Insights into Structure-Activity Relationships of 3-Arylhydrazonoindolin-2-One Derivatives for Their Multitarget Activity on β-Amyloid Aggregation and Neurotoxicity

Despite the controversial outcomes of clinical trials executed so far, the prevention of β-amyloid (Aβ) deposition and neurotoxicity by small molecule inhibitors of Aβ aggregation remains a target intensively pursued in the search of effective drugs for treating Alzheimer’s disease (AD) and related neurodegeneration syndromes. As a continuation of previous studies, a series of new 3-(2-arylhydrazono)indolin-2-one derivatives was synthesized and assayed, investigating the effects of substitutions on both the indole core and arylhydrazone moiety. Compared with the reference compound 1, we disclosed equipotent derivatives bearing alkyl substituents at the indole nitrogen, and fairly tolerated bioisosteric replacements at the arylhydrazone moiety. For most of the investigated compounds, the inhibition of Aβ40 aggregation (expressed as pIC50) was found to be correlated with lipophilicity, as assessed by a reversed-phase HPLC method, through a bilinear relationship. The N1-cyclopropyl derivative 28 was tested in cell-based assays of Aβ42 oligomer toxicity and oxidative stress induced by hydrogen peroxide, showing significant cytoprotective effects. This study confirmed the versatility of isatin in preparing multitarget small molecules affecting different biochemical pathways involved in AD.


Introduction
Alzheimer's disease (AD) is the most common cause of age-related neurodegenerative pathologies. AD represents a serious challenge for health systems, physicians and caregivers, because of its disabling course and the limited efficacy of pharmacological therapies [1]. The treatments approved for AD, namely the restoration of cholinergic transmission by means of acetylcholinesterase inhibitors [2], and the neuroprotection from glutamate excitotoxicity exerted by memantine [3], are only symptomatic and do not meet the clinical need for effective disease-modifying drugs.
A typical feature of AD consists in the deposition of extracellular β-amyloid (Aβ) peptide aggregates (amyloid plaques), starting from cholinergic neurons of hippocampus and then A typical feature of AD consists in the deposition of extracellular β-amyloid (Aβ) peptide aggregates (amyloid plaques), starting from cholinergic neurons of hippocampus and then progressively extending to the whole brain cortex [4]. Aβ peptide derives from proteolysis of amyloid precursor protein (APP) catalyzed by β-and γ-secretases, and is formed as 40-(Aβ40) or 42-mer (Aβ42). Aβ monomers in AD brains aggregate to soluble oligomers, that in turn lead to intermediate protofibrils and finally to the deposition of amyloid plaques [5]. The formation of such oligomeric and prefibrillar species is correlated with the neurotoxicity in the AD brain, which represents a major causal factor of the cognitive impairment and the synaptic loss in AD patients [6][7][8].
The amyloidogenesis of Aβ occurs since the early stages of the disease insurgence, so that preventing the oligomerization and/or fibrillization process could represent a promising disease-modifying treatment for AD. Despite the number of research findings in this field, only a very small number of molecules have reached the preclinical stage, and no one entered therapy so far [9]. However, many small molecules acting as disruptors of protein-protein interactions have demonstrated potential in inhibiting Aβ aggregation [10]. Among them, indole derivatives such as melatonin [11], fluorinated indoles [12], hydroxyindoles [13] (Figure 1) displayed the structural features for an efficient antiaggregating activity. Particularly, they act as intercalators in hydrophobic interactions between Aβ side chains, including aromatic π-stacking interactions [14][15][16][17]. As recently shown by our structure-activity relationship (SAR) studies [16,18], these hydrophobic interactions taking place between Aβ and many classes of small molecules could be reinforced by polar interactions and/or hydrogen bond (HB) formation. From previous studies [19][20][21][22], 5-methoxyisatin 3-(4-isopropylphenyl)hydrazone (1, Figure 1) was identified