Design and Synthesis of Caffeine-Based Derivatives with Antioxidant and Neuroprotective Activity: In Vitro Evaluation and SwissADME Profiling

TL;DR

Researchers designed new caffeine derivatives that show strong antioxidant and neuroprotective effects in lab tests, potentially useful for treating neurodegenerative diseases.

Contribution

The study introduces novel caffeine-based compounds with enhanced antioxidant and neuroprotective activity compared to caffeine and standard antioxidants.

Findings

Several caffeine derivatives showed higher antioxidant activity than caffeine and Trolox at low concentrations.

Compound AL-7 provided up to 85% protection against glutamate-induced excitotoxicity in human neuroblastoma cells.

SwissADME analysis predicted good oral bioavailability but limited blood–brain barrier permeability for the derivatives.

Abstract

Oxidative stress and excitotoxicity are key contributors to neuronal damage in various neurodegenerative diseases. Caffeine, a widely used neuroactive compound with moderate antioxidant properties, may benefit from structural modifications to enhance its neuroprotective potential. In this study, a series of novel caffeine derivatives was synthesized and evaluated for antioxidant and potential neuroprotective relevance using in vitro models of oxidative stress and glutamate-induced excitotoxicity in SH-SY5Y human neuroblastoma cells. Antioxidant capacity was assessed using ABTS•+ radical cation decolorization and DPPH radical scavenging assays. Most derivatives exhibited strong free radical scavenging activity, surpassing both caffeine and the reference antioxidant Trolox at low concentrations (5 µM). Notably, compounds AL-7, AL-8, AL-9, and AL-10 demonstrated particularly high activity. Cytotoxicity evaluation using the MTT assay revealed low toxicity for all compounds, with calculated IC50 values above 500 µM. Intracellular reactive oxygen species (ROS) levels measured by the DCFH-DA assay showed that several derivatives, especially AL-4, significantly reduced H2O2-induced oxidative stress. In neuroprotection assays, compounds AL-0, AL-1, and AL-4 markedly protected against hydrogen peroxide-induced damage, restoring cell viability up to 73%, while AL-7 achieved up to 85% protection against L-glutamate-induced excitotoxicity, outperforming caffeine. In silico SwissADME analysis indicated favorable oral bioavailability, with predicted gastrointestinal absorption and limited blood–brain barrier permeability. Overall, these findings highlight structurally modified caffeine derivatives as promising antioxidant and neuroprotective agents warranting further mechanistic and therapeutic investigation.

Figures

Tables

Table 6. Drug likeness of the compounds AL0–AL10.

Molecule	Lipinski	Ghose	Veber	Egan	Muegge	Bioavailability Score
AL0	0	No; 1 violation: WLOGP < −0.4)	No; 1 violation: TPSA > 140	No; 1 violation: TPSA > 131.6	0	0.55
AL1	0	0	0	0	0	0.55
AL2	0	No; 1 violation: MR > 130	0	0	0	0.55
AL3	0	0	No; 1 violation: TPSA > 140	No; 1 violation: TPSA > 131.6	No; 1 violation: TPSA > 150	0.55
AL4	0	0	No; 1 violation: TPSA > 140	No; 1 violation: TPSA > 131.6	No; 1 violation: TPSA > 150	0.55
AL5	0	0	0	0	0	0.55
AL6	0	No; 1 violation: MW > 480	0	0	0	0.55
AL7	0	No; 1 violation: MW > 480	0	0	0	0.55
AL8	Yes; 1 violation: NorO > 10	0	No; 1 violation: TPSA > 140	No; 1 violation: TPSA > 131.6	No; 1 violation: TPSA > 150	0.55
AL9	No; 2 violations: MW > 500, NorO > 10	No; 2 violations: MW > 480, MR > 130	No; 1 violation: TPSA > 140	No; 1 violation: TPSA > 131.6	No; 1 violation: TPSA > 150	0.17
AL10	0	0	0	0	0	0.55

Full text

1. Introduction

Neurodegenerative diseases are a heterogeneous group of progressive disorders of the central nervous system (CNS), characterized by selective neuronal loss and dysfunction of specific brain regions [1]. While Alzheimer’s disease and Parkinson’s disease are the most well-known and extensively studied forms of neurodegeneration, several other clinically significant conditions—such as amyotrophic lateral sclerosis (ALS), multiple system atrophy (MSA), spinocerebellar ataxia (SCA), frontotemporal dementia (FTD), Huntington’s disease (HD), and various mitochondrial encephalopathies—also impose a substantial societal and medical burden [2]. A common hallmark among these diseases is impaired cellular homeostasis, prominently featuring elevated oxidative stress and compromised antioxidant defense mechanisms [3]. Oxidative stress results from an imbalance between the generation of reactive oxygen species (ROS) and the capacity of endogenous antioxidant systems to neutralize them. In the CNS, this imbalance is particularly detrimental due to the brain’s high oxygen consumption, its abundance of polyunsaturated fatty acids, and relatively low levels of antioxidant enzymes [4]. Excessive ROS production leads to oxidative damage of lipids, proteins, and nucleic acids, ultimately promoting neuronal dysfunction and apoptosis. In many neurodegenerative conditions beyond Alzheimer’s and Parkinson’s diseases, distinct oxidative alterations have been documented [5]. For instance, in ALS, elevated superoxide radical activity and accumulation of lipid peroxides have been observed [6]. In Huntington’s disease, mitochondrial dysfunction, increased hydrogen peroxide (H_2_O_2_) production, and reduced activity of antioxidant enzymes such as glutathione peroxidase and catalase are well established [3]. Similar oxidative mechanisms are evident in SCA, where ROS-mediated damage contributes to cerebellar degeneration and progressive motor impairment [7].

To investigate the molecular mechanisms underlying oxidative stress and to evaluate the protective potential of various compounds, in vitro models of chemically induced oxidative damage are frequently employed. Among the most widely used agents are hydrogen peroxide (H_2_O_2_) and L-glutamate [8,9]. Hydrogen peroxide itself is not a reactive oxygen species but is a relatively stable, membrane-permeable molecule that can readily diffuse across cell membranes. In the presence of transition metal ions (e.g., Fe^2+^), H_2_O_2_ participates in Fenton-type reactions, leading to the generation of highly reactive hydroxyl radicals (•OH), which are responsible for oxidative cellular damage [10]. Exposure to H_2_O_2_ in neuronal cell cultures leads to mitochondrial dysfunction, altered expression of redox-sensitive genes, and apoptotic cell death. Furthermore, L-glutamate—the principal excitatory neurotransmitter in the CNS—exerts toxic effects at elevated concentrations through a process known as excitotoxicity [11]. This process involves overstimulation of N-methyl-D-aspartate (NMDA) and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors, excessive calcium influx, and subsequent mitochondrial dysfunction [12]. In addition, high extracellular glutamate inhibits the cysteine/glutamate antiporter, leading to depletion of intracellular cysteine and reduced glutathione synthesis, thereby weakening cellular antioxidant capacity and increasing susceptibility to oxidative damage [13].

In recent years, certain natural compounds and their derivatives have attracted considerable interest as potential therapeutic agents for neurodegenerative diseases [14]. Among them, caffeine—a natural purine alkaloid widely consumed worldwide—is well known for its stimulatory effects on the CNS [15]. Beyond its psychostimulant properties, caffeine exhibits diverse biochemical activities relevant to oxidative stress and neurodegeneration [16]. Notably, caffeine displays a dual role, acting either as an antioxidant or a pro-oxidant depending on factors such as concentration, cellular context, and interactions with other compounds [17]. As an antioxidant, caffeine can scavenge ROS, stabilize mitochondrial membranes, and enhance intracellular glutathione levels [18]. Its major metabolites, including paraxanthine and theophylline, also contribute to these protective effects [19,20]. In addition, caffeine inhibits phosphodiesterase activity and antagonizes adenosine receptors, mechanisms that may exert antioxidant effects in cellular models relevant to neurodegenerative conditions through modulation of calcium signaling and suppression of neuroinflammation [21]. Conversely, under certain conditions—particularly at high concentrations—caffeine may promote ROS formation or interact unfavorably with metal ions, thereby exerting pro-oxidant effects. Chemical modification of the caffeine scaffold offers opportunities to generate novel derivatives with enhanced bioactivity, improved selectivity, and more favorable pharmacokinetic profiles. Structural modifications can influence properties such as lipophilicity, blood–brain barrier permeability, and affinity for specific molecular targets. Recent studies have demonstrated that selected caffeine derivatives exhibit superior antioxidant activity compared to the parent compound while maintaining low cytotoxicity [22], highlighting their potential as neuroprotective agents in disorders where oxidative stress and mitochondrial dysfunction play central roles.

