Pharmacophore-driven design and biological evaluation of heterocyclic PDE4B inhibitors for Alzheimer's disease Authors: Karthik G. Pujar, Dipen Purohit, Richa Prabhakar Shetty, Parthasarathi S, Krishna Yallappa Kolachi, Aishwarya Bharathi H. M., Pooja Gandharvachari Achar, Prabitha P. Published: 2026 ABSTRACT Alzheimer's disease (AD) is a complex neurodegenerative condition marked by a gradual loss of cognitive function, impaired synaptic signalling, and sustained neuroinflammatory responses. Among the molecular pathways associated with disease progression, cyclic adenosine monophosphate (cAMP) signalling, plays a crucial factor in disruption of cellular homeostasis. Phosphodiesterase-4 (PDE4), especially the PDE4B isoform, controls intracellular cAMP levels in neuronal and glial cells, which affects inflammatory signalling and synaptic plasticity in the brain. Therefore, selectively targeting PDE4B is a novel approach to identify the small molecule's therapeutic activity to restore the cAMP signalling, with parallel regulation of the dose-limiting side effects. GSK356278 and Rolipram are the known PDE4B modulators, which have been considered for the current study to generate significant pharmacophoric features and execute drug design as CNS-active PDE4B inhibitors. Six heterocyclic derivatives were designed and synthesised by altering the core scaffold in order to maximise hydrophobicity, CNS permeability, and Q-pocket engagement without depending on direct Zn2+ or Mg2+ coordination. Among these, molecule VCHH-19 and VCPH-19 demonstrated significant CNS compatibility and good binding stability. PC12 cell lines were chosen to verify the in vitro neuroprotective response, and the Drosophila model for in vivo for AD assessment. All of these results indicated the significance of pyridine-based heterocycles molecules as potential candidates for PDE4B-targeted AD treatment. Introduction Alzheimer's disease (AD), the most common neurodegenerative illness and a major cause of dementia, affects over 50 million people globally. Amyloid-β (Aβ) plaque deposition, neurofibrillary tangles, synaptic impairment, and progressive neuronal loss are all revealed by neuropathological examination.1 From clinical perspective, memory, cognitive function, and behaviour gradually deteriorate as a result of this illness. Approved treatments for AD, such as the NMDA receptor antagonist memantine and acetylcholinesterase inhibitors (donepezil, rivastigmine, and galantamine), only alleviate symptoms and not halt the disease's progression.2 This therapeutic gap highlights the need for innovative approaches that directly address the mechanisms causing neurodegeneration rather than merely providing temporary symptom relief. A crucial molecular pathway for both neuronal survival and cognitive function is cyclic adenosine monophosphate (cAMP) signalling. In the central nervous system, cAMP serves as a second messenger that regulates synaptic plasticity and long-term potentiation.3 When Protein Kinase A (PKA) is activated, the cAMP response element-binding protein (CREB), which controls the transcription of genes essential for neuronal survival and synaptic maintenance, including brain-derived neurotrophic factor (BDNF), when phosphorylated. The pathology of AD has been closely linked to the disruption of this pathway, which results in increased neuroinflammatory responses and synaptic dysfunction. Phosphodiesterases (PDEs) use enzymatic hydrolysis to control intracellular levels of cAMP and cGMP. PDE4, one of the eleven known PDE families, is highly expressed in the brain and are specific to cAMP signalling. PDE4B in particular is highly expressed in immune-responsive of the brain and microglia cells, where it regulates inflammatory signalling via NF-κB pathways.4 In AD, excessive PDE4B activity accelerates neuronal damage by lowering cAMP levels and encouraging neuroinflammation. Although early PDE4 inhibitors, like Rolipram with neuroprotective, cognitive, and dose-limiting emesis has been primarily caused by PDE4B inhibition, restricted to their clinical use. The development of novel PDE4B-selective inhibitors with better cellular tolerance developed a significant approach due to the clinical constraints.5,6 A brain-penetrant PDE4 inhibitor that is presently in phase I trials, GSK356278, shows decreased emetogenic liability and supports the viability of selective PDE4B targeting pathway. Strong inhibitors bind primarily through hydrophobic and hydrogen-bonding interactions rather than metal chelation in a distinct catalytic pocket with nearby Q1 and Q2 subpockets, according to structural analysis of the PDE4B–Rolipram co-crystal structure. Based on these discoveries, pharmacophore model of GSK356278 determined its aromatic and hydrogen-bonding characteristics in an small molecule, essential for PDE4B inhibition.7 Pyridine is a favourable scaffold for CNS activity, as demonstrated by superimposition studies that also showed overlap between the pyridine moieties of GSK356278 and the CNS medication donepezil.8 On this basis, the present study integrates ligand-based drug design strategies to develop novel pyridine-based heterocyclic PDE4B inhibitors, which were assessed using an in vitro tests and in vivo Drosophila model to find potential candidates for AD treatment. Materials and methods In silico studies 3D structural alignment and RMSD analysis of ligand designed For RMSD evaluation, the ligands, including three reference molecules, were energy minimized using the OPLS4 force field on Schrödinger Maestro Suite. The ligands were validated for 3D structural orientation and alignment using RMSD scoring to assess optimal fit. Three references selected are Donepezil, a CNS-active medication with neuroprotective relevance; GSK356278, a clinically evaluated PDE4B inhibitor;8 and Rolipram, the co-crystal ligand defining the canonical PDE4B binding mode. These reference ligands were superimposed for spatial 3D overlay and RMSD calculation using Maximum Common Substructure (MCS) method, focusing on the orientation of heterocyclic scaffolds, other aromatic rings, and PDE4B-binding substituents. During scaffold selection, recurrent structural elements—including a six-membered aromatic ring system, a pyridine-containing heterocycle, and consistent methoxy substituents—were used as references. GSK356278 was chosen as the primary template since it showed superior RMSD scores and 3D alignment. Before molecular docking, all novel ligands were validated for structural orientation and was then proceeded for in silico studies. Pharmacophore modelling and