Hydroxypyridinone-Diamine Hybrids as Potential Neuroprotective Agents in the PC12 Cell-Line Model of Alzheimer’s Disease

There is an urgent need to propose effective treatments for Alzheimer’s disease (AD). Although the origin of the disease is poorly understood, several therapeutic options have been proposed. The new therapeutic approaches targeting biometal-mediated neurodegenerative pathways appear to be interesting ones. As a continuation of our preceding studies, two novel series of advanced glycation endproducts (AGE)/advanced lipid peroxidation endproducts (ALE) inhibitors have been developed as multifunctional scavengers. This extended work allowed us to highlight the new hydroxypyridinone-diamine hybrid IIa-3 bearing a C4 alkyl linker between the two pharmacophores. This derivative exhibited preserved potent capacities to trap reactive carbonyl species (vicinal diamine function) as well as reactive oxygen species and transition metals (hydroxypyridinone moiety) in comparison with previously described lead compound 1. In addition, its good predicted absorption, distribution, metabolism and excretion (ADME) properties were correlated with a better efficacy to inhibit in vitro methylglyoxal-induced apoptosis in neuronal-like PC12 cells. This new promising agent revealed improved druglikeness and ability to prevent biometal-mediated oxidative and carbonyl stress amplification involved in AD pathogenesis.


Introduction
The development of drugs that can either alleviate or suppress Alzheimer's disease (AD) onset and progression continues to be an urgent healthcare challenge due to the social and economic impacts of this pathology. New therapeutic approaches targeting biometal-mediated neurodegenerative pathways appear to be promising ones. In recent years, it has been assumed that the contribution of transition metals, such as iron and copper, as AD pathogenesis initiators supersedes the amyloid β (Aβ) hypothesis [1,2]. In AD brains, iron and copper overload stimulates not only APP (amyloid precursor protein) cleavage by β-secretase but also Aβ aggregation and toxicity. Indeed, redox-active metal ion-Aβ complexes catalyze ROS (reactive oxygen species) production leading to strong oxidative damage on biomolecules [3]. Furthermore, biometals are implicated in the formation of tau-associated neurofibrillary tangles (NFTs) that constitute the second histopathological hallmarks of AD. Fe 3+ and Cu 2+ promote tau phosphorylation disturbing tau-microtubule assembly and stabilization. Interaction

Chemical Synthesis
As shown in Table 1 and Figure 1, our previously described synthetic strategy based on a key pseudopeptidic coupling step between diamine building blocks 2-6 and hydroxypyridinone (HOPO) ligands 9-10 was first extended to allow a decrease in size of the amide or piperazine-1,4-diamide linker and afforded the derivatives Ia-1 to Ia-4 or Ib-1 to Ib-3 respectively Table 1. Diamine building blocks and HOPO ligands (or other related precursors) implicated in the key coupling step of our AGE/ALE inhibitor synthetic strategy.
Since the synthesis of the diamine building blocks 3, 4 and 6 as well as HOPO ligands 9a-b, 10a-b, 3,2-HOPO and maltol benzyl ethers have been previously described, we focused here only on the preparation of the new precursors 2, 5, 7 and 8.
2.1.1. Synthesis of New Diamine Building Blocks 2, 5, 7 and 8 As shown in Scheme 1 and according to previously described reaction conditions [8], a new azide precursor 2 was prepared starting from commercially available N-carboxybenzyl L-asparagine (N-Cbz-Asn-OH). The dehydration of its side-chain amide function using trifluoroacetic anhydride in pyridine first provided the cyano intermediate 11 in 92% yield. The corresponding N-Boc amino derivative 12 was subsequently obtained after NaBH4 reduction of nitrile function in the presence of NiCl2.6H2O as a catalyst and (Boc)2O. The conversion of acid function into primary alcohol followed by mesylate formation and treatment with NaN3 afforded the azide intermediate 15 in good yields. Removal of Boc-protecting group was performed to furnish the attempted C2 alkyl side-chain amino ligand 2 as hydrochloride salt.  Diamine Building Blocks 1 HOPO Ligands 1,2 1 Diamine building blocks 3, 4 and 6 as well as HOPO ligands 9b and 10b have been described in our previous publication [8]. 2 HOPO ligands 9a and 10a as well as 3,2-HOPO and maltol benzyl ethers have been previously described in the literature [10][11][12][13].
Since the synthesis of the diamine building blocks 3, 4 and 6 as well as HOPO ligands 9a-b, 10a-b, 3,2-HOPO and maltol benzyl ethers have been previously described, we focused here only on the preparation of the new precursors 2, 5, 7 and 8.
2.1.1. Synthesis of New Diamine Building Blocks 2, 5, 7 and 8 As shown in Scheme 1 and according to previously described reaction conditions [8], a new azide precursor 2 was prepared starting from commercially available N-carboxybenzyl L-asparagine (N-Cbz-Asn-OH). The dehydration of its side-chain amide function using trifluoroacetic anhydride in pyridine first provided the cyano intermediate 11 in 92% yield. The corresponding N-Boc amino derivative 12 was subsequently obtained after NaBH4 reduction of nitrile function in the presence of NiCl2.6H2O as a catalyst and (Boc)2O. The conversion of acid function into primary alcohol followed by mesylate formation and treatment with NaN3 afforded the azide intermediate 15 in good yields. Removal of Boc-protecting group was performed to furnish the attempted C2 alkyl side-chain amino ligand 2 as hydrochloride salt.  Diamine Building Blocks 1 HOPO Ligands 1,2 1 Diamine building blocks 3, 4 and 6 as well as HOPO ligands 9b and 10b have been described in our previous publication [8]. 2 HOPO ligands 9a and 10a as well as 3,2-HOPO and maltol benzyl ethers have been previously described in the literature [10][11][12][13].
Since the synthesis of the diamine building blocks 3, 4 and 6 as well as HOPO ligands 9a-b, 10a-b, 3,2-HOPO and maltol benzyl ethers have been previously described, we focused here only on the preparation of the new precursors 2, 5, 7 and 8.