The present study aims to synthesize and characterize novel caffeine derivatives and to evaluate their antioxidant and neuroprotective potential using in vitro models of oxidative stress and excitotoxicity induced by H_2_O_2_ and L-glutamate, two widely applied experimental paradigms that recapitulate key aspects of neurodegenerative pathology [23,24].

2. Materials and Methods

2.1. Chemistry

2.1.1. Chemicals and Reagents

All chemicals and solvents were of synthetic grade and purchased from Acros Organics (Madrid, Spain). Thin-layer chromatography (TLC) was performed on Kieselgel 60 F254 plates (Merck, Darmstadt, Germany). The mobile phase consisted of 25% NH_4_OH:acetone:CHCl_3_:n-butanol (1:3:3:4, v/v). TLC was used to monitor reaction progress and to assess the purity of the synthesized compounds.

2.1.2. Physical Measurements

Melting points were determined using a Büchi 535 electrothermal apparatus (Büchi Labortechnik AG, Flawil, Switzerland) and are reported uncorrected. UV spectra were recorded on a Jenway 6715 UV/VIS spectrophotometer (Jenway, Stone, UK). Infrared (IR) spectra were obtained with a Nicolet iS10 FT-IR spectrometer equipped with a Smart iTR accessory (Thermo Scientific, Waltham, MA, USA).

2.1.3. NMR Spectroscopy

^1^H and ^13^C NMR spectra were recorded in DMSO-d_6_ using a Bruker 600 WM NMR spectrometer (Bruker BioSpin GmbH, Rheinstetten, Germany) operating at 600 MHz and 75 MHz, respectively. Tetramethylsilane (TMS) was used as an internal standard. Chemical shifts (δ) are reported in parts per million (ppm).

2.1.4. Elemental Analysis

Elemental analyses (C, H, N) were performed on a EuroEA3000 Single Analyzer (EuroVector S.p.A., Pavia, Italy). Obtained values were within ±0.4% of theoretical predictions.

2.1.5. General Synthesis Methodology and Nomenclature

Chemical names were generated using the structure-to-name algorithm included in ChemBioDraw Ultra 11.0 (CambridgeSoft, Cambridge, MA, USA).

General procedure for syntheses of hydrazide-hydrazones of 3-[(1,3,7-trimethyl-2,6- dioxo-2,3,6,7-tetrahydro-1H-purin-8-yl)thio]-propanoic acid (AL1–AL10)

To a solution of AL0 (0.00064 mol) in 5 mL of hot ethanol, the corresponding ketone (0.00064 mol) was added. The reaction mixture was stirred under reflux on an electromagnetic stirrer until the starting hydrazide was exhausted (TLC control). Then the solvent was removed under reduced pressure and the dry residue was recrystallized from ethanol/water (1:1). The corresponding yields and reaction times of the purified products are presented in Table 1.

Synthesis of N′-(1-(p-tolyl)ethylidene)-3-((1,3,7-trimethyl-2,6-dioxo-2,3,6,7-tetrahydro-1H-purin-8-yl)thio)propanehydrazide (AL1).

M.p.: 272 °C (dec.). FTIR (ATR, cm^−1^): 3202 (νNH); 1708 (νCO—xanthine); 1679 (νCO—xanthine); 1651 with shoulder at 1663 (νCO—amide I); 1602, 1537 (νC=N, νC=C—xanthine, δNH—amide II). 1H NMR (DMSO-d6) δ, ppm: 7.80 (d, 2H, 2-Ph, 6-Ph, J = 1.8 Hz), 7.25 (d, 2H, 3-Ph, 5-Ph, J = 1.8 Hz), 3.67 (s,3H, N7-CH3), 3.37–3.53 (m,8H, N1-CH3, N3-CH3, S-CH2), 3.19 (t, 2H, CH2, J = 8.0 Hz), 2.35 (s, 3H, 4-Ph-CH3), 2.25 (s, 3H,N=C-CH3). 13C NMR (75 MHz, DMSO-d6) δ, ppm: 168.9 (-CO), 154.9 (C6=O), 151.4 (C2=O), 159.0 (C8), 150.2 (C4), 156.9 (−C=), 138.1 (C1arom), 127.6 (C2, C6arom), 128.4 (C3, C5arom), 142.8 (C4arom), 110.7 (C5), 34.5 (−CH2−), 32.4 (N7-CH3), 32.0 (−CH2−S), 29.7 (N3-CH3), 28.1 (N1-CH3), 15 (CH3−CH=). For C20H24N6O3S (Mm = 428.51) calculated: C 56.05% H 5.65% N 19.61% S 7.48%; found: C 56.01%, H 5.24%, N 19.52%, S 7.42%.

Synthesis of N′-(1-(4-isobutylphenyl)ethylidene)-3-((1,3,7-trimethyl-2,6-dioxo-2,3,6,7-tetrahydro-1H-purin-8-yl)thio)propanehydrazide (AL2).

M.p.: 230 °C (dec.). FTIR (ATR, cm^−1^): 3205 (νNH); 1708 (νCO—xanthine); 1648 with shoulders at 1668 and 1651 (νCO—xanthine, νCO—amide I); 1604, 1532 (νC=N, νC=C—xanthine, δNH—amide II). 1H NMR (DMSO-d6) δ, ppm: 7.68 (d, 2H, 2-Ph, 6-Ph, J = 1.8 Hz), 7.17 (d, 2H, 3-Ph, 5-Ph, J = 1.8 Hz), 3.70 (s,3H, N7-CH3), 3.50–3.65 (m,8H, N1-CH3, N3-CH3, S-CH2), 3.20 (t, 2H, CH2, J = 8.0 Hz), 2.21 (s, 3H,N=C-CH3), 1.88 (s, 2H, 4-Ph-CH2-), 0.87 (m, 7H, -CH(CH3)2). 13C NMR (75 MHz, DMSO-d6) δ, ppm: 169.1 (-CO), 155.1 (C6=O), 151.4 (C2=O), 159.0 (C8), 150.3 (C4), 156.9 (−C=), 138.1 (C1arom), 130.4 (C2, C6arom), 129.3 (C3, C5arom), 147.4 (C4arom), 110.7 (C5), 45.4 (Ar-CH2),34.5 (−CH2−), 32.4 (N7-CH3), 32.3 (−CH2−S), 30.1 (CH(CH3)2), 29.5 (N3-CH3), 28.1 (N1-CH3), 22.4 (CH(CH3)2), 14.9 (CH3−CH=). For C23H30N6O3S (Mm = 470.59) calculated: C 58.70% H 6.43% N 17.86% S 6.81%; found: C 58.32%, H 6.05%, N 17.52%, S 6.79%.

Synthesis of N′-(1-(4-aminophenyl)ethylidene)-3-((1,3,7-trimethyl-2,6-dioxo-2,3,6,7-tetrahydro-1H-purin-8-yl)thio)propanehydrazide(AL3).