as a promising inhibitor of Aβ aggregation and acetylcholinesterase (AChE), with IC50s in the submicromolar and low μM range, respectively. Herein, we synthesized a number of congeners of compound 1, and evaluated the in vitro activity as inhibitors of Aβ40 aggregation. With the aim of extending the SAR exploration, the new derivatives were designed in order to undertake a more systematic exploration ( Figure 2) of: (i) the 3-arylhydrazone moiety, by introducing a number of diverse substituents and/or modifying the aryl substituent (compounds 4-11, Scheme 1; compounds 13-18, Scheme 2); (ii) the length and the chemical nature of the linker (compounds 19 and 21-23, Scheme 3); (iii) the substitution either on the nitrogen (compounds 24, 26, 28, 32-34; Scheme 4) and the benzene moiety of the indole ring (compounds 36-39, Scheme 5). SARs, including correlation of antiaggregating activity with lipophilicity, were investigated, and for one of the most potent Aβ40 aggregation inhibitors (compound 28) cytoprotection from toxic Aβ42 oligomers in a cell-based assay was evaluated. Moreover, taking into account the role of oxidative stress in AD and other neurodegenerative diseases [23], the antioxidant property of 28 against hydrogen peroxide insult in SH-SY5Y cells was also tested. From previous studies [19][20][21][22], 5-methoxyisatin 3-(4-isopropylphenyl)hydrazone (1, Figure 1) was identified as a promising inhibitor of Aβ aggregation and acetylcholinesterase (AChE), with IC 50 s in the submicromolar and low µM range, respectively. Herein, we synthesized a number of congeners of compound 1, and evaluated the in vitro activity as inhibitors of Aβ 40 aggregation. With the aim of extending the SAR exploration, the new derivatives were designed in order to undertake a more systematic exploration ( Figure 2) of: (i) the 3-arylhydrazone moiety, by introducing a number of diverse substituents and/or modifying the aryl substituent (compounds 4-11, Scheme 1; compounds 13-18, Scheme 2); (ii) the length and the chemical nature of the linker (compounds 19 and 21-23, Scheme 3); (iii) the substitution either on the nitrogen (compounds 24, 26, 28, 32-34; Scheme 4) and the benzene moiety of the indole ring (compounds 36-39, Scheme 5). SARs, including correlation of antiaggregating activity with lipophilicity, were investigated, and for one of the most potent Aβ 40 aggregation inhibitors (compound 28) cytoprotection from toxic Aβ 42 oligomers in a cell-based assay was evaluated. Moreover, taking into account the role of oxidative stress in AD and other neurodegenerative diseases [23], the antioxidant property of 28 against hydrogen peroxide insult in SH-SY5Y cells was also tested.

Inhibition of Amyloid Aggregation
In vitro inhibition of Aβ aggregation was assessed through a ThT fluorescence-based method [18], with the use of 2% 1,1,1,3,3,3-hexafluoroisopropanol (HFIP) as aggregation enhancer. In this medium-throughput assay, we preferred using Aβ40 peptide, being more manageable than Aβ42 and less prone to the formation of preaggregates [18]. Samples of Aβ were co-incubated with test molecules in PBS at 100 μM concentration and the antiaggregating activities were measured after 2 h

Inhibition of Amyloid Aggregation
In vitro inhibition of Aβ aggregation was assessed through a ThT fluorescence-based method [18], with the use of 2% 1,1,1,3,3,3-hexafluoroisopropanol (HFIP) as aggregation enhancer. In this medium-throughput assay, we preferred using Aβ 40 peptide, being more manageable than Aβ 42 and less prone to the formation of preaggregates [18]. Samples of Aβ were co-incubated with test molecules in PBS at 100 µM concentration and the antiaggregating activities were measured after 2 h of incubation at 25 • C. For compounds showing >80% Aβ 40 aggregation inhibition, IC 50 s were determined.
An early objective of the present study was to replace the 4-iPr substituent with phenyl groups bearing polar and apolar electron-withdrawing substituents (compounds 4-6, Table 1). Such modifications determined a drop of the activity, and only the 3 ,5 -bis(trifluoromethyl) derivative 6 showed IC 50 in the low micromolar range (9.9 µM), which resulted however 25-fold less potent than compound 1.
