Screening of Some Sulfonamide and Sulfonylurea Derivatives as Anti-Alzheimer’s Agents Targeting BACE1 and PPARγ

Department of Neurology, e ird Hospital of Jinan, Shandong, Jinan 250132, China Department of Neurology, People’s Hospital of Danyang, Jiangsu, Danyang 212300, China Department of Neurology, No. 1 Hospital, Handan 056002, Hebei, China Department of Neurology, Beijing Shijitan Hospital, Capital Medical University, Beijing 100038, China Department of Neurology, Emergency Medical Center of Chongqing, Chongqing 400014, China Department of Neurology, e First People’s Hospital of Taizhou, Zhejiang, Taizhou 318020, China Pharmaceutical Medicinal Chemistry & Drug Design Department, Faculty of Pharmacy (Boys), Al-Azhar University, Cairo 11884, Egypt Zoology Department, Faculty of Science, Al-Azhar University, Cairo 11884, Egypt Pharmaceutical Chemistry Department, College of Pharmacy, Taif University, Taif 21974, Saudi Arabia


Introduction
Alzheimer's disease (AD) is one of the most well-known neurodegenerative diseases, and it is characterized by a series of various mental conditions such as memory loss and other cognitive impairments [1]. Today, 50 million people are suffering from AD, and 5.4 million of them are Americans. Currently, a new case appear every 67 seconds, and by 2050, each new case of AD is expected to appear every 33 seconds, with an approximate prevalence range between 11 million and 16 million patients [2]. AD is a major health issue for all communities.
Physiological deregulations associated with disease progression have been identified, such as loss of synapses and synaptic activity, structural and functional mitochondrial defects, inflammatory responses, development of extracellular neuritis plaques, and neuronal losses [3]. e neurological disorders of AD occur extracellular, essentially in amyloid-β peptide (Aβ), or intracellular by aggregation of phosphorylated tau proteins.
ere are many AD treatments comprising 105 drugs, of which twenty-five are in phase I, fifty-two are in phase II, and twenty-eight are in phase III [4]. e approved drug treatments include the following: (i) cholinesterase inhibitors as donepezil, rivastigmine, and galantamine [5,6] and (ii) memantine that uncompetitively blocks the NMDA receptor and can act as a neuroprotective [7]. Additionally, there are many potential future drug treatments: (i) antiamyloid treatment, targeting amyloid-β (Aβ), gamma-secretase, or beta-secretase (as AZD3293 and MK-8931) [8]; (ii) immunization against Aβ as AN 1792, which was stopped because of meningoencephalitis in 6% of subjects [9]; (iii) monoclonal antibodies as bapineuzumab, which is a monoclonal antibody targeting Aβ, examined in patients with mild to moderate AD during phase III clinical trial [7]; (iv) tau-targeted therapy currently in clinical trials as methylthioninium (phase II clinical trial), and it has showed advantages for patients with AD after fifty weeks therapy [10]. e β-secretase enzyme, which is known as β-site amyloid precursor protein cleaving enzyme 1 (BACE1), initiates the production of the toxic amyloid-β (Aβ) that plays an important role in Alzheimer's disease [11]. BACE1 is a crucial therapeutic target for reducing cerebral Aβ concentrations in AD because of its vital role in the generation of Aβ. Nevertheless, BACE1 also acts as a housekeeping enzyme, and it is implicated in the treating of many other proteins that are responsible for appropriate neuronal tissue function [12]. In addition, recent studies on multitarget directed ligands have provided insight and hope on the use of natural products in targeting AD via BACE-beta amyloid mechanism [13].
Studies have shown that an altered insulin pathway can affect the deposition of amyloid-β protein and the phosphorylation of tau protein, both leading factors in the development of AD [14]. Hence, drugs used for type 2 diabetes mellitus (T2DM) as sulfonylurea receptor SUR agonists and peroxisome proliferator-activated receptors (PPARɣ) agonists may represent a promising candidate to fight against AD [15].
Sulfonylureas (SUs) are one of the most well-known groups of antidiabetic drugs that stimulate insulin secretion by interacting with the pancreatic ATP-sensitive potassium channels (KATP) [16], which are also found in neurons [17]. It was reported that glibenclamide treatment of mice decreased hippocampal A., inhibited neuronal apoptosis, and improved synaptic plasticity of the hippocampus [18]. Also, a combination of metformin and sulfonylurea has been reported to decrease the risk of dementia by 35% over eight years [19]. Recently, it has been reported that the incidence of dementia is decreased in T2DM patients when receiving SUs treatments [20,21].
New sulfonamides were designed and synthesized aiming to produce multifunctional agents against Alzheimer's disease. eir investigation resulted in the identification of promising leads that can help in further development of new promising candidates [22].
PPARc plays an essential role in the metabolism of glucose and in the processing of fatty acids, making it a key target for the production of antihyperglycemic agents [23]. ey also enhance skeletal muscle sensitivity, inhibit hepatic gluconeogenesis, boost glycemic control, and decrease circulating insulin levels [24]. It has been reported that PPARc agonists can suppress proinflammatory molecules in peripheral immune cells and resident glial cells; hence, PPARc agonists are considered to be active against Alzheimer's disease. In addition, these agonists showed significant effects in neurodegenerative CNS disorders in animals [25]. PPARc agonists were also reported to have a role in neurons inflammatory disorders because of their potential to suppress NFkB-mediated signals at numerous sites [26].
Moreover, it has been confirmed before that sulfonylureas and sulfonamides have the PPARc agonistic activity [27,28]. Additionally, it was reported that some sulfonamide derivatives showed bioactivity against BACE1, indicating their expected potential as promising anti-AD active agents [29].
In the normal physiological conditions, PPARc can be expressed in the brain at low levels. A full analysis of gene expression has lately shown that mRNA levels are high in patients with AD [30]. is indicates that PPARc plays a vital role in the modulation of the AD pathophysiology. e drugs currently used are primarily aimed at symptomatic treatment. Such agents have a limited therapeutic efficacy over rather short periods. erefore, the development of new therapeutic approaches is critical [31].
In this work, our rationale focused on investigating some recorded sulfonylureas and sulfonamides on PPARc and BACE1. Accordingly, we have performed virtual screening using docking studies for tens of sulfonamides and sulfonylureas against PPARc and BACE1. e most promising compounds were further tested in vitro against PPARc and BACE1.

