregated Aβ species, as well. Atomic force microscopy was also applied to
monitor the morphology of Aβ deposits. The compounds also possessed the predicted antioxidant properties; approximately
30% of the synthesized compounds showed a radical scavenging effect equal to or better than that of resveratrol or ascorbic acid.
he formation of misfolded, amyloid-like protein assemblies
in cells and tissues is observed in many aging-related
of primary importance,13,14 although it is not the exclusive factor
in governing amyloid formation.15 The literature about small
organic molecule inhibitors is less systematic, focusing on their
biopharmaceutical properties rather than their mechanism of
action.4,16,17
Oxidative stress is believed to contribute to neurodegeneration
in AD. Because in vivo studies indicate elevated levels of oxidative
stress in the AD-affected brain,18 including antioxidant properties
in the design of Aβ self-assembly inhibitor compounds appears to
be desirable.19,20 The precise relationship among Aβ self-
assembly, neurotoxicity, and oxidative stress is still somewhat
unclear. Aβ and some of its derivatives generate free radicals
spontaneously upon oligomerization and fibrillogenesis, most
likely with the contribution of metal ions.21−23 Formation of free
radicals during the disassembly of preformed Aβ fibrils24 and the
T
diseases such as Alzheimer’s disease (AD). The major
constituent of these protein aggregates in the case of AD is the
amyloid β (Aβ) peptide.1,2 Traditionally, the insoluble Aβ fibrils
were proposed to be one of the major causes of AD; however,
more recently, soluble oligomeric species of Aβ were shown to
exhibit even stronger and more potent neurotoxicity.3 While the
application of small molecule fibril and oligomer formation
inhibitors has been a popular treatment strategy,4 relatively few
studies address the development of compounds that affect
multiple toxic processes.5,6
Many inhibitors of Aβ self-assembly have been identified,
including small organic molecules, peptides, peptidomimetics,
and proteins.7−9 These compounds have been categorized as
anti-fibril or anti-oligomer compounds. Oligomer structures
were generally detected with conformation-specific antibod-
ies.10−12 Peptide-based inhibitors have been frequently used to
investigate the driving forces responsible for self-assembly, and
π−π stacking between aromatic residues has been identified to be
Received: September 4, 2012
Revised: December 1, 2012
Published: January 24, 2013
© 2013 American Chemical Society
1137
dx.doi.org/10.1021/bi3012059 | Biochemistry 2013, 52, 1137−1148

Biochemistry
Article
free radical scavenging capacity of Aβ itself have also been
observed.25 Regardless of whether oxidative stress precedes
amyloid assembly or the level of reactive oxygen species (ROS)
increases as a consequence of changes in the oligomeric state of
Aβ, free radicals negatively affect cellular function and
survival.26,27 Optimally, small molecule agents targeting Aβ
self-assembly and/or disassembly should not induce the
formation of ROS, and they should scavenge any ROS present.
Dietary antioxidants, especially plant-derived polyphenols, may
provide beneficial effects in AD through multiple mecha-
nisms.28−30 Although they can protect against the effects of
ROS, most of the natural antioxidants are poor drug candidates
because of a lack of metabolic stability, oral bioavailability, or
brain penetration.31
Herein, we describe the synthesis and evaluate the structure−
activity relationship of a new class of multifunctional compounds
that interfere with the self-assembly of Aβ into fibrils and
oligomers and also are able to combat the effects of harmful free
radicals. A diverse group of diaryl-hydrazones were synthesized
and tested in this study. While a number of useful therapeutic
agents are hydrazones and/or hydrazines, including central
nervous system penetrant drugs,32 such compounds have been
infrequently used in AD-related studies.33,34
Biochemical Assays. Sodium dihydrogenphosphate, diso-
dium hydrogenphosphate, sodium azide, sodium hydroxide,
sodium chloride, glycine, dimethyl sulfoxide, and thioflavin-T
were purchased from Sigma-Aldrich. 1,1,1,3,3,3-Hexafluoro-2-
propanol (HFIP), dimethyl sulfoxide (DMSO), fluorescamine,
ultrapure Tween 20, tetramethylbenzidine (free base), N,N-
dimethylacetamide, tetrabutylammonium borohydride, and 30%
(w/w) H2O2 were also obtained from Sigma-Aldrich. Lyophi-
lized synthetic Aβ(1−40) and Aβ(1−42) peptides (purity of
>95%) and N-α-biotinyl-Aβ(1−42) (bio-Aβ42) were from
AnaSpec. Fatty acid-free fraction V bovine serum albumin was
purchased from Boehringer-Mannheim. Streptavidin-bound
horseradish peroxidase (SA-HRP) was from Rockland. Neu-
trAvidin (NA) was obtained from Pierce. High-binding 9018
enzyme-linked immunosorbent assay (ELISA) plates were
purchased from Costar. The standards used in the size exclusion
chromatography−high-performance liquid chromatography
(SEC−HPLC) investigations and the reference compounds
(resveratrol and ascorbic acid) used in radical scavenging were
purchased from Sigma-Aldrich. β-Nicotinamide adenine dinu-
cleotide (NADH), phenazine methosulfate (PMS), and nitro-
blue tetrazolium (NBT) were from VWR, while 2,2-bis(4-tert-
octylphenyl)-1-picrylhydrazine (DPPH) was obtained from
Calbiochem.
Determination of Inhibitor Activity in Aβ Fibrillo-
genesis and Fibril Disassembly by a Thioflavin-T
Fluorescence Assay. Inhibition of Fibril Formation. The
assay was conducted using a standard published procedure.35−37
The synthetic lyophilized Aβ(1−40) or Aβ(1−42) peptide was
dissolved in 100 mM NaOH to a concentration of 40 mg/mL.
This solution was diluted in 10 mM HEPES, 100 mM NaCl,
0.02% NaN3 (pH 7.4) buffer to a final peptide concentration of
100 μM. The use of a strong base (NaOH) avoided the
isoelectric point of Aβ, preventing low-pH-assisted fibril or
amorphous aggregate formation, providing the peptide in a
monomeric state.38,39 The hydrazones were dissolved (10 mM)
in DMSO and added to the Aβ samples in HEPES buffer in a 1:1
molar ratio of inhibitor to Aβ during the initial assays. This ratio
was varied between 1 and 0.025 (1, 0.75, 0.5, 0.25, 0.125, 0.05,
and 0.025) during the EC50 determinations. The mixtures were
vortexed for 30 s, and the solutions were incubated at 37 °C (77
rpm shaking) and then samples withdrawn for analysis after the
desired incubation time.
MATERIALS AND METHODS
■
Synthesis. The substituted hydrazines, the benzaldehydes,
and the 19F nuclear magnetic resonance (NMR) reference
compound CFCl3 were purchased from Aldrich. DMSO-d6 and
CDCl3 used as a solvent (99.8%) for the NMR studies were
Cambridge Isotope Laboratories products. Other solvents used
in synthesis with a minimal purity of 99.5% were from Fisher.
The mass spectrometric identification of the products was
conducted with an Agilent 6850 gas chromatograph-5973 mass
spectrometer system (70 eV electron impact ionization) using a
30 m DB-5 column (J&W Scientific). An Agilent high-
performance liquid chromatography−mass spectrometry
(HPLC−MS) instrument (Series 1200 HPLC-6130 Quadrupole
MS) was also used for the identification of certain compounds
that appeared to be thermally unstable above 250 °C, the injector
temperature for gas chromatography and mass spectrometry
1
(GC−MS). The H, 13C, and 19F NMR spectra were recorded
with a 300 MHz superconducting Varian Gemini 300 NMR
spectrometer, in DMSO-d6 and CDCl3 with tetramethylsilane
and CFCl3 as internal standards.