The replacement of the hydrazone phenyl ring with a number of other aromatic or heteroaromatic rings was carried out for investigating the effects of bulkiness and/or additional π-π interactions on the anti-aggregating potency. The bioisosteric replacement of the phenyl group with pyrid-2-yl (7) resulted in a much lower activity (IC 50 90 µM), whereas the 1-naphthyl moiety in compound 8 did recover a fair potency (IC 50 28 µM), indicating that additional aromatic interactions may improve the antiaggregating potency. In contrast, quinolyl analogues 9 and 10, (similarly to the pyrid-2-yl congener 7), showed lower potency, and the 7-chloroquinolin-4-yl derivative 11 resulted a very weak inhibitor. The replacement of the hydrazone phenyl ring with a number of other aromatic or heteroaromatic rings was carried out for investigating the effects of bulkiness and/or additional π-π interactions on the anti-aggregating potency. The bioisosteric replacement of the phenyl group with pyrid-2-yl (7) resulted in a much lower activity (IC50 90 μM), whereas the 1-naphthyl moiety in compound 8 did recover a fair potency (IC50 28 μM), indicating that additional aromatic interactions may improve the antiaggregating potency. In contrast, quinolyl analogues 9 and 10, (similarly to the pyrid-2-yl congener 7), showed lower potency, and the 7-chloroquinolin-4-yl derivative 11 resulted a very weak inhibitor.
Further information on the bioisosteric replacement of phenylhydrazone ring was achieved with thiazolylhydrazones 13-18 (Table 2). An exploration of the substituents on the thiazole moiety clearly indicated a preference for 4-phenylthiazole derivative 15, which indeed proved to be a potent inhibitor of Aβ aggregation (IC50 1.2 μM), apparently due to additional hydrophobic/aromatic interactions attained by the phenyl group at the C4 position of the thiazole ring.
Further information on the bioisosteric replacement of phenylhydrazone ring was achieved with thiazolylhydrazones 13-18 (Table 2). An exploration of the substituents on the thiazole moiety clearly indicated a preference for 4-phenylthiazole derivative 15, which indeed proved to be a potent inhibitor of Aβ aggregation (IC 50 1.2 µM), apparently due to additional hydrophobic/aromatic interactions attained by the phenyl group at the C4 position of the thiazole ring.  The 3 -substituted congeners of 4-phenylthiazol-2-yl derivatives 16-18, regardless the physicochemical feature of the meta substituent, retained antiaggregating activity in the micromolar range, but resulted 8-to-30-fold less potent than 15, thereby suggesting critical steric requirements for these derivatives. In contrast, smaller alkyl substituents, namely methyl or chloromethyl in compounds 13 and 14, respectively, displayed contrasting effects, with the 4-chloromethyl derivative 13 retaining a fair anti-aggregating potency (IC 50 13 µM), and 4,5-dimethyl analogue resulting a very weak inhibitor.
According to our investigation strategy (Figure 2), the following step was aimed at exploring the effects on the inhibition of Aβ aggregation of a few variations of the linker (length and the chemical nature) between the two structural moieties (Table 3).  The introduction of alkyl and phenylalkyl groups at the indole N and methylation of the hydrazone NH in the hit structure 1 were then investigated (Table 4). Inhibition data highlighted that small and bulky alkyls on the 1H-indole nitrogen of 1 were tolerated, so that the N-methyl (26), N-cyclopropyl (28) and N-phenylalkyl (32-34) derivatives resulted almost equipotent with the hit compound 1 in the submicromolar range (IC50s ranging from 0.53 to 0.83 μM). A nonlinear relation between the antifibrillogenic potency and lipophilicity of the N-alkyl groups exists for these compounds, the N-benzyl derivative 32 being the most potent one within the subset (IC50 0.53 μM). In contrast, methylation of the hydrazone nitrogen (24) resulted in a more than 10-fold decrease in the inhibition potency compared with the respective non-methylated compound (26), which proves that N-methylation of hydrazone N may prevent IMHB formation with carbonyl O at 2-position [22] and/or may introduce a steric effect that hinders hydrazone NH from acting as HB-donor (HBD) in the interaction with the Aβ peptide.  