analysis BIOVIA Discovery Studio (Dassault Systèmes) was used for pharmacophore modelling, which has a feature of “Auto Pharmacophore Generation”. It generates the pharmacophore output based on highest selectivity predicted by Genetic Function Approximation (GFA) model. This GFA model defines the selectivity of pharmacophore built from training set of 500 models, i.e. from MiniMaybridge inbuilt database. The six designed compounds (VCU, VCT, VCHH, VCPH, VCDNP, and VCHA) that were shortlisted from the virtual screening of 108 molecules were imported in a three-dimensional format and energy-minimised using default parameters, along with 3 reference ligands (GSK356278, Rolipram, and Donepezil). The following parameters were set i.e. maximum features = 6; minimum feature distance = 2; maximum pharmacophores conformers = 10; which were used to assign features like aromatic rings, hydrogen bond acceptors (HBA), hydrogen bond donors (HBD), and hydrophobic, cationic, and anionic moieties. In order to assess feature conservation and spatial overlap, the pharmacophore alignment protocol was used to align the generated pharmacophore models of the designed compounds with those of the reference ligands.9 The alignment of aromatic cores, pyridine-associated HBA features, and hydrophobic regions that corresponded to PDE4B's Q1 and Q2 sub-pockets of Rolipram was the main focus of the analysis. In line with structural evidence that PDE4B inhibition does not require any direct coordination of Zn2+ or Mg2+ hence those pharmacophoric features were excluded but the features contributed to non-metal bindings at the pocket are the methoxy groups and five-membered groups.7 The identified pharmacophore features were correlated with structural information derived from the PDE4B–Rolipram co-crystal structure (PDB ID: 1XN0) to incorporate coherence between observations based on ligands and structures. Molecular docking and molecular dynamics simulations were then supported by the verified pharmacophore models.10 Molecular docking and virtual screening studies From the Protein Data Bank (PDB) the crystal structure of the PDE4B catalytic domain co-crystallised with Rolipram (PDB ID: 1XN0) was obtained. The Schrödinger Maestro Suite's Protein Preparation Wizard was used to prepare proteins in accordance with a procedure that has been documented in computational studies. Appropriate protonation states were created, bond orders were assigned, and missing hydrogen atoms were added. Crystallographic water molecules farther than 5 Å from the active site were eliminated, and hydrogen-bonding networks were optimised.11 The OPLS4 force field restrained heavy atoms for minimising the energy of the protein structure until an RMSD convergence of 0.30 Å was attained. The default van der Waals scaling parameters were used to define the docking cartesian grid around the native rolipram-binding pocket. To create three-dimensional structures and ionisation states indicative of physiological pH, all ligands were created using ChemDraw Professional 16.0 and LigPrep/Epik. For the comparison study, the reference compounds Donepezil, Rolipram, and GSK356278 were also docked. The OPLS4 force field was used to generate and minimise the energy of thirty conformers generated for each ligand at the time of docking. The analysis of the Glide docking module's Standard Precision (SP) mode was performed according to the ranking of the docked poses and Glide scores; they were observed for binding and pose orientation in 3D with key interactions as those of known PDE4B compounds.11,12 Molecular dynamics simulation To examine the stability and dynamic behaviour of PDE4B–ligand complexes, molecular dynamics (MD) simulations were performed using the Desmond molecular dynamics engine within the Schrödinger Maestro suite (v14.2). In order to ensure a minimum buffer distance of 10 Å from the protein surface, each complex was embedded in an orthorhombic cartesian box 10 Å × 10 Å × 10 Å solvated with TIP3P water molecules using the Desmond System Builder. Standard ionic strength conditions were used, and the systems were neutralised by adding the proper Na+ and Cl− counterions. The OPLS4 force field was used for energy minimisation before simulation to eliminate unfavourable contacts.13,14 The minimised systems were then equilibrated under NPT conditions at 300 K and 1 atm using the Martyna–Tobias–Klein barostat for pressure coupling and the Nose–Hoover chain thermostat for temperature control. A stepwise molecular dynamics simulation strategy was employed to evaluate ligand stability within the PDE4B active site. Initially, the six selected ligands, together with the reference compounds GSK356278, Donepezil, and Rolipram, were subjected to short 1 ns MD simulations to study unstable binding conformations. Complexes that maintained stable binding during this initial phase were subsequently progressed to 10 ns MD simulations for preliminary assessment of dynamic behaviour and interaction persistence. The VCPH-19 complex and the three reference compounds were advanced to 50 ns production simulations based on the findings. The most promising systems, GSK356278–PDE4B and VCPH-19–PDE4B, were then subjected to extended 100 ns and 200 ns MD simulations to assess their long-timescale stability. Trajectories were recorded, which were carried out under the NPT ensemble at 300 K and 1 atm. Long-range electrostatic interactions were treated using the Particle Mesh Ewald (PME) method, and van der Waals interactions were truncated at a cutoff distance of 9 Å. The default Desmond integration was used to apply bond constraints involving hydrogen atoms. The root means square deviation (RMSD), root mean square fluctuation (RMSF), radius of gyration (Rg), and hydrogen bond occupancy were all evaluated through trajectory analysis using Simulation Interaction Diagram (SID) tools in order to evaluate complex stability and interaction persistence. ADMET prediction and CNS multiparameter optimisation (MPO) assessment Using QikProp (Schrödinger Suite), in silico ADMET and CNS multiparameter optimisation (MPO) analyses were carried out for the six designed compounds. GSK356278, for which experimental in vivo and cellular ADMET data are available, was used to first validate the predictive system.8 Model reliability was evaluated by comparing reported values with predicted parameters. ADMET and CNS-MPO descriptors were created for the designed compounds after validation. To allow for contextual comparison, the co-crystal ligand Rolipram and the conventional Donepezil were also studied.14 Synthetic