Disassembly of Preformed Fibrils.35 Fibrils were grown as
described above. The growth of the fibrils was followed by
thioflavin-T (THT) fluorescence measurements until the THT
fluorescence intensity leveled off. The solution was then divided
into aliquots for the disassembly studies. Ten millimolar stock
compound solutions were prepared by dissolving the inhibitor
compounds in DMSO, and this stock solution was added to the
fibril samples to yield an inhibitor concentration of 100 μM (1%
DMSO) in the initial assays. For the EC50 determinations, the
concentration of the test compounds was varied between 2.5 and
100 μM (inhibitor:Aβ molar ratios of 1, 0.75, 0.5, 0.25, 0.125,
0.05, and 0.025) while the peptide concentration was kept
constant at 100 μM. After being vigorously vortexed for 30 s, the
solutions were re-incubated at 37 °C while being gently shaken
(77 rpm), and the decrease in the amount of fibrils in each sample
was measured after 4 days by THT fluorescence. To ensure that
the inhibitors did not displace THT from the fibrils to decrease
fluorescence intensity and thus cause false positive results, several
experiments were conducted with compounds that showed
significant inhibition of fibril formation or disassembly. The
Synthesis of 1-Benzylidene-2-phenylhydrazine. In a 15 mL
Erlenmeyer flask, 0.106 g (1 mmol) of benzaldehyde and 0.108 g
(1 mmol) of phenylhydrazine were dissolved in 2 mL of
dichloromethane. The reaction mixture was kept at room
temperature for 10 min and then placed in a freezer (−20 °C) for
30 min. During this period, the product slowly crystallized from
the mixture. The crystalline 1-benzylidene-2-phenylhydrazine
was filtered and the product air-dried for 12 h. The purity was
verified using GC−MS, liquid chromatography and mass
spectrometry (LC−MS), and NMR. Impurities were removed
by recrystallization or preparative TLC to yield at least 98% pure
product.
Synthesis of the Substituted Hydrazones. The method
described above was used for the synthesis of all other
hydrazones studied in this work, applying a variety of substituted
hydrazines and benzaldehydes. The structure of the products was
confirmed using mass spectrometry and 1H, 19F (when
applicable), and 13C NMR. The spectral data of the compounds
are listed in the Supporting Information.
1138
dx.doi.org/10.1021/bi3012059 | Biochemistry 2013, 52, 1137−1148

Biochemistry
Article
inhibitor-free Aβ control sample was labeled with THT, and the
fluorescence intensity was measured. Then an inhibitor (all
compounds were tested separately) was added to the sample, and
the fluorescence measurement was repeated. No significant
difference was obtained in the observed intensities, indicating
that the inhibitors did not affect the fluorescence assay.
Determination of the Amount of Fibrils by THT
Fluorescence.40−42 The relative amount of fibrils was
determined by THT fluorescence spectroscopy. The fluores-
cence intensities of the inhibitor-containing samples were
compared to those without any inhibitor (control). The
fluorescence measurements were taken using a Hitachi F-2500
fluorescence spectrophotometer. The incubated peptide sol-
utions were briefly vortexed before each measurement, and then
3.5 μL aliquots of the suspended fibrils were withdrawn and
added to 700 μL of 5 μM THT prepared freshly in 50 mM
glycine-NaOH (pH 8.5) buffer. The maximal fluorescence
intensity of these mixtures was measured at an emission
wavelength of 484 5 nm with a preset excitation wavelength
of 435 nm. None of the inhibitor compounds showed
fluorescence intensity in this region. For the purposes of a
screening assay, the fibril signal generated under the conditions
of the assay in the presence of 1% DMSO (solvent control) and
in the absence of compound is taken to be 100%. The EC50 values
of potent compounds were determined as described pre-
viously.35,43
dispensed into each well of a polypropylene 96-well plate and
100 μL of PBS containing the desired concentration of test
compound and 1% DMSO added to initiate oligomer formation
at room temperature. After 30 min, 50 μL of 0.3% (v/v) Tween
20 was added to stop oligomer assembly. Fifty microliters of this
mixture was then assayed for oligomer content by a single-site
streptavidin-based assay.
Biotinyl-Aβ(1−42) Single-Site Streptavidin-Based Assay for
the Measurement of Biotinyl-Aβ(1−42) Oligomer Con-
tent.45,46 Fifty microliters of 1 μg/mL NA in 10 mM NaPi
(pH 7.5) was coated per well overnight at 4 °C on Costar 9018
high-binding ELISA plates sealed with adhesive plastic film. The
plates were blocked by the addition of 200 μL of phosphate-
buffered saline (PBS) [10 mM sodium phosphate, 150 mM NaCl
(pH 7.5), and 0.1% (v/v) Tween 20] at room temperature for 1−
2 h and stored at 4 °C. After removal of the blocking solution, a
sample containing a mixture of oligomers and monomers of the
biotinylated peptide (50 μL containing up to 10 nM biotinyl-Aβ)
was allowed to bind for 2 h at room temperature. The wells were
washed three times with TBST [20 mM Tris-HCl, 34 mM NaCl
(pH 7.5), and 0.1% (v/v) Tween 20] on a Biotek EL × 50
automated plate washer. After the sample had been washed, 50
μL of a 1:20000 SA-HRP dilution in PBS and 0.1% (v/v) Tween
20 was added, the plate was sealed, and the incubation was
continued for 1 h at room temperature. The plate was washed
again with TBST; 100 μL of a tetramethylbenzidine/H2O2
substrate solution was added, and the plate was incubated at
room temperature for 5−10 min. The OD450 was determined on
a Biotech Synergy HT plate reader after the reaction had been
stopped with 100 μL of 1% (v/v) H2SO4. For the purposes of a
screening assay, the oligomer signal generated under the
conditions of the assay in the presence of 1% DMSO (solvent
control) and in the absence of compound was taken to be 100%.
Assay for Aβ Oligomer Disassembly. Preparation of
Preformed Biotinyl-Aβ(1−42) Oligomers. Biotinyl-Aβ(1−42)
was disaggregated as described for the assembly assay and
DMSO added to the dried film to produce an 8 μg/mL bio42
stock solution. After 10 min, the disaggregated peptide in DMSO
was diluted 50-fold into PBS [20 mM sodium phosphate and 145
mM NaCl (pH 7.5)] in a polypropylene container to a monomer
concentration of 33.7 nM (0.16 μg/mL). After 1 h at room
temperature, an equal volume of PBS with 0.6% (v/v) Tween 20
was added to stop oligomer formation and stabilize the
oligomers. This mixture of oligomeric and monomeric biotinyl-
Aβ(1−42) was either used immediately or stored at −75 °C in
polypropylene tubes for up to 6 months. These oligomers are
>70 kDa, and their size distribution determined by size exclusion
chromatography is similar to that of Aβ oligomers from AD
brain.47
Atomic Force Microscopy. The morphologies of the
incubated peptide samples were studied using atomic force
microscopy (AFM) as described previously.43 Aliquots (2 μL)
were spotted on freshly cleaved mica sheets and air-dried. The
buffer salts were washed off with deionized water. The
measuements were taken using a Bruker-Innova SPM instru-
ment.
Size Exclusion Chromatography.44 Twenty microliters of
day 0, 4, and 6 samples were centrifuged at 16100g (∼13200
rpm) (5415D Eppendorf centrifuge) for 35 min to sediment long
insoluble fibrils, and the supernatant was stored at 4 °C between
injections to slow any further fibril growth. Ten microliters of the
supernatant was injected. Size exclusion chromatography was
performed on the HPLC system (Jasco PU-20895, quaternary
gradient pump attached to a Hewlett-Packard Series 1050 UV
detector) equipped with a TOSOH TSK-G3000SWXL column
[30 cm × 7.8 mm (inside diameter), 5 μm particle size] with a
separation range of 1−500 kDa (globular proteins). The mobile
phase was 0.1 M phosphate buffer [0.05% sodium azide and 0.1
M Na2SO4 (pH 6.7)] for the detection of monomeric and
oligomeric species. The flow rate was adjusted to 0.5 mL/min,
and the elution peaks were detected using a UV−vis detector at
254 nm (λ).
Determination of the Inhibitor Activity of Aβ Oligomer
Formation and Oligomer Disassembly by Biotinyl-
Streptavidin Assays.45,46 Because Aβ(1−40) is the most
abundant form of the peptide and readily forms fibrils, we
decided to use it in the assays described above. However, Aβ(1−
42) was used for anti-oligomer assays because Aβ(1−40) forms
oligomers poorly unless a stimulant is applied at the low peptide
concentration (10 nM) used to avoid fibril formation.
Inhibition of Aβ Oligomer Assembly. Biotinyl-Aβ(1−42)
stored as a 1 mg/mL solution in HFIP at −75 °C was dried with a
nitrogen stream, treated with neat trifluoroacetic acid for 10 min
at room temperature to disaggregate the peptide, and dissolved
to a concentration of 500 nM (50×) in DMSO as described
previously.45,46 Two microliters of monomeric peptide was
Oligomer Dissociation. Twenty-five microliters of 16.8 nM
preformed bio42 oligomers in PBS and 0.3% Tween 20 was
pipetted into wells of a polypropylene (wide well) 96-well plate
[0.5 mL well capacity (Fisher catalog no. 12565502)] followed
by 125 μL of PBS containing compound and 1% (v/v) DMSO.