Rational of Molecular Design.
Studying the structureactivity relationships of PPARc agonists revealed that they have five basic structural features for binding to PPARc. e features include an acidic head, a linker attached to an aromatic scaffold (spacer group), a linker, and a heteroaromatic lipophilic tail [32,33] (Figure 1).
iazolidinediones (TZDs) class is the most famous class of compounds reported as high-affinity agonists to PPARc [34,35].
Also, it was reported that sulfonylureas and sulfonamides have PPARc agonistic activity. Accordingly, in this study, moieties of sulfonylurea and sulfonamide serve as acid heads required for the agonistic action of PPARc. In our compounds, the sulfonyl (SO 2 ) group acts as a single atom spacer between an acid head and an aromatic group. e para-disubstituted phenyl group acts as an aromatic spacer, which is essential for ideal PPARc agonism [36]. Many different linkers between the lipophilic tail and an aromatic spacer have been utilized in our compounds. ese linkers are important for the agonistic action of PPARc. Eventually, various heteroaromatic nuclei were used to serve as the lipophilic tail necessary for PPARc.
Based on the above considerations and to obtain new anti-AD agents, thirty-five reported sulfonamide and sulfonylurea derivatives [37][38][39] having the same essential pharmacophoric features of the reported PPARc agonists ( Figure 2) have been screened using virtual screening via docking studies (Figure 3). Based on docking studies, the most promising candidates were subjected to further in silico and in vitro studies. e in silico studies include docking against BACE1, ADMET, and physiochemical properties. e in vitro studies comprise the PPARc-ligand binding assay and BACE1 inhibitory activities.

Preparation of the Target Molecules.
e target molecules, PPARc (PDB ID: 1FM6 resolution 2.1Å, https:// www.rcsb.org/structure/1FM6) and BACE1 (PDB ID: 2qk5, resolution 2.2, https://www.rcsb.org/structure/2QK5), were downloaded from the Protein Data Bank. MOE software was used in the performance of the docking analysis [40]. In this procedure, the free energies and binding modes of the analyzed compounds against PPARc and BACE1were detected. e active sites of both targets were prepared by a sequence of several steps including deleting water molecules, protonation of each amino acid chain, and hiding the hydrogen atoms.