2.1.1. Synthesis of New Diamine Building Blocks 2, 5, 7 and 8 As shown in Scheme 1 and according to previously described reaction conditions [8], a new azide precursor 2 was prepared starting from commercially available N-carboxybenzyl L-asparagine (N-Cbz-Asn-OH). The dehydration of its side-chain amide function using trifluoroacetic anhydride in pyridine first provided the cyano intermediate 11 in 92% yield. The corresponding N-Boc amino derivative 12 was subsequently obtained after NaBH4 reduction of nitrile function in the presence of NiCl2.6H2O as a catalyst and (Boc)2O. The conversion of acid function into primary alcohol followed by mesylate formation and treatment with NaN3 afforded the azide intermediate 15 in good yields. Removal of Boc-protecting group was performed to furnish the attempted C2 alkyl side-chain amino ligand 2 as hydrochloride salt.  Diamine Building Blocks 1 HOPO Ligands 1,2 1 Diamine building blocks 3, 4 and 6 as well as HOPO ligands 9b and 10b have been described in our previous publication [8]. 2 HOPO ligands 9a and 10a as well as 3,2-HOPO and maltol benzyl ethers have been previously described in the literature [10][11][12][13].
Since the synthesis of the diamine building blocks 3, 4 and 6 as well as HOPO ligands 9a-b, 10a-b, 3,2-HOPO and maltol benzyl ethers have been previously described, we focused here only on the preparation of the new precursors 2, 5, 7 and 8.
2.1.1. Synthesis of New Diamine Building Blocks 2, 5, 7 and 8 As shown in Scheme 1 and according to previously described reaction conditions [8], a new azide precursor 2 was prepared starting from commercially available N-carboxybenzyl L-asparagine (N-Cbz-Asn-OH). The dehydration of its side-chain amide function using trifluoroacetic anhydride in pyridine first provided the cyano intermediate 11 in 92% yield. The corresponding N-Boc amino derivative 12 was subsequently obtained after NaBH4 reduction of nitrile function in the presence of NiCl2.6H2O as a catalyst and (Boc)2O. The conversion of acid function into primary alcohol followed by mesylate formation and treatment with NaN3 afforded the azide intermediate 15 in good yields. Removal of Boc-protecting group was performed to furnish the attempted C2 alkyl side-chain amino ligand 2 as hydrochloride salt.  Diamine Building Blocks 1 HOPO Ligands 1,2 1 Diamine building blocks 3, 4 and 6 as well as HOPO ligands 9b and 10b have been described in our previous publication [8]. 2 HOPO ligands 9a and 10a as well as 3,2-HOPO and maltol benzyl ethers have been previously described in the literature [10][11][12][13].
Since the synthesis of the diamine building blocks 3, 4 and 6 as well as HOPO ligands 9a-b, 10a-b, 3,2-HOPO and maltol benzyl ethers have been previously described, we focused here only on the preparation of the new precursors 2, 5, 7 and 8.
2.1.1. Synthesis of New Diamine Building Blocks 2, 5, 7 and 8 As shown in Scheme 1 and according to previously described reaction conditions [8], a new azide precursor 2 was prepared starting from commercially available N-carboxybenzyl L-asparagine (N-Cbz-Asn-OH). The dehydration of its side-chain amide function using trifluoroacetic anhydride in pyridine first provided the cyano intermediate 11 in 92% yield. The corresponding N-Boc amino derivative 12 was subsequently obtained after NaBH4 reduction of nitrile function in the presence of NiCl2.6H2O as a catalyst and (Boc)2O. The conversion of acid function into primary alcohol followed by mesylate formation and treatment with NaN3 afforded the azide intermediate 15 in good yields. Removal of Boc-protecting group was performed to furnish the attempted C2 alkyl side-chain amino ligand 2 as hydrochloride salt.  Diamine Building Blocks 1 HOPO Ligands 1,2 1 Diamine building blocks 3, 4 and 6 as well as HOPO ligands 9b and 10b have been described in our previous publication [8]. 2 HOPO ligands 9a and 10a as well as 3,2-HOPO and maltol benzyl ethers have been previously described in the literature [10][11][12][13].
Since the synthesis of the diamine building blocks 3, 4 and 6 as well as HOPO ligands 9a-b, 10a-b, 3,2-HOPO and maltol benzyl ethers have been previously described, we focused here only on the preparation of the new precursors 2, 5, 7 and 8.
2.1.1. Synthesis of New Diamine Building Blocks 2, 5, 7 and 8 As shown in Scheme 1 and according to previously described reaction conditions [8], a new azide precursor 2 was prepared starting from commercially available N-carboxybenzyl L-asparagine (N-Cbz-Asn-OH). The dehydration of its side-chain amide function using trifluoroacetic anhydride in pyridine first provided the cyano intermediate 11 in 92% yield. The corresponding N-Boc amino derivative 12 was subsequently obtained after NaBH4 reduction of nitrile function in the presence of NiCl2.6H2O as a catalyst and (Boc)2O. The conversion of acid function into primary alcohol followed by mesylate formation and treatment with NaN3 afforded the azide intermediate 15 in good yields. Removal of Boc-protecting group was performed to furnish the attempted C2 alkyl side-chain amino ligand 2 as hydrochloride salt.

Synthesis of New Multifunctional Scavengers of Series Ia-b and IIa-b by Coupling of Diamine Building Blocks with HOPO Ligands
A hydroxypyridinone moiety was introduced as a well-known iron chelator pharmacophore to complete the hybridization of our two new series of AGE/ALE inhibitors [18]. Different key coupling steps were designed to build various linkers between the two scavenging cores. Herein, the series  A hydroxypyridinone moiety was introduced as a well-known iron chelator pharmacophore to complete the hybridization of our two new series of AGE/ALE inhibitors [18]. Different key coupling steps were designed to build various linkers between the two scavenging cores. Herein, the series Ia-b was characterized by an amide or piperazine-1,4-diamide spacer whereas the series IIa-b was based on an alkyl or piperazine-1,4-dialkyl linker, respectively.