M.p.: 284 °C (dec.). FTIR (ATR, cm^−1^): 3189 (νNH); 1707 (νCO—xanthine); 1646 with shoulder at 1666 and 1656 (νCO—xanthine, νCO—amide I); 1604, 1533 (νC=N, νC=C—xanthine, δNH—amide II). 1H NMR (DMSO-d6) δ, ppm: 7.38 (d, 2H, 2-Ph, 6-Ph, J = 1.8 Hz), 6.48 (d, 2H, 3-Ph, 5-Ph,J = 1.07 Hz), 3.73 (s,3H, N7-CH3), 3.50–3.62 (m,8H, N1-CH3, N3-CH3, S-CH2), 3.40 (t, 2H, CH2, J = 8.0 Hz), 2.12 (s, 3H,N=C-CH3). 13C NMR (75 MHz, DMSO-d6) δ, ppm: 169.0 (-CO), 154.7 (C6=O), 151.6 (C2=O), 158.8 (C8), 150.4 (C4), 159.9 (−C=), 138.1 (C1arom), 130.4 (C2, C6arom), 113.4 (C3, C5arom), 151.5 (C4arom), 110.7 (C5), 34.5 (−CH2−), 32.3 (N7-CH3), 32.3 (−CH2−S), 29.7 (N3-CH3), 28.1 (N1-CH3), 15 (CH3−CH=). For C19H23N7O3S (Mm = 429.50) calculated: C 53.13% H 5.40% N 22.83% S 7.46%; found: C 52.92%, H 5.18%, N 22.44%, S 7.43%.

Synthesis of N′-(1-(3-aminophenyl)ethylidene)-3-((1,3,7-trimethyl-2,6-dioxo-2,3,6,7-tetrahydro-1H-purin-8-yl)thio)propanehydrazide (AL4).

M.p.: 231.5 °C (dec.). FTIR (ATR, cm^−1^): 3193 (νNH); 1705 (νCO—xanthine); 1670, 1648 (νCO—xanthine, νCO—amide I); 1602, 1550 with shoulder at 1540 (νC=N, νC=C—xanthine, δNH—amide II). 1H NMR (DMSO-d6) δ, ppm: 7.15 (d, 2H, 2-Ph, J = 1.8 Hz), 6.80–6.90 (m, 3H, 3-Ph, 5-Ph, 6-Ph), 3.67 (s,3H, N7-CH3), 3.37–3.54 (m,8H, N1-CH3, N3-CH3, S-CH2), 3.20 (t, 2H, CH2, J = 8.0 Hz), 2.17 (s, 3H,N=C-CH3). 13C NMR (75 MHz, DMSO-d6) δ, ppm: 169.0 (-CO), 154.7 (C6=O), 151.6 (C2=O), 158.8 (C8), 150.4 (C4), 156.9 (−C=), 131.6 (C1arom), 115.3 (C2arom), 129.1 (C5arom), 118.4 (C4arom), 129 (C6arom) 110.7 (C5), 34.5 (−CH2−), 32.3 (N7-CH3), 32.3 (−CH2−S), 29.7 (N3-CH3), 28.1 (N1-CH3), 14.9 (CH3−CH=). For C19H23N7O3S (Mm = 429.50) calculated: C 53.13% H 5.40% N 22.83% S 7.46%; found: C 52.92%, H 5.18%, N 22.44%, S 7.43%.

Synthesis of N′-(1-(4-chlorophenyl)ethylidene)-3-((1,3,7-trimethyl-2,6-dioxo-2,3,6,7-tetrahydro-1H-purin-8-yl)thio)propanehydrazide (AL5).

M.p.: 275.5 °C (dec.). FTIR (ATR, cm^−1^): 3199 (νNH); 1699 (νCO—xanthine); 1660 with shoulder at 1668 and 1664 (νCO—xanthine, νCO—amide I); 1607, 1533 (νC=N, νC=C—xanthine, δNH—amide II). 1H NMR (DMSO-d6) δ, ppm: 7.92 (d, 2H, 3-Ph, 5-Ph,, J = 1.3 Hz), 7.52 (d, 2H, 2-Ph, 6-Ph, J = 1.8 Hz), 3.68 (s,3H, N7-CH3), 3.37–3.52 (m,8H, N1-CH3, N3-CH3, S-CH2), 3.19 (t, 2H, CH2, J = 8.0 Hz), 2.27 (s, 3H,N=C-CH3). 13C NMR (75 MHz, DMSO-d6) δ, ppm: 169.1 (-CO), 155.1 (C6=O), 151.4 (C2=O), 159.1 (C8), 150.5 (C4), 156.9 (−C=), 138.2 (C1arom), 128.5 (C2, C6arom), 129.4 (C3, C5arom), 138.5 (C4arom), 110.5 (C5), 34.6 (−CH2−), 32.3 (N7-CH3), 33.2 (−CH2−S), 29.5 (N3-CH3), 28.2 (N1-CH3), 14.9 (CH3−CH=). For C19H21ClN6O3S (Mm = 448.93) calculated: C 50.83% H 4.71% Cl 7.90% N 18.72% S 7.14%; found: C 50.52%, H 4.45%, Cl 7.79%, N 18.54%, S 7.25%.

Synthesis of N′-(1-(4-bromophenyl)ethylidene)-3-((1,3,7-trimethyl-2,6-dioxo-2,3,6,7-tetrahydro-1H-purin-8-yl)thio)propanehydrazide (AL6).

M.p.: 264.5 °C (dec.). FTIR (ATR, cm^−1^): 3203 (νNH); 1697 (νCO—xanthine); 1655 with shoulder at 1635 and 1599 (νCO—xanthine, νCO—amide I); 1538 with shoulder at 1546 (νC=N, νC=C—xanthine, δNH—amide II). 1H NMR (DMSO-d6) δ, ppm: 7.86 (d, 2H, 3-Ph, 5-Ph,, J = 1.48 Hz), 7.65 (d, 2H, 2-Ph, 6-Ph, J = 1.85 Hz), 3.67 (s,3H, N7-CH3), 3.37 (m,8H, N1-CH3, N3-CH3, S-CH2), 3.18 (t, 2H, CH2, J = 8.0 Hz), 2.27 (s, 3H,N=C-CH3). 13C NMR (75 MHz, DMSO-d6) δ, ppm: 169.1 (-CO), 155.5 (C6=O), 151.3 (C2=O), 159.1 (C8), 150.3 (C4), 156.9 (−C=), 138.2 (C1arom), 128.7 (C2, C6arom), 130.9 (C3, C5arom), 127.7 (C4arom), 111.5 (C5), 34.6 (−CH2−), 32.3 (N7-CH3), 33.2 (−CH2−S), 29.9 (N3-CH3), 28.2 (N1-CH3), 15.1 (CH3−CH=). For C19H21BrN6O3S (Mm = 493.38) calculated: C 46.25% H 4.29% Br 16.20% N 17.03% S 6.50%; found: C 45.95%, H 4.45%, Br 16.22%, N 16.85%, S 6.25%.

Synthesis N′-(1-(2,4-dichlorophenyl)ethylidene)-3-((1,3,7-trimethyl-2,6-dioxo-2,3,6,7-tetrahydro-1H-purin-8-yl)thio)propanehydrazide (AL7).

M.p.: 209–210 °C. FTIR (ATR, cm^−1^): 3200 (νNH); 1700 (νCO—xanthine); 1656 with shoulder at 1670 (νCO—xanthine, νCO—amide I); 1603, 1537 with shoulder at 1506 (νC=N, νC=C—xanthine, δNH—amide II). 1H NMR (DMSO-d6) δ, ppm: 7.73 (d, 1H, 6-Ph, J = 1.4 Hz), 7.57 (d, 2H, 3-Ph, 5-Ph, J = 1.8 Hz), 3.70 (s,3H, N7-CH3), 3.50–3.35 (m,8H, N1-CH3, N3-CH3, S-CH2), 3.20 (t, 2H, CH2, J = 8.0 Hz), 2.16 (s, 3H,N=C-CH3). 13C NMR (75 MHz, DMSO-d6) δ, ppm: 169.1 (-CO), 155.4 (C6=O), 151.6 (C2=O), 159.5 (C8), 150.3 (C4), 156.9 (−C=), 123.6 (C1arom), 133.9 (C2arom), 130.2 (C3arom), 137.5 (C4arom), 127.0 (C5arom), 131.9 (C6arom), 111.3 (C5), 34.6 (−CH2−), 32.3 (N7-CH3), 33.2 (−CH2−S), 29.5 (N3-CH3), 28.2 (N1-CH3), 14.6 (CH3−CH=). For C19H20Cl2N6O3S (Mm = 483.37) calculated: C 47.21% H 4.17% Cl 14.67% N 17.39% S 6.63%; found: C 47.52%, H 4.15%, Cl 14.79%, N 17.10%, S 6.25%.

