Novel Hybrid Indole-Based Caffeic Acid Amide Derivatives as Potent Free Radical Scavenging Agents: Rational Design, Synthesis, Spectroscopic Characterization, In Silico and In Vitro Investigations

Antioxidant small molecules can prevent or delay the oxidative damage caused by free radicals. Herein, a structure-based hybridization of two natural antioxidants (caffeic acid and melatonin) afforded a novel hybrid series of indole-based amide analogues which was synthesized with potential antioxidant properties. A multiple-step scheme of in vitro radical scavenging assays was carried out to evaluate the antioxidant activity of the synthesized compounds. The results of the DPPH assay demonstrated that the indole-based caffeic acid amides are more active free radical scavenging agents than their benzamide analogues. Compared to Trolox, a water-soluble analogue of vitamin E, compounds 3a, 3f, 3h, 3j, and 3m were found to have excellent DPPH radical scavenging activities with IC50 values of 95.81 ± 1.01, 136.8 ± 1.04, 86.77 ± 1.03, 50.98 ± 1.05, and 67.64 ± 1.02 µM. Three compounds out of five (3f, 3j, and 3m) showed a higher capacity to neutralize the radical cation ABTS•+ more than Trolox with IC50 values of 14.48 ± 0.68, 19.49 ± 0.54, and 14.92 ± 0.30 µM, respectively. Compound 3j presented the highest antioxidant activity with a FRAP value of 4774.37 ± 137.20 μM Trolox eq/mM sample. In a similar way to the FRAP assay, the best antioxidant activity against the peroxyl radicals was demonstrated by compound 3j (10,714.21 ± 817.76 μM Trolox eq/mM sample). Taken together, compound 3j was validated as a lead hybrid molecule that could be optimized to maximize its antioxidant potency for the treatment of oxidative stress-related diseases.


Introduction
Since the majority of our biological activities only take place in the presence of oxygen, it is necessary for life. However, the oxidation reaction may lead to cell damage, leading to the degradation of various oxygen substrates, proteins, lipids, and DNA. This may cause numerous diseases including inflammation, obesity, diabetes, arthritis, etc. [1,2]. The oxygen paradox contributes to the generation of free radicals which possess a single electron on an oxygen or nitrogen atom, known as reactive oxygen species (ROS) or reactive nitrogen species (RNS) [3,4]. Free radicals are highly unstable. They could be valuable when they are involved in physiological functions or harmful if there is no balance between the defense systems and ROS/RNS, or when the organism is incapable of restricting the destruction triggered by the free radicals, which is known by the oxidative stress [5].
The potent antioxidant activity of both chemical scaffolds (indole and caffeic acid) as shown in Figure 1 encourages our team to design a hybrid scaffold that could have potential antioxidant power. Consequently, development of an efficient indole-caffeic pharmacophore could be enormously important. As illustrated in Figure 2, the methoxy group of melatonin was substituted with an amino group, followed by amide formation by reacting with various caffeic acid analogues to generate the desired amide derivatives (3a-m). The selection process of the functional groups in compounds 3a-m was inspired by their existence in neuroprotectant compounds which showed powerful activities against oxidative stress [47][48][49]. For this objective, a straightforward synthetic approach was used to synthesize the new hybrid indole-caffeic amide analogues 3a-m, and their therapeutic potential against ROS was preliminary assessed (Figure 2).

Chemical Reagents, Purification, and Instrumentation
The general protocols utilized for the chemical synthesis, structure elucidation, and purity of the newly synthesized indole-caffeic acid hybrids are provided in the Supplementary File.  The reaction was carried out at room temperature (25 • C) for 2 h. The excess acetonitrile was evaporated. Work-up was performed using ethyl acetate (EA) and water. The organic solution was evaporated, dried, and purified via flash column chromatography (SiO 2 , n-hexane: EA = 10:1) to obtain the indole-caffeic amide derivatives in suitable yields (Table 1).

Ferric Reducing Power (FRAP) Assay
A stock solution of Trolox (3 mM in methanol) was made, and the following dilutions were prepared at the concentrations of 1500, 1000, 800, 400, 200, 100, and 50 µM. Samples were initially dissolved in DMSO to obtain a 40 mM concentration (depending on the molecular weight of each compound). Then, they were diluted to reach the concentration of 0.2 mM with methanol. The assay was performed as reported [52], with minor modifications. For details, please refer to the Supplementary Materials.

Oxygen Radical Absorbance Capacity (ORAC Assay)
A stock solution of Trolox (2 mM in MeOH) was prepared, and the following dilutions were prepared: 1200, 900, 600, 500, 400, 300, 200, 100, and 50 µM. Samples were initially dissolved in DMSO at concentrations of 40 mM according to the provided molecular weights. Then, samples were diluted with methanol until reaching the concentration of 0.1 mM. The assay was carried out as reported [53], with modifications. Further details are provided in the Supplementary Materials.