As shown in Scheme 4, the other new AGE/ALE inhibitors that possess a simple C4 or C9 alkyl or piperazine-1,4-dialkyl linker between the two pharmacophores were successfully synthesized to constitute the series IIa-b. The 2-methyl-3,4-HOPO derivatives 24d-e were prepared carrying out a condensation of maltol benzyl ether with amino starting blocks 4 and 5 following the protocol of Santos et al. [19]. Furthermore, a nucleophilic substitution of the iodo group of diamine derivatives 7 and 8 by 3,2-HOPO benzyl ether was achieved using sodium hydride in DMF to give hybrid products 27c-d in 72 and 62% yield, respectively [13]. A novel piperazine-1,4-dialkyl intermediate 27e was also prepared by reduction of the piperazine-1,4-diamide spacer of the corresponding analog 27b using borane-methyl sulfide complex following the protocol of Boschelli et al. [20]. The attempted compounds IIa-1 to IIa-4 and IIb-1 were finally provided via an O-debenzylation and diamine deprotection sequence as a promising novel series of hydroxypyridinone-diamine hybrids. Various new final products have been synthesized in 7-11 steps and constitute two novel series of HOPO-diamine hybrids: i) the series I was obtained in 13-39% overall yields and can be arranged in Ia with compounds Ia-1 to Ia-4 bearing an amide linker and Ib with compounds Ib-1 to Ib-3 bearing a piperazine-1,4-diamide linker and ii) the series II was prepared in 2-57% overall yields and can be arranged in IIa with compounds IIa-1 to IIa-4 possessing an alkyl spacer and IIb with compound IIb-1 possessing a piperazine-1,4-dialkyl spacer. The impact of these different pharmacomodulations on their ADME, physicochemical and biological properties as multifunctional AGE/ALE inhibitors, with respect to lead compound 1, in particular, have then been investigated.
Various new final products have been synthesized in 7-11 steps and constitute two novel series of HOPO-diamine hybrids: i) the series I was obtained in 13-39% overall yields and can be arranged in Ia with compounds Ia-1 to Ia-4 bearing an amide linker and Ib with compounds Ib-1 to Ib-3 bearing a piperazine-1,4-diamide linker and ii) the series II was prepared in 2-57% overall yields and can be arranged in IIa with compounds IIa-1 to IIa-4 possessing an alkyl spacer and IIb with compound IIb-1 possessing a piperazine-1,4-dialkyl spacer. The impact of these different pharmacomodulations on their ADME, physicochemical and biological properties as multifunctional AGE/ALE inhibitors, with respect to lead compound 1, in particular, have then been investigated.

QikProp-Predicted ADME Properties
A prediction of the ADME properties of newly synthesized products was performed using QikProp, a Schrödinger software. As shown in Figure 2, logP o/w (octanol/water partition coefficient) values calculated for derivatives of series IIa appeared much superior to that of lead compound 1 ( Table 2: clogP o/w of −1.88). Indeed, a rise of one to three logarithmic units was found for clogP o/w of novel HOPO-alkyl-diamine hybrids IIa-1 to IIa-4. There was no significant difference in terms of clogP o/w between 3,2-HOPO derivatives IIa-3 and IIa-4 and their 2-methyl-3,4-HOPO analogs IIa-1 and IIa-2, respectively. However, the lipophilicity strongly increased in this series with the length of  (Table 2: Ia-3 vs. Ia-2 with clogP o/w of −2.10 vs. −1.54, respectively), but they remained in the range of that of lead compound 1. In the series I, the introduction of a piperazine cycle inside the spacer greatly decreased the clogP o/w for 3,2-HOPO derivatives whereas the nature of HOPO moiety seemed to have no particular impact ( Figure 2: Ib-2 vs. 1 and Ib-1). However, the reduction of piperazine-1,4-diamide spacer of Ib-3 into a piperazine-1,4-dialkyl spacer led to an increase of more than one logarithmic unit for clogP o/w value of IIb-1 ( Table 2: clogP o/w of −2.52 and −1.31, respectively). As expected, the nature and the length of the spacer significantly affected the lipophilicity of HOPO-diamine hybrids. In particular, the introduction of an alkyl linker in the series IIa greatly improved it. This was correlated with an important increase of logBB (brain/blood partition coefficient) value calculated for the product IIa-3 compared to that of lead compound 1 (Table 2: clogBB of −0.86 and −1.77, respectively). In addition, enhanced apparent Caco-2 cell (a gutblood barrier model) and MDCK cell (a blood-brain barrier model) permeabilities to this new alkyl derivative were found ( It is also particularly interesting to note that this product exhibited the most important central nervous system activity on a −2 (inactive) to +2 (active) QikProp-established scale ( Table 2: IIa-3 vs. 1 with CNS activity of −1 vs. −2). Finally, in view of these predicted ADME properties and especially its improved capacity to cross the blood brain barrier, the HOPO-alkyl-diamine hybrid IIa-3 appears as a novel promising drug candidate for treating AD. Now, a further screening is to be carried out to confirm its good RCS trapping, antioxidant and Cu 2+ -chelating capacities.