The plate was sealed with a plastic sheet (Nunc 236366) and
shaken (150 rpm) at room temperature for 16−18 h. The
amount of biotinyl-Aβ(1−42) oligomers remaining was
measured by transferring 100 μL from each well to an NA-
coated (50 ng/well) well of a Costar 9018 ELISA plate and
quantified as described for oligomer assembly.
Determination of the Antioxidant Properties of the
Compounds. Scavenging of the DPPH Radical by the
Inhibitors.48 A 2,2-diphenyl-1-picrylhydrazyl radical (DPPH)
1139
dx.doi.org/10.1021/bi3012059 | Biochemistry 2013, 52, 1137−1148

Biochemistry
Article
stock solution was prepared in ethanol at a concentration of
105.3 μM. The test compound stock solutions were prepared by
a cosolvent method to enhance their solubility and keep the final
DMSO concentration below 0.1% in the assay. First, the
compounds were dissolved in DMSO at a concentration of 10
mM and then diluted with ethanol to a concentration of 0.2 mM.
Ninety-five microliters of the DPPH stock solution was
dispensed into each well of a 96-well microplate, and then 5
μL of each test compound stock solution was added. Thus, the
final concentrations of DPPH and test compounds in the assays
were 100 and 10 μM, respectively. The plate was wrapped in
aluminum foil and kept at 37 °C for 30 min. After incubation, the
decrease in DPPH absorption at 519 nm was measured by a
Versamax microplate reader at 30 and 60 min. The percent
scavenging was calculated using the expression [(Absc − Abstc)/
Absc] × 100, where Absc is absorption of the control sample that
contained no inhibitor and Abstc is the absorption measured in
the presence of the test compounds.
Scavenging of the Superoxide Radical by the
Inhibitors .49 β-Nicotinamide adenine dinucleotide (NADH,
1.5 mM), 0.03 mM phenazine methosulfate (PMS), and 0.375
mM nitroblue tetrazolium (NBT) solutions were prepared in
100 mM phosphate buffer (pH 7.4). The test compounds were
dissolved in DMSO (10 mM). Ten microliters of this compound
solution was diluted with 74 μL of 100 mM phosphate buffer (pH
7.4) and 76 μL of acetonitrile, yielding a 0.625 mM solution
containing <0.1% DMSO. One hundred seventy-one microliters
of 100 mM phosphate buffer (pH 7.4) was dispensed into each
well of the microtiter plate along with 9 μL of 0.625 mM test
compound solutions. Fifteen microliters of 0.375 mM NBT and
15 μL of 1.5 mM NADH were rapidly added to each well. Lastly,
15 μL of a 0.03 mM PMS solution was added to each well, except
the blank. The final concentrations of PMS, NBT, NADH, and
test compounds in the assay were 2, 25, 100, and 25 μM,
respectively. The production of reduced NBT was monitored by
a Versamax microplate reader at 560 nm. The absorbances were
measured after 0, 1, 6, 8, 10, 12, 14, and 15 min.49 The percent
superoxide scavenging activity was calculated using the
expression [(Absc − Abstc)/Absc] × 100, where Absc is the
absorption of the sample that contained no inhibitor and Abstc is
the absorption measured in the presence of the test compounds.
proposed compounds are diaryl-hydrazones that were inspired
by the structure of resveratrol. Through systematic modification,
their structure can readily be altered to fine-tune their effects on
Aβ assembly and antioxidant activity and possibly improve their
pharmacokinetic properties. These compounds are likely
antioxidants and could also affect π−π stacking interactions
between Aβ units. The design of these compounds was based on
the hypothesis that the tautomeric equilibrium between forms I
and II (Figure 1) would provide continuous conjugation and thus
a nearly uniform π-electron delocalization for the compounds.
The rapid tautomerism would result in an electronic structure
(III) in which all three atoms (C−N−N) would be of at least
partially sp2 character and could contribute to a conjugated
electron flow between the two aromatic rings. The general
structure of the hydrazones and their similarity to resveratrol are
illustrated in Figure 1.
On the basis of the reasoning described above, a variety of
diaryl-hydrazones were synthesized from commercially available
benzaldehydes and arylhydrazines. The basic synthetic proce-
dure for the preparation of these compounds is summarized in
Figure 2.
Figure 2. Synthesis of diaryl hydrazones.
The starting materials for the synthesis were selected to ensure
that hydrazones with varied substituents could be prepared. The
aryl groups in the products possess either electron-donating or
electron-withdrawing substituents, including fluorine-containing
substituents that are commonly thought to increase the
lipophilicity of the compounds.51,52 A combinatorial approach
was used to synthesize the compound library, possibly making
every possible product from the building block pool. Overall, 44
compounds were synthesized. The general structure of the
compounds is summarized in Figure 3.
Effect of Diaryl-Hydrazone-Based Inhibitors on Aβ
Self-Assembly. After completion of the syntheses and
structural validation, the compounds were subjected to several
biochemical assays. Because of the neurotoxicity associated with
both the fibrillar aggregates and soluble oligomeric species, these
assays included tests to evaluate the inhibitory potential of the
compounds in the self-assembly of the Aβ peptide (fibril and
oligomer formation) as well as the disassembly of the preformed
fibrils and oligomers. The well-known quantitative THT
fluorescence spectroscopy assay was applied for the determi-
nation the antifibrillogenic activity of the compounds.40−42 The
calculated intensity values were based on the background
(fluorescence of the THT alone)-corrected maximal fluores-
cence intensities in the 480−490 nm region of the emission
spectra after incubation for 4 days. The measurements were
taken when the fibril assembly reached a plateau in the control
sample. All data were normalized to inhibitor-free controls.
These initial assays were conducted with a 1:1 Aβ:inhibitor ratio
at 100 μM Aβ; thus, the original inhibitor concentration was 100
μM. The compounds were also screened for anti-oligomer
activity using the quantitative biotinyl-Aβ(1−42) single-site
streptavidin-based assay,45,46 at the initial Aβ:inhibitor ratio of
RESULTS AND DISCUSSION
■
Design of the Proposed Diaryl-Hydrazone-Based
Inhibitors. Resveratrol (Figure 1), a natural product antiox-
idant, served as a starting point in our inhibitor design. This
compound is commonly used in AD-related studies as an Aβ self-
assembly inhibitor as well as a radical scavenger.29,30,50 Our
Figure 1. General structures of hydrazones and resveratrol.
1140
dx.doi.org/10.1021/bi3012059 | Biochemistry 2013, 52, 1137−1148

Biochemistry
Article
most compounds, 35 and 37 caused the formation of much
shorter and nonuniform deposits, indicating the presence of
protofibrils and amorphous protein deposits. This suggests that
these compounds favored different pathways as compared to the
unaltered control and most modulators or allowed the
combination of pathways and remodeling.
Both Aβ(1−40) and Aβ(1−42) are found in fibrillar
aggregates in AD brain.1−3 Because Aβ(1−40) is the most
abundant form of the peptide and readily forms fibrils, it was used
in the fibrillogenesis assays. Aβ(1−42) was used for anti-
oligomer assays because Aβ(1−40) forms oligomers poorly
unless a stimulant, which introduces interpretation issues, is
applied at the low peptide concentration (10 nM) used to avoid
fibril formation. Because the two peptide variants may behave
differently in the fibril formation assays, experiments were
conducted with both peptides under identical conditions in the
presence and absence of selected promoter and inhibitor
compounds. Fibril growth was investigated as a function of
time and expressed as the intensity of the measured fluorescence.
The data are shown in Figure 5.
The graphs are in agreement with previous observations of
Aβ(1−40) and Aβ(1−42) fibril formation. The curves show that
Aβ(1−40) fibril growth in the modulator-free sample (control)
requires a longer time to reach maximal fluorescence (5 days)
than Aβ(1−42), which displays a much faster initial rate reaching
maximal fluorescence after 3 days. THT fluorescence decreases
after the maximal ITHT values, most likely because of aging of
separate fibrils into bundles that lose β-sheet structure or that the
THT cannot penetrate. The major differences between the two
peptides are kinetic issues: Aβ(1−42) exhibits a higher initial rate
of fibrillogenesis, while Aβ(1−40) forms fibril bundles faster.