The introduction of alkyl and phenylalkyl groups at the indole N and methylation of the hydrazone NH in the hit structure 1 were then investigated (Table 4). Inhibition data highlighted that small and bulky alkyls on the 1H-indole nitrogen of 1 were tolerated, so that the N-methyl (26), N-cyclopropyl (28) and N-phenylalkyl (32-34) derivatives resulted almost equipotent with the hit compound 1 in the submicromolar range (IC50s ranging from 0.53 to 0.83 μM). A nonlinear relation between the antifibrillogenic potency and lipophilicity of the N-alkyl groups exists for these compounds, the N-benzyl derivative 32 being the most potent one within the subset (IC50 0.53 μM). In contrast, methylation of the hydrazone nitrogen (24) resulted in a more than 10-fold decrease in the inhibition potency compared with the respective non-methylated compound (26), which proves that N-methylation of hydrazone N may prevent IMHB formation with carbonyl O at 2-position [22] and/or may introduce a steric effect that hinders hydrazone NH from acting as HB-donor (HBD) in the interaction with the Aβ peptide.  The introduction of alkyl and phenylalkyl groups at the indole N and methylation of the hydrazone NH in the hit structure 1 were then investigated (Table 4). Inhibition data highlighted that small and bulky alkyls on the 1H-indole nitrogen of 1 were tolerated, so that the N-methyl (26), N-cyclopropyl (28) and N-phenylalkyl (32)(33)(34) derivatives resulted almost equipotent with the hit compound 1 in the submicromolar range (IC50s ranging from 0.53 to 0.83 μM). A nonlinear relation between the antifibrillogenic potency and lipophilicity of the N-alkyl groups exists for these compounds, the N-benzyl derivative 32 being the most potent one within the subset (IC50 0.53 μM). In contrast, methylation of the hydrazone nitrogen (24) resulted in a more than 10-fold decrease in the inhibition potency compared with the respective non-methylated compound (26), which proves that N-methylation of hydrazone N may prevent IMHB formation with carbonyl O at 2-position [22] and/or may introduce a steric effect that hinders hydrazone NH from acting as HB-donor (HBD) in the interaction with the Aβ peptide.  The introduction of alkyl and phenylalkyl groups at the indole N and methylation of the hydrazone NH in the hit structure 1 were then investigated (Table 4). Inhibition data highlighted that small and bulky alkyls on the 1H-indole nitrogen of 1 were tolerated, so that the N-methyl (26), N-cyclopropyl (28) and N-phenylalkyl (32-34) derivatives resulted almost equipotent with the hit compound 1 in the submicromolar range (IC50s ranging from 0.53 to 0.83 μM). A nonlinear relation between the antifibrillogenic potency and lipophilicity of the N-alkyl groups exists for these compounds, the N-benzyl derivative 32 being the most potent one within the subset (IC50 0.53 μM). In contrast, methylation of the hydrazone nitrogen (24) resulted in a more than 10-fold decrease in the inhibition potency compared with the respective non-methylated compound (26), which proves that N-methylation of hydrazone N may prevent IMHB formation with carbonyl O at 2-position [22] and/or may introduce a steric effect that hinders hydrazone NH from acting as HB-donor (HBD) in the interaction with the Aβ peptide.  The introduction of alkyl and phenylalkyl groups at the indole N and methylation of the hydrazone NH in the hit structure 1 were then investigated (Table 4). Inhibition data highlighted that small and bulky alkyls on the 1H-indole nitrogen of 1 were tolerated, so that the N-methyl (26), N-cyclopropyl (28) and N-phenylalkyl (32-34) derivatives resulted almost equipotent with the hit compound 1 in the submicromolar range (IC50s ranging from 0.53 to 0.83 μM). A nonlinear relation between the antifibrillogenic potency and lipophilicity of the N-alkyl groups exists for these compounds, the N-benzyl derivative 32 being the most potent one within the subset (IC50 0.53 μM). In contrast, methylation of the hydrazone nitrogen (24) resulted in a more than 10-fold decrease in the inhibition potency compared with the respective non-methylated compound (26), which proves that N-methylation of hydrazone N may prevent IMHB formation with carbonyl O at 2-position [22] and/or may introduce a steric effect that hinders hydrazone NH from acting as HB-donor (HBD) in the interaction with the Aβ peptide. A comparison of the data in Table 3 with that of the hit compound 1 revealed a 100-fold decrease of potency (from 0.43 to 41 µM), when linker was changed in the 1H-indole derivative 19, likely due to the lack of intramolecular hydrogen bond (IMHB) between the hydrazone NH and the carbonyl O at position 2 of isatin, which favors co-planarity between the two aromatic moieties. The inhibitory activity was maintained in the low micromolar range by the aza-derivatives 21 (IC 50 6.7 µM), that is a strict analog of 1, and 22 (IC 50 11 µM), that is the 3-chloro congener of 21, whereas a sharp drop of activity was observed for the carboxyhydrazide derivative 23. The antiaggregating potency of compounds 21-23 appears to be related to their lipophilicity, as assessed by logP values calculated with the ACDLab software (4.97, 4.43 and 3.53 for 21, 22 and 23, respectively).
