2.2.3. Candidusin A (**4**) and 4"-dehydroxycandidusin A (**5**)

Candidusin A (**4**) scavenged 32% of radicals in DPPH assay at a concentration 100 μM. 4"-Dehydroxycandidusine A (**5**) was more effective and scavenged 49% of radicals (Table 2). Candidusin B (C-3" hydroxylated analogue of candidusin A) showed 59% DPPH radical scavenging at a concentration of 100 μg/mL [34], i.e., at a molar concentration of the compound that was more than 2.7 times higher.

Earlier, candidusin A (**4**) showed cytotoxic activities against HL-60 (IC50 77.56 μM), A-549 (IC50 19.34 μM), and P-388 (IC50 46.83μM) tumor cells [35], while 4-dehydroxycandidusin A (**5**) exhibited cytotoxic activities against KB (IC50 22.66 μM) and A 549 (IC50 34.01 μM) tumor cells [36]. In our experiments, candidusin A (**4**) and 4"-dehydroxycandidusin A (**5**) had low cytotoxicity against neuroblastoma Neuro2a cells with an IC50 of 75.7 and 78.9 μM, respectively (Table 2). These compounds were used at a non-toxic concentration of 10 μM for the treatment of cells in PD models.

Candidusin A (**4**) did not have any effect on ROS formation in the 6-OHDA-treated cells (Figure 2). 4"-Dehydroxycandidusin A (**5**) decreased the ROS level in 6-OHDA-treated cells by 34%, being more effective as a radical scavenger. Candidusin A (**4**) had no effect on the viability of cells when it was added 1 h before treatment with 6-OHDA, and increased 6-OHDA-treated cell viability by 24% at a concentration of only 10 μM, when it was added 1 h after 6-OHDA (Figure 3c). 4"-Dehydroxycandidusin A (**5**) increased cell viability by more than 80% (when the compound was added 1 h before 6-OHDA) and 62% (when it was added 1 h after 6-OHDA). These effects were observed at a concentration of 10 μM only. When the concentration of compound **5** was reduced tenfold, its neuroprotective effect was not preserved.

In the PQ-induced model, compound **4** decreased ROS formation by 27% at a concentration of 10 μM. Compound **5** was more effective and statistically significantly decreased ROS formation by 19% and 40%, at concentrations of 1 and 10 μM, respectively (Figure 4). Nevertheless, compound **4** did not have any effect on the viability of the PQ-treated cells, but its 4"-dehydroxylated derivative (**5**) statistically significantly increased the viability of these cells by 17% at concentration of 10 μM only (Figure 5).

Thus, the presence of hydroxy groups at C-3" and C-4" in candidusins decreases the radical scavenging activity of these compounds in the cell-free assay. Moreover, the hydroxylation at C-4" results in the significant decreasing their neuroprotective effect.

#### 2.2.4. Mactanamide (**6**)

Mactanamide (**6**) was not cytotoxic against neuroblastoma Neuro2a cells up to 100 μM.

Compound **6** scavenged 15% DPPH radicals in cell-free assays at a concentration of 100 μM (Table 2). In cell experiments, this compound (**6**) demonstrated a significant antiradical effect: at 10 μM, it inhibited ROS formation in the 6-OHDA-treated neuronal cells by 30% (Figure 2).

Mactanamide (**6**) increased the 6-OHDA-treated cell viability by 42% at 10 μM in the 6-OHDA-induced PD model when the compound was added 1 h before neurotoxin. When its concentration was reduced tenfold, its neuroprotective effect was preserved (Figure 3d).

In the PQ-treated cells, mactanamide (**6**) decreased ROS formation by 32% and 37%, at concentrations of 1 and 10 μM, respectively (Figure 4). However, compound **6** did not show any neuroprotective activity on the viability of the PQ-treated cells (Figure 5).

Earlier, mactanamide showed fungistatic activity against *Candida albicans*, and an influence on osteoclast differentiation without any cytotoxicity [31,37]. Antioxidant and neuroprotective properties of mactanamide were demonstrated, for the first time, in this investigation.

Thus, compounds **1**, **2**, **3**, and **6** were non-cytotoxic for Neuro2a cells up to a concentration of 100 μM. Compounds **4** and **5** demonstrated low cytotoxicity with an IC50 of 75.7 and 78.9 μM, respectively. This allowed investigating the neuroprotective activity of all compounds in non-toxic concentrations of 1 and 10 μM. Neuroprotective effects of the compounds were studied in two PD in vitro models using 6-OHDA and PQ as inducers of neuronal cell damage.

Melatonin-like compound **1** demonstrated an effect in increasing cell viability in both models, but the effect on PQ-treated cells was more pronounced. In both cases, neuroprotective effects were accompanied with a decrease of ROS formation in the 6-OHDA- and PQ-treated cells. Melatonin (**1a**) decreased ROS formation in both PD models, but it increased cell viability in the PQ-induced model only.

Polyketides **2** and **3** demonstrated ROS-decreasing effects in both PD models. Nevertheless, these compounds increased cell viability in the 6-OHDA-induced model only.