Other than this, the two peptides behaved similarly under our
conditions. The addition of either promoters (3 and 16) or
inhibitors (35 and 37) of fibril formation also affected the two
peptides similarly. The percentile values of fibrillogenesis
promotion or inhibition at the maximum of control {3 days
[Aβ(1−42)] vs 5 days [Aβ(1−40)]} and the shape of the curves
are close, indicating that both Aβ variants behaved similarly
under our assay conditions.
Figure 3. Structure of diaryl hydrazones used in this study.
0.0002 at 0.01 μM biotinyl-Aβ(1−42) (50 μM inhibitor). The
intensities of the inhibitor-containing samples (Isample) were
normalized to that of the control sample (Icontrol) containing Aβ
and solvent control. The inhibition data expressed as percentile
values with which a compound decreased the expected signal
(control) are listed in Table 1:
Many of the hydrazones possessed significant activity in
fibrillogenesis inhibition and disassembly. Eleven compounds
(9−11, 13, 17−19, and 35−38) showed better than 50% activity
in the fibril assembly inhibition and disassembly of preformed
fibrils. Two compounds (17 and 37) exhibited practically
quantitative inhibition in these assays. However, the most
pronounced effect was observed for inhibition of oligomer
formation and disassembly of the preformed oligomers. Of the
43 compounds, 30 exhibited >50% inhibition (36 showed better
then 40%). A similar success rate was observed in the disassembly
of the preformed oligomers. Many inhibitors (19 of 43)
completely disassembled (90−100%) the oligomers, while
overall 22 compounds showed >50% disassembly. The basic
scaffold appears to be a suitable fit for inhibitor design, as a broad
variety of hydrazones exhibited good to complete inhibitory
activity in the initial assays.
Comparing the results of the assays described above, we found
the hydrazones commonly appear to act preferentially either as
an anti-fibril agent or as an anti-oligomer agent. Compounds 1−8
demonstrate this very clearly; while nearly completely inhibiting
oligomer formation, they do not inhibit fibrillogenesis or even
promote it. When compounds exhibit anti-fibril activity, they
both inhibit assembly and promote disassembly. A similar
phenomenon was observed with anti-oligomer compounds. This
Isample
inhibition (%) = 100 −
× 100
Icontrol
In certain cases when compounds promoted self-assembly, Isample
> Icontrol; therefore, negative inhibition percentile values were
obtained.
After the quantitative THT fluorescence assays, an independ-
ent method, atomic force microscopy (AFM), was used to
confirm the THT data and observe the morphology of the
samples obtained with and without (control) hydrazones.
Illustrative AFM images of a control sample and samples
prepared in the presence of inhibitor and promoter compounds
are shown in Figure 4
The AFM images are in good correlation with the THT
fluorescence data (Table 1). The control sample (Figure 4,
Control) shows the expected, dense network of fibrils. Images
obtained in the presence of compounds 1, 3, and 16 indicate even
more extensive fibril formation, very similar to that of the control.
Considering that these compounds showed a mild fibril growth-
promoting effect (Table 1), the images are in agreement with the
quantitative data. In contrast, compounds 13, 19, and 35 visibly
decreased the density of the fibrils and 37 resulted in an almost
complete disappearance of fibrils. While similar fibrillar
morphology was observed in the control and in samples with
1141
dx.doi.org/10.1021/bi3012059 | Biochemistry 2013, 52, 1137−1148

Biochemistry
Article
a
Table 1. Inhibition of the Aβ Self-Assembly by the Diaryl Hydrazones
b
c
d
e
compd
log P
inhibitionfibril (%)
disassemblyfibril (%)
inhibitionoligomer (%)
disassemblyoligomer (%)
1
3.03
3.16
4.25
3.98
3.58
3.98
3.45
4.50
3.48
4.50
4.50
2.33
3.40
4.25
4.34
4.34
3.13
4.50
3.78
4.41
3.95
4.09
5.17
5.26
5.17
5.17
3.82
4.15
5.11
5.11
4.15
3.82
5.02
4.62
3.58
4.36
2.88
3.52
2.88
2.88
2.85
3.90
3.70
3.42
−37 24.6
20 12.3
−21 16.0
−34 17.3
−12 15.4
33 29.8
−38 20.6
−
48 1.7
35 1.7
41 6.0
28 7.2
33 7.7
31 3.7
22 19.0
33 8.3
41 15.2
56 11.1
63 11.8
55 4.3
26 16.4
−12 11.6
−16 15.2
−14 16.4
91 4.2
12 11.0
59 2.1
27 10.8
17 2.9
27 1.4
34 15.4
29 14.0
20 8.2
27 6.2
47 5.7
−
98 1.5
100 0.2
93 1.0
95 0.8
97 1.0
96 0.5
84 4.2
95 1.6
76 6.6
27 5.1
34 15.8
43 8.2
−
100 0.5
100 2.3
100 0.9
100 0.2
100 0.6
100 0.2
42 6.8
2
3
4
5
6
7
8
4
6
3
0
9.3
9
52 11.8
53 16.0
83 0.8
6.5
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
11.1
5.2
23 9.8
19 7.6
54 30.2
−8.2 32.9
16 7.6
−
88 6.7
98 4.1
93 0.2
13 10.2
75 8.2
99 2.3
95 3.9
99 0.3
−15 4.6
−16 7.8
−15 5.6
−26 18.6
87 28.7
56 18.8
54 18.6
39 31.6
−18.7 85.0
−17 10.5
37 1.5
7
8.8
100 1.1
55 14.2
82 2.5
32 9.3
43 7.4
99 0.8
41 6.4
28 5.0
73 6.8
31 14.1
60 48.5
85 1.5
0
7.8
15 2.9
91 1.9
54 12.5
16 5.4
100 0.4
31 6.4
−5 1.6
−11 2.9
−4 4.5
23 5.8
100 0.1
10 0.8
100 0.4
98 2.3
97 0.7
98 0.7
−13 6.2
15 14.7
11 1.8
−31 25.9
3
40.0
−32 27.0
14 15.4
−50 9.6
−28 26.8
10 9.8
44 16.1
41 10.5
40 15.3
−8 49.8
59 4.6
−46 13.3
−29 30.5
21 17.0
38 33.7
63 12.9
53 29.5
81 19.3
78 0.7
4
0.6
57 18.9
80 6.3
1
11.1
75 1.5
24 7.8
99 1.4
75 4.0
41 1.0
50 11.9
45 17.8
23 17.3
63 8.2
55 18.2
87 1.4
89 2.0
−25 40.3
72 7.1
41 16.8
37 5.4
47 14.2
36 15.3
100 1.1
64 16.6
53 28.0
96 3.7
4
3.9
−43 17.7
−18 3.9
19 17.2
15 0.2
96 0.4
57 1.5
97 0.7
100 0.5
a
b
The values represent means the standard deviation of the data (n = 2−7). Inhibition of the Aβ fibrillogenesis by the compound at a 1:1
c
d
inhibitor:Aβ ratio, with 100 μM Aβ. Disassembly of preformed Aβ fibrills by the compound at a 1:1 inhibitor:Aβ ratio, with 100 μM Aβ. Inhibition
of the Aβ oligomer formation by the compound at an Aβ:inhibitor ratio of 0.0002, with 10 nM Aβ. Disassembly of preformed Aβ oligomers by the
e
compound at an Aβ:inhibitor ratio of 0.0002, with 10 nM Aβ.
is consistent with general observations in the literature10,11 as
well as our own recent work with organofluorine compounds.43
Compounds 35, 36, 38, 43, and 44 exhibited good to complete
inhibition in all anti-fibril and anti-oligomer assays.
The data also indicated that 24 compounds (1−7, 14−16, 22,
23, 25, 26, 30, 31, 33−36, and 41−44) appeared to be effective
against oligomer assembly and in oligomer disassembly. Given
the different time frame of the assays (30 min for inhibition of
oligomer assembly vs overnight for disassembly), it is possible
that certain compounds that are rapid disassembly agents would
appear to be assembly inhibitors. To confirm whether the
compounds are dual oligomer formation inhibitors and
disassembly agents as opposed to rapid-acting disassembly
agents, we performed short time dissociation assays. These assays
were conducted for 2 h, binding to the NA plate at three
concentrations for each compound spanning the range of EC50
values for the dissociation assay (50, 25, and 12.5 μM). No
dissociation was observed for those compounds during that
period of time compared to a compound (2,5-dihydroxybenzoic
acid) that is a rapid dissociator and does result in disassembly
over that period of time.44 These observations support the
contention that the oligomer assembly inhibitors acted by
1142
dx.doi.org/10.1021/bi3012059 | Biochemistry 2013, 52, 1137−1148

Biochemistry
Article
Figure 4. Atomic force microscopy images of Aβ(1−40) samples incubated without (control) and with the modulators for 6 days. The compound
numbers denote the inhibitor compounds as they are shown in Figure 3.
listed above are dual assembly inhibitors and dissociators of
Aβ(1−42) oligomers.