Energy Minimization.
e energy of the target receptors was minimized using the energy minimization algorithm of the MOE tool. e following parameters were utilized for energy minimization: gradient: 0.05, force field: MMFF94X + solvation, and chiral constraint: current geometry. Energy minimization was terminated when the root mean square gradient falls below 0.05. e initial and final energy of protein were calculated (in kcal/mol) by GizMOE using MMFF94X force field with the conjugant gradient method. e minimized structure was used as the template for docking. en, the binding active sites of the target receptors were defined. Journal of Chemistry

Calculating the Active Site Sequence.
Active sites present in the target receptors were identified from the following parameters: compute, surfaces and maps, molecular surface, create, and isolate.

Preparation of Ligand
Molecule. e next step involves the preparation of the tested compound for docking. is step starts by drawing these compounds using ChemBio-Draw Ultra 14.0. e created file was saved as MDL-SD format. en, the saved file was opened using MOE. Next, several processes were carried out including protonation of the 3D structures of the tested compounds and rosiglitazone, minimization of the potential energy, and validation process.

Docking Process.
e binding of the ligand molecule with the target molecules was carried out using MOE program to find the correct conformation (with the rotation of bonds, the structure of molecule is not rigid) and configuration (with the rotation of whole molecule, the structure of the molecule remains rigid) of the ligand, so as to obtain minimum energy structure. e parameters used for docking were as follows: total runs � 30, gradient � 0.01, no. of return poses � 1000, iteration limit � 500, potential energy grid: ON, rescoring1: London dG, and refinement: force field. e output from MOE was further analyzed with Discovery Studio 4.0 software [41,42].

In Silico ADMET Analysis.
e ADMET model of Discovery Studio software version 4.0 was used to access the drug-like properties of the tested compounds. Lipinski's rule of five was used as a standard [43]. is rule explains different properties of a drug molecule, as well as its solubility, absorption, interaction, metabolism, excretion, and toxicity were also predicted. In this study, the protocol of ADMET descriptor module of the small molecules was applied as following steps.

Ligand Preparation.
e tested compounds were drawn using ChemBioDraw Ultra 14.0 and saved as MDL-SD format.
is file was opened using Discover Studio software version 4.0. e ligand geometry was minimized by applying the force field algorithm, and the ligand preparation protocol was achieved using the following parameters: change ionization: true, generate tautomer: true, generate isomer: true, fixed bad valences: true, generate coordinates: 3D, parallel processing: false, and then run [44].

Drug-Likeness Prediction.
From the protocol of small molecules, ADMET descriptors were used. en, the prepared ligands in the previous step were applied. Also, the different descriptors including aqueous solubility, bloodbrain barrier penetration, cytochrome P450 2D6 enzyme inhibition, hepatotoxicity, and plasma protein binding level were applied [44].

BACE1 Assay.
e inhibition activity of tested compounds to BACE1 was determined via manufacturer's protocol from Invitrogen Kit (L0724) [45]. IC 50 (concentration of compound required to displace 50% of titrated ligand) was calculated, and each experiment was repeated twice.

PPARc-Ligand Binding Assay.
e binding affinity of the tested compounds to PPARc was determined via the fluorescence polarization assay technique [46]. e Polar Screen ™ PPARc-competitor assay kit (Invitrogen, Carlsbad, CA) was used in this technique. According to the manufacturer's instructions, the procedure of determination of binding affinity was conducted. EC 50 (concentration of compound required to displace 50% of titrated ligand) was calculated [47], and each experiment were repeated twice.

Computational Determination of the Essential Physicochemical Properties.
e essential physicochemical properties of the tested compounds (1, 2, 3, 4, and 5) were determined using Discovery Studio 4.0 software. First, the compounds were prepared by the application of force fields via CHARMM and MMFF94. en, the different molecular descriptors were predicted from the general purpose protocol [38,44,48].