The introduction of alkyl and phenylalkyl groups at the indole N and methylation of the hydrazone NH in the hit structure 1 were then investigated (Table 4). Inhibition data highlighted that small and bulky alkyls on the 1H-indole nitrogen of 1 were tolerated, so that the N-methyl (26), N-cyclopropyl (28) and N-phenylalkyl (32-34) derivatives resulted almost equipotent with the hit compound 1 in the submicromolar range (IC 50 s ranging from 0.53 to 0.83 µM). A nonlinear relation between the antifibrillogenic potency and lipophilicity of the N-alkyl groups exists for these compounds, the N-benzyl derivative 32 being the most potent one within the subset (IC 50 0.53 µM). In contrast, methylation of the hydrazone nitrogen (24) resulted in a more than 10-fold decrease in the inhibition potency compared with the respective non-methylated compound (26), which proves that N-methylation of hydrazone N may prevent IMHB formation with carbonyl O at 2-position [22] and/or may introduce a steric effect that hinders hydrazone NH from acting as HB-donor (HBD) in the interaction with the Aβ peptide. Finally, compounds 36-39, bearing diverse lipophilic and hydrophilic substituents on the indole ring (Table 5), were tested for the effects on Aβ aggregation. The introduction of n-butyl (36) or Br (37) at 5-and 7-position, respectively, resulted detrimental, leading in the latter case to complete loss of activity. The evidence of an enhancing role of hydrophilic substituents (particularly 5,6-dihydroxy derivatives) for an efficient inhibition of Aβ40 aggregation, emerging from our previous works [20,21], was endeavored by introducing a net negative charge with the sulfonate salt 38 and a protecting group for catechol derivatives, i.e., the 5,6-methylenedioxy substituent, in 39. Both modifications resulted poorly effective, and the two compounds showed inhibitory potencies much lower than that of compound 1. Finally, compounds 36-39, bearing diverse lipophilic and hydrophilic substituents on the indole ring (Table 5), were tested for the effects on Aβ aggregation. The introduction of n-butyl (36) or Br (37) at 5-and 7-position, respectively, resulted detrimental, leading in the latter case to complete loss of activity. The evidence of an enhancing role of hydrophilic substituents (particularly 5,6-dihydroxy derivatives) for an efficient inhibition of Aβ 40 aggregation, emerging from our previous works [20,21], was endeavored by introducing a net negative charge with the sulfonate salt 38 and a protecting group for catechol derivatives, i.e., the 5,6-methylenedioxy substituent, in 39. Both modifications resulted poorly effective, and the two compounds showed inhibitory potencies much lower than that of compound 1. Finally, compounds 36-39, bearing diverse lipophilic and hydrophilic substituents on the indole ring (Table 5), were tested for the effects on Aβ aggregation. The introduction of n-butyl (36) or Br (37) at 5-and 7-position, respectively, resulted detrimental, leading in the latter case to complete loss of activity. The evidence of an enhancing role of hydrophilic substituents (particularly 5,6-dihydroxy derivatives) for an efficient inhibition of Aβ40 aggregation, emerging from our previous works [20,21], was endeavored by introducing a net negative charge with the sulfonate salt 38 and a protecting group for catechol derivatives, i.e., the 5,6-methylenedioxy substituent, in 39. Both modifications resulted poorly effective, and the two compounds showed inhibitory potencies much lower than that of compound 1. To place on a more quantitative basis the above SAR trends, the lipophilicity of a number of isatin arylhydrazone derivatives was measured by an RP-HPLC method [34,35]. The polycratic capacity factors (log k' w ) were determined (details in the Experimental Section). Lipophilicity descriptors (i.e., 1-octanol-water partition coefficients) were also calculated (CLOGPs) using two computational tools (Bio-Loom software, v. 1.7 and ACDLabs software, release 10.0). The ACDLabs software was used also for calculating descriptors of polarizability and size, namely molar refractivity (CMR) and molar volume (MV). The experimental lipophilicity parameters (log k' w ), along with CMR and MV, are reported in Table 6 for 18 compounds achieving finite IC 50 s (<100 µM) in the Aβ aggregation assay.  Figure S1). The linear equation correlating log k' w and CLOGP has a slope close to one (+1.03) and an intercept of about +0.70, the latter revealing a systematic positive deviation from linearity, likely due to the so-called "silanophilic" interactions on the silica-based C18 stationary phase [36][37][38].
A biphasic relationship was observed for the majority of examined (thirteen out of eighteen) compounds between pIC 50 and log k' w ( Figure 3).