Candidusin A (**4**) and 4"-dehydrocandidusin A (**5**) have minimal differences between their chemical structures but this has a significant effect on their neuroprotective activity. In the 6-OHDA and PQ models, compound **5** produced a significant increase of cell viability, whereas compound **4** did not demonstrate any effect in the PQ model, and low increased cell viability on the 6-OHDA-treated cells. A similar influence of both compounds on ROS formation in the 6-OHDA- and PQ-treated cells was observed. Compound **4** had no significant effect on ROS formation in the 6-OHDA-treated cells, and decreased ROS formation in the PQ-treated cells, at a concentration of only 10 μM. By contrast, compound **5** was very effective in the 6-OHDA-induced PD model, and decreased ROS formation in the PQ-treated cells at concentrations 1 and 10 μM.

Mactanamide (**6**) demonstrated a significant decrease of ROS formation in both the 6-OHDAand PQ-induced PD models. However, this 2,5-diketopiperasine alkaloid increased viability of the 6-OHDA-treated cells only, and did not have any statistically significant effects on viability of the PQ-treated cells.

It should be noted that DPPH radical scavenging activity was shown for all compounds in varying degrees, and decreasing of ROS formation in 6-OHDA- and PQ-treated cells could be the result of radical scavenging by these compounds. However, differences between the effects of these compounds, on ROS formation and cell viability in different PD models, were observed.

In our investigation, two PD-like cell models, induced by neurotoxin 6-OHDA and pesticide paraquat, were used. Neurotoxins and pesticides share a common mechanism to induce damage to dopaminergic neurons that is correlated with an increased oxidative status caused by high levels of ROS, anions, and free radicals [6]. However, the effect of each of the inducers, on neurons, has the same differences.

6-OHDA has a specific neurotoxic effect on neurons containing dopamine, serotonin, and norepinephrine receptors. The structure of 6-OHDA is similar to dopamine and norepinephrine, and, therefore, this neurotoxin uses the same catecholaminergic transport system (the dopamine and norepinephrine transporters), and causes specific degeneration of dopaminergic and noradrenergic neurons [6]. Inside neurons, 6-OHDA is rapidly autooxidized to hydrogen peroxide and paraquinone, which are both highly toxic to mitochondria, by specifically affecting complex I. This process results in an increase of ROS generation and cell death [38]. Moreover, it was reported that 6-OHDA induces oxidative stress both during its autoxidation to *p*-quinone and, also, during one-electron reduction of *p*-quinone to *p*-semiquinone, catalyzed by flavoenzymes that transfer one electron [39]. In addition to these effects, 6-OHDA-induced cell death is dependent of such intracellular processes as neuroinflammation, mitochondria dysfunction, endoplasmic reticulum stress, and autophagy [40].

Paraquat causes oxidative stress in neuronal cells by another pathway. Divalent paraquat ion (PQ2+) is reduced to monovalent paraquat ion (PQ+) by NADPH-oxidase of mitochondrial complex I. Subsequently, PQ+ accumulates in dopaminergic neurons and reestablishes a new redox reaction intracellularly, leading to the generation of intracellular free radicals, such as superoxide and dopamine-reactive substances. This will eventually lead to dopaminergic neuron cell death [41]. Moreover, PQ toxicity correlates with DNA fragmentation, caspase-3 cascade modulation, and dysregulation of autophagy [5,42].

Metabolic investigations of the molecular mechanisms associated with 6-OHDA and PQ toxicity were carried out by NMR spectroscopy and mass spectrometry. It was shown that PQ selectively upregulated the pentose phosphate pathway (PPP) to increase NADPH reducing equivalents, and stimulate paraquat redox cycling, oxidative stress, and cell death. PQ also stimulated an increasing in glucose uptake, the translocation of glucose transporters to the plasma membrane, and adenosine monophosphate-activated protein kinase activation. In the contract, 6-OHDA did not demonstrate an influence on PPP. In addition, while paraquat induced a reduction in glucose-dependent glutamate-derived glutathione synthesis, 6-OHDA treatment increased this process [43,44].

In this study, we observed time differences between 6-OHDA and PQ effects on ROS formation in Neuro2a cells (Figure S16). 6-OHDA caused an increase of ROS level of 30% for 30 min after addition to the cell suspension. The effect of PQ on ROS formation was insignificant after 30 min, and an increase of ROS levels in cells, by 39%, was observed 1 h after adding of PQ to the cell suspension.

Compounds **2**, **3**, and **6** demonstrated neuroprotective effects in the 6-OHDA-induced PD model only. For this reason, they could protect Neuro2a cells against the damaging influence of products of 6-OHDA autooxidation, due to their antioxidant properties. Compound **5** increased the viability of 6-OHDA-treated cells by 80%, but it increased viability of PQ-treated cells by 17% only. This suggests the same mechanism of action.

On the other hand, compounds **1** and **1a** were more effective in the PQ-induced model, and increased cell viability by 40% and 24%, respectively, whereas in the 6-OHDA-induced model, compound **1** increased cell viability by 23% only, and melatonin (**1a**) was ineffective.

It was earlier published that melatonin and some related compounds demonstrated antioxidant activity in cell-free assays [45], and different neuroprotective effects in the in vitro experiments [46–48]. Pre-treating of PC12 cells with melatonin for 3 h increased viability of the cells, and prevented apoptosis in the 6-OHDA-induced PD model [49,50]. In addition, it was reported that melatonin diminished caspase-3 enzyme activity, cleavage of DNA fragmentation factor 45, and DNA fragmentation observed in the MPTP-treated neuroblastoma cells [46]. For this reason, melatonin-related compound **1** could influence on viability of the PQ-treated cells in a similar manner.