procedures General synthetic strategy The heterocyclic analogues were synthesised as depicted in Fig. 1via Claisen–Schmidt condensation to obtain the key intermediate – vanillin chalcone (VC-19),15 followed by cyclisation with hydrazine (HH), phenyl hydrazine (PH), 2,4-dinitrophenyl hydrazine (DNP), urea (U), thiourea (T), and hydroxylamine (HA) to obtain molecules VCHH-19, VCPH-19, VCDNP-19, VCU-19, VCT-19, and VCHA-19, respectively. Scheme of synthesis of designed molecules. (A), (B) and (D) 2–3 conc. HCl & refluxed at 70 °C for 6–8 h, (E), (F) 40% NaOH & stirred at room temperature for 2–3 h, (C) 0.1 mol sodium acetate refluxed at 70 °C for 5 h. Synthesis of chalcone intermediate (VC-19) Equimolar concentration of 3-acetylpyridine and vanillin were dissolved in ethanol, followed by dropwise addition of 40% aqueous KOH solution. The reaction was stirred for 3 h at room temperature and kept overnight for crystallization. The crude solid was filtered, washed with ethanol, and recrystallised to yield pure VC-19.15 Synthesis of pyrazole ring (VCHH-19, VCPH-19 & VCDNP-19) VC-19 (0.01 mol) and hydrazine hydrate (0.01 mol) were dissolved in ethanol (25 mL), followed by the addition of a few drops of concentrated HCl. The mixture was refluxed at 70 °C for 6–8 h, and reaction progress was monitored by TLC (silica gel). Upon completion, the reaction was poured onto crushed ice to precipitate the product, which was filtered, dried, and recrystallised from rectified spirit to obtain VCHH-19. The compound VCPH-19 and VCDNP-19 were synthesised using phenylhydrazine and 2,4-dinitrophenylhydrazine, respectively, using a similar process.16,17 Synthesis of pyridine ring (VCT-19 & VCU-19) VC-19 (0.01 mol) and thiourea (0.01 mol) were dissolved in a 40% ethanolic solution of NaOH. The mixture was stirred at room temperature for 2–3 h, then poured into 400 mL of cold water and stirred for 1 h. The solution was refrigerated at 4 °C for 24–28 h to complete precipitation. The product was filtered, washed, and recrystallised from ethanol to give VCT-19. A similar procedure was adopted to synthesise VCU-19 using urea.16 Synthesis of iso-oxazoline ring (VCHA-19) VC-19 (0.01 mol), hydroxylamine hydrochloride (0.01 mol), and sodium acetate (0.01 mol) were refluxed in absolute ethanol (15 mL) at 70 °C for 5 h. The reaction was monitored by using TLC. After completion, the mixture was poured into ice-cold water to precipitate the product, which was filtered, dried, and recrystallised.18 2-Methoxy-4-(3-(pyridin-3-yl)-4,5-dihydro-1H-pyrazol-5-yl) phenol (VCHH-19) Melting point 185 ± 3; IR (cm−1): 3445 (O–H), 3230 (N–H), 3050 (Ar–C–H), 2935 (C–H), 1605, 1580, 1508 (C Created by potrace 1.16, written by Peter Selinger 2001-2019 C/CN), 1245, 1170 (C–O); 1H-NMR (400 MHz, CHLOROFORM-D) δ 8.82 (dd, J = 2.1, 0.8 Hz, 1H), 8.55 (dd, J = 4.7, 1.7 Hz, 1H), 8.02 (dt, J = 7.9, 1.9 Hz, 1H), 7.31 (dd, J = 8.0, 4.8 Hz, 1H), 6.92 (d, J = 2.0 Hz, 1H), 6.90–6.87 (m, 1H), 6.83 (dd, J = 8.1, 2.0 Hz, 1H), 6.09 (s, 1H), 4.91 (t, J = 10.2 Hz, 1H), 3.88 (s, 3H), 3.45 (dd, J = 16.4, 10.8 Hz, 1H), 3.01 (dd, J = 16.4, 9.5 Hz, 1H); 13C-NMR (101 MHz, CHLOROFORM-D) δ 149.6, 148.5, 147.4, 147.1, 145.6, 134.1, 133.1, 129.1, 123.7, 119.5, 114.7, 108.7, 77.5, 77.2, 76.8, 64.6, 56.1, 41.2; MS: m/z (M + 1) = 270.16, calculated 269.12; (% yield – 70%). 2-Methoxy-4-(1-phenyl-3-(pyridin-3-yl)-4,5-dihydro-1H-pyrazol-5-yl) phenol (VCPH-19) Melting point 180 ± 3; IR (cm−1): 3433 (N–H), 3055 (Ar–C–H), 2926, 2854 (C–H), 1598, 1556, 1506 (CC/CN), 1433, 1330 (C–N), 1217, 1157 (N–N), 1064, 1028 (C–N), 922, 871 (Ar–C–H bending), 748 (Ar–H); 1H-NMR (400 MHz, DMSO-D6) δ 8.94 (s, 1H), 8.90 (d, J = 1.8 Hz, 1H), 8.54 (dd, J = 4.7, 1.5 Hz, 1H), 8.11 (dt, J = 8.0, 1.8 Hz, 1H), 7.45 (dd, J = 8.0, 4.8 Hz, 1H), 7.18–7.14 (m, 2H), 7.05 (d, J = 7.8 Hz, 2H), 6.90 (d, J = 1.6 Hz, 1H), 6.76–6.65 (m, 3H), 5.37 (dd, J = 12.1, 7.1 Hz, 1H), 3.89 (dd, J = 17.5, 12.3 Hz, 1H), 3.71 (s, 3H), 3.15 (dd, J = 17.6, 7.1 Hz, 1H); 13C-NMR (101 MHz, DMSO-D6) δ 149.7, 148.5, 147.3, 146.4, 145.4, 144.8, 133.7, 133.1, 129.3, 124.3, 119.5, 118.7, 116.3, 113.8, 110.6, 79.7, 64.0, 56.1, 40.7, 40.5, 40.3, 40.1, 39.9, 39.6, 39.4; MS: m/z (M + 1) = 344.17, calculated. 345.15; (% yield – 72%). 4-(1-(2,4-Dinitrophenyl)-3-(pyridin-3-yl)-4,5-dihydro-1H-pyrazol-5-yl)-2-methoxyphenol (VCDNP-19) Melting point 205 ± 3; IR (cm−1): 3051 (Ar–C–H), 2924, 2851 (C–H), 1656 (CN), 1593, 1570, 1518 (CC/CN), 1469, 1438 (C–N), 1383 (NO2), 1288 (C–N), 1219, 1190, 1123 (C–O/C–N), 1016 (N–N), 979 (Ar–C–H bending), 837, 813, 750 (Ar–H), 703 (C–N), 630, 563 (NO2 bending).; 1H-NMR (400 MHz, chloroform-D) δ 9.22 (d, J = 1.8 Hz, 1H), 8.80 (dd, J = 4.4, 1.5 Hz, 1H), 8.28 (dt, J = 7.9, 1.9 Hz, 1H), 7.79 (d, J = 15.6 Hz, 1H), 7.49–7.44 (m, 1H), 7.15 (d, J = 2.0 Hz, 1H), 7.04–6.96 (m, 2H), 6.83 (d, J = 8.3 Hz, 1H), 6.56 (d, J = 12.8 Hz, 1H), 5.95 (s, 1H), 3.98 (s, 3H); 13C NMR (δ): 40.9, 56.1, 63.5, 109.2, 115.5, 119.0, 123.2–130.6, 136.1, 137.8, 138.3–138.6, 146.8, 148.2–148.6, 152.5; MS: m/z (M − 1) = 432.49, calculated 433.1; (% yield – 75%). 4-(4-Hydroxy-3-methoxyphenyl)-6-(pyridin-3-yl) pyrimidine-2(1H)-thione (VCT-19) Melting point 176 ± 3; IR (cm−1): 3465–3392 (O–H), 3346 (N–H), 2928, 2854 (C–H), 1675 (CS), 1613, 1599, 1516 (CC/CN), 1452, 1402 (C–N), 1348 (C–N), 1243, 1168, 1022 (C–O/C–N), 842 (Ar–C–H bending), 775 (Ar–H), 515 (C–S).1H-NMR (400 MHz, D2O) δ 8.95 (s, 1H), 8.76 (dd, J = 3.7, 1.4 Hz, 1H), 8.24 (dd, J = 8.9, 1.4 Hz, 1H), 7.73 (dd, J = 7.7, 4.9 Hz, 1H), 7.24 (s, 1H), 7.15 (s, 2H), 5.09 (dd, J = 10.7, 4.5 Hz, 1H), 4.11 (s, 3H), 2.60 (ddd, J = 14.6, 4.4, 0.9 Hz, 1H), 2.51–2.45 (m, 1H); 13C-NMR (101 MHz, D2O) δ 182.0, 181.6, 175.3, 171.2, 148.8, 147.7, 146.0, 145.0, 134.8, 131.5, 124.1, 119.8, 115.7, 110.8, 80.1, 56.0, 52.7, 23.4; MS: m/z (M + 1) = 314.10, calculated. 313.09; (% yield – 77%). 4-(4-Hydroxy-3-methoxyphenyl)-6-(pyridin-3-yl) pyrimidin-2(1H)-one (VCU-19) Melting point 173 ± 3; IR (cm−1): 3682–3580 (O–H), 3390, 3288, 3188 (N–H), 2933 (C–H), 1739 (CO), 1616, 1564, 1507 (CC/CN), 1468, 1419 (C–N), 1342, 1267 (C–N), 1201, 1159, 1091, 1024 (C–O/C–N), 931, 842 (Ar–C–H bending), 731 (Ar–H); 1H-NMR (400 MHz, DMSO-D6) δ 9.23 (d, J = 1.7 Hz, 1H), 8.72 (dd, J = 4.7, 1.7 Hz, 1H), 8.35 (dt, J = 7.9, 1.9 Hz, 1H), 7.64 (d, J = 15.0 Hz, 1H), 7.54–7.50 (m, 1H), 7.38–7.50 (1H), 7.32 (t, J = 0.6 Hz, 1H), 7.17–7.14 (m, 1H), 6.51 (s, 1H), 5.35 (s, 2H), 3.76 (s, 3H); 13C-NMR (101 MHz, DMSO-D6) δ 188.1, 153.1, 150.1, 149.6, 148.0, 145.7, 135.8, 133.2, 126.0, 124.7, 123.9, 118.3, 115.6, 111.7, 55.8, 40.1, 39.9, 39.9, 39.7, 39.5, 39.3, 39.2, 39.1, 38.9; MS: m/z (M + 1) = 297, calculated 296.1; (% yield – 65%). 