Chemical Synthesis
As sketched in Scheme 1, a series of indole-based benzamide and caffeic acid amide analogues 3a-m was synthesized. The amide formation reaction was accomplished via reacting 5-aminoindole (1) in acetonitrile solvent with a variety of commercially available benzoic or caffeic acid derivatives (2). The coupling reagents EDCI and HOBt were used, in addition to DIPEA as an organic base. As shown in Table 1, a variety of indole-based benzamide and caffeic acid amide analogues possessing different chemical substituents were acquired in acceptable yields.

Structure Elucidation of the Newly Synthesized Amide Derivatives 3a-m
The chemical structures of the newly synthesized indole-based benzamide and caffeic acid amide analogues (3a-m) were elucidated using different spectroscopic techniques. The purity of compounds 3a-m was found to be more than 96%. The 1 H NMR spectra of all the synthesized analogues were characterized by two major singlet peaks with high chemical shifts (>9.00 ppm); the proton of the NH group of the indole scaffold and the proton of the amide group (CONH).
As provided in the Supplementary File, the 1 H NMR spectrum of the indole-based benzamide analogue 3a displayed the free para NH 2 group at 5.09 ppm as a broad peak (br) representing the two protons of the amino group. In addition, the amide carbon (CO) appeared clearly in the 13 C NMR spectrum of compound 3a at 165.59 ppm, which confirmed the formation of the amide group. Indeed, 13 C NMR spectra of all the newly synthesized indole-based benzamide and caffeic acid amide analogues (3a-m) showed signals resonating around 163.00-168.00 ppm (CO group of the amide moiety). For compound 3b, the two methoxy groups were found at 3.82 and 55.57 ppm in the 1 H and 13 C NMR, respectively. The 1 H NMR chart of compound 3c was characterized by a singlet long aromatic peak at 6.93 ppm representing the two para phenyl protons of the 4-bromo-3,5-dihydroxy benzoyl moiety. In addition, the protons of the two hydroxyl moieties were detected as a singlet broad peak at 10.29 ppm. In its 13 C NMR chart, compound 3c showed a long peak at 155.60 attributable to the two carbons carrying the two hydroxyl moieties. The 1 H NMR spectrum of the benzamide derivative 3d was characterized presence of three singlet peaks with chemical shifts higher than 9.00 ppm (the indole NH, the amide NH, and the free phenolic OH groups).
The first synthesized caffeic acid amide ((E)-3-(4-hydroxyphenyl)-N-(1H-indol-5yl)acrylamide, 3e) showed these three protons in the range of 9.85-10.98 ppm. In the meantime, one vinylic proton was successfully detected with its characteristic trans J coupling constant of 16.00 Hz at 6.64 ppm. The carbon of the amide moiety appeared at 163.98 in the 13 C NMR spectrum of the caffeic acid amide 3e, while the carbon holding the free phenolic OH group was detected at 159.32 ppm. Similarly, the 1 H NMR spectrum confirmed the synthesis and the final chemical structure of the second caffeic acid amide derivative in this series (3f) by the presence of three singlet peaks in the range of 9.44-10.98 ppm representing the indole NH, amide NH, and phenolic OH groups, in addition to the three protons of the meta methoxy group (m-OCH 3 ) in the aliphatic region (3.83 ppm). Its 13 C NMR chart showed the amide carbon peak at 163.92 ppm, the two carbons bearing the free OH and the methoxy groups at 148.82 and 148.29 ppm, in addition to the aliphatic carbon of the methoxy group at 55.93 ppm.
The 1 H NMR spectrum of compound 3g was characterized by a long peak in the aliphatic region at 3.85 representing the six protons of the two methoxy groups. Meanwhile, its 13 C NMR chart showed the amide carbon chemical shift at 165.01 ppm, a long peak at 147.86 ppm attributable to the two carbons that hold the two methoxy groups, and the characteristic peak of the carbon bearing the free OH at 139.08 ppm. Similarly, 3,4-dihydroxy-N-(1H-indol-5-yl)benzamide (3h) was characterized by the two common singlet peaks at 10.98 and 9.75 ppm representing the NH protons of the amide linkage and the indole ring. In addition, a broad singlet peak of 2H was found at 9.35 ppm, attributable to the two free OH phenolic moieties. The two carbons bearing these phenolic hydroxyl groups were detected in its 13 C NMR spectrum at 148.82 and 145.20 ppm. The methyl group (CH 3 ) of analogue 3i was represented by a singlet peak (3H) in the aliphatic region at 2.34 ppm and a peak at 20.29 ppm ( 13 C NMR).
In addition to the two common singlet peaks of the NH groups of the amide linker and the indole ring at 10.99 and 9.87 ppm, the third caffeic amide analogue 3j also showed a broad peak of 2H representing the protons of the two hydroxyl groups at 9.28 ppm and a doublet peak at 6.58 with a coupling constant of 12.00 Hz attributable to a vinylic proton. The amide CO group was represented at 163.95 ppm ( 13 C NMR), while the two carbons bearing the two hydroxyl groups were represented by two peaks at 147.88 and 146.01 ppm. A vinylic carbon of compound 3j was also successfully detected at 140.11 ppm. In the 13 C NMR spectrum of compound 3k, the amide carbon and the carbon atom holding the free OH group were displayed above 160.00 ppm. In compound 3l, the amide carbon appeared at 165.10 ppm, while the two carbons bearing the free OH and the methoxy groups were displayed at 150.00 and 147.58 ppm. The chemical shifts of the methoxy group in 3l were represented by a singlet peak in the aliphatic region at 56.12 ppm ( 13 C NMR) and 3.86 ppm ( 1 H NMR).
Finally, the chemical structure of the final indole-based caffeic acid amide analogue 3m was confirmed by detecting three singlet peaks with chemical shifts of more than 8.00 ppm representing the three protons of the free OH, the amide NH, and the indole NH. Moreover, the two vinylic protons were clearly identified as two doublet peaks with J values of 16.00 Hz at 7.48 and 6.71 ppm. In addition, the six protons of the two methoxy groups were found as a long singlet peak at 3.83 ppm. The amide carbon appeared at 163.84 ppm, the two carbons bearing the two methoxy groups appeared as a long peak at 148.52 ppm, and the two aliphatic carbons of the methoxy groups showed a singlet peak at 56.37 ppm. These data proved and confirmed the formation and purity of the desired amide derivatives 3a-m.