Synthesis of N′-(1-(4-nitrophenyl)ethylidene)-3-((1,3,7-trimethyl-2,6-dioxo-2,3,6,7-tetrahydro-1H-purin-8-yl)thio)propanehydrazide (AL8).

M.p.: 220–221 °C. FTIR (ATR, cm^−1^): 3207 (νNH); 1697 with shoulder at 1704 (νCO—xanthine); 1655 with shoulder at 1637 and 1619 (νCO—xanthine, νCO—amide I); 1597, 1538 with shoulder at 1550 (νC=N, νC=C—xanthine, δNH—amide II). 1H NMR (DMSO-d6) δ, ppm: 8.31 (d, 2H, 3-Ph, 5-Ph,, J = 1.5 Hz), 8.18 (d, 2H, 2-Ph, 6-Ph, J = 1.8 Hz), 3.67 (s,3H, N7-CH3), 3.35–3.67 (m,8H, N1-CH3, N3-CH3, S-CH2), 3.19 (t, 2H, CH2, J = 8.0 Hz), 2.32 (s, 3H,N=C-CH3). 13C NMR (75 MHz, DMSO-d6) δ, ppm: 169.1 (-CO), 155.5 (C6=O), 151.3 (C2=O), 159.2 (C8), 150.4 (C4), 156.9 (−C=), 138.2 (C1arom), 128.2 (C2, C6arom), 120.9 (C3, C5arom), 164.3 (C4arom), 111.2 (C5), 34.5 (−CH2−), 32.3 (N7-CH3), 33.2 (−CH2−S), 30.0 (N3-CH3), 28.3 (N1-CH3), 15.0 (CH3−CH=). For C19H21N7O5S (Mm = 459.48) calculated: C 49.67% H 4.61% N 21.34% S 6.98%; found: C 49.35%, H 4.32%, N 20.95%, S 6.75%.

Synthesis of N′-(1-(3,4,5-trimethoxyphenyl)ethylidene)-3-((1,3,7-trimethyl-2,6-dioxo-2,3,6,7-tetrahydro-1H-purin-8-yl)thio)propanehydrazide (AL9).

M.p.: 227–227.5 °C. FTIR (ATR, cm^−1^): 3172 (νNH); 1704 (νCO—xanthine); 1686 with shoulder at 1672 and 1661 (νCO—xanthine, νCO—amide I); 1605, 1557, 1538 (νC=N, νC=C—xanthine, δNH—amide II). 1H NMR (DMSO-d6) δ, ppm: 7.18 (d, 2H, 2-Ph, 6-Ph, J = 2.3 Hz), 3.69 (s,3H, N7-CH3), 3.32–3.42 (m,8H, N1-CH3, N3-CH3, S-CH2), 3.19 (t, 2H, CH2, J = 8.0 Hz), 2.24 (s, 3H,N=C-CH3). 13C NMR (75 MHz, DMSO-d6) δ, ppm: 169.1 (-CO), 155.2 (C6=O), 151.4 (C2=O), 159.2 (C8), 150.1 (C4), 156.9 (−C=), 131.6 (C1arom), 110.3 (C2arom), 153.0 (C3, C5arom), 142.4 (C4arom), 110.3 (C6arom), 111.7 (C5), 60.8 (OCH3, Ar-4), 56.3 (OCH3, Ar-3,5), 34.6 (−CH2−), 32.3 (N7-CH3), 33.2 (−CH2−S), 29.5 (N3-CH3), 28.2 (N1-CH3), 14.9 (CH3−CH=). For C22H28N6O6S (Mm = 504.56) calculated: C 52.37% H 5.59% N 16.66% S 6.35%; found: C 51.90%, H 5.15%, N 16.32%, S 6.27%.

Synthesis of N′-(1-(4-fluorophenyl)ethylidene)-3-((1,3,7-trimethyl-2,6-dioxo-2,3,6,7-tetrahydro-1H-purin-8-yl)thio)propanehydrazide (AL10).

M.p.: 221.5–222.5 °C. FTIR (ATR, cm^−1^): 3193 (νNH); 1698 (νCO—xanthine); 1661 with shoulder at 1635 and 1611 (νCO—xanthine, νCO—amide I); 1596, 1538 with shoulder at 11,546 (νC=N, νC=C—xanthine, δNH—amide II). 1H NMR (DMSO-d6) δ, ppm: 7.47 (d, 2H, 2-Ph, 6-Ph, J = 1.8 Hz), 7.07 (d, 2H, 3-Ph, 5-Ph,, J = 1.3 Hz), 3.68 (s,3H, N7-CH3), 3.37–3.52 (m,8H, N1-CH3, N3-CH3, S-CH2), 3.19 (t, 2H, CH2, J = 8.0 Hz), 2.24 (s, 3H,N=C-CH3). 13C NMR (75 MHz, DMSO-d6) δ, ppm: 169.1 (-CO), 155.4 (C6=O), 151.3 (C2=O), 159.1 (C8), 150.3 (C4), 156.9 (−C=), 138.1 (C1arom), 130.4 (C2, C6arom), 115.6 (C3, C5arom), 164.9 (C4arom), 111.2 (C5), 34.6 (−CH2−), 32.3 (N7-CH3), 33.2 (−CH2−S), 29.5 (N3-CH3), 28.2 (N1-CH3), 15.1 (CH3−CH=). For C19H21FN6O3S (Mm = 432.47) calculated: C 52.77% H 4.89% F 4.39% N 19.43% S 7.41%; found: C 52.55%, H 4.75%, F 4.15%, N 19.24%, S 7.33%.

2.2. Antioxidant Assays

All antioxidant assays were performed across a range of micromolar concentrations, selected based on preliminary screening experiments and previously published literature data, where “compounds are typically tested in the low micromolar range to assess intracellular antioxidant activity while maintaining cell viability” [25].

2.2.1. DPPH Radical Scavenging Assay

The antioxidant activity of the tested caffeine derivatives was evaluated using the 2,2-diphenyl-1-picrylhydrazyl (DPPH) radical scavenging assay, based on a modified protocol originally described by Brand-Williams et al. [26]. A working solution of DPPH (440 µg/mL) was prepared in ethanol. The compounds were dissolved in DMSO and tested at various micromolar concentrations. Reaction mixtures were incubated in the dark at 25 °C for 30 min. After incubation, absorbance was measured at 517 nm using a microplate reader. Trolox, a well-known antioxidant, was used as a positive control. All measurements were performed in triplicate (n = 3). Results are presented as the percentage inhibition of the DPPH radical relative to the control.

2.2.2. ABTS Radical Scavenging Assay

The ABTS (2,2′-azinobis(3-ethylbenzothiazoline-6-sulfonic acid)) radical scavenging activity of the tested caffeine derivatives was evaluated using a modified method based on Re et al. [27]. The ABTS•^+^ radical cation was generated by mixing equal volumes of an aqueous ABTS solution (14 mmol/L) and potassium persulfate (4.9 mmol/L). The mixture was kept in the dark at room temperature for 12–16 h to allow complete radical formation. The resulting ABTS•^+^ solution was then diluted 20-fold with ethanol to prepare a 5% working solution. For the assay, 100 µL of the ABTS•^+^ working solution was combined with 100 µL of the test compound (dissolved in ethanol) in a 96-well microplate. After incubation in the dark for 6–10 min at room temperature, absorbance was measured at 734 nm using a microplate reader. Trolox was used as a reference antioxidant. All experiments were performed in triplicate (n = 3), and results are expressed as percentage inhibition of the ABTS•^+^ radical relative to the control.

2.3. Cell Culture and Treatment Conditions

The human neuroblastoma cell line SH-SY5Y was obtained from Sigma-Aldrich (ECACC). Cells were maintained in RPMI 1640 medium supplemented with 10% heat-inactivated fetal bovine serum (FBS), 2 mmol/L L-glutamine, and 1% penicillin/streptomycin antibiotic solution. Cultures were kept in a humidified incubator at 37 °C with 5% CO_2_. The culture medium was refreshed every 2–3 days to maintain optimal growth conditions. For experimental assays, cells were seeded into 96-well microplates at a density of 2.5 × 10^4^ cells per well and allowed to adhere for 24 h prior to treatment with the test compounds.