2-Methoxy-4-(3-(pyridin-3-yl)-4,5-dihydroisoxazol-5-yl) phenol (VCHA-19) Melting point 270.28; IR (cm−1): 3456 (O–H), 3050 (Ar–C–H), 2920 (C–H), 1742 (CO), 1681, 1601 (CN/CC), 1574, 1524, 1475 (aromatic CC), 1425 (C–N), 1333, 1316 (C–N), 1244, 1204 (C–O), 1149, 1064, 1034 (C–O/C–N), 920 (Ar–C–H bending); 1H NMR (δ): 2.95 (2H, dd), 3.79 (3H, s), 5.79 (1H, dd), 6.59–6.77 (2H), 7.02 (1H, dd), 7.44 (1H, ddd), 7.73 (1H, ddd), 8.53 (1H, dt), 9.06 (1H, ddd); 13C NMR (δ): 43.4, 56.1, 82.3, 109.2, 115.5, 123.4, 126.9, 128.1, 130.6, 137.8, 146.8, 148.2–148.6, 155.7; MS: m/z (M + 1) = 271.41, calculated. 270.14; (% yield – 73%). In vitro cytotoxicity (MTT assay) The cytotoxicity of the synthesised compounds was evaluated using the MTT assay using PC12 cell lines, cultured as adherent monolayers in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 2% foetal calf serum and antibiotics (penicillin, streptomycin, and kanamycin) at 37 °C in a humidified 5% CO2 environment. Cells were washed with PBS, detached using 0.2% trypsin-EDTA, resuspended in complete medium, and seeded at 100 µL per well in 96-well plates. After 48 hours of proliferation, cells were treated with test compounds and incubated for the time specified. Cell viability was calculated as (absorbance of treated cells/absorbance of control) × 100.19 Intracellular cAMP quantification The intracellular cAMP levels were measured using a luciferase biosensor assay. PC12 cells were seeded and allowed to attach for 24 hours before treatment. Following exposure to the test compounds, luminescence was measured according to the manufacturer's protocol, and cAMP concentrations were estimated using a standard calibration curve. The assays were run in triplicate, and statistical significance was determined by comparing the results to those of the vehicle controls.20 In vivo evaluation in Drosophila melanogaster Transgenic Drosophila melanogaster stains (UAS-GFP and elav-Gal4), procured from the Department of Zoology, Mysore University, were used to develop AD-like progeny using standard genetic crosses. Flies were kept at 25 ± 2 °C and 60–65% relative humidity in media containing 80 g of semolina, 70 g of jaggery, 9 g of agar, and 4.4 mL of propionic acid per litre. Before transfer, 15 g of yeast powder was added. Vials typically held 50–100 adults for maintenance.21 Acute toxicity studies were carried out by exposing cohorts of 120 wild (Oregon-K) flies divided evenly into four dose groups. The flies were fed the test compounds and observed at 0, 24, 48, and 72 hours to determine per cent mortality and estimate the LC50 for toxicity benchmarking purposes. These findings determined the low and high dose ranges for subsequent behavioural testing. A negative geotaxis assay was used to evaluate locomotor performance. Unanaesthetised flies from the Normal, Diseased, Standard, T1, and T2 groups were gently tapped at the base of a 10-cm vertical climbing apparatus. To evaluate locomotor function, we measured the percentage of flies that crossed a 10 cm line within 10 seconds. We ran this test three separate times on days 0, 4, and 7 after treatment and used the results to calculate a climbing height for each group. Comparing these heights allowed us to see whether the treatment improved movement. After seven days of treatment, we also measured acetylcholinesterase (AChE) activity using a spectrophotometric assay. First, we anaesthetised and sacrificed the flies, then homogenised them in phosphate buffer. After centrifugation, we collected the clear supernatant. To determine enzyme activity, we incubated the supernatant with DTNB and an acetylcholine substrate in a 96-well plate, then measured absorbance at 421 nm. Each assay was performed twice, and the results were compared to the control group.22 Results & discussion Rationale for target, reference compound, and scaffold selection Phosphodiesterase-4B (PDE4B) was chosen as a receptor target because of its significant role in regulating neuronal cyclic adenosine monophosphate (cAMP) signalling, synaptic plasticity, and neuroinflammatory processes, which are one of the causes to Alzheimer's disease pathology. Structural examination of the PDE4B catalytic domain (PDB ID: 1XN0) reveals a highly conserved active site dominated by a hydrophobic clamp formed by Phe414 and Phe446, along with a critical hydrogen-bonding interaction mediated by the invariant Gln443 glutamine switch, which plays an essential role in ligand orientation and PDE4 isoform selectivity. Because Donepezil has been shown to have neuroprotective and neurotrophic effects in PC-12 neuronal cell models, including mitigation of Aβ-induced toxicity, reduction of oxidative stress, and enhancement of neurite outgrowth, it was included as a reference compound for biological testing. Despite not directly inhibiting PDE4, Donepezil's capacity to indirectly improve cAMP/PKA/CREB signalling through cholinergic mechanisms offers a useful benchmark for assessing neuroprotective effects. Because of its proven PDE4B potency, exposure to the central nervous system, and favourable therapeutic index in preclinical studies, the clinically evaluated PDE4 inhibitor GSK356278 was included as a structural and pharmacological benchmark. The co-crystal ligand Rolipram, on the other hand, has a limited therapeutic window and poor cell tolerance despite having a well-defined binding mode involving π–π stacking with Phe446 and interaction with Gln443. A conserved overlap of pyridine-containing aromatic regions within the PDE4B Q-pocket was discovered through structure-based superimposition of Donepezil, GSK356278, and Rolipram. The choice of a pyridine-centred scaffold for ligand design was influenced by this observation. In silico studies RMSD-based structural alignment of designed ligands Structural overlay analysis clearly distinguished the reference ligands with design complexes. Rolipram and donepezil showed higher structural deviation, with RMSD values exceeding 3.0 Å. This indicated noticeable differences in both aromatic core orientation and substituent positioning. While Rolipram's co-crystal structure offers useful insights into the PDE4B binding pocket, its small scaffold and limited options for functional modification make it a less ideal template for design. Similarly, though donepezil is CNS-active and biologically relevant, its structural alignment is not optimal for achieving PDE4B-selective binding. In contrast, GSK356278 consistently gave low RMSD values all below 2.0 Å, indicating better spatial alignment and a