Physicochemical Evaluations: Oxygen Radical Absorbance Capacity (ORAC), RCS Trapping and Cu 2+ -Chelating Assays
In order to evaluate the antioxidant activity of new synthesized HOPO-diamine hybrids, an oxygen radical absorbance capacity assay using fluorescein (ORACFL) was performed [21,22]. The fluorescein (FL) fluorescence decay induced by AAPH reagent, a peroxyl radical generator was followed in time at 37 °C. The measurement of the area under the curve (AUC) of the sample in comparison with the control corresponding to an absence of antioxidant allowed us to determine the  vs. 1.73 and PMDCK of 4.27 vs. 1.07 nm/sec, respectively). It is also particularly interesting to note that this product exhibited the most important central nervous system activity on a −2 (inactive) to +2 (active) QikProp-established scale ( Table 2: IIa-3 vs. 1 with CNS activity of −1 vs. −2). Finally, in view of these predicted ADME properties and especially its improved capacity to cross the blood brain barrier, the HOPO-alkyl-diamine hybrid IIa-3 appears as a novel promising drug candidate for treating AD. Now, a further screening is to be carried out to confirm its good RCS trapping, antioxidant and Cu 2+ -chelating capacities. model) permeabilities to this new alkyl derivative were found ( . It is also particularly interesting to note that this product exhibited the most important central nervous system activity on a −2 (inactive) to +2 (active) QikProp-established scale ( Table 2: IIa-3 vs. 1 with CNS activity of −1 vs. −2). Finally, in view of these predicted ADME properties and especially its improved capacity to cross the blood brain barrier, the HOPO-alkyl-diamine hybrid IIa-3 appears as a novel promising drug candidate for treating AD. Now, a further screening is to be carried out to confirm its good RCS trapping, antioxidant and Cu 2+ -chelating capacities. important increase of logBB (brain/blood partition coefficient) value calculated for the product IIa-3 compared to that of lead compound 1 (Table 2: clogBB of −0.86 and −1.77, respectively). In addition, enhanced apparent Caco-2 cell (a gutblood barrier model) and MDCK cell (a blood-brain barrier model) permeabilities to this new alkyl derivative were found ( . It is also particularly interesting to note that this product exhibited the most important central nervous system activity on a −2 (inactive) to +2 (active) QikProp-established scale ( Table 2: IIa-3 vs. 1 with CNS activity of −1 vs. −2). Finally, in view of these predicted ADME properties and especially its improved capacity to cross the blood brain barrier, the HOPO-alkyl-diamine hybrid IIa-3 appears as a novel promising drug candidate for treating AD. Now, a further screening is to be carried out to confirm its good RCS trapping, antioxidant and Cu 2+ -chelating capacities. Po/w represent the octanol/water partition coefficient. QikProp-recommended values for logPo/w are ranged from −2 to 6.5. 2 BB represent the brain/blood partition coefficient for orally delivered drugs. QikProp-recommended values for logBB are ranged from −3 to 1.2. 3 PCaco represent the apparent Caco-2 cell permeability in nm/sec. Caco-2 cells constitute a gutblood barrier model. QikProp-recommended PCaco values are significant for non-active transport. PCaco values <25 correspond to a poor permeability whereas PCaco values >500 correspond to a great permeability. 4 PMDCK represent the apparent MDCK cell permeability in nm/sec. MDCK cells constitute a bloodbrain barrier model. QikProp-recommended PMDCK values are significant for non-active transport. PCaco values <25 correspond to a poor permeability whereas PMDCK values >500 correspond to a great permeability. 5 CNS activity represent the QikProp-predicted central nervous system activity of drug candidate on a −2 (inactive) to +2 (active) scale. introduction of an alkyl linker in the series IIa greatly improved it. This was correlated with an important increase of logBB (brain/blood partition coefficient) value calculated for the product IIa-3 compared to that of lead compound 1 (Table 2: clogBB of −0.86 and −1.77, respectively). In addition, enhanced apparent Caco-2 cell (a gutblood barrier model) and MDCK cell (a blood-brain barrier model) permeabilities to this new alkyl derivative were found ( . It is also particularly interesting to note that this product exhibited the most important central nervous system activity on a −2 (inactive) to +2 (active) QikProp-established scale ( Table 2: IIa-3 vs. 1 with CNS activity of −1 vs. −2). Finally, in view of these predicted ADME properties and especially its improved capacity to cross the blood brain barrier, the HOPO-alkyl-diamine hybrid IIa-3 appears as a novel promising drug candidate for treating AD. Now, a further screening is to be carried out to confirm its good RCS trapping, antioxidant and Cu 2+ -chelating capacities. Po/w represent the octanol/water partition coefficient. QikProp-recommended values for logPo/w are ranged from −2 to 6.5. 2 BB represent the brain/blood partition coefficient for orally delivered drugs. QikProp-recommended values for logBB are ranged from −3 to 1.2. 3 PCaco represent the apparent Caco-2 cell permeability in nm/sec. Caco-2 cells constitute a gutblood barrier model. QikProp-recommended PCaco values are significant for non-active transport. PCaco values <25 correspond to a poor permeability whereas PCaco values >500 correspond to a great permeability. 4 PMDCK represent the apparent MDCK cell permeability in nm/sec. MDCK cells constitute a bloodbrain barrier model. QikProp-recommended PMDCK values are significant for non-active transport. PCaco values <25 correspond to a poor permeability whereas PMDCK values >500 correspond to a great permeability. 5 CNS activity represent the QikProp-predicted central nervous system activity of drug candidate on a −2 (inactive) to +2 (active) scale. introduction of an alkyl linker in the series IIa greatly improved it. This was correlated with an important increase of logBB (brain/blood partition coefficient) value calculated for the product IIa-3 compared to that of lead compound 1 (Table 2: clogBB of −0.86 and −1.77, respectively). In addition, enhanced apparent Caco-2 cell (a gutblood barrier model) and MDCK cell (a blood-brain barrier model) permeabilities to this new alkyl derivative were found ( . It is also particularly interesting to note that this product exhibited the most important central nervous system activity on a −2 (inactive) to +2 (active) QikProp-established scale ( Table 2: IIa-3 vs. 1 with CNS activity of −1 vs. −2). Finally, in view of these predicted ADME properties and especially its improved capacity to cross the blood brain barrier, the HOPO-alkyl-diamine hybrid IIa-3 appears as a novel promising drug candidate for treating AD. Now, a further screening is to be carried out to confirm its good RCS trapping, antioxidant and Cu 2+ -chelating capacities. Po/w represent the octanol/water partition coefficient. QikProp-recommended values for logPo/w are ranged from −2 to 6.5. 2 BB represent the brain/blood partition coefficient for orally delivered drugs. QikProp-recommended values for logBB are ranged from −3 to 1.2. 3 PCaco represent the apparent Caco-2 cell permeability in nm/sec. Caco-2 cells constitute a gutblood barrier model. QikProp-recommended PCaco values are significant for non-active transport. PCaco values <25 correspond to a poor permeability whereas PCaco values >500 correspond to a great permeability. 