2.4. Cytotoxicity Assessment

The cytotoxic potential of the synthesized caffeine derivatives was evaluated using the MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay, following the protocol described by Mosmann [28]. SH-SY5Y cells were exposed to various concentrations of the compounds (0.1, 1, 5, 10, 50, 100, 250, and 500 µM) for 24 h. After treatment, 100 µL of MTT solution (5 mg/mL in culture medium) was added to each well, and the plates were incubated for 3 h at 37 °C to allow viable cells to form formazan crystals. The crystals were then solubilized by adding 100 µL of DMSO per well. Absorbance was measured at 570 nm with a background correction at 690 nm using a Synergy 2 microplate reader (BioTek Instruments, Inc., Winooski, VT, USA). Each concentration was tested in six replicates (n = 6), and cell viability was expressed relative to untreated controls.

2.5. DCFH-DA Assay for Intracellular ROS Detection

Intracellular reactive oxygen species (ROS) levels were quantified using the fluorescent probe 2′,7′-dichlorofluorescein diacetate (DCFH-DA), adapting the method of Wang and Joseph [29]. SH-SY5Y cells were seeded at a density of 3 × 10^4^ cells/well in 96-well plates and incubated for 24 h at 37 °C in a 5% CO_2_ atmosphere. The cells were then treated with the test compounds at concentrations of 0.1, 1, 5, 10, 20, and 50 µM for 2 h. For consistency and clarity of data presentation, selected representative concentrations (1, 10, and 50 µM) were used for graphical visualization of the results. Subsequently, 60 µM DCFH-DA (prepared in serum-free medium) was added, and cells were incubated for 1 h at 37 °C in the dark to allow intracellular oxidation to the fluorescent DCF. After washing with PBS containing Ca^2+^ and Mg^2+^, oxidative stress was induced by applying 500 µM hydrogen peroxide for 30 min. Following a second PBS wash, fluorescence was immediately measured at excitation/emission wavelengths of 485/20 nm and 528/20 nm using a Synergy 2 microplate reader (BioTek Instruments, USA). Each concentration was tested in six replicates (n = 6). ROS levels were expressed as a percentage relative to the H_2_O_2_-treated control group. An untreated control group (cells + DCFH-DA only) was included to indicate basal ROS levels.

2.6. Hydrogen Peroxide-Induced Oxidative Stress Model

To simulate oxidative stress, SH-SY5Y cells were seeded at 3.5 × 10^4^ cells/well in 96-well plates and incubated for 24 h at 37 °C with 5% CO_2_. Cells were pretreated with the caffeine derivatives at concentrations of 0.1, 1, 5, 10, 20, and 50 µM for 90 min. Representative concentrations (1, 10, and 50 µM) were selected for data visualization, while the full concentration range was included in the experimental evaluation. Subsequently, hydrogen peroxide (1 mM), prepared in culture medium, was applied for 15 min under standard culture conditions. The oxidative stimulus was then removed by replacing the medium with fresh culture medium, and the cells were incubated for an additional 24 h. Cell viability was assessed by the MTT assay as described above. Protective effects of the compounds were calculated as a percentage, with 0% representing cell damage caused by H_2_O_2_ alone and 100% corresponding to viability of untreated cells. Experiments were performed in six replicates (n = 6).

2.7. L-Glutamate-Induced Neurotoxicity Model

The neuroprotective potential of the caffeine derivatives was further assessed using an L-glutamate-induced excitotoxicity model. SH-SY5Y cells were seeded at 3.5 × 10^4^ cells/well in 96-well plates and incubated for 24 h under standard conditions. After medium removal, cells were pretreated with test compounds at concentrations of 0.1, 1, 5, 10, 20, and 50 µM for 30 min. For graphical presentation of the results, representative concentrations (1, 10, and 50 µM) were selected. Treatment solutions were then discarded, and cells were washed with PBS before exposure to 65.55 mM L-glutamate for 24 h to induce neurotoxicity. The L-glutamate stock solution (65.55 mM) was prepared by initial dissolution of L-glutamate in hydrochloric acid, followed by careful adjustment of the pH to 7.4 using sodium hydroxide under continuous monitoring with a calibrated pH meter. The final concentration was achieved by adjusting the volume with phosphate-buffered saline (PBS). Cell viability was assessed by the MTT assay. All treatments were performed in six replicates (n = 6). Control groups included untreated cells (intact control) and cells treated with L-glutamate alone (damaged control), which served as reference points for evaluating neuroprotection.

2.8. Statistical Analysis

Statistical analysis was performed using GraphPad Prism software (version 8.0, GraphPad Software, La Jolla, CA, USA). Cell-based experiments were conducted in six replicates (n = 6), and data are expressed as mean ± SD. Antioxidant assays (DPPH and ABTS) were performed in triplicate (n = 3). Group comparisons were conducted using one-way analysis of variance (ANOVA), followed by Dunnett’s post hoc test for multiple comparisons versus control. Differences were considered statistically significant at * p ≤ 0.05, ** p ≤ 0.01, and *** p ≤ 0.001.

3. Results

3.1. Synthesis of the Target Hydazide-Hydrazones

The synthesis of the target compounds was carried out according to the sequence presented in Scheme 1. The preparation of 8-bromocaffeine (1), 3-(1,3,7-xanthinyl-8-thio)propanoic acid (2), and its methyl ester (3) followed procedures previously described by [24].

Hydrazinolysis of methyl ester 3 afforded the hydrazide intermediate AL0, according to [30], which was subsequently condensed with selected carbonyl compounds. The methyl ester was selected as the starting substrate due to its high yield, ease of isolation and purification, and the facility with which the methanol fragment can be displaced by nucleophiles.

Hydrazinolysis proceeded efficiently under reflux in an alcoholic medium. TLC monitoring (systems 1 and 2) showed complete conversion within 2 h, providing hydrazide AL0 in yields of approximately 88%.

Direct hydrazinolysis of the free acid (2) with hydrazine hydrate was unsuccessful. Below 70 °C, no reaction occurred, whereas higher temperatures led to rapid darkening of the reaction mixture, forming a resinous mass above 100 °C. TLC analysis (system 2) revealed only trace amounts of the desired hydrazide along with numerous unidentified by-products.

The second step of the synthesis—the condensation of hydrazide AL0 with carbonyl partners—was performed using a 1:1 molar ratio (hydrazide:carbonyl compound) in boiling ethanol. Reaction times (Table 1) were determined by TLC monitoring of the disappearance of the starting hydrazide.

3.2. Evaluation of the In Vitro Antioxidant Activity of the Tested Caffeine Derivatives

3.2.1. Evaluation of the ABTS•+ Radical Scavenging Activity of the Tested Caffeine Derivative

The antioxidant activity of the synthesized caffeine derivatives was assessed using the ABTS•^+^ radical scavenging assay at concentrations of 5, 50, and 100 µM. Compound 0 and caffeine were used as structural references, while Trolox served as the antioxidant standard (Figure 1). All tested derivatives exhibited pronounced ABTS•^+^ scavenging activity, with inhibition values generally exceeding 95%. Notably, compounds AL8, AL9, and AL10 demonstrated strong activity even at the lowest tested concentration (5 µM), exhibiting inhibition levels higher than that of Trolox (109%, 106%, and 107%, respectively, compared to 103% for Trolox). This enhanced activity was maintained at 50 µM, with inhibition values remaining above 100%, indicating high radical scavenging efficiency at relatively low concentrations.

At 100 µM, most derivatives retained high antioxidant capacity, with inhibition values close to or exceeding those of Trolox. In contrast, compounds AL2, AL3, and AL6 showed comparatively lower activity at this concentration (96%, 91%, and 93%, respectively), although their scavenging capacity remained substantial. Compound AL7 exhibited the highest activity at 100 µM, reaching 115% inhibition, while AL10 and Trolox showed comparable effects (104% and 101%, respectively).