scaffold well-suited to the PDE4B catalytic pocket. Overlap analysis identified key structural features in this compound: a pyridine core, a six-membered aromatic ring, and a favourably oriented methoxy substituent. When we aligned our designed ligands against GSK356278, RMSD values remained low (maximum ≈1.38 Å). This confirmed that the core scaffold geometry was successfully preserved and validated the GSK compound guided design strategy (Fig. 2). RMSD-based structural alignment of three reference compounds. Pharmacophore analysis In line with the structural and physicochemical requirements of the PDE4B catalytic pocket, a recurrent combination of aromatic rings, hydrogen bond acceptors (HBA), hydrogen bond donors (HBD), and hydrophobic features were seen throughout the models. The pharmacophore of GSK356278 (Fig. 3) showed several aromatic features aligned along the central scaffold among the reference ligands. These features correspond to planar ring systems that support π–π stacking interactions within the hydrophobic clamp of Phe414–Phe446. The Gln443 glutamine switch, a crucial component controlling PDE4B selectivity, is in a good position to interact with the HBA moieties, which are primarily linked to heteroatoms of the pyridine moiety and were oriented towards the Q-pocket. Furthermore, favourable van der Waals interactions with nonpolar residues lining the catalytic cavity are reflected in hydrophobic features at the terminal regions, which support binding stability. In contrast, the Rolipram pharmacophore (Fig. 3) showed a simpler feature pattern dominated by hydrophobic and aromatic elements. This likely reflects its limited hydrogen-bond acceptor/donor (HBA/HBD) representation and efficient anchoring within the binding pocket. Pharmacophore alignment and feature mapping of reference and designed PDE4B ligands. Green indicates a hydrogen bond donor (HBD) region, pink/magenta indicates a hydrogen bond acceptor (HBA) region, orange represents an aromatic ring feature, and cyan/blue highlights a hydrophobic feature important for nonpolar interactions within the binding site. The Donepezil pharmacophore (Fig. 3), on the other hand, displayed aromatic features that aid CNS penetration, along with distinct HBA and HBD groups that enable polar interactions and contribute to neuroprotective activity. Even though Donepezil itself doesn't directly inhibit PDE4B, its partial spatial overlay with GSK356278 underscores the value of Donepezil-like motifs in designing CNS-targeted ligands. When compared to the reference compounds, pharmacophore analysis of the designed ligands (VCDNP-19, VCHA-19, VCHH-19, VCPH-19, VCT-19, and VCU-19) showed a strong conservation of important features (Fig. 3). The aromatic cores consistently aligned spatially with those of Rolipram and GSK356278 throughout the series, suggesting that the π-interaction framework necessary for PDE4B recognition was preserved. The pharmacophoric patterns seen in the reference ligands were closely reflected in the designed molecules, which showed a balanced arrangement of hydrogen bond acceptor (HBA) sites close to heterocyclic regions and hydrophobic features at peripheral positions. Interestingly, the designed series' aromatic characteristics closely matched those of GSK356278, suggesting that the Phe414–Phe446 hydrophobic clamp may have π–π stacking interactions. Simultaneously, HBD and HBA features were positioned to interact similarly to the reference inhibitor with the Gln443 selectivity region. Similar to GSK356278, compounds like VCPH-19 and VCHH-19 showed a dense clustering of hydrophobic and aromatic features, indicating successful retention of the core binding framework. HBA/HBD features that extend into the Q1 and Q2 subpockets were introduced by additional peripheral modifications, allowing for increased interaction versatility. Overall, the successful transfer of binding-relevant motifs from validated PDE4B inhibitors into the newly designed scaffolds is confirmed by the observed correspondence between reference and designed pharmacophore features. Molecular docking To evaluate biologically relevant binding orientations within the PDE4B catalytic site, reference-guided pose selection was used during molecular docking. All thirty conformations were docked and assessed against reference molecules (Table 1). The co-crystal ligand Rolipram, has successfully replicated its experimentally observed binding pose and produced a Glide SP score of −7.367 kcal mol−1, and was used to first validate the docking protocol. In accordance with the crystallographic binding mode and proper placement in relation to the catalytic metal region, the docked orientation maintained important interactions, such as direct hydrogen bonds within 3.0 Å and π–π stacking interactions within 6.0 Å. With a Glide score of −5.849 kcal mol−1, the clinically tested inhibitor GSK356278 docked with a stable orientation marked by π–π stacking interactions within 6.0 Å and proper positioning of its aromatic scaffold within the Phe414–Phe446 hydrophobic clamp. With a docking score of −6.425 kcal mol−1 and a direct hydrogen bond with Gln443 (≤3.0 Å), the standard compound Donepezil demonstrated partial engagement with residues that determine PDE4B selectivity (Fig. 4). Table 1Molecular docking data of the shortlisted and reference moleculesCompoundDocking score (kcal mol−1)Hydrogen bond interactionsπ–π/Hydrophobic interactionsRemarkRolipram (co-crystal)−7.367Gln443, Thr407Phe414, Phe446, Trp406Reference for binding pose alignmentGSK356278−5.849Tyr233, Phe414Phe446, Ile410Experimental comparatorDonepezil−6.425Gln443Phe414, Phe446Good for scoring comparisonVCHH-19−6.951Gln443, Thr407, water-mediatedPhe414, Trp406, Ile410Better key interactionsVCT-19−6.650Gln443Phe414, Met347, Trp406Better key interactionsVCHA-19−6.362Gln443Phe414, Phe446Moderate key interactionsVCPH-19−6.256Water-mediated (Asp392)Phe446, Leu393Moderate key interactionsVCU-19−5.847Weak/noneHydrophobic contacts onlyWeak key interactionsVCDNP-19−5.359NonePeripheral hydrophobicWeak key interactions Molecular docking data of the shortlisted and reference molecules The 3D and 2D binding interactions of reference ligands in the PDE4B active site (PDB: 1XN0). Showing specific amino acid interaction with PHE 446 and GLN 443. (a) Rolipram in green (b) GSK356278 in pink (c) donepezil in cyan. For the designed compounds, each generated conformation for the designed compounds was assessed separately, and the most representative pose was chosen based on how closely it resembled the GSK356278 binding mode. Among the criteria