Compounds exhibiting >50% inhibition in the screening
assays were titrated to determine their EC50 values. The EC50
calculations followed a previously described method.34,42
Fluorescence intensity versus molar ratio functions were used
to determine the relative potency of inhibitors like the analysis of
the Michaelis−Menten kinetics or ligand binding to macro-
molecules35,43
ECmax
EC50 + P
P
ITHT = 100 −
where ITHT is the fluorescence intensity of the inhibitor-
containing sample (as a percent of control), P is the inhibitor:Aβ
molar ratio, EC50 is the median inhibitor constant, and ECmax is
the maximal level of inhibition. The double-reciprocal plot of the
formula allowed the determination of EC50. Because inhibitor:Aβ
molar ratios (P) were applied in the formula, the EC50 values
were obtained as a ratio as well. Multiplying this ratio by the Aβ
concentration provided the values in units of concentration
(micromolar). The EC50 values of the most active compounds
obtained in the different assays are listed in Table 2.
The data in Table 2 confirm the single-concentration
screening observations described above. The large proportion
of highly active compounds indicate that the diaryl-hydrazone
skeleton is a promising scaffold for the synthesis of potent Aβ
self-assembly inhibitors. The overall activity of the compounds,
in particular in oligomer-related assays, is significant; the low
micromolar EC50 values indicate that these compounds are
potential leads for further inhibitor development. Most
importantly, three compounds (34, 35, and 38) interfered with
both stages of Aβ assembly, desirably affecting all of the fibril and
oligomer assembly and disassembly processes.
Figure 5. Effect of incubation time on the amount of fibrils formed
(expressed as fluorescence intensity, ITHT) in (A) Aβ(1−40) and (B)
Aβ(1−42) samples incubated without (control) and with the effector
compounds for 6 days. The compounds numbers denote the inhibitor
◆
compounds as they are shown in Figure 3: ( ) control and in the
■
▲
●
presence of ( ) 3, ( ) 16, (×) 35, and ( ) 37. The values represent
means the standard deviation of the data (n = 3).
inhibiting the formation of oligomers rather than acting as rapid
dissociators in the assembly assay. Therefore, the compounds
1143
dx.doi.org/10.1021/bi3012059 | Biochemistry 2013, 52, 1137−1148

Biochemistry
Article
The first (top) chromatogram clearly indicates the significant
amount of soluble oligomers (both high and low molecular
masses at ∼11 and 16 min retention times) in the uninhibited
control and the decreasing amount or complete lack of such
species in the presence of the selected compounds. For example,
compound 34 appears to be one of the best oligomer inhibitors
in the studied group based on the data listed in Tables 1 and 2.
Accordingly, the chromatogram of the sample prepared with 34
shows a minimal amount of soluble oligomeric species. For our
purposes, these chromatograms suggest that while the
compounds exhibit strong fibril inhibition they do not promote
the formation or stabilization of oligomers. In certain cases, e.g.,
37, the chromatogram indicates a relatively low peptide
concentration. This suggests that a significant amount of peptide
was removed during centrifugation. From the AFM image of the
same sample [37 (Figure 4)], sampled before centrifugation, the
morphology of the peptide fibrils contained much shorter units
as well as nonuniformly shaped, possibly amorphous, Aβ deposits
as opposed to an extended fibrillar network (Figure 4, Control).
These deposits are not fibrillar aggregates and do not fluoresce in
the presence of THT. These observations explain the low
fluorescence values obtained, as well as the low recovery of
peptide observed in the SEC−HPLC chromatograms.
Antioxidant Properties of Diaryl-Hydrazones. Because
the roles of oxidative stress and harmful free radicals also appear
to be important in the development of AD,18−20 incorporating
antioxidant properties in the design of anti-AD compounds could
prove to be useful. Therefore, two commonly applied assays were
used to evaluate the radical scavenging activity of the hydrazones
mentioned above. One assay utilizes 2,2-diphenyl-1-picrylhy-
drazyl (DPPH), a stable free radical, and the decrease in the free
radical concentration is measured directly by the absorbance of
DPPH at a wavelength (λ) of 519 nm.48 The other assay involves
the generation of superoxide radicals in a nicotinamide adenine
dinucleotide (NADH)/phenazinemethosulfate (PMS) system
and the measurement of the absorbance of nitroblue tetrazolium
(NBT), an indicator molecule that turns blue when reduced by
superoxide. The free radical scavenging activities of the test
compounds are determined by their ability to inhibit the
reduction of NBT by superoxide.49 The data, illustrated in
Figures 7 and 8, are compared to those obtained with reference
compounds ascorbic acid53 and resveratrol,29,30,49 both of which
are well-known antioxidants. All compounds were tested for their
radical scavenging ability in these two assays.
Table 2. EC50 Values of Selected Diaryl-Hydrazones in the
Inhibition of Aβ Self-Assembly
a
b
c
d
e
EC50‑F
EC50‑DF
(μM)
EC50‑O
EC50‑DO
(μM)
compd log P
(μM)
(μM)
1
3.03
3.16
4.25
3.98
3.58
3.98
3.45
4.50
3.48
4.50
4.50
2.33
4.25
4.34
4.34
3.13
4.50
3.78
4.41
4.09
5.17
5.17
3.82
4.15
4.15
5.02
4.62
3.58
4.36
2.88
3.52
2.85
3.70
3.42
−
−
−
−
−
−
−
−
16
68
33
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
62.5
−
−
−
−
−
−
−
−
−
−
−
64
−
−
−
−
95
−
10.6
6.5
5.4
5.4
7
21
3
2
3
7
4
6.4
5.2
7.3
36
−
−
−
−
−
2.3
14.5
12.5
−
−
−
5
6
3.6
18
20
80
90
54
43
4.7
5.8
7.8
−
7
8
9
10
11
12
14
15
16
17
18
19
20
22
23
25
27
28
31
33
34
35
36
37
38
41
43
44
−
10
−
80
−
−
−
34
−
5
−
39
58
2.4
−
22
32
6.3
−
−
43
−
−
−
95
15
−
30
14
−
36
−
−
7.7
2.1
2.1
12
11
−
45
4.6
8
1.8
1.4
22
17
−
24
7.4
60
6.6
−
−
5.4
a
Dashes indicate no inhibition or an EC50 of >100 μM for fibril
inhibition and no inhibition or an EC50 of >50 μM for oligomer
b
c
inhibition. EC50 value of the inhibition of Aβ fibrillogenesis. EC50
d
value of the disassembly of preformed Aβ fibrils. EC50 value of the
inhibition of Aβ oligomer formation. EC50 value of the disassembly of
e
preformed Aβ oligomers.
In both assays, the percent reduction in the absorbance is
directly proportional to the amount of free radical scavenged and
serves as a quantitative measure of the antioxidant ability. The
hydrazone scaffold possesses significant antioxidant activity. In
the DPPH scavenging assay, ∼30% of the compounds showed
activity better than or equal to that of either ascorbic acid or
resveratrol. In several instances (29, 30, and 35), the hydrazones
scavenged more than 65% of the radicals, twice the activity of the
reference scavengers (ascorbic acid and resveratrol) at the same
concentration (10 μM in DPPH and 25 μM in superoxide
scavenging assays).
The hydrazones also exhibited activity against the superoxide
radical. Approximately 27% of the compounds possessed activity
equal to or greater than that of the reference compounds;
moreover, three compounds (19, 31, and 41) showed 2 times
and three compounds (39, 40, and 42) more than 3 times the
activity compared to that of ascorbic acid and resveratrol at the
same concentration.