Caffeine, used as a reference compound, showed lower ABTS•^+^ scavenging activity than most derivatives at 5 and 50 µM (65% and 75%, respectively), but approached near-maximal inhibition at 100 µM (98%), comparable to the least active derivatives. Importantly, none of the tested compounds demonstrated a clear dose-dependent trend within the investigated concentration range, suggesting that their antioxidant activity reaches a plateau at relatively low concentrations.

Overall, these results indicate that structural modification of caffeine did not compromise its antioxidant potential and, in several cases, led to a marked enhancement of ABTS•^+^ radical scavenging activity.

3.2.2. DPPH Radical Scavenging Assay of the Tested Caffeine Derivatives

The caffeine-derived compounds were evaluated for their ability to scavenge the stable DPPH radical at concentrations of 5, 50, and 100 µM (Figure 2). All tested compounds demonstrated high radical scavenging activity, with most reaching a plateau effect already at lower concentrations, indicating strong antioxidant efficiency.

Among the tested compounds, AL7 exhibited the highest DPPH scavenging activity, reaching 115% inhibition at 100 µM. In contrast, AL2 and AL6 showed comparatively lower activity at this concentration, with inhibition values of 88% and 93%, respectively. The majority of the derivatives, including AL9, AL10, as well as the antioxidant standard Trolox, maintained consistent inhibition levels close to or exceeding 100% across the tested concentration range.

Caffeine, used as a reference compound, displayed a concentration-dependent DPPH scavenging effect, with inhibition increasing from 56% at 5 µM to 93% at 100 µM, indicative of moderate antioxidant activity that improves with increasing concentration. The observed variability in the radical scavenging performance of the caffeine derivatives in the DPPH assay likely reflects differences in their reactivity toward the DPPH radical, which is known to favor hydrogen atom–donating antioxidants. These findings complement the results obtained from the ABTS assay and further support the enhanced antioxidant potential of the synthesized caffeine derivatives.

3.3. Cytotoxicity Evaluation of the Caffeine Derivatives on SH-SY5Y Neuroblastoma Cells

In the next phase of our research, we were focused on evaluating the cytotoxic profile of the newly synthesized caffeine derivatives to determine their safety and biocompatibility in neuronal cells. This step is essential for the preclinical validation of any candidate molecule intended for neuroprotective or therapeutic use. The cytotoxic potential of these compounds was assessed using the MTT assay in SH-SY5Y neuroblastoma cells, treated with concentrations ranging from 0.1 to 500 µM. After 24 h of exposure, all tested compounds exhibited low cytotoxicity, with calculated IC_50_ values exceeding 500 µM (Table 2), indicating a favorable safety profile under the experimental conditions applied.

An exception was observed for compound AL9, which showed a dose-dependent reduction in cell viability, with an estimated IC_50_ of 135 µM. All other derivatives, including compounds AL0–AL8, and AL10, demonstrated high IC_50_ values, indicating minimal cytotoxic effects under the experimental conditions. Caffeine, used as a reference compound, also exhibited negligible cytotoxicity in SH-SY5Y cells, further supporting its safety and biocompatibility in neuronal models. The results are presented in Table 2.

3.4. Effect of the Compounds on Intracellular ROS Levels in SH-SY5Y Cells

The intracellular antioxidant capacity of the tested caffeine derivatives was evaluated using the DCFH-DA assay in SH-SY5Y neuroblastoma cells exposed to 500 µM hydrogen peroxide (H_2_O_2_), a commonly used inducer of oxidative stress. Cells were pretreated with the compounds at concentrations ranging from 0.1 to 50 µM, and the levels of reactive oxygen species (ROS) were quantified and expressed as the percentage of reduction compared to the H_2_O_2_-treated control group. For clarity of data presentation, representative concentrations (1, 10, and 50 µM) were selected for graphical visualization, while all tested concentrations were included in the experimental analysis. Several derivatives exhibited strong intracellular antioxidant activity, with a clear reduction in ROS levels observed at the selected representative concentrations. Compounds AL0, AL3, AL7, AL9, and AL10 demonstrated the most pronounced effects, showing substantial ROS reduction at 10 µM, which was maintained or further enhanced at 50 µM. These results indicate efficient intracellular ROS scavenging within a biologically relevant concentration range. Compounds AL1, AL4, and AL6 also produced notable, though comparatively moderate, reductions in ROS levels, which nevertheless reached statistical significance, suggesting measurable antioxidant capacity at the cellular level. In contrast to the trends observed in the previous dataset, none of the tested derivatives exhibited negligible activity under the selected experimental conditions. Even compounds displaying moderate effects contributed to a detectable attenuation of H_2_O_2_-induced oxidative stress. Caffeine, used as a reference compound, showed weaker intracellular antioxidant activity compared to most of the synthesized derivatives, with limited ROS reduction at 1 µM and a moderate increase in activity at higher concentrations. However, its overall effectiveness remained lower than that of several caffeine derivatives, underscoring the impact of structural modifications on enhancing antioxidant performance. Taken together, these findings demonstrate that multiple caffeine-derived compounds effectively attenuate intracellular oxidative stress in SH-SY5Y cells, there by supporting their potential neuroprotective properties. The data are summarized in Figure 3.

3.5. Protective Effects of the Derivatives in a Model of H2O2-Induced Oxidative Damage

The antioxidant effects in cellular models relevant to neurodegenerative conditions of the synthesized caffeine derivatives were evaluated in SH-SY5Y neuroblastoma cells exposed to oxidative stress induced by hydrogen peroxide (1 mM H_2_O_2_, 15 min). Cells were pretreated with the test compounds at concentrations ranging from 0.1 to 50 µM for 1.5 h, followed by a short exposure 15 min to 500 µM H_2_O_2_ to induce oxidative damage. After 24 h, cell viability was measured using the MTT assay. Results were expressed as percentage protection, with two reference points: untreated control cells representing 100% protection and H_2_O_2_-treated cells (without compound pretreatment) representing 0% protection. For improved clarity of data visualization, representative concentrations (1, 10, and 50 µM) were selected for graphical presentation, while all tested concentrations were included in the experimental evaluation.

AL0 exhibited the most robust antioxidant effects in cellular models relevant to neurodegenerative conditions across the tested concentration range, showing consistently high levels of protection already at 10 µM, which were maintained or slightly enhanced at 50 µM. Compounds AL1 and AL4 also demonstrated strong and reproducible protective effects, particularly at higher concentrations, where protection values exceeded approximately 70%, indicating pronounced neuroprotective efficacy. Compounds AL7 and AL8 provided comparable levels of protection, exceeding approximately 65% at the highest tested concentration, further supporting their neuroprotective potential.

Compound AL3 displayed moderate neuroprotective efficacy, conferring partial but statistically significant protection across the tested concentrations. In contrast, compounds AL2, AL5, and AL6 showed weaker protective effects, with protection levels remaining below 50%, suggesting limited efficacy under the applied experimental conditions. Compound AL10 exhibited variable responses, without a clear concentration-dependent trend, indicating heterogeneity in its protective capacity.

Caffeine, included as a reference compound, showed a concentration-dependent neuroprotective effect, with modest protection at 1 µM and a maximal effect of approximately 60% at 50 µM. Although caffeine displayed measurable efficacy, it was consistently outperformed by several of its synthetic derivatives, most notably AL0, AL1, and AL4. These findings indicate that specific structural modifications of caffeine can markedly enhance potential neuroprotective relevance, potentially through improved antioxidant efficiency and/or enhanced cellular uptake. The full set of neuroprotection results is presented in Figure 4.

3.6. Neuroprotective Activity of the Derivatives in a Glutamate-Induced Toxicity Model

The neuroprotective potential of the tested caffeine derivatives was evaluated in SH-SY5Y cells subjected to excitotoxicity induced by L-glutamate. Cells were pretreated with the compounds at six concentrations (0.1–50 µM) for 30 min, followed by exposure to 65.55 mM L-glutamate, corresponding to its IC_50_ value, to induce excitotoxic injury. For clarity of data presentation, representative concentrations (1, 10, and 50 µM) were selected for graphical visualization, while all tested concentrations were included in the experimental analysis. Cell viability was assessed after 24 h (Figure 5).