used to choose the pose were favourable accommodation within the hydrophobic pocket, preservation of significant hydrogen-bond donor and acceptor interactions within allowable distance thresholds (≤3.0 Å), and similar orientation of the pyridine and aromatic rings. Only poses that replicated interaction patterns consistent with the reference ligands were retained, and the corresponding docking scores were used for comparative evaluation. All of the shortlisted compounds adopted binding orientations that were nearly identical to the reference ligands by positioning their aromatic cores inside the Phe414–Phe446 hydrophobic clamp and orienting heteroatoms towards the Gln443 selectivity region. Docking scores for the designed series ranged from −6.951 to −5.359 kcal mol−1. For VCHH-19 and VCT-19, the most favourable binding scores were −6.951 kcal mol−1 and −6.650 kcal mol−1, respectively. Both compounds maintained conserved π–π stacking interactions within 6.0 Å and hydrogen-bond contacts within 3.0 Å, similar to the interaction pattern of the co-crystal ligand. While retaining a partial hydrophobic engagement within the active site, VCPH-19 (−6.256 kcal mol−1) and VCHA-19 (−6.362 kcal mol−1) demonstrated a moderate binding affinity. On the other hand, VCU-19 (−5.847 kcal mol−1) and VCDNP-19 (−5.359 kcal mol−1) showed weaker binding, which is in line with the absence of close hydrogen-bonding interactions and π–π stacking, indicating reduced polar and aromatic complementarity within the PDE4B catalytic pocket (Fig. 5). VCPH-19 2D and 3D binding interactions inside PDE4B's active site (PDB: 1XN0). Showing specific amino acid interaction with PHE 446 and GLN 443. Molecular dynamics The conserved Gln443 glutamine switch largely controls PDE4B selectivity, whereas Phe414 and Phe446 form a hydrophobic clamp that stabilises aromatic heterocycles through π–π interactions. The persistence of ligand contacts with Gln443 and Phe446 was thus the main focus of molecular dynamics (MD) analyses, especially for ligands with pyridine-based scaffolds (Table 2). Table 2Timescale stability data of the shortlisted and reference moleculesLigandTime (ns)Gln443 (%)Phe446 (%)Binding characterGSK35627819967Direct + π–πGSK356278106457StableGSK35627850None38π–πGSK356278100None47StableGSK356278200None61Highly stableCo-crystal108953Water-mediatedDonepezil10047PoorVCPH-19105766Direct + π–πVCPH-19502257StableVCPH-191001039ModerateVCPH-19200None25Stable Timescale stability data of the shortlisted and reference molecules Short-time MD screening (1 ns and 10 ns) An extremely stable early binding profile was shown by the reference inhibitor GSK356278. Strong interactions with Gln443 (99%) and Phe446 (67%) were seen at 1 ns, successfully anchoring the pyridine moiety inside the active site. At 10 ns, these interactions were mostly maintained (Gln443: 64%; Phe446: 57%), suggesting quick stabilisation inside the PDE4B catalytic pocket. At 1 ns, the co-crystal ligand rolipram made strong contacts with Gln443 (95%) and Phe446 (53%). By 10 ns, however, the interaction with Gln443 was primarily water-mediated (89%), indicating indirect stabilisation as opposed to direct anchoring. With a brief Phe446 contact at 1 ns and a total loss of Gln443 interaction by 10 ns, the standard compound donepezil exhibited weak and irregular binding, which is consistent with low PDE4B selectivity. VCPH-19 showed the most promising early behaviour among the designed compounds. Effective engagement of the pyridine scaffold within the hydrophobic clamp was demonstrated by the significant improvement of moderate interactions with Phe446 (29%) and Gln443 (28%) at 1 ns by 10 ns (Phe446: 66%; Gln443: 57%). The profiles of the other analogues were worse: VCHH-19 maintained moderate Gln443 contact but weak Phe446 interaction; VCU-19 primarily engaged Gln443 through water mediation; VCT-19 was unable to maintain interactions with either residue; VCHA-19 demonstrated strong Phe446 contact without Gln443 engagement; and VCDNP-19 displayed weak and inconsistent interactions. Only GSK356278, rolipram, donepezil, and VCPH-19 were moved on to longer simulations based on these findings. Intermediate MD evaluation (50 ns) GSK356278 sustained stable aromatic anchoring through Phe446 (38%) at 50 ns. In addition to direct Phe446 contact (52%), rolipram continued to show strong interaction with Gln443 (89%), primarily water-mediated. Donepezil demonstrated poor stabilisation, with no discernible engagement of Gln443 and a reduction in Phe446 interaction to 11%. VCPH-19, on the other hand, maintained balanced interactions and sustained positioning within the PDE4B active site by remaining in contact with Phe446 (57%) and Gln443 (22%). As a result, for long-timescale simulations, only GSK356278 and VCPH-19 were chosen. Long-timescale stability (100 ns and 200 ns) GSK356278 demonstrated remarkable stability during prolonged simulations, with Phe446 interaction rising from 47% at 100 ns to 61% at 200 ns, indicating strong long-term pyridine ring anchoring. Additionally, VCPH-19 showed dual interactions at 100 ns (Phe446: 39%; Gln443: 10%) and sustained stability. Through Phe446 interaction, VCPH-19 maintained significant hydrophobic stabilisation at 200 ns (25%). Both complexes had minimal conformational drift and limited flexibility within the active site, according to RMSD and RMSF analyses. The 200 ns simulations as a whole show that long-term stability is primarily determined by persistent Phe446 engagement, and they identify VCPH-19 as the most promising designed analogue for additional optimisation (Fig. 6). Molecular dynamics analysis of PDE4B (1XN0) protein–ligand complexes (GSK356278 & VCPH-19). (A) RMSD plots showing overall stability of the complexes during simulation. (B) RMSF plots indicating residue-wise flexibility of the protein. (C) Ligand RMSF plots showing atomic fluctuations of each ligand. (D) Interaction profiles summarizing key contacts formed throughout the simulation. (E) 2D interaction diagrams highlighting major ligand–protein binding interactions. ADMET and CNS multiparameter optimization (MPO) assessment The central nervous system drug-likeness of the designed compounds was evaluated using ADMET and CNS multiparameter optimisation (MPO) analyses, using GSK356278 as a verified reference ligand. The reliability of the prediction framework was supported by the close agreement between the predicted physicochemical properties of GSK356278 and published experimental data. The CNS-MPO score framework calculations, were as per Pfizer Worldwide Research and Development. The