In Silico Druggability Studies of the Newly Synthesized Amide Derivatives 3a-m
There is no guarantee that a small molecule that possesses a potent interaction with its target protein could be a successful therapeutic candidate. Poor absorption, distribution, metabolism, and excretion (ADME) characteristics may be the reason for this failure. Thus, many promising small molecules fail during the drug discovery process. Moreover, the drug development process is expensive. Accordingly, the pharmacokinetic (PK) characteristics of the indole-based benzamide and caffeic acid amide analogues (3a-m) were evaluated via the SwissADME platform by using distance/pharmacophore models coded as graphbased marks [54]. Using this platform, numerous crucial characteristics can be anticipated including the solubility of the final compounds, their gastrointestinal absorption, and brain entry abilities. During the different steps of the new drug development, these PK factors would constitute the foundation stone of the outcome's anticipation [55].
Another major PK property is the topological polar surface area (TPSA) of a compound which refers to the surface sum over the entire polar atoms, mainly nitrogen and oxygen, together with their associated hydrogen atoms. The TPSA is obtained by subtracting from the molecular surface the area of carbon atoms, halogens, and hydrogen atoms bonded to carbon atoms (i.e., nonpolar hydrogen atoms). TPSA is considered a great metric to improve the ability of a drug to penetrate the cells, which could enhance the efficacy of the synthesized drug candidate. Molecules possessing TPSA > 140 Å 2 are predicted to not be able to cross cell membranes. On the other hand, a TPSA value < 90 Å 2 was found to be essential for a drug candidate to cross the blood-brain barrier (BBB) [56]. Furthermore, compliance with the Lipinski rule of five [57] is another important guide on whether a compound can be taken orally. The outcomes of the in silico PK study are presented in Table 2 and Figure 3.  The predicted PK properties of all the newly synthesized indole-based benzamide and caffeic acid amide analogues (3a-m) revealed that all compounds would have high gastrointestinal absorption. In addition, all compounds displayed compliance to the Lipinski rule of five, indicating their great potential to be promising drug candidates with acceptable PK characteristics. Most compounds showed TPSA < 90 Å 2 , suggesting a potential antioxidant effect also in the brain to battle different neurodegenerative diseases, as hypothesized. These results suggest the PK stability of the indole-based benzamide and caffeic acid amide series.