Several derivatives, including AL0, AL1, AL4, AL7, AL9, and AL10, demonstrated pronounced antioxidant effects in cellular models relevant to neurodegenerative conditions, particularly at the selected representative concentrations of 10 and 50 µM, where substantial protection against glutamate-induced excitotoxicity was observed. Among these compounds, AL7 exhibited the highest neuroprotective efficacy, reaching protection levels of approximately 80–85%, indicating a strong ability to counteract glutamate-induced neuronal injury.

Other derivatives, such as AL0, AL1, AL4, AL9, and AL10, also conferred significant and reproducible protection, although to a slightly lesser extent than AL7, suggesting robust but compound-specific neuroprotective profiles.

In contrast, compounds AL2 and AL6 displayed weaker or less consistent neuroprotective responses, with lower protection levels across the tested concentrations, indicating limited efficacy in this excitotoxicity model.

Caffeine, used as a reference compound, exhibited moderate and concentration-dependent antioxidant effects in cellular models relevant to neurodegenerative conditions, with protection values increasing from approximately 30% at lower concentrations to around 50% at 50 µM. However, its overall efficacy remained inferior to that of several caffeine-derived compounds, further supporting the notion that chemical modification of the caffeine scaffold enhances resistance to glutamate-induced neuronal damage. Importantly, treatment with caffeine or the tested derivatives alone did not significantly affect cell viability under the applied experimental conditions, confirming that the observed effects are attributable to protection against L-glutamate-induced toxicity.

3.7. SwissADME Predictions of Pharmacokinetics and Drug-Likeness Properties of the Target Caffeine Derivatives

In this study, we used SwissADME to perform a comprehensive in silico pharmacokinetic evaluation of the study compounds. The corresponding necessary for the calculations 3D structures of the studied compounds are presented on Figure 6.

The results provide insight into the drug-like nature and suitability for further optimization of the study compounds, contributing to the broader field of drug development using artificial intelligence and rational pharmacokinetic modeling. The corresponding physicochemical properties of the compounds AL0–AL10 were calculated and the results are presented in Table 3.

The bioavailability radar plot (Figure 7) confirms that the majority of the compounds studied fall comfortably within the optimal region for oral bioavailability. The bioavailability radar summarizes six physicochemical properties relevant for oral drug-likeness: lipophilicity (LIPO), size (SIZE), polarity (POLAR), solubility (INSOLU), saturation (INSATU) and flexibility (FLEX). The pink area in each plot delineates the optimal range for these descriptors and compounds whose radar lies entirely or predominantly within this colored region are predicted to have suitable properties for oral bioavailability.

The bioavailability radar plot (Figure 7) confirms that the studied compounds are mostly within the optimal range for oral bioavailability. All six physicochemical parameters—lipophilicity, size, polarity, solubility, saturation and flexibility—fall within the optimal drug similarity, which enhances its potential for oral administration.

The BOILED-Egg model (Figure 8) further supports this conclusion.

In terms of metabolic interactions, the compounds did not show predicted inhibitory activity against the major cytochrome P450 isoforms (CYP1A2, CYP2C19 and CYP2D6), indicating a low potential for metabolic drug interactions. Inhibitory activity was predicted only for CYP2C9 and CYP3A4.

The solubility profile classifies the molecules as “soluble” or “moderately soluble.” Only the unsubstituted hydrazide is predicted to be “very soluble (e.g., ESOL solubility: 25.8 mg ml^−1^) (Table 4).

Key descriptors including gastrointestinal absorption(HIA), blood–brain barrier permeability (BBB) and efflux/retention by P-glycoprotein (Pgp) and the possible cytochrome P450 interactions were evaluated (Table 5).

4. Discussion

This study aimed to evaluate the antioxidant and neuroprotective potential of a series of novel caffeine derivatives through comprehensive in vitro assays, including ABTS•^+^ and DPPH radical scavenging tests, intracellular ROS measurements, and neurotoxicity models based on hydrogen peroxide (H_2_O_2_) and L-glutamate exposure in SH-SY5Y neuroblastoma cells.

The ABTS•^+^ and DPPH assays revealed that all caffeine derivatives possessed strong radical scavenging activity, often exceeding that of the parent compound caffeine and the standard antioxidant Trolox. Notably, compounds AL8, AL9, and AL10 displayed remarkable efficacy even at the lowest concentration tested (5 µM), with radical inhibition surpassing 100% in the ABTS•^+^ assay. The plateau effect observed at higher concentrations suggests that these compounds rapidly achieve maximal radical scavenging capacity, a phenomenon commonly reported with antioxidants that reach saturation in their interaction with free radicals [31]. The improved antioxidant performance of the derivatives compared with caffeine suggests that specific structural modifications enhance electron-donating capacity or radical stabilization, potentially by altering the electronic environment or introducing additional radical-reactive sites. Consistent with these findings, the DPPH assay confirmed that AL7 exhibited the highest scavenging activity at 100 µM, reinforcing its potent antioxidant capability. The slight variability between the ABTS•^+^ and DPPH results likely reflects differences in the physicochemical properties and radical species targeted by each assay; ABTS•^+^ involves a charged radical cation soluble in aqueous media, whereas DPPH is a neutral radical soluble in organic solvents [25].

Our observations showed that the appearance of halogen atoms, as electron-withdrawing groups, enhanced the antioxidant activity of the compounds. Moreover, antioxidant activity was influenced by the size and polarizability of the halogen substituents on the phenyl ring, consistent with our findings. A clear trend was observed in which antioxidant activity decreased with increasing electron-withdrawing capacity of the halogens (F > Cl > Br). Notably, the dichlorinated compound AL7, bearing ortho- and para-substituents, exhibited higher antioxidant activity (115%) than the monochlorinated compound AL5 (100%) in both radical scavenging assays at a concentration of 100 μM. Furthermore, compound AL10, containing a fluorine substituent, showed greater antioxidant activity than its bromine-substituted analog AL6, indicating that the electronegativity of phenyl ring substituents (F > Br) plays a significant role in modulating the antioxidant activity of the newly synthesized caffeine Schiff base derivatives. These findings are supported by a current publication from Zhu et al., evaluating the influence of halogen substituents on the radical scavenging assay of Schiff base compounds [32].

These complementary assays strengthen the conclusion that the synthesized caffeine derivatives possess broad-spectrum antioxidant activity.

The MTT cytotoxicity assays confirmed the general safety of the derivatives in neuronal cells, with IC_50_ values exceeding 500 µM for most compounds. AL9 was a notable exception, exhibiting a moderate dose-dependent cytotoxicity (IC_50_ = 135 µM), indicating that while chemical modification can enhance antioxidant properties, it may also introduce cytotoxic liabilities depending on the specific molecular features. Nevertheless, the low cytotoxicity profiles observed for the remaining derivatives, including caffeine, support their suitability for further neuropharmacological evaluation, where preservation of neuronal viability is critical.

The moderate dose-dependent cytotoxicity of AL9 may be due to the introduction in the structure of three methoxy groups, which may be related to the enhanced toxicity compared to halogen-substituted caffeine Schiff bases because the methoxy groups are strong electron-donating substituents that increase the electron density of the aromatic ring, lower O–H (or analogous bond) dissociation enthalpies, and stabilize intermediate radicals, which promotes redox activity and formation of reactive species that can cause oxidative stress and cellular damage, whereas halogens (electron-withdrawing groups) tend to raise bond dissociation energies, reduce radical formation, and make the ring less reactive biologically; mechanistic studies on substituted phenols show that electron-donating groups stabilize phenoxyl radicals and facilitate subsequent toxic pathways such as quinone formation, while electron-withdrawing groups slow radical formation and change the dominant toxicity mechanism, often resulting in lower radical-mediated toxicity at comparable concentrations [33,34,35].

Intracellular ROS levels were measured using the DCFH-DA assay, which relies on the cell-permeant probe 2′,7′-dichlorofluorescein diacetate. Once inside the cell, DCFH-DA is hydrolyzed to non-fluorescent DCFH, which is oxidized by ROS into the highly fluorescent dichlorofluorescein (DCF), providing a sensitive measure of oxidative stress [36]. Several caffeine derivatives demonstrated pronounced intracellular antioxidant activity, as evidenced by a clear reduction in ROS accumulation induced by 500 µM H_2_O_2_. Compounds including AL0, AL3, AL4, AL7, AL9, and AL10 exhibited substantial ROS reduction at the representative concentrations, confirming their ability to exert antioxidant effects within the cellular environment.