Considering the exceptionally complex nature of Aβ self-
assembly and the possible stabilization formation of multiple
intermediates, e.g., neurotoxic oligomers, it is essential to know
whether compounds stabilize soluble Aβ species in the process of
inhibiting or reversing the formation of fibrils. For antifibrillo-
genic compounds, whether the inhibitor halts the self-assembly
at the Aβ monomer level or allows the formation of soluble
oligomers while preventing the formation of insoluble fibrils is
crucially important. As the THT fluorescence detects and
measures the insoluble fibrillar assemblies but not generally
soluble oligomers, the supernatant of the selected samples, after
centrifugation to remove large aggregates or fibrils, was subjected
to size exclusion chromatography (SEC−HPLC) to determine
the distribution of monomeric and oligomeric species that exist
in the inhibited solutions.44 One fibril promoter (16) and three
inhibitors (34, 36, and 37) were selected for these studies. The
chromatograms, including molecular mass calibration with
globular proteins, are shown in Figure 6.
1144
dx.doi.org/10.1021/bi3012059 | Biochemistry 2013, 52, 1137−1148

Biochemistry
Article
Figure 6. TSK G3000SWXL SEC−HPLC chromatograms of Aβ samples after centrifugation at 16100g (to remove fibrils) with or without selected
hydrazone inhibitors. The compound numbers refer to structures shown in Figure 3 and listed in Table 1. The molecular mass globular protein standards
used to calibrate the column elution profile were β-galactosidase (132 kDa), bovine serum albumin (66 kDa), superoxide dismutase (32 kDa), and
lysozyme (14.7 kDa). The intense peaks at ∼24 min represent DMSO used as a solvent during the preparation of the inhibitor solutions.
Figure 7. Radical scavenging activity of selected hydrazones (>10% activity) illustrated as the percent reduction of the absorbance of DPPH at 519 nm
(λ) after incubation for 60 min. The compound numbers refer to structures shown in Figure 3 and Table 1. Ascorbic acid (aa) and resveratrol were used
as reference compounds. The values represent means the standard deviation of the data (n = 3−4).
Conjugated electron structure and potentially oxidizable
atoms are associated with strong antioxidant properties,
characteristics that our hydrazones possess. As depicted in
Figure 1, the proposed tautomeric equilibrium via an 1,3-H shift
ensures that the compounds would have an extended largely
delocalized electron structure and the N atom is commonly
observed to react with high-energy oxygen species. While the
activities of the compounds in the two radical scavenging assays
generally correlate, a few compounds that showed strong activity
in the DPPH assay performed poorly in superoxide scavenging.
Similarly, some strong superoxide scavengers exhibited only
moderate activity against DPPH. The most likely explanation of
1145
dx.doi.org/10.1021/bi3012059 | Biochemistry 2013, 52, 1137−1148

Biochemistry
Article
Figure 8. Superoxide scavenging activity of selected hydrazones (>2% activity) illustrated as the percent reduction of the absorbance of NBT at 560 nm
(λ) after incubation for 10 min. The compound numbers refer to structures shown in Figure 3 and Table 1. Ascorbic acid (aa) and resveratrol were used
as reference compounds. The values represent means the standard deviation of the data (n = 3−4).
this phenomenon lies in the significantly different size of the two
radicals; the superoxide radical is small compared to the sterically
bulky DPPH radical. As superoxide is the physiologically relevant
radical, superoxide scavenging is considered to be of primary
importance.
The extended conjugated electron structure could also
contribute to the inhibition of the self-assembly of the Aβ
peptide most likely through aromatic π−π interactions. The
possible correlation between the radical scavenging effect and the
self-assembly inhibition is unclear at this level of research. While
there are several hydrazones that are strong inhibitors as well as
scavengers, there are compounds that show excellent inhibition
but poor radical scavenging or strong radical scavenging but poor
self-assembly inhibition. Thus, at present, it appears that the
radical scavenging and self-assembly activities are not linked
together.
Nonetheless, in the studied set of compounds sharing the same
core structure, a variety of activities were found from self-
assembly inhibitors to promoters. The significant changes in the
activity suggest that the substituents of the individual compounds
have a significant effect on the overall characteristics of the
hydrazones in the assays. Over the years, a variety of compounds
have been studied for their potential Aβ self-assembly inhibition
properties.4,7−13 Several recent papers were focused on the
mechanism of action of selected examples.54−59 While the
primary goal of this work is to identify hydrazone derivatives that
are active against Aβ self-assembly and can also scavenge free
radicals, analyzing the structure−activity relationship of
hydrazones could provide valuable information about the
druglike properties or the mode of action of these compounds.
The first important aspect that we considered was the
hydrophobicity, which would primarily predict the potential
bioavailability of the compounds and also affects their membrane
permeability. Therefore, the log P values, commonly used to
describe these features, were estimated using BioChemDraw
(version 9.0, CambridgeSoft) (Table 1). It was observed that the
log P values of the hydrazones studied ranged from 2.33 to 5.17.
According to the Lipinski rules,60 compounds for which 2 < log P
< 5 are considered potentially bioavailable. Thus, while there are
significant differences among the compounds, all can be
considered to possess appropriate log P values. Because the log
P values of the large majority of the compounds listed above fall
within the range of 3−5, no clear relationship can be found
between activity and hydrophobicity. Given the nature of the
hydrazones (strongly basic compounds) and the assay conditions
(slightly basic pH), the hydrazones will exist in their nonionized,
basic form and, thus, likely they will not act as surfactants.54,56 If
one considers the effect of substituents on the different activities,
several observations can be made, although an unambiguous
conclusion cannot be drawn.
Compounds that possess H as an R2 group appear to act as
strong fibrillogenesis promoters and excellent oligomer inhib-
itors (see compounds 1−7, 14, and 16). As several other
compounds with considerably small R2 groups (e.g., CF3; 21, 22,
25, 28, 30, 31, and 34−36) exhibit strong oligomer inhibition,
this phenomenon is probably linked to the size of the
substituents, rather than their chemical nature. The same effect
is even more pronounced for oligomer disassembly agents. A
majority of the 19 compounds (1−6, 14−16, 22, 25, 31, 33−36,
41, 43, and 44) that showed close to 100% disassembly in the
initial assays (Table 1) have H as R2 (11 of 19 compounds). This
ratio increases to 16 of 19 compounds if one considers the also
small CF3 group as R2. These compounds either are fibrillo-
genesis promoters or do not significantly affect the fibril
formation. Thus, on the basis of this, we can conclude that
when small substituents (especially H) are in the R2 position, a
potent selective anti-oligomer effect, in both assembly inhibition
and oligomer disassembly, can be expected. On the other side of
the scaffold, when R1 is a strongly polar group, such as COOH or
OH (e.g., 1, 21, 28, 31, 32, 41, and 42), the compound will most
likely promote fibril formation. In contrast, many of the strong
fibril inhibitors contain a NO2 substituent, mostly in the R1
position.
Compounds with NO2 and/or COOH groups dominate the
strong superoxide scavengers. All hydrazones that exhibited
>25% superoxide scavenging (10, 13, 19, 20, 31, 36, and 39−42)
contain either NO2, COOH, or, in a few cases, both. Not
surprisingly, two of the three most active superoxide scavengers
(>60% scavenging activity) (39 and 40) possess both NO2 and
COOH. The enhanced radical scavenging effect of these two
1146
dx.doi.org/10.1021/bi3012059 | Biochemistry 2013, 52, 1137−1148

Biochemistry
Article
groups can be explained by their delocalized electron structure
fostering a stronger and further extended electron delocalization,
a key factor in radical scavenging.
REFERENCES
■
(1) Wetzel, R., Ed. (1999) Amyloid, Prions, and Other Protein
Aggregates. Methods in Enzymology, Vol. 309, Academic Press, San
Diego.
(2) Chiti, F., and Dobson, C. M. (2008) Amyloid formation by globular
proteins under native conditions. Nat. Chem. Biol. 5, 15−22.
(3) Walsh, D. M., and Selkoe, D. J. (2007) Aβ oligomers: A decade of
discovery. J. Neurochem. 101, 1172−1184.
This discussion indicates that the hydrazone scaffold serves as
a useful foundation for the design of multifunctional anti-AD
compounds. We showed that the substituents on this backbone
play an important role in determining their activity in the
different assays. More hydrazones with a broad group of
substituents, including bulky (e.g., tert-butyl or another ring)
or electron-donating ones (e.g., any alkyl), need to be tested to
draw an accurate conclusion. In addition, the synthesis of further
derivatives will be extended to aryl-alkyl (e.g., acetophenones)
and aryl-aryl (e.g., benzophenones) ketones to observe the effect
of other groups connected to the chain.