DPPH Radical Scavenging Activity
First, all the synthesized analogues (3a-m) were initially screened for their scavenging effects on the DPPH radical. Trolox was used as a reference (IC 50 = 33.84 ± 1.01 µM). As illustrated in Table 3, our goal was to discover the antioxidant properties of incorporating the phenolic OH group(s) in addition to some other moieties such as methoxy, bromo, and amino groups in different positions on the phenyl ring, which is attached to position 5 of the antioxidant indole core via two different linkers (depending on the amide type, benzamide or caffeic acid amide). While compounds 3b-e, 3g, 3i, 3k, and 3l showed moderate to weak activity as compared to Trolox, compounds 3a, 3f, 3h, 3j, and 3m were found to have excellent radical scavenging activities with IC 50 values of 95.81 ± 1.01, 136.8 ± 1.04, 86.77 ± 1.03, 50.98 ± 1.05, and 67.64 ± 1.02 µM. It was noted that compounds possessing the 3,4-dihydroxyphenyl moiety exhibited promising activities (the benzamide derivative 3h and the caffeic acid amide derivative 3j). On the other hand, the 4-amino-3-hydroxyphenyl moiety was only able to demonstrate its free radical scavenging effect in compound 3a, which has the 4-amino-3-hydroxy phenyl moiety. The 4-hydroxy-3-methoxyphenyl and 4-hydroxy-3,5-dimethoxyphenyl moieties were only able to show their activities in the caffeic acid amide analogues 3f and 3m, respectively. It was also noticed that the majority of the synthesized caffeic acid amide derivatives (three out of four) were able to show higher free radical scavenging activity compared to their benzamide analogues. It could be the double-bond moiety in these caffeic acid amide derivatives that may increase the capacity of the molecule to interact with the free radicals via enhancing the electron conjugation effect in the whole chemical structure so that they do not engage in a destructive biochemical reaction. Based on this primary screening, the position and nature of substitutions on the phenyl moiety and the presence of the double bond in the middle of the structure were found to be essential factors directly affecting the free radical scavenging activity of these new indole-based amides. Consequently, the most potent derivatives (3a, 3f, 3h, 3j, and 3m) were further evaluated.  3a, 3f, 3h, 3j, and 3m). The ABTS values of the five samples are presented in Table 3. While compound 3a (with the 4-amino-3-hydroxy phenyl moiety) was able to scavenge the radical cation ABTS •+ with an IC 50 value of 33.33 ± 1.96 µM, which is almost the similar potency of the standard Trolox (29.62 ± 1.86 µM), compound 3h possessing 3,4-dihydroxy phenyl moiety showed a higher IC 50 value of 39.98 ± 0.92 µM. Interestingly, three compounds out of the five (3f, 3j, and 3m) showed higher capacities to neutralize the radical cation ABTS •+ than Trolox with IC 50 values of 14.48 ± 0.68, 19.49 ± 0.54, and 14.92 ± 0.30 µM, respectively.

FRAP Assay
The three highly potent analogues (3f, 3j, and 3m) were considered for FRAP and ORAC assays. The FRAP assay assesses the antioxidant properties of the tested compound based on its reducing ability. The values obtained, shown in Table 3, were consistent with the DPPH and ABTS assays. In this study, compound 3j (the caffeic acid derivative possessing 3,4-dihydroxyphenyl moiety) presented the highest antioxidant capacity with a FRAP value of 4774.37 ± 137.20 µM Trolox eq/mM sample, followed by compounds 3m (4-hydroxy-3,5-dimethoxyphenyl moiety containing caffeic acid derivative) and 3f (4-hydroxy-3-methoxyphenyl moiety containing caffeic acid derivative) with values of 2308.7 ± 73.73 and 1951.45 ± 75.97 µM Trolox eq/mM sample, respectively. Based on these findings, it could be concluded that caffeic amide analogue 3j possessing the two phenolic OH groups not only offered the top free radical scavenging capability, but also the strongest reducing power among the tested compounds. Indeed, the antioxidant activity of a small molecule largely depends on both the chemical structure of the compound and the test system. Accordingly, it cannot be fully assessed by one single technique due to the various mechanisms of antioxidant action. As a result, the ORAC test was chosen to be the next further test for these three promising analogues (3f, 3j, and 3m).

ORAC Assay
Through the ORAC test, the antioxidant capacity was investigated of the three highly active compounds (3f, 3j, and 3m) that had demonstrated high antioxidant activity with the previous DPPH, ABTS, and FRAP tests. The ORAC test was intended to validate the results obtained with the previous approaches and extend the activity profile for each tested derivative. All tested compounds exhibited a dynamic ability to reduce the oxidative degradation of the fluorescent molecule, caused by peroxyl radicals. Compounds 3m and 3f showed very high ORAC antioxidant power (9253.47 ± 806.00 and 7293.46 ± 208.48 µM Trolox eq/mM sample, respectively). In a similar way to the previous assay (FRAP), the best antioxidant capacity against the peroxyl radicals was observed for compound 3j (10,714.21 ± 817.76 µM Trolox eq/mM sample).
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/metabo13020141/s1, chemical reagents, purification, and instrumentation details, 1 HNMR, 13 CNMR, purity, and HRMS data of the compounds reported in this study, in addition to the detailed calibration curves of Trolox used in RFAP and ORAC assays.