These findings indicate that multiple derivatives are capable of efficiently scavenging or neutralizing intracellular reactive oxygen species before oxidative damage occurs. The observed discrepancies between cell-free antioxidant capacity and intracellular ROS modulation emphasize that chemical radical scavenging does not always directly translate into cellular efficacy, which may depend on factors such as cellular uptake, stability, and intracellular localization.

Hydrogen peroxide is a reactive oxygen species capable of crossing cell membranes and generating highly reactive hydroxyl radicals via Fenton chemistry, causing oxidative damage to proteins, lipids, and DNA, ultimately triggering cell death [37]. The H_2_O_2_-induced oxidative stress model demonstrated that several derivatives—most notably AL0, AL1, and AL4—significantly improved cell survival after oxidative insult, achieving protection levels of approximately 70% in the most effective derivatives. The variability in protective efficacy among the derivatives further underscores that neuroprotection is a multifactorial process, influenced not only by antioxidant activity but also by compound bioavailability, reactivity, and interactions with intracellular signaling pathways.

Excitotoxicity induced by excessive L-glutamate exposure represents a major pathological mechanism in neurodegeneration [38]. Glutamate overstimulation of NMDA and AMPA receptors leads to excessive calcium influx, triggering mitochondrial dysfunction, ROS overproduction, and activation of cell death pathways [39]. In this study, the glutamate toxicity model revealed that derivatives such as AL7, AL0, AL1, and AL4 conferred significant neuroprotection, with AL7 achieving approximately 80–85% protection at the highest tested representative concentration (50 µM). These results indicate that selected derivatives are capable of counteracting excitotoxic injury, possibly through antioxidant mechanisms that mitigate glutamate-induced oxidative stress and/or modulation of intracellular signaling cascades related to calcium homeostasis and apoptotic regulation. The superior protective efficacy of the synthetic derivatives compared with caffeine itself suggests that rational molecular modification can enhance biological performance, potentially through improved target interactions or pharmacokinetic properties.

Despite the pronounced antioxidant and neuroprotective effects observed in the applied cellular models, the present study has certain limitations. In particular, the activity of endogenous antioxidant enzymes, such as catalase (CAT), superoxide dismutase (SOD), and glutathione reductase (GR), was not directly assessed. While the current work was designed as an initial in vitro screening to evaluate cellular antioxidant capacity and neuroprotection, inclusion of enzymatic activity measurements would provide deeper mechanistic insight into the molecular pathways involved. Future studies will therefore focus on investigating the modulation of endogenous antioxidant defense systems in advanced in vitro models and in vivo settings to further elucidate the mechanisms underlying the observed neuroprotective effects.

Collectively, these findings demonstrate that the synthesized caffeine derivatives combine potent in vitro antioxidant activity with biologically relevant neuroprotective effects in cellular models of oxidative and excitotoxic stress. The dual capacity to scavenge free radicals and protect neuronal cells from damage induced by H_2_O_2_ and L-glutamate supports their identification as promising lead compounds rather than definitive neuroprotective agents, warranting further investigation.

Given that antioxidant and neuroprotective efficacy alone is insufficient to predict translational potential, an in silico assessment of pharmacokinetic and drug-likeness properties was performed using SwissADME. Predicting the ADME profile of newly synthesized xanthine hydrazide-hydrazones is essential for understanding their pharmacokinetic behavior, optimizing dosing strategies, and minimizing toxicity risks in potential therapeutic settings. Key descriptors, including gastrointestinal absorption (HIA), blood–brain barrier (BBB) permeability, P-glycoprotein (P-gp) efflux liability, cytochrome P450 interactions, and physicochemical parameters, were analyzed using the BOILED-Egg model and related predictive tools [40].

From a physicochemical perspective, the studied molecules exhibit molecular weights within the acceptable range (below 500 g·mol^−1^), moderate flexibility (5–7 rotatable bonds), and topological polar surface area (TPSA) values generally compatible with oral drug-likeness. Five compounds (AL0, AL3, AL4, AL8, and AL9) exceeded the TPSA threshold of 130 Å^2^, placing them in the borderline region of the BOILED-Egg diagram. For the remaining derivatives, these parameters favor effective membrane permeability while remaining within the optimal range for passive diffusion. The consensus logP_o_/𝓌 values (1.18–3.00) indicate balanced lipophilicity, avoiding extremes that could compromise oral bioavailability. As expected, the unsubstituted hydrazide AL0 displayed increased hydrophilicity. Three-dimensional structural analysis revealed relatively compact and planar conformations, supported by the low fraction of sp^3^-hybridized carbon atoms (0.32–0.45), consistent with the aromatic-rich topology imposed by the xanthine core and phenyl-containing side chains.

SwissADME predictions further indicated that the novel compounds are likely to be passively absorbed from the gastrointestinal tract but are not expected to cross the blood–brain barrier. While limited BBB permeability may reduce the risk of central nervous system side effects in non-neurological indications, it also suggests that further structural optimization or targeted delivery strategies may be required to fully exploit their neuroprotective potential in CNS disorders. Importantly, compounds AL0–AL2, AL5–AL7, and AL10 were predicted not to be substrates of P-glycoprotein, suggesting reduced efflux susceptibility and more favorable systemic retention compared with AL3, AL4, AL8, and AL9.

Caffeine is metabolized predominantly in the liver by the cytochrome P450 enzyme system to the major metabolite paraxanthine and the minor metabolites theobromine and theophylline, primarily via CYP1A2 [41]. Microsomal studies have shown that CYP2C8, CYP2C9, and CYP3A4 may also contribute to caffeine metabolism [42,43]. Human CYP2C9 is among the most abundant drug-metabolizing cytochrome P450 enzymes, with hepatic expression levels comparable to CYP3A4 [44], and plays a key role in the metabolism of numerous clinically used drugs [45,46]. In the present study, the synthesized compounds were not predicted to inhibit major CYP isoforms such as CYP1A2, CYP2C19, or CYP2D6, indicating a low risk of clinically relevant metabolic drug–drug interactions. Predicted inhibitory activity was limited to CYP2C9 and CYP3A4, which are not primary enzymes involved in xanthine metabolism. This relative metabolic neutrality further enhances the appeal of these derivatives as candidates for repositioning or combination therapy, particularly in polypharmacy settings.

5. Conclusions

A series of ten novel caffeine-based hydrazone derivatives was synthesized in moderate to good yields. The structures of the new molecules were elucidated using appropriate spectroscopic methods. The synthesized caffeine derivatives demonstrated enhanced antioxidant and neuroprotective activities compared with the parent compound caffeine, effectively reducing oxidative stress and glutamate-induced toxicity in neuronal cells.

In vitro experimental data indicated that selected derivatives possess strong radical scavenging capacity and confer significant protection against oxidative and excitotoxic injury, while maintaining low intrinsic cytotoxicity. These findings support the notion that rational chemical modification of the caffeine scaffold can improve its biological efficacy without substantially increasing toxicity.

In silico pharmacokinetic predictions suggested that the novel compounds are likely to be passively absorbed from the gastrointestinal tract but are not expected to cross the blood–brain barrier. In addition, compounds AL0–AL2, AL5–AL7, and AL10 were predicted not to be substrates of P-glycoprotein (P-gp), suggesting reduced efflux susceptibility and more favorable systemic retention compared with AL3, AL4, AL8, and AL9.

Regarding metabolic interactions, the compounds did not exhibit predicted inhibitory activity against major cytochrome P450 isoforms, including CYP1A2, CYP2C19, and CYP2D6, indicating a low potential for clinically relevant metabolic drug–drug interactions. This favorable metabolic profile further supports their suitability for continued pharmacological development.

Overall, these results highlight the potential of structural modification to enhance the antioxidant and neuroprotective properties of caffeine-derived compounds. The identified derivatives represent promising lead candidates for further mechanistic investigations and in vivo studies aimed at evaluating their therapeutic relevance in neurodegenerative conditions associated with oxidative stress and excitotoxicity.