4 PMDCK represent the apparent MDCK cell permeability in nm/sec. MDCK cells constitute a bloodbrain barrier model. QikProp-recommended PMDCK values are significant for non-active transport. PCaco values <25 correspond to a poor permeability whereas PMDCK values >500 correspond to a great permeability. 5 CNS activity represent the QikProp-predicted central nervous system activity of drug candidate on a −2 (inactive) to +2 (active) scale. introduction of an alkyl linker in the series IIa greatly improved it. This was correlated with an important increase of logBB (brain/blood partition coefficient) value calculated for the product IIa-3 compared to that of lead compound 1 (Table 2: clogBB of −0.86 and −1.77, respectively). In addition, enhanced apparent Caco-2 cell (a gutblood barrier model) and MDCK cell (a blood-brain barrier model) permeabilities to this new alkyl derivative were found ( . It is also particularly interesting to note that this product exhibited the most important central nervous system activity on a −2 (inactive) to +2 (active) QikProp-established scale ( Table 2: IIa-3 vs. 1 with CNS activity of −1 vs. −2). Finally, in view of these predicted ADME properties and especially its improved capacity to cross the blood brain barrier, the HOPO-alkyl-diamine hybrid IIa-3 appears as a novel promising drug candidate for treating AD. Now, a further screening is to be carried out to confirm its good RCS trapping, antioxidant and Cu 2+ -chelating capacities. Po/w represent the octanol/water partition coefficient. QikProp-recommended values for logPo/w are ranged from −2 to 6.5. 2 BB represent the brain/blood partition coefficient for orally delivered drugs. QikProp-recommended values for logBB are ranged from −3 to 1.2. 3 PCaco represent the apparent Caco-2 cell permeability in nm/sec. Caco-2 cells constitute a gutblood barrier model. QikProp-recommended PCaco values are significant for non-active transport. PCaco values <25 correspond to a poor permeability whereas PCaco values >500 correspond to a great permeability. 4 PMDCK represent the apparent MDCK cell permeability in nm/sec. MDCK cells constitute a bloodbrain barrier model. QikProp-recommended PMDCK values are significant for non-active transport. PCaco values <25 correspond to a poor permeability whereas PMDCK values >500 correspond to a great permeability. 5 CNS activity represent the QikProp-predicted central nervous system activity of drug candidate on a −2 (inactive) to +2 (active) scale. introduction of an alkyl linker in the series IIa greatly improved it. This was correlated with an important increase of logBB (brain/blood partition coefficient) value calculated for the product IIa-3 compared to that of lead compound 1 (Table 2: clogBB of −0.86 and −1.77, respectively). In addition, enhanced apparent Caco-2 cell (a gutblood barrier model) and MDCK cell (a blood-brain barrier model) permeabilities to this new alkyl derivative were found ( . It is also particularly interesting to note that this product exhibited the most important central nervous system activity on a −2 (inactive) to +2 (active) QikProp-established scale ( Table 2: IIa-3 vs. 1 with CNS activity of −1 vs. −2). Finally, in view of these predicted ADME properties and especially its improved capacity to cross the blood brain barrier, the HOPO-alkyl-diamine hybrid IIa-3 appears as a novel promising drug candidate for treating AD. Now, a further screening is to be carried out to confirm its good RCS trapping, antioxidant and Cu 2+ -chelating capacities. Po/w represent the octanol/water partition coefficient. QikProp-recommended values for logPo/w are ranged from −2 to 6.5. 2 BB represent the brain/blood partition coefficient for orally delivered drugs. QikProp-recommended values for logBB are ranged from −3 to 1.2. 3 PCaco represent the apparent Caco-2 cell permeability in nm/sec. Caco-2 cells constitute a gutblood barrier model. QikProp-recommended PCaco values are significant for non-active transport. PCaco values <25 correspond to a poor permeability whereas PCaco values >500 correspond to a great permeability. 4 PMDCK represent the apparent MDCK cell permeability in nm/sec. MDCK cells constitute a bloodbrain barrier model. QikProp-recommended PMDCK values are significant for non-active transport. PCaco values <25 correspond to a poor permeability whereas PMDCK values >500 correspond to a great permeability. 5 CNS activity represent the QikProp-predicted central nervous system activity of drug candidate on a −2 (inactive) to +2 (active) scale.

Physicochemical Evaluations: Oxygen Radical Absorbance Capacity (ORAC), RCS Trapping and Cu 2+ -Chelating Assays
In order to evaluate the antioxidant activity of new synthesized HOPO-diamine hybrids, an oxygen radical absorbance capacity assay using fluorescein (ORAC FL ) was performed [21,22]. The fluorescein (FL) fluorescence decay induced by AAPH reagent, a peroxyl radical generator was followed in time at 37 • C. The measurement of the area under the curve (AUC) of the sample in comparison with the control corresponding to an absence of antioxidant allowed us to determine the protective effect of tested compounds (Figure 3a). Trolox, a water-soluble vitamin E analog was used as a standard for the calculation of ORAC FL values at 10 µM expressed as µmol trolox equivalent (TE)/µmol of tested compound with respect to the linear equation of its calibration curve (Net AUC vs. concentration) [23]. As shown in Figure 3b, the majority of newly designed products appeared to be efficient ROS scavengers (ORAC FL values ≥0.8 µmol TE/µmol, except for IIb-1 and Ib-1), especially compared to the very weak protective effect of carnosine, reported as a good antioxidant (Table 3: ORAC FL value of 0.08 µmol TE/µmol) [24]. The HOPO moiety previously revealed to be responsible for the radical trapping activity [8]. Nevertheless, the lead compound 1 remains the most potent HOPO-diamine hybrid ( Table 3: ORAC FL value of 1.96 µmol TE/µmol). In the two novel series, the 3,2-HOPO derivatives exhibited slightly better antioxidant properties than their 2-methyl-3,4-HOPO analogs (Figure 3b: Ib-2 vs. Ib-1 and IIa-3 vs. IIa-1). Furthermore, some pharmacomodulations carried out on the linker could affect this protective effect. In the series I, the compound Ia-4 turned out to be almost Pharmaceuticals 2019, 12, 162 9 of 28 as active as lead compound 1 whereas its analog Ia-3 bearing a one-carbon shorter amide linker provided a less interesting result (Table 3: ORAC FL values of 1.67 and 0.81 µmol TE/µmol, respectively). The simple removal of the carbonyl group from HOPO moiety greatly improved its reactivity. However, the introduction of a piperazine cycle inside the spacer significantly decreased ROS scavenging capacities (Figure 3b: 1 vs. Ib-2 and Ia-2 vs. Ib-1). In the series II, there was a negative impact of the reduction of piperazine-1,4-diamide linker into piperazine-1,4-dialkyl spacer (Table 3: ORAC FL values of 0.99 and 0.57 µmol TE/µmol for Ib-3 and IIb-1, respectively). Finally, the 3,2-HOPO derivative IIa-3 possessing a C4 alkyl spacer appeared slightly more efficient than its analog IIa-4 bearing a C9 alkyl linker (Table 3: ORAC FL values of 0.98 and 0.84 µmol TE/µmol, respectively). Judging from its good conserved antioxidant activity and with its promising Qikprop-predicted ADME properties in mind, this HOPO-alkyl-diamine hybrid could be considered as a new lead compound. Its RCS trapping and Cu 2+ -chelating capacities have been evaluated according to previously described conditions [8].