(4) Estrada, L. D., and Soto, C. (2007) Disrupting β-amyloid for
Alzheimer’s disease treatment. Curr. Top. Med. Chem. 7, 115−126.
(5) Van der Schyf, C. J., Mandel, S., Geldenhuys, W. J., Amit, T.,
Avramovich, Y., Zheng, H., Fridkin, M., Gal, S., Weinreb, O., Am, O. B.,
Sagi, Y., and Youdim, M. B. H. (2007) Novel multifunctional anti-
Alzheimer drugs with various CNS neurotransmitter targets and
neuroprotective moieties. Curr. Alzheimer Res. 4, 522−536.
(6) Cavalli, A., Bolognesi, M. L., Capsoni, S., Andrisano, V., Bartolini,
M., Margotti, E., Cattaneo, A., Recanatini, M., and Melchiorre, C. (2007)
A small molecule targeting the multifactorial nature of Alzheimer’s
disease. Angew. Chem., Int. Ed. 46, 3689−3692.
CONCLUSIONS
■
The analysis of the data presented herein indicates that the diaryl-
hydrazone skeleton is a promising scaffold for the design and
synthesis of multifunctional compounds against Aβ self-
assembly. Characteristically, the compounds in the synthesized
hydrazone library exhibited strong activity either against fibril or
oligomer formation as well as the disassembly of the preformed
Aβ assemblies. This observation is in agreement with earlier
suggestions that not all stable oligomers are obligatory precursors
to the fibrils and that the two processes can occur in distinct
pathways10,11 as we have observed in previous investigations.42
However, a few compounds in our hydrazone library exhibited
desirable effects against both fibril and oligomer self-assembly, as
well as disassembling preformed fibrils and oligomers that form
during Aβ self-assembly without stabilizing potentially toxic
intermediates. Furthermore, ∼30% of the studied hydrazones
showed significant antioxidant behavior, matching or surpassing
the relevant activity of the well-known antioxidants ascorbic acid
and resveratrol.
(7) Stains, C., Mondal, K., and Ghosh, I. (2007) Molecules that target
β-amyloid. ChemMedChem 2, 1674−1692.
(8) Torok, B., Dasgupta, S., and Torok, M. (2008) Chemistry of small
̈
̈
̈
̈
molecule inhibitors of Alzheimer’s amyloid-β fibrillogenesis. Curr.
Bioact. Compd. 4, 159−174.
(9) Liu, T., and Bitan, G. (2012) Modulating self-assembly of
amyloidogenic proteins as a therapeutic approach for neurodegenerative
diseases: Strategies and mechanisms. ChemMedChem 7, 359−374.
(10) Necula, M., Kayed, R., Milton, S., and Glabe, C. G. (2007) Small
molecule inhibitors of aggregation indicate that amyloid β oligomeriza-
tion and fibrillization pathways are independent and distinct. J. Biol.
Chem. 282, 10311−10324.
(11) Glabe, C. G. (2008) Structural classification of toxic amyloid
oligomers. J. Biol. Chem. 283, 29639−29643.
(12) Nerelius, C., Johansson, J., and Sandegren, A. (2009) Amyloid β-
peptide aggregation. What does it result in and how can it be prevented?
Front. Biosci. 14, 1716−1729.
(13) Sciaretta, K. L., Gordon, D. J., and Meredith, S. C. (2006) Peptide-
based inhibitors of amyloid assembly. Methods Enzymol. 413, 273−312.
(14) Armstrong, A. H., Chen, J., McKoy, A. F., and Hecht, M. H.
(2011) Mutations that replace aromatic side chains promote
aggregation of the Alzheimer’s Aβ peptide. Biochemistry 50, 4058−4067.
(15) Gazit, E. (2005) Mechanisms of amyloid fibril self-assembly and
inhibition. Model short peptides as a key research tool. FEBS J. 272,
5971−5978.
On the basis of our investigations, three compounds (34, 35,
and 38) with combined Aβ self-assembly modulation and
antioxidant activity were identified, providing promising starting
points for more potent druglike inhibitor design based on the
hydrazone scaffold.
(16) Gerrard, J. A., Hutton, C. A., and Perugini, M. A. (2007) Inhibiting
protein-protein interactions as an emerging paradigm for drug
discovery. Mini-Rev. Med. Chem. 7, 151−157.
(17) LeVine, H., III (2007) Small molecule inhibitors of Aβ-assembly.
Amyloid 14, 185−197.
ASSOCIATED CONTENT
■
S
* Supporting Information
Synthesis and spectral characterization of the hydrazones used in
this study. This material is available free of charge via the Internet
at http://pubs.acs.org.
(18) Barnham, K. J., Masters, C. L., and Bush, A. I. (2004)
Neurodegenerative diseases and oxidative stress. Nat. Rev. 3, 205−214.
(19) Cavalli, A., Bolognesi, M. L., Minarini, A., Rosini, M., Tumiatti, V.,
Recanatini, M., and Melchiorre, C. (2008) Multi-target-directed ligands
to combat neurodegenerative diseases. J. Med. Chem. 51, 347−372.
(20) DeToma, S. A., Choi, J.-S., Braymer, J. J., and Lim, M. H. (2011)
Myricetin: A naturally occurring regulator of metal-induced amyloid-β
aggregation and neurotoxicity. ChemBioChem 12, 1198−1201.
(21) Varadarajan, S., Yatin, S., Aksenova, M., and Butterfield, D. A.
(2000) Review: Alzheimer’s amyloid β peptide-associated free radical
oxidative stress and neurotoxicity. J. Struct. Biol. 130, 184−208.
(22) Atwood, C. S., Obrenovich, M. E., Liu, T., Chan, H., Perry, G.,
Smith, M. A., and Martins, R. N. (2003) Amyloid-β: A chameleon
walking in two worlds. A review of the trophic and toxic properties of
amyloid-β. Brain Res. Rev. 43, 1−16.
AUTHOR INFORMATION
■
Corresponding Author
*Department of Chemistry, University of Massachusetts Boston,
100 Morrissey Blvd., Boston, MA 02125-3393. Telephone: (617)
287-6159. Fax: (617) 287-6030. E-mail: marianna.torok@umb.
edu.
Funding
Financial support provided by the University of Massachusetts
Boston and the National Institutes of Health (Grant R-15
AG025777-03A1 to B.T. and M.T. and Grant R21AG028816-01
to H.L.) is gratefully acknowledged.
(23) Huang, X., Moir, R. D., Tanzi, R. E., Bush, A. I., and Rogers, J. T.
(2004) Redox-active metals, oxidative stress, and Alzheimer’s disease
pathology. Ann. N.Y. Acad. Sci. 1012, 153−163.
(24) Shoval, H., Weiner, L., Gazit, E., Levy, M., Pinchuk, I., and
Lichtenberg, D. (2008) Polyphenol-induced dissociation of various
Notes
The authors declare no competing financial interest.
1147
dx.doi.org/10.1021/bi3012059 | Biochemistry 2013, 52, 1137−1148

Biochemistry
Article
amyloid fibrils results in a methionine-independent formation of ROS.
Biochim. Biophys. Acta 1784, 1570−1577.
(45) LeVine, H., III (2006) Biotin-avidin interaction-based screening
assay for Alzheimer’s β-peptide oligomer inhibitors. Anal. Biochem. 356,
265−272.
(25) Baruch-Suchodolsky, R., and Fisher, B. (2009) Aβ40 either
soluble or aggregated, is a remarkably potent antioxidant in cell-free
oxidative systems. Biochemistry 48, 4354−4370.
(46) LeVine, H., III, Ding, Q., Walker, J. A., Voss, R. S., and Augelli-
Szafran, C. E. (2009) Clioquinol and other hydroxyquinoline derivatives
inhibit Aβ(1−42) oligomer assembly. Neurosci. Lett. 465, 99−103.
(47) LeVine, H., III (2004) Alzheimer’s β-peptide oligomer formation
at physiologic concentrations. Anal. Biochem. 335, 81−90.
(48) (a) Huang, D., Ou, B., and Prior, R. L. (2005) The chemistry
behind antioxidant capacity assays. J. Agric. Food Chem. 53, 1841−1856.
(b) Sanchez-Moreno, C. (2002) Review: Methods used to evaluate the
free radical scavenging activity in foods and biological systems. Food Sci.
Technol. Int. 8, 121−129. (c) Brand-Williams, W., Cuvelier, M. E., and
Berset, C. (1995) Use of a free radical method to evaluate antioxidant
activity. Lebensm.-Wiss. Technol. (1968-2004) 28, 25−30.