Docking Studies against PPARc.
e receptor-based drug design (docking) approach [49][50][51][52][53][54][55][56] was used to analyse the binding mode of the reported compounds with the PPARc. According to the literature survey, the PPARc cavity consists of three main parts: an entrance, arm I, and arm II. Arm I contains four polar residues such as Ser289, Tyr473, His323, and His449 involved in hydrogen bonding. Arm II comprises of Ile281, Ile 341, Leu353, and Val339, while the entrance consists of Leu330 Leu333, Arg288, and Ser342 [57] ( Figure 4). e results of virtual screening using docking studies revealed that some of the analyzed compounds exhibited similar orientations inside the putative binding sites of PPARc. e binding energies of these compounds against PPARc are illustrated in Table 1. e results indicated that there are five compounds that have promising binding affinity toward PPARc. en, the binding mode of the most promising members was discussed in detail as follows.
Rosiglitazone (the co-crystalized ligand) showed a binding mode with an affinity value equal to −24.44 kcal/ mol. iazolidinedione nucleus was directed into the polar site of the receptor. iazolidinedione formed two hydrogen bonding interactions with Ser289 and His449. e pyridine moiety formed hydrophobic interactions with Arg288 and Val339. e central phenyl ring formed hydrophobic interactions with Leu330 and Cys285 (Figures 5 and 6).
Compound 1 showed a binding mode like that of the cocrystalized ligand (rosiglitazone), with binding energy of −18.05 kcal/mol. e sulfonamide group was directed into the polar part of PPARc forming two hydrogen bonds with Cys285 and Ser289. 3-Phenylquinazolin-4(3H)-one moiety was oriented in the hydrophobic pocket forming hydrophobic interactions with Leu333, Glu343, Arg288, Cys285, and Ile341. Also, it has formed a hydrogen bond with Ser342. e phenyl spacer formed hydrophobic interaction with Ile326, Leu330, Cys285, and Arg288. Moreover, the amide moiety formed a hydrogen bond with Arg288 (Figures 7  and 8). e binding mode of compound 2 (affinity value of −28.62 kcal/mol) was virtually like that of the co-crystalized ligand where the sulfonamide group was oriented in the hydrophilic region, forming three hydrogen bonding interactions with Tyr473, His449, and Ser289. e 3-(4-bromophenyl) quinazolin-4(3H)-one moiety was oriented in the hydrophobic region to form six hydrophobic interactions and two hydrogen bonds with Glu343 and Ser342 (Figures 9 and 10). e mapping surface technique was carried out to show compound 2 occupying the active pocket of PPARc ( Figure 11).   (Figures 12-17).

Docking Studies against BACE1.
e best five compounds that showed good affinity to PPARc were selected to be subjected for further docking studies on BACE1. Results of docking studies revealed that the tested compounds have a good binding mode inside the pocket of BACE1 with binding energies ranging from −20.33 to −25.16 (Table 2). e co-crystalized ligand showed a binding mode with an affinity value equal to −25.60 kcal/mol. It has formed five hydrogen bonds with Asp93, Gly95, Asp289, Gly291, and r293. In addition, it has showed three hydrophobic interactions with Tyr132, Phe169, and Tyr259 (Figures 18 and  19).
Compound 3 (as a representative example, all figures for the other docked compounds are provided in supplementary data) showed a binding mode like that of co-crystalized ligand with a binding energy of −25.16 kcal/mol. It has formed three hydrogen bonds with Gly95, Arg189, and r133. In addition, it has formed four hydrophobic interactions with Tyr132, Trp176, Leu91, and Ile179 (Figures 20 and 21). e mapping surface technique was carried out to show compound 3 occupying the active pocket of BACE1 ( Figure 22).
With regard to the binding mode of compounds 1, 2, 4, and 5 against BACE, these compounds exhibited binding modes like that of co-crystalized ligand with a binding energy of −25. 16

In Silico ADMET Analysis.
ADMET studies are techniques used for the prediction of the pharmacokinetic behavior of new chemical agents [44,48]. In this procedure, the analyzed compounds and the reference drug (pioglitazone) were subjected for computational determination of absorption level, cytochrome P450 2D6 enzyme inhibition, hepatotoxicity, plasma protein binding level, and solubility level. Discovery Studio 4.0 software was used in this technique (Table 3 and Figure 23). e results revealed that most of the tested compounds have moderate intestinal absorption. On the other hand, compound 3 has good intestinal absorption, while compound 4 has poor intestinal absorption. e CYP2D6 score predicts the inhibitory and noninhibitory behavior of particular compounds on cytochrome P450 2D6 enzyme. e results showed that all the tested compounds are CYP2D6 noninhibitors. Also, the hepatotoxicity prediction values were in the range from 0.256 to 0.4567. Accordingly, liver dysfunction, the most common side effect of CYP2D6 inhibitors, is unexpected upon administration of such members. e plasma protein binding model predicts the binding ability of a compound to plasma proteins. All the tested compounds have a PPB level of more than 95%. It is widely known that many drug candidates did not reach the final phase of clinical trials due to problems related to their absorption properties [58]. Depending on the computation analysis of the synthesized compounds, it was found that most of the compounds have an ADME aqueous solubility logarithmic level equal to 1 or 2, indicating low-to-moderate aqueous solubility.