A kinetic study of RCS adduct formation was assessed by LCMS after the incubation of new key diamine compounds IIa-3 and Ia-3 with MGO or MDA at 37 • C for 24 h (Figure 4). Indeed, the appearance of already described major pyrazine or 1,4-diazepine adducts could be followed on UV chromatogram at 190 nm, respectively [8]. AUC of the total peak of adducts was compared with the remaining free scavenger peak for samples collected at regular time intervals. As shown in Table 3, the reactivity of primary vicinal diamine function of IIa-3 and Ia-3 with dicarbonyl compounds remains as interesting as that of lead compound 1. Newly designed derivatives were found to be very potent MGO scavengers. In particular, 100% of MGO adduct formation was detected in 15 min of incubation with IIa-3 vs. only 46% with carnosine used as reference [24]. However, the MDA trapping capacity of this HOPO-alkyl-diamine hybrid revealed weaker than that of Ia-3 and lead compound 1 both bearing an amide linker (Table 3: 40% vs. 83% and 81% MDA adduct formation, respectively, in 1 h of incubation). The modification of the spatial geometry of the hybrid structure induced by the introduction of an alkyl spacer seems to lead to a lower reactivity of diamine function with bulky RCS like α,β-unsaturated aldehydes. and 0.84 μmol TE/μmol, respectively). Judging from its good conserved antioxidant activity and with its promising Qikprop-predicted ADME properties in mind, this HOPO-alkyl-diamine hybrid could be considered as a new lead compound. Its RCS trapping and Cu 2+ -chelating capacities have been evaluated according to previously described conditions [8].
(a) (b)    A kinetic study of RCS adduct formation was assessed by LCMS after the incubation of new key diamine compounds IIa-3 and Ia-3 with MGO or MDA at 37 °C for 24 h (Figure 4). Indeed, the appearance of already described major pyrazine or 1,4-diazepine adducts could be followed on UV chromatogram at 190 nm, respectively [8]. AUC of the total peak of adducts was compared with the remaining free scavenger peak for samples collected at regular time intervals. As shown in Table 3, the reactivity of primary vicinal diamine function of IIa-3 and Ia-3 with dicarbonyl compounds remains as interesting as that of lead compound 1. Newly designed derivatives were found to be very potent MGO scavengers. In particular, 100% of MGO adduct formation was detected in 15 min of incubation with IIa-3 vs. only 46% with carnosine used as reference [24]. However, the MDA trapping capacity of this HOPO-alkyl-diamine hybrid revealed weaker than that of Ia-3 and lead compound 1 both bearing an amide linker (Table 3: 40% vs. 83% and 81% MDA adduct formation, respectively, in 1 h of incubation). The modification of the spatial geometry of the hybrid structure induced by the introduction of an alkyl spacer seems to lead to a lower reactivity of diamine function with bulky RCS like α,β-unsaturated aldehydes.  The evaluation of Cu 2+ -chelating properties of derivatives IIa-3 and Ia-3 representing the two new series of AGE/ALE inhibitors was performed using murexide as complexometric indicator. Tested compounds at different concentrations were incubated with CuSO4.5H2O for 10 min at rt. Murexide was subsequently introduced into the mixture and reacted with remaining free Cu 2+ for additional 1 min. The absorbance ratio A485/A520 (λmax of Cu 2+ /murexide complex: 485 nm and λmax of free murexide: 520 nm) furnished the remaining free Cu 2+ concentration with respect to calibration curves (A485/A520 vs. Cu 2+ concentration). Knowing the total quantity of metal ions added into the reaction mixture (control conditions without tested product), proportion of Cu 2+ chelation by HOPO-diamine hybrids was estimated by difference. Newly synthesized compounds IIa-3 and Ia-3 appeared to be much more potent than carnosine in complexing Cu 2+ ( Figure 5). In the hybrid structure, it has been previously demonstrated that the role of HOPO moiety is more essential than that of the diamine function as far as the Cu 2+ -chelating activity is concerned [8]. As shown in Table   0  The evaluation of Cu 2+ -chelating properties of derivatives IIa-3 and Ia-3 representing the two new series of AGE/ALE inhibitors was performed using murexide as complexometric indicator. Tested compounds at different concentrations were incubated with CuSO 4. 5H 2 O for 10 min at rt. Murexide was subsequently introduced into the mixture and reacted with remaining free Cu 2+ for additional 1 min. The absorbance ratio A 485 /A 520 (λ max of Cu 2+ /murexide complex: 485 nm and λ max of free murexide: 520 nm) furnished the remaining free Cu 2+ concentration with respect to calibration curves (A 485 /A 520 vs. Cu 2+ concentration). Knowing the total quantity of metal ions added into the reaction mixture (control conditions without tested product), proportion of Cu 2+ chelation by HOPO-diamine hybrids was estimated by difference. Newly synthesized compounds IIa-3 and Ia-3 appeared to be much more potent than carnosine in complexing Cu 2+ ( Figure 5). In the hybrid structure, it has been previously demonstrated that the role of HOPO moiety is more essential than that of the diamine function as far as the Cu 2+ -chelating activity is concerned [8]. As shown in Table 3, the 2-methyl-3,4-HOPO derivative Ia-3 showed 83% Cu 2+ chelation at 200 µM and revealed almost as efficient as the positive standard EDA (ethylenediamine) whereas the 2,3-HOPO one IIa-3 exhibited its chelating capacity in only 60%. This provides novel evidence that demonstrates the slightly weaker biometal complexing properties of the 2,3-HOPO moiety vs. the 2-methyl-3,4-HOPO one. However, judging from similar interesting results obtained with IIa-3 and lead compound 1, the nature of the linker seems to have no particular impact in Cu 2+ chelation ( Figure 5). 3, the 2-methyl-3,4-HOPO derivative Ia-3 showed 83% Cu 2+ chelation at 200 μM and revealed almost as efficient as the positive standard EDA (ethylenediamine) whereas the 2,3-HOPO one IIa-3 exhibited its chelating capacity in only 60%. This provides novel evidence that demonstrates the slightly weaker biometal complexing properties of the 2,3-HOPO moiety vs. the 2-methyl-3,4-HOPO one. However, judging from similar interesting results obtained with IIa-3 and lead compound 1, the nature of the linker seems to have no particular impact in Cu 2+ chelation ( Figure 5). indicator reacting with remaining free Cu 2+ , was added and the mixture was incubated for additional 1 min at rt. The absorbance ratio A485/A520 (λmax of Cu 2+ /murexide complex: 485 nm and λmax of free murexide: 520 nm) provided remaining free Cu 2+ concentration with respect to calibration curves (A485/A520 vs. Cu 2+ concentration). Knowing the total quantity of metal ions introduced in the reaction mixture (control conditions without tested product), percentage of Cu 2+ chelation by tested compounds was calculated by difference. Data are presented as means ± SEM of triplicates.