(49) Liu, F., Ooi, V. E. C., and Change, S. T. (1997) Free radical
scavenging activity of mushroom polysaccharide extracts. Life Sci. 60,
763−771.
(50) Delmas, D., Aires, V., Limagne, E., Dutartre, P., Mazue, F.,
Ghiringhelli, F., and Latruffe, N. (2011) Transport, stability, and
biological activity of resveratrol. Ann. N.Y. Acad. Sci. 1215, 48−59.
(51) Hiyama, T., Ed. (2000) Organofluorine Compounds, Springer,
Berlin.
(52) Ojima, I., McCarthy, J. R., and Welch, J. T., Eds. (1996)
Biomedical Frontiers of Fluorine Chemistry, American Chemical Society,
Washington, DC.
(26) Siegel, S. J., Bieschke, J. E., Powers, T., and Kelly, J. W. (2007) The
oxidative stress metabolite 4-hydroxynonenal promotes Alzheimer
protofibril formation. Biochemistry 46, 1503−1510.
(27) Starkov, A. A., and Beal, F. M. (2008) Portal to Alzheimer’s
disease. Nat. Med. 14, 1020−1021.
(28) Shoval, H., Lichhtenberg, D., and Gazit, E. (2007) The molecular
mechanisms of the anti-amyloid effects of phenols. Amyloid 14, 73−87.
(29) Rossi, L., Mazzitelli, S., Archiello, M., Capo, C. R., and Rotilio, G.
(2008) Benefits from dietary polyphenols for brain aging and
Alzheimer’s disease. Neurochem. Res. 33, 2390−2400.
(30) Ono, K., Condron, M. M., Ho, L., Wang, J., Zhao, W., Pasinetti, G.
M., and Teplow, D. B. (2008) Effects of grape seed-derived polyphenols
on amyloid β-protein self-assembly and cytotoxicity. J. Biol. Chem. 283,
32176−32187.
(31) Xu, J. Z., Yeung, S. Y., Chang, Q., Huang, Y., and Chen, Z. Y.
(2004) Comparison of antioxidant activity and bioavailability of tea
epicatechins with their epimers. Br. J. Nutr. 91, 873−881.
(32) Rollas, S., and Kucu̧ kguzel, G. (2007) Biological activities of
̈ ̈ ̈
hydrazone derivatives. Molecules 12, 1910−1939.
(33) Gemma, S., Colombo, L., Forloni, G., Savini, L., Fracasso, C.,
Caccia, S., Salmona, M., Brindisi, M., Joshi, V., Tripaldi, P., Giorgi, G.,
Taglialatela-Scafati, O., Novellino, E., Fiorini, I., Campiani, G., and
Butini, S. (2011) Pyrroloquinoxaline hydrazones as fluorescent probes
for amyloid fibrils. Org. Biomol. Chem. 9, 5137−5148.
(53) Balsano, C., and Alisi, A. (2009) Antioxidant effects of natural
bioactive compounds. Curr. Pharm. Des. 15, 3063−3073.
(54) Feng, B. I., Toyama, B. H., Wille, H., Colby, D. W., Collins, S. R.,
May, B. C. H., Prusiner, S. B., Weissman, J., and Shoichet, B. K. (2008)
Small-molecule aggregates inhibit amyloid polymerization. Nat. Chem.
Biol. 4, 197−199.
(34) Alptuzuen, V., Prinz, M., Horr, V., Scheiber, J., Radacki, K.,
Fallarero, A., Vuorela, P., Engels, B., Braunschweig, H., Erciyas, E., and
Holzgrabe, U. (2010) Interaction of (benzylidene-hydrazono)-1,4-
dihydropyridines with β-amyloid, acetylcholine, and butyrylcholine
esterases. Bioorg. Med. Chem. 18, 2049−2059.
(55) Lendel, C., Bertoncini, C. W., Cremades, N., Waudby, C. A.,
Vendruscolo, M., Dobson, C. M., Schenk, D., Christodoulou, J., and
Toth, G. (2009) On the mechanism of nonspecific inhibitors of protein
aggregation: Dissecting the interactions of α-synuclein with congo red
and lacmoid. Biochemistry 48, 8322−8334.
(35) Torok, M., Abid, M., Mhadgut, S. C., and Torok, B. (2006)
̈
̈
̈
̈
Organofluorine inhibitors of amyloid fibrillogenesis. Biochemistry 45,
5377−5383.
(56) Lendel, C., Bolognesi, B., Wahlstrom, A., Dobson, C. M., and
̈
(36) Sood, A., Abid, M., Hailemichael, S., Foster, M., Torok, B., and
Torok, M. (2009) Effect of chirality of small molecule organofluorine
inhibitors of amyloid self-assembly on inhibitor potency. Bioorg. Med.
Chem. Lett. 19, 6931−6934.
(37) Sood, A., Abid, M., Sauer, C., Hailemichael, S., Foster, M., Torok,
B., and Torok, M. (2011) Disassembly of preformed amyloid β fibrils by
small organofluorine molecules. Bioorg. Med. Chem. Lett. 21, 2044−
̈
̈
Graslund, A. (2010) Detergent-like Interaction of Congo Red with the
̈
̈
̈
Amyloid β Peptide. Biochemistry 49, 1358−1360.
(57) Lamberto, G. R., Torres-Monserrat, V., Bertoncini, C. W.,
Salvatella, X., Zweckstetter, M., Griesinger, C., and Fernandez, C.
(2011) Toward the discovery of effective polycyclic inhibitors of α-
synuclein amyloid assembly. J. Biol. Chem. 286, 32036−32044.
(58) Ryan, T. M., Griffin, M. D. W., Teoh, C. L., Ooi, J., and Howlett,
G. J. (2011) High-affinity amphipathic modulators of amyloid fibril
nucleation and elongation. J. Mol. Biol. 406, 416−429.
̈
̈
̈
̈
2047.
(38) Fezoui, Y., Hartley, D. M., Harper, J. D., Khurana, R., Walsh, D.
M., Condron, M. M., Selkoe, D. J., Lansbury, P. T., Fink, A. L., and
Teplow, D. B. (2000) An improved method of preparing the amyloid β-
protein for fibrillogenesis and neurotoxicity experiments. Amyloid 7,
166−178.
(59) Abelein, A., Bolognesi, B., Dobson, C. M., Graslund, A., and
̈
Lendel, C. (2012) Hydrophobicity and conformational change as
mechanistic determinants for nonspecific modulators of amyloid β self-
assembly. Biochemistry 51, 126−137.
(39) Bourhim, M., Kruzel, M., Srikrishnan, T., and Nicotera, T. (2007)
Linear quantitation of Aβ aggregation using thioflavin T: Reduction in
fibril formation by colostrinin. J. Neurosci. Methods 160, 264−268.
(40) Naiki, H., Higuchi, K., Hosokawa, M., and Takeda, T. (1989)
Fluorometric determination of amyloid fibrils in vitro using the
fluorescent dye, thioflavine T. Anal. Biochem. 177, 244−249.
(41) LeVine, H., III (1993) Thioflavin T interaction with synthetic
Alzheimer’s disease β-amyloid peptides: Detection of amyloid
aggregation in solution. Protein Sci. 2, 404−410.
(60) Lipinski, C. A., Lombardo, F., Dominy, B. W., and Feeney, P. J.
(1997) Experimental and computational approaches to estimate
solubility and permeability in drug discovery and development settings.
Adv. Drug Delivery Rev. 23, 3−25.
(42) Nilsson, M. R. (2004) Techniques to study amyloid fibril
formation in vitro. Methods 34, 151−160.
(43) Torok, B., Sood, A., Bag, S., Kulkarni, A., Borkin, D., Lawler, E.,
̈
̈
Dasgupta, S., Landge, S. M., Abid, M., Zhou, W., Foster, M., LeVine, H.,
III, and Torok, M. (2012) Structure-activity relationship of organo-
̈
̈
fluorine inhibitors of amyloid-β self-assembly. ChemMedChem 7, 910−
919.
(44) Walsh, D. M., Lomakin, A., Benedek, G. B., Condron, M. M., and
Teplow, D. B. (1997) Amyloid β-protein fibrillogenesis: Detection of a
protofibrillar intermediate. J. Biol. Chem. 272, 22364−22372.
1148
dx.doi.org/10.1021/bi3012059 | Biochemistry 2013, 52, 1137−1148