In the above equation, β is the nonlinear term of the Kubinyi's equation; n represents the number of data points, r 2 the coefficient of determination (squared correlation coefficient), and s the standard deviation of the regression equation; 95% confidence intervals of the regression coefficients Omitting from the regression analysis the strong outliers 6, 8, and 17-19, the following bilinear Equation (1) was obtained according to the Kubinyi's model [39]: pIC 50 = 0.65 (±0.11) log k' w − 0.95 (±0.33) log (βk' w + 1) + 2.32 n = 13, r 2 = 0.900, s = 0.271, logβ = −6.47, optimum log k' w = 6.80 (1) In the above equation, β is the nonlinear term of the Kubinyi's equation; n represents the number of data points, r 2 the coefficient of determination (squared correlation coefficient), and s the standard deviation of the regression equation; 95% confidence intervals of the regression coefficients are given in parentheses. The optimum value of log k' w , as calculated from the above equation, is 6.80. Similar, but statistically slightly poorer, bilinear correlation was obtained with CMR (r 2 = 0.704; Supplementary Materials Figure S3).
The pIC 50 values of the outliers 6, 8, 17-19 resulted 1.2-1.5 log units lower than those expected by the bilinear relationship with lipophilicity. These deviations may be reasonably explained as follows: (i) compound 19 is the only phenylhydrazone of 1H-indole-3-carbaldehyde, and not of isatin, whose carbonyl O at position 2 may play a critical role either in forming IMHB with the hydrazone NH or in interacting with counterparts of the amyloid peptides through dipole-dipole interactions or HBs; (ii) compounds 6 and 8 (as well as the quinoline-bearing compounds 9-11), 17 and 18, regardless of lipophilicity, may undergo detrimental steric effects when approaching and interacting into the binding site of Aβ peptide because of bulky moieties around the phenylhydrazone ( Figure 3). As for the 4-phenylthiazol-2-yl-containing derivatives, compounds 17 and 18, bearing OMe or Br in meta position of phenyl, resulted less active than the respective unsubstituted analog 15 and the less bulky 3 -OH congener 16, which in fact fit the bilinear equation model.
Taken together, our data suggested that both the 5-methoxy substituent on the indolin-2-one moiety and the 4-isopropyl substituent on the phenylhydrazone moiety play an essential role in maintaining the Aβ 40 antiaggregating potency in the submicromolar range. N-methylation of hydrazone NH, elimination of 2-carbonyl, introduction of other substituents at positions 5, 6 and 7 of the indolin-2-one nucleus and replacement of the isopropyl group with (hetero)aromatic moieties proved unfavorable, whereas alkylation of the indole nitrogen was tolerated at least up to an optimal lipophilicity corresponding, in our case, to the N 1 -benzyl derivative 32.
Based on data in Tables 1-5, compound 28 was selected for further biological and biophysical studies, aimed at elucidating the mechanism of disruption of Aβ aggregation, and evaluating the activity in biochemical assays related to neurodegeneration, and cytoprotection in a cell model of Aβ toxicity. Starting from the evidence that a bulky substituent led to a net improvement of antiaggregating activity (Table 5), we chose the N 1 -cyclopropyl derivative 28 as the best compromise between in vitro activity and aqueous solubility. As expected, solubility of 28 resulted lower than lead compound 1 (21 vs. 71 µM [22] in Tris/HCl buffer, pH 7.4 at 25 • C), but sufficient to allow performing the co-incubation assays where 20 µM represents the upper concentration threshold to test.