Finally, in view of its particularly promising predicted ADME properties and its good conserved RCS, ROS and biometal scavenging efficiency, the newly designed 3,2-HOPO-alkyl-diamine IIa-3 can be selected as a novel lead compound.

Cell Viability
Cytotoxicity of the novel HOPO-diamine hybrids was evaluated on an in vitro neuronal-like cell model, the rat pheochromocytoma cell-line PC12 widely used to study neurodegenerative diseases [25,26]. A sensitive colorimetric CCK-8 (cell counting kit-8) assay was performed to determine cell viability [27]. In the cell, mitochondrial dehydrogenases reduced the WST-8 reagent in the presence of 1-methoxy PMS used as an electron carrier to give a water-soluble formazan dye. The number of living cells was estimated by medium absorbance measurement after incubation with different concentrations of tested compounds for 48 h. A positive cytotoxicity standard was assessed by adding 10% DMSO and cell viability was expressed as % of control conditions (non-treated cells). Except for compounds IIa-2 and IIa-4 bearing a C9 alkyl linker and to a lesser extent for compounds Ia-1 and Ia-2, no cytotoxicity of new multifunctional diamine derivatives was detected in PC12 cells after 48 h of treatment at 10 μM as well as at 100 μM ( Figure 6). The increased cellular toxicity of hybrids IIa-2 and IIa-4 could be explained by the well-known ability of flexible long alkyl chains to induce membrane disruption and cytolysis slotting into the phospholipid bilayer [28]. , used as a complexometric indicator reacting with remaining free Cu 2+ , was added and the mixture was incubated for additional 1 min at rt. The absorbance ratio A 485 /A 520 (λ max of Cu 2+ /murexide complex: 485 nm and λ max of free murexide: 520 nm) provided remaining free Cu 2+ concentration with respect to calibration curves (A 485 /A 520 vs. Cu 2+ concentration). Knowing the total quantity of metal ions introduced in the reaction mixture (control conditions without tested product), percentage of Cu 2+ chelation by tested compounds was calculated by difference. Data are presented as means ± SEM of triplicates.
Finally, in view of its particularly promising predicted ADME properties and its good conserved RCS, ROS and biometal scavenging efficiency, the newly designed 3,2-HOPO-alkyl-diamine IIa-3 can be selected as a novel lead compound.

Cell Viability
Cytotoxicity of the novel HOPO-diamine hybrids was evaluated on an in vitro neuronal-like cell model, the rat pheochromocytoma cell-line PC12 widely used to study neurodegenerative diseases [25,26]. A sensitive colorimetric CCK-8 (cell counting kit-8) assay was performed to determine cell viability [27]. In the cell, mitochondrial dehydrogenases reduced the WST-8 reagent in the presence of 1-methoxy PMS used as an electron carrier to give a water-soluble formazan dye. The number of living cells was estimated by medium absorbance measurement after incubation with different concentrations of tested compounds for 48 h. A positive cytotoxicity standard was assessed by adding 10% DMSO and cell viability was expressed as % of control conditions (non-treated cells). Except for compounds IIa-2 and IIa-4 bearing a C9 alkyl linker and to a lesser extent for compounds Ia-1 and Ia-2, no cytotoxicity of new multifunctional diamine derivatives was detected in PC12 cells after 48 h of treatment at 10 µM as well as at 100 µM ( Figure 6). The increased cellular toxicity of hybrids IIa-2 and IIa-4 could be explained by the well-known ability of flexible long alkyl chains to induce membrane disruption and cytolysis slotting into the phospholipid bilayer [28].

In Vitro MGO-Induced Apoptosis Inhibition in the Model AD Cell-Line PC12
With the aim of assessing their ability to limit MGO-induced apoptosis in AD, a pre-treatment of PC12 cells with compounds Ia-4, IIa-3 and IIa-4 representing the two new series of AGE/ALE inhibitors were carried out. After incubation in the presence of MGO at 37 °C, apoptosis level was evaluated by ELISA detection of DNA fragmentation and reflected by a measurement of optical density (OD) [29,30]. Cells incubated without MGO represented control conditions and a positive apoptosis standard was constituted in the presence of MGO, but in the absence of the tested product. As shown in Figure 7, the apoptosis level was greatly decreased in the presence of HOPO-alkyl-diamine derivatives IIa-3 and IIa-4 at 100 μM whereas there was no significant effect of the compound Ia-4 bearing an amide linker neither at 10 μM nor at 100 μM. Furthermore, the new multifunctional scavengers IIa-3 and IIa-4 revealed to be more efficient than lead compound 1 to inhibit MGO-induced apoptosis in PC12 cells (21% of apoptosis inhibition with 1 at 100 μM (previously reported result [8]) vs. 39% and 60% with IIa-3 and IIa-4, respectively). Considering their similar trapping abilities, these results can be explained by the increased lipophilicity of IIa-3 and IIa-4 associated with the enhanced QikProp-predicted cellular permeability to these compounds (Table 2). However, the long-chain alkyl hybrid IIa-4 must be ruled out for further investigations because of its detected cytotoxicity in PC12 cells. Finally, together with its RCS, ROS and biometal scavenging efficiency and its good QikProp-predicted ADME properties, HOPO-alkyl-diamine derivative IIa-3 exhibited an in vitro potent biological activity. These new data constitute a promising evidence of its improved druglikeness and capacity to prevent biometal-mediated oxidative and carbonyl stress progression in the PC12 cell-line model of AD.