Biological Evaluation
3.3.1. BACE1 Assay. Compounds 1-5 were examined to determine their abilities to inhibit BACE1 according to the manufacturer's protocol from Invitrogen (L0724), and the assay was performed to investigate the activity of the selected compounds against BACE1.
e results of inhibition are reported in Table 4      Journal of Chemistry

PPARc-Ligand Binding Assay.
Compounds with promising computational PPARc affinities (1, 2, 3, 4, and 5) were examined to determine their binding affinities to PPARc. e fluorescence polarization assay technique [46] was carried out to assess binding affinities of the selected compounds with PPARc-LBD. Rosiglitazone, as one of the most active PPARc agonist, was used as a positive control. e results of binding are reported in Table 5 as EC 50 values. Compound 2 exhibited a significant binding affinity to PPARc with an EC 50 value of 0.289 μM when compared with rosiglitazone (EC 50 � 0.292 μM). Compounds 1, 3, 4, and 5 showed high PPARc binding affinities of 0.399, 0.462, 0.473, and 0.426, respectively. Interestingly, the PPARc binding affinity of the most active compounds was consistent with that of computational PPARc affinity.

Computational Determination of the Essential Physicochemical Properties.
Calculated partition coefficient (c log P), molecular polar surface area, molecular solubility, molecular volume, molecular weight, hydrogen bond acceptors,  Gly 284 Cys 285 Gln 286 Arg 288 Ser 289 His 323 Ile 326 Tyr 327 Leu 330 Leu 333 Val 339 Leu 340 Ile 341 Ser 342 His 449 Leu 469 Tyr 473  ese values may explain the variation in their biological activity compared with their lipophilicity. Interestingly, the c log P values for the most active compounds lied in the ideal range of lipophilicity [59], which facilitate BBB penetration and consequently can treat AD disease. Based on these results, we noted a correlation between the PPARc affinity of the target compounds and their lipophilic characters.

Conclusion
irty-five sulfonamide and sulfonylurea derivatives have been screened using docking studies for their inhibition to PPARc, and then, the promising molecules were also docked into BACE1 as a potential target for anti-AD agents. Five compounds showed promising affinities against BACE1 and PPARc with binding energies ranging from −17.96 to −28.62 Kcal/mol. e ADMET studies were tested in silico using Discovery Studio 4.0 software.
e results revealed that the tested compounds have a CYP2D6 noninhibitory effect, moderate aqueous solubility, and intestinal absorption.
e physicochemical properties were determined in silico. It was found that the c log P values for the tested compounds range from 1.927 to 4.325. ese values may explain the variation in their biological activity compared with their lipophilicity. Interestingly, the c log P values for the most active compounds lied in the ideal range of lipophilicity, which facilitate BBB penetration and consequently can treat AD disease. Moreover, the compounds were in consistence with the Lipinski rule of five, which indicates their oral bioavailability. Finally, additional in vitro studies were carried out for these compounds to estimate their activity on both PPARc and BACE1. e results revealed the promiscuity of these compounds (1, 2, 3, 4, and 5) to target both PPARc and BACE1, which are potential targets in treatment of AD. Compound 2 showed a good activity on both targets, and its EC 50 value is 0.289 μM against PPARc and EC 50 value is 1.24 μM against BACE1. Finally, we can say that the most active candidates may serve as useful lead compounds in search for powerful anti-AD agents.

AD:
Alzheimer's disease ADMET: Absorption, distribution, metabolism, excretion, and toxicity Aβ: Amyloid-β peptide BACE1: Beta-secretase 1 c log P: Calculated partition coefficient CHARMM: Chemistry at harvard macromolecular mechanics CNS: Central nervous system EC 50 : e concentration of a drug which induces a response halfway between the baseline and maximum after a specified exposure time IC 50  Data Availability e data used in this study are given within the manuscript.

Conflicts of Interest
e authors declare that they have no conflicts of interest.

Authors' Contributions
Ning Li and Yan Wang are the first authors and contributed equally to this work.

Supplementary Materials
Only for reviewing consideration and not for publication the manuscript contains all the essential figures. However, the supplementary data file comprises the 2D structures of the most promising compounds docked into the active site of BACE1. (Supplementary Materials)