Kinetics of Aβ 42 Fibrillization
For ThT and circular dichroism (CD) time-course measures of Aβ aggregation (50 µM), alone or in the presence of 28 (20 µM), samples were prepared in PBS containing 2% v/v ethanol as the cosolvent, and incubated at 37 • C. Aβ 42 peptide was preferred, being the principal responsible of amyloid burden in AD brain. Kinetics of amyloid aggregation of Aβ 42 was followed by means of ThT fluorescence and CD absorption. In the latter case, we monitored the increase of the negative band at 215 nm, probing the random coil to β-sheet transition in folding peptide. Results depicted in Figure 4 show a fast fibrillization of self-aggregating peptide (dotted black line), with fibrils already detectable in large amount in the first 48 h of incubation. Soluble β-rich species are in turn massively detected after two days (full black line), suggesting that the fibrillization process takes place by quickly recruiting oligomer intermediates. On the other hand, samples co-incubated with 28 showed a sharply reduced amount either of β-sheet arranged and ThT-stained aggregates (red lines), reaching only after one week a fibril content of about 30% compared with that of control peptide. amyloid burden in AD brain. Kinetics of amyloid aggregation of Aβ42 was followed by means of ThT fluorescence and CD absorption. In the latter case, we monitored the increase of the negative band at 215 nm, probing the random coil to β-sheet transition in folding peptide. Results depicted in Figure 4 show a fast fibrillization of self-aggregating peptide (dotted black line), with fibrils already detectable in large amount in the first 48 h of incubation. Soluble β-rich species are in turn massively detected after two days (full black line), suggesting that the fibrillization process takes place by quickly recruiting oligomer intermediates. On the other hand, samples co-incubated with 28 showed a sharply reduced amount either of β-sheet arranged and ThT-stained aggregates (red lines), reaching only after one week a fibril content of about 30% compared with that of control peptide.

Protection Assays Against Oxidative and Cytotoxic Effects
Protection from Aβ42-induced cytotoxicity was measured for compound 28 in a conventional cell-based assay, measuring cell viability by means of MTT reduction [40]. Antioxidant activity of 28 was assessed in an H2O2-induced oxidation cell model. Reactive oxygen species (ROS) production was detected by means of a spectrofluorometric measure of the fluorescent probe 2′,7′-dichlorofluorescein (DCF), formed by oxidation of 2′,7′-dichlorodihydrofluorescein (DCFH) [41]. In both assays, human SH-SY5Y neuroblastoma cell line was used.
The aggregates formed by 5 μM Aβ42 in cultured SH-SY5Y cells produced around 50% of cell death within two days, while cells co-incubated with equimolar 5 μM Aβ42 and 28 were fully viable

Protection Assays Against Oxidative and Cytotoxic Effects
Protection from Aβ 42 -induced cytotoxicity was measured for compound 28 in a conventional cell-based assay, measuring cell viability by means of MTT reduction [40]. Antioxidant activity of 28 was assessed in an H 2 O 2 -induced oxidation cell model. Reactive oxygen species (ROS) production was detected by means of a spectrofluorometric measure of the fluorescent probe 2 ,7 -dichlorofluorescein (DCF), formed by oxidation of 2 ,7 -dichlorodihydrofluorescein (DCFH) [41]. In both assays, human SH-SY5Y neuroblastoma cell line was used.
The aggregates formed by 5 µM Aβ 42 in cultured SH-SY5Y cells produced around 50% of cell death within two days, while cells co-incubated with equimolar 5 µM Aβ 42 and 28 were fully viable in the same time frame (Figure 5, top). This result agreed with the observed activity in vitro and confirmed that previously observed in a close set of congeners [21,22]. Furthermore, to better investigate the cytoprotection exerted by this class of compounds, the antioxidant activity of compound 28 was investigated. Figure 5 (bottom) shows the radical-scavenging effects of increasing concentrations (0 to 20 µM) of 28 against oxidation induced by 100 µM hydrogen peroxide. Quercetin, a well-known natural antioxidant, was used as reference compound. ROS scavenging was exerted by 28 even at lower concentrations and reached a maximum effect at 20 µM concentration, which is about 50% of the effect shown by 100 µM quercetin. Thus, the antioxidant activity of 28 is supposed to be comparable to that of quercetin in the assay conditions herein used. compound 28 was investigated. Figure 5 (bottom) shows the radical-scavenging effects of increasing concentrations (0 to 20 μM) of 28 against oxidation induced by 100 μM hydrogen peroxide. Quercetin, a well-known natural antioxidant, was used as reference compound. ROS scavenging was exerted by 28 even at lower concentrations and reached a maximum effect at 20 μM concentration, which is about 50% of the effect shown by 100 μM quercetin. Thus, the antioxidant activity of 28 is supposed to be comparable to that of quercetin in the assay conditions herein used.