In Vitro MGO-Induced Apoptosis Inhibition in the Model AD Cell-Line PC12
With the aim of assessing their ability to limit MGO-induced apoptosis in AD, a pre-treatment of PC12 cells with compounds Ia-4, IIa-3 and IIa-4 representing the two new series of AGE/ALE inhibitors were carried out. After incubation in the presence of MGO at 37 • C, apoptosis level was evaluated by ELISA detection of DNA fragmentation and reflected by a measurement of optical density (OD) [29,30]. Cells incubated without MGO represented control conditions and a positive apoptosis standard was constituted in the presence of MGO, but in the absence of the tested product. As shown in Figure 7, the apoptosis level was greatly decreased in the presence of HOPO-alkyl-diamine derivatives IIa-3 and IIa-4 at 100 µM whereas there was no significant effect of the compound Ia-4 bearing an amide linker neither at 10 µM nor at 100 µM. Furthermore, the new multifunctional scavengers IIa-3 and IIa-4 revealed to be more efficient than lead compound 1 to inhibit MGO-induced apoptosis in PC12 cells (21% of apoptosis inhibition with 1 at 100 µM (previously reported result [8]) vs. 39% and 60% with IIa-3 and IIa-4, respectively). Considering their similar trapping abilities, these results can be explained by the increased lipophilicity of IIa-3 and IIa-4 associated with the enhanced QikProp-predicted cellular permeability to these compounds (Table 2). However, the long-chain alkyl hybrid IIa-4 must be ruled out for further investigations because of its detected cytotoxicity in PC12 cells. Finally, together with its RCS, ROS and biometal scavenging efficiency and its good QikProp-predicted ADME properties, HOPO-alkyl-diamine derivative IIa-3 exhibited an in vitro potent biological activity. These new data constitute a promising evidence of its improved druglikeness and capacity to prevent biometal-mediated oxidative and carbonyl stress progression in the PC12 cell-line model of AD.

Generalities
All reagents and solvents were obtained from commercial suppliers and were used without further purification. Anhydrous solvents were dried with the PureSolv ® MD 5 solvent purification system. Column chromatography was performed over Merck Silica gel 60 Å (40-63 μm) or using a Grace Davison Purification Reveleris ® Flash system. Routine monitoring of reactions was carried out using Merck thin layer chromatography (TLC) Silica gel 60 F254 plates, visualized under UV light (254 nm) and revealed using a phosphomolybdic acid 95% solution in ethanol. Melting points (mp) were determined on a Stuart SMP3 device. IR measurements were performed on a Jasco FT/IR-4200 system fitted with an ATR-golden gate and were reported using the frequency of absorption (cm −1 ). nuclear magnetic resonance (NMR) spectra were recorded on a Bruker 400 (or 300) cryosonde NMR instrument. Chemical shifts () are expressed in parts per million (ppm) downfield from tetramethylsilane as an internal standard and were referenced to the deuterated solvent. 1 H, 13 C and Q-DEPT data were reported in the order of chemical shift, multiplicity (s = singlet, br s = broad singlet, d = doublet, t = triplet, q = quartet, quint = quintet and m = multiplet), integration, coupling constants in Hertz (Hz). Negative peaks (np) are quoted for Q-DEPT NMR spectra. High-resolution mass spectra (HRMS) were obtained from a Micromass-Waters Q-TOF Ultima spectrometer, in electrospray ionization (ESI) mode (positive or negative). For liquid chromatography coupled with mass spectrometry (LCMS), UV chromatograms and mass spectra were obtained from a Shimadzu LCMS-2020 system, at 190 nm and by positive ESI-MS interface (detection mode: scan, interface voltage: tuning file, DL voltage: 100 V, Q-array DC: 40 V, Q-array RF: 40 V). The gradient elution was performed on a Phenomenex Kinetex ® HPLC C18 column using an injection volume of 1-2 μL and a mobile phase composed of water/acetonitrile (solvent A/solvent B) with 0.1% formic acid (98:2 during 2 min, 55:45 during 2 min and 45:55 during 3 min with a flow of 0.3 mL/min at 40 °C).

Generalities
All reagents and solvents were obtained from commercial suppliers and were used without further purification. Anhydrous solvents were dried with the PureSolv ® MD 5 solvent purification system. Column chromatography was performed over Merck Silica gel 60 Å (40-63 µm) or using a Grace Davison Purification Reveleris ® Flash system. Routine monitoring of reactions was carried out using Merck thin layer chromatography (TLC) Silica gel 60 F254 plates, visualized under UV light (254 nm) and revealed using a phosphomolybdic acid 95% solution in ethanol. Melting points (mp) were determined on a Stuart SMP3 device. IR measurements were performed on a Jasco FT/IR-4200 system fitted with an ATR-golden gate and were reported using the frequency of absorption (cm −1 ). nuclear magnetic resonance (NMR) spectra were recorded on a Bruker 400 (or 300) cryosonde NMR instrument. Chemical shifts (δ) are expressed in parts per million (ppm) downfield from tetramethylsilane as an internal standard and were referenced to the deuterated solvent. 1 H, 13 C and Q-DEPT data were reported in the order of chemical shift, multiplicity (s = singlet, br s = broad singlet, d = doublet, t = triplet, q = quartet, quint = quintet and m = multiplet), integration, coupling constants in Hertz (Hz). Negative peaks (np) are quoted for Q-DEPT NMR spectra. High-resolution mass spectra (HRMS) were obtained from a Micromass-Waters Q-TOF Ultima spectrometer, in electrospray ionization (ESI) mode (positive or negative). For liquid chromatography coupled with mass spectrometry (LCMS), UV chromatograms and mass spectra were obtained from a Shimadzu LCMS-2020 system, at 190 nm and by positive ESI-MS interface (detection mode: scan, interface voltage: tuning file, DL voltage: 100 V, Q-array DC: 40 V, Q-array RF: 40 V). The gradient elution was performed on a Phenomenex Kinetex ® HPLC C18 column using an injection volume of 1-2 µL and a mobile phase composed of water/acetonitrile (solvent A/solvent B) with 0.1% formic acid (98:2 during 2 min, 55:45 during 2 min and 45:55 during 3 min with a flow of 0.3 mL/min at 40 • C).

Synthesis of Precursors 7 and 8
General procedure: Compound 17 or 22 (1-1.3 mmol) was dissolved in acetone (15-20 mL) and sodium iodide (5 equiv) was added. The mixture was stirred at rt for 10-16 h. The solvent was removed and the residue was taken up in EtOAc (3 × 30-40 mL). The organic phase was washed with brine (3 × 30-40 mL), dried over Na 2 SO 4 and concentrated in vacuo to give compound 7.
For compound 9a and 10a, synthetic pathways and NMR spectra were in full accordance with those reported in the literature [10,11].