majority of the designed molecules had polar surface area (PSA) values within the desired range for CNS penetration (<90 Å2) and molecular weights below 350 Da. With log P/log D values in the range of roughly 1.2–3.2, compounds VCHH-19, VCT-19, VCHA-19, and VCPH-19 demonstrated balanced lipophilicity and obtained favourable CNS-MPO scores (4.2–4.5) that were comparable to the reference ligands. VCU-19 and VCDNP-19, on the other hand, displayed lower MPO scores and higher PSA values, indicating less suitability for CNS exposure. Overall, the CNS compatibility of some pyridine-based derivatives, especially VCPH-19 and VCHH-19, is supported by the ADMET and MPO profiles (Table 3). Table 3ADMET and CNS multiparameter optimization (MPO) assessment data of the selected and reference moleculesCompoundMW (Da)PSA (Å2)HBAHBDLog PLog DpKaCNS MPOGSK356278(exp.)439.5<9071NANA7.20NAGSK356278(pre.)439.53495.995712.032.217.164.5Rolipram275.164.63412.342.3010.234.3Donepezil439.595.99812.032.217.164.6VCHH-19269.373.46322.112.438.814.5VCT-19313.374.22411.251.598.664.2VCHA-19270.354.45412.182.398.714.5VCPH-19345.454.45312.893.178.644.4VCU-19299.395.83410.911.108.694.0VCDNP-19433.4143.69513.053.218.603.2 ADMET and CNS multiparameter optimization (MPO) assessment data of the selected and reference molecules Structural characterization The corresponding Scheme's reaction sequences were followed in the synthesis of the pyridine-based heterocyclic derivatives. Melting point determination and thin-layer chromatography (TLC) using silica gel-G plates with n-hexane : ethyl acetate (3 : 1, v/v) as the mobile phase were used to first evaluate the purity of the obtained compounds. When necessary, additional purification was accomplished by recrystallisation from ethanol or rectified spirit. A variety of spectroscopic methods, such as 1H NMR, 13C NMR, FT-IR, and LC-MS, were used to confirm the structures of the synthesised compounds (VC-19, VCHH-19, VCPH-19, VCT-19, VCU-19, VCDNP-19, and VCHA-19). The presence of the anticipated functional groups and heterocyclic frameworks was confirmed by the spectral data, which were in agreement with the suggested molecular structures. In vitro cytotoxicity evaluation (MTT assay) Following a 24-hour exposure, the cytotoxicity profiles of the produced heterocyclic analogues were assessed in PC12 neuronal cells. At the lowest tested concentration (25–50 µg mL−1), all compounds-maintained viability responses of more than 93%, demonstrating high initial biocompatibility. At higher doses, a slow, dose-dependent decrease in mitochondrial activity was seen, which is consistent with ISO-10993-5 MTT cytotoxicity. Table 4 provides a summary of the entire viability profile. Table 4Percent cell viability of synthesized analogues in PC12 cells (MTT assay)Compound25 µg50 µg100 µg150 µg200 µg250 µg300 µgIC50 (µg mL−1)VCPH-1994.4883.3572.3962.3050.8634.2720.04200VCHH-1993.4983.6171.8153.1341.5733.2017.63150VC-1994.0083.3374.9253.0036.9224.6617.16150VCDNP-19NANA71.5764.0849.6728.4213.76200VCU-19NANA65.0852.4339.5130.9619.84150VCT-19NANA74.7466.2951.5446.9535.89200Donepezil96.892.186.579.462.253.545.1250 Percent cell viability of synthesized analogues in PC12 cells (MTT assay) With cell viability continuously maintained at >70% up to 100 µg mL−1 (Table 4), all tested compounds demonstrated favourable neuronal compatibility at lower concentrations, indicating minimal basal cytotoxicity. Similar dose–response profiles were shown by the chalcone precursor VC-19, the pyrazole derivative VCHH-19, and VCU-19, each of which had an IC50 of 150 µg mL−1, indicating moderate cytotoxicity only at higher concentrations. On the other hand, with IC50 values of 200 µg mL−1, VCPH-19, VCDNP-19, and VCT-19 showed enhanced cytocompatibility. Given its proven neuronal safety, donepezil, the reference compound, demonstrated the highest IC50 (250 µg mL−1). Cell viability ranged from 13% to 36% at the highest concentration tested, consistent with the concentration-dependent decrease in mitochondrial activity typical of MTT assays. Interestingly, microscopic inspection showed no signs of acute cytotoxicity or morphological abnormalities. When we combined computational results with biological data, two compounds VCPH-19 and VCHH-19 emerged as the most promising candidates. At lower concentrations, both compounds exhibited clear dose-dependent responses and high cellular tolerance, maintaining >93% viability at 25 µg mL−1. Based on ADMET predictions, VCPH-19 showed the best safety profile, with an IC50 of 200 µg mL−1. This supports the idea that phenyl substitution improves cellular compatibility. VCHH-19, although it had a lower IC50, still maintained acceptable viability up to 100 µg mL−1, indicating sufficient neuronal safety. Taken together, these results point to VCPH-19 as the primary lead compound, with VCHH-19 serving as a structurally instructive supporting analogue for further PDE4B-focused research. Intracellular cAMP assay Following a 24-hours treatment with VCPH-19, VCHH-19, and Donepezil (standard control), intracellular cAMP levels in PC-12 cells were measured using a luciferase-based biosensor assay in compliance with ISO-10993-5 cytocompatibility guidelines. The basal luminescence of untreated cells was 0.243 RLU, which served as the fold-change analysis reference. Intracellular cAMP levels were clearly elevated by both synthesised compounds in a concentration-dependent manner (Fig. 7), which is consistent with inhibition of PDE4B-mediated cAMP degradation. Luminescence was gradually increased by VCPH-19 from 0.271 RLU (100 µg) to 0.791 RLU (300 µg), which is a 3.25-fold increase over basal levels. Among the synthesised analogues, VCHH-19 produced the strongest response, rising from 0.347 RLU (50 µg) to 0.866 RLU (250 µg), a 3.56-fold enhancement. Donepezil validated assay sensitivity by demonstrating a similar dose-dependent response, reaching 0.894 RLU. Overall, VCPH-19 demonstrated strong but marginally reduced activation, in line with docking and molecular dynamics predictions, whereas VCHH-19 showed cAMP elevation similar to donepezil. These results validate that both compounds effectively inhibit PDE4B and encourage further testing of these compounds in models of neurodegenerative diseases. Effect of VCPH-19, VCHH-19, and standard donepezil on concentration-dependent modulation of intracellular cAMP levels in PC12 cells measured using a luciferase-based biosensor assay, indicating dose-dependent elevation of intracellular cAMP levels. Data are expressed as mean ± SEM (n = 3). Statistical analysis was performed using two-way