Chemistry
Commercial reagents and solvents were purchased from Sigma-Aldrich (Milan, Italy). Melting points (mp) were determined by the capillary method on a Stuart SMP3 electrothermal apparatus (Bibby Scientific, Milan, Italy). IR spectra were recorded using potassium bromide disks on a Spectrum One FT-IR spectrophotometer (Perkin Elmer, Milan, Italy); only the most significant IR absorption bands are reported. 1 H-NMR spectra were recorded in DMSO-d6 on a Mercury 300 spectrometer (Varian, Cernusco sul Naviglio, Italy). Chemical shifts are expressed in δ (ppm) and the coupling constants J in Hz. The following abbreviations were used: s, singlet; d, doublet; t, triplet; qn, quintuplet; sx, sextuplet; ep, septuplet; dd, double doublet; m, multiplet; br s, broad

Chemistry
Commercial reagents and solvents were purchased from Sigma-Aldrich (Milan, Italy). Melting points (mp) were determined by the capillary method on a Stuart SMP3 electrothermal apparatus (Bibby Scientific, Milan, Italy). IR spectra were recorded using potassium bromide disks on a Spectrum One FT-IR spectrophotometer (Perkin Elmer, Milan, Italy); only the most significant IR absorption bands are reported. 1 H-NMR spectra were recorded in DMSO-d 6 on a Mercury 300 spectrometer (Varian, Cernusco sul Naviglio, Italy). Chemical shifts are expressed in δ (ppm) and the coupling constants J in Hz. The following abbreviations were used: s, singlet; d, doublet; t, triplet; qn, quintuplet; sx, sextuplet; ep, septuplet; dd, double doublet; m, multiplet; br s, broad singlet. Chromatographic separations were performed on silica gel 63-200 (Merck, Milan, Italy). ESI-MS was performed with an electrospray interface and an ion trap mass spectrometer (1100 Series LC/MSD Trap System, Agilent, Palo Alto, CA, USA). The sample was infused via a KD Scientific syringe pump at a rate of 10 mL/min. The pressure of the nebulizer gas was 15 psi. The drying gas was heated to 350 • C at a flow of 5 L/min. Full-scan mass spectra were recorded in the mass/charge (m/z) range of 50-800 amu. For some representative compounds, HRMS experiments were performed with a dual electrospray interface (ESI) and a quadrupole time-of-flight mass spectrometer (Q-TOF, Agilent 6530 Series Accurate-Mass Quadrupole Time-of-Flight LC/MS, Agilent Technologies Italia S.p.A., Cernusco sul Naviglio, Italy).

Conclusions
The SAR studies on the indoline-2-one 3-arylhydrazone derivatives reported herein warranted further gain of knowledge in our ongoing study of inhibition of Aβ aggregation. Rooting on previous results [19][20][21] and keeping phenylhydrazone 1 [22] as a hit compound prone to optimization, the focused exploration of substituents on the indole scaffold, the modifications on the hydrazone linker, and the isosteric replacement of the phenyl of hydrazone moiety with thiazole, highlighted the essential role of 5-methoxy substituent, the requisite of an optimal lipophilicity, and the opportunity of alkylating the indole nitrogen for maintaining an efficient inhibition of Aβ aggregation. In this regard, a handful of derivatives (thiazolyl hydrazone 15 and N 1 -substituted indolin-2-ones 26, 28, 32-34) were found almost equipotent in vitro with the hit compound 1. Among them, compound 28, i.e., the N 1 -cyclopropyl derivative of 1, emerged from cell-based assays as a multitarget agent with strong cytoprotective effects in both Aβ 42 oligomer toxicity and oxidative stress. These results were in full agreement with those already reported for congeners of 28, particularly for cytoprotection from amyloid insult that resulted well comparable with reference compound 1 [22]. Aβ antifibrillogenic data have been analyzed to disclose reliable QSARs, and a bilinear correlation between pIC 50 and log k' w determined by RP-HPLC was derived which hold for those compounds that do not undergo detrimental steric effects in binding Aβ peptides.
The present work adds new information to the potential of indolin-2-one 3-arylhydrazones as privileged structures for the inhibition of Aβ aggregation and neurotoxicity, and pleiotropic agents for AD therapy. Despite the claims of disengagement of pharmaceutical industry from AD research [42], because of the controversial results so far obtained in clinical trials, many programs from research institutions continue to bet in the so-called amyloid hypothesis. In this light, the multitarget activity of 28 could deserve for this compound a preliminary pharmacological investigation in animal models of Aβ-dependent neurodegeneration, provided that further efforts in ameliorating aqueous solubility would be beneficial. The potency and versatility of indolin-2-one derivatives may still encounter the interest of research and development of small molecules as pharmacological tools for treating AD and related neurodegenerative syndromes.