ANOVA followed by Tukey's multiple comparison test. SEM values are depicted in graph. In vivo evaluation of AChE activity and behavioral restoration in Drosophila melanogaster During the 0th, 4th and 7th day of the treatment, locomotor performance (negative geotaxis) and after 7th day of the treatment acetylcholinesterase (AChE) activity were used to assess the therapeutic efficacy of VCPH-19 in an in vivo Alzheimer's disease model in transgenic Drosophila melanogaster Previous acute toxicity evaluations were used to determine the dose levels for the treatment groups (T1: low dose; T2: high dose). Negative geotaxis (locomotor performance): to track behavioural recovery, locomotor function was evaluated on Days 0, 4, and 7 (Fig. 8). Over the course of the study, disease-controlled flies showed a progressive decline in their locomotor function. On the other hand, climbing ability significantly improved in both VCPH-19 treatment groups. In comparison to the diseased controls, T1 and T2 showed a distinct recovery by Day 4, suggesting early functional rescue. By Day 7, T2-treated flies were getting close to the donepezil standard's performance, whereas T1 showed a steady but modest improvement. Effective mitigation of AD-like motor deficits is supported by the time- and dose-dependent recovery. Behavioural assessment of treatment effects across experimental groups. Protective effects of VCPH-19 against the transgenic AD model of Drosophila melanogaster (locomotor activity). Values are presented as, mean ± SEM, (n = 3), Two-way ANOVA was used with Tukey's multiple comparison. ###p < 0.001, significant when compare to normal group, *p < 0.05 significant when compare to treatment 1 and 2 treated groups, **p < 0.01, significant when compare to standard control group, ***p < 0.001 significant when compare to treatment 1, 2 and standard treated groups. Acetylcholinesterase activity: On Day 7 (Fig. 9), diseased flies had significantly higher levels of AChE activity (0.553 µmol per min per mg protein) in comparison with normal group, which is consistent with cholinergic dysfunction. Whereas treatment VCPH-19 (T2) showed a significant reduction in with AChE level, which was very similar to the donepezil treated group (0.379 µmol per min per mg protein). Improved locomotor performance is correlated with the restoration of enzymatic activity, indicating that VCPH-19 modulates cholinergic signalling by inhibiting PDE4B. Biochemical assessment of treatment effects across experimental groups. Protective effect of VCPH-19 in transgenic AD model on total Acetylcholinesterase level in transgenic Drosophila melanogaster flies. All values are expressed as mean ± SEM, (n = 3) ###p < 0.001, compared to normal control; ***p < 0.001,***p < 0.001, as compared with Disease control flies. (One-way ANOVA followed by Tukey's multiple comparison tests at significance level p < 0.05 confidence interval). Conclusion In this study, a pharmacophore-guided design strategy was employed to identify pyridine-based heterocyclic scaffolds as selective PDE4B inhibitors for Alzheimer's disease. A validated five-feature pharmacophore model enabled efficient virtual screening and scaffold selection, leading to the synthesis of seven representative analogues. Docking against the PDE4B catalytic domain (1XN0) revealed strong binding affinities for the pyrazole derivatives VCPH-19 and VCHH-19, supported by key interactions with Gln443 and Phe446. Molecular dynamics simulations (200 ns) confirmed stable ligand–protein complexes with low RMSD values and sustained hydrogen-bond occupancy. All synthesized compounds exhibited acceptable neuro-cytocompatibility (ISO-10993-5), with VCPH-19 and VCHH-19 producing significant intracellular cAMP elevation. In vivo studies further showed that treatment with VCPH-19 improved locomotor behaviour and partially restored AChE activity. PDE4B inhibition elevates intracellular cAMP levels, leading to activation of the PKA/CREB signalling cascade. Since these pathological processes contribute to cholinergic dysfunction a major pathological hallmark in Alzheimer's disease and elevated AChE activity, PDE4B inhibition may indirectly restore cholinergic neurotransmission by normalizing AChE activity and enhancing acetylcholine availability. Overall, VCPH-19, closely followed by VCHH-19, emerges as a promising scaffold for further optimization targeting PDE4B. Author contributions Karthik G. Pujar conceived and designed the study, performed the computational investigations, analysed the data, and drafted the manuscript. Richa Prabhakar Shetty assisted with computational analysis. Parthasarathi carried out spectral characterization. Krishna Yallappa Kolachi and Aishwarya Bharathi H. M. assisted in manuscript preparation and data organization. Prabitha P. supervised the work, critically reviewed the manuscript, and served as the corresponding author. All authors read and approved the final manuscript. Pooja Gandharvachari Achar assisted in the statistical analysis. Conflicts of interest The authors declare that there are no conflicts of interest regarding the publication of this manuscript. Abbreviations Alzheimer's disease Amyloid-beta Acetylcholinesterase Absorption, distribution, metabolism, excretion and toxicity Brain-derived neurotrophic factor Cyclic adenosine monophosphate Cyclic guanosine monophosphate Central nervous system cAMP response element-binding protein Density functional theory Dulbecco's modified eagle medium Embryonic lethal abnormal vision – GAL4 Ionization and tautomer prediction tool (Schrödinger) Fourier transform infrared spectroscopy Standard precision/extra precision docking (Schrödinger) Hydrogen bond acceptor Hydrogen bond donor Lethal concentration Liquid chromatography-mass spectrometry Distribution coefficient Partition coefficient Molecular dynamics MultiParameter optimization 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide Nuclear magnetic resonance Constant number of particles, pressure, and temperature ensemble Optimized potentials for liquid simulations 4 (force field) Pheochromocytoma-12 Protein data bank Phosphodiesterase-4 Phosphodiesterase-4B Protein kinase A Particle Mesh Ewald Polar surface area Research collaboratory for structural bioinformatics Radius of gyration Relative light units Root mean square deviation Root mean square fluctuation Structure–activity relationship Simulation interaction diagram Thin-layer chromatography Three-site transferable intermolecular potential water model Upstream activating sequence – green fluorescent protein Microgram Supplementary Material