MB has been studied sporadically for over 100 years with its initial biological activity uncovered in the 1890s 
. Recently, discovery of its cognitive enhancing and neuroprotective effects has reinvigorated research into MB. MB and TB’s oxygen enhancing effects were initially observed in aerobic metabolism 
. Although the initial results were promising, research into the oxygen enhancing properties of MB did not continue until the 1960s, at which time MB’s actions on the electron transport chain were identified and MB was shown to accept electrons from NADH and transfer them to cytochrome c independent of coenzyme Q10 in isolated live mitochondria 
. Recently, we have demonstrated MB’s neuroprotective action and its relationship to MBs electron shunt 
. To elucidate the structural characteristics necessary for MBs mechanisms, we have compared MB to a selected group of MB related compounds.
Our results indicate that the MB related compounds can be divided into four groups based on their structure-activity relationships in neuroprotective and bioenergetics assays. The first group consists of compounds containing only the phenothiazine nucleus (phenothiazine and 2-chlorophenothiazine). These compounds were highly efficacious and potent in the IAA, glutamate, and rotenone neurotoxicity assays, but had no effect on anaerobic glycolysis, cellular oxygen consumption or the complex I-III shunt. The second group of compounds are those with amine side chains attached to the 3, 7 carbons of the phenothiazine nucleus (MB and TB). Both MB and TB had high potencies and efficacies in the neurotoxicity assays, coupled with their ability to enhance cellular oxygen consumption and decrease anaerobic glycolysis. Both compounds were also unique in their ability to act as an intermediate between complex I and cytochrome c. The third group was made up of the compounds with a side chain attached to the 10 nitrogen of phenothiazine (chlorpromazine and promethazine). Promethazine and chlorpromazine were less potent in the glutamate, IAA, and rotenone neurotoxicity assays. In addition, promethazine and chlorpromazine are less efficacious in the IAA assay as compared to the two previous groups with the exposed nitrogen motif in the phenothiazine nucleus (MB and phenothiazine). Besides being less potent, neither promethazine nor chlorpromazine had any effect on cellular oxygen consumption, anaerobic glycolysis, or the complex I-III shunt. The fourth group contained only one compound, NR. NR has a substitution of a nitrogen in place of the 5 sulfur yielding a phenazine nucleus with side chains on the 3, 7 carbons. NR had decreased neuroprotective potency relative to MB in the glutamate, rotenone and IAA assays as well as a decreased efficacy in the IAA assay. However, NR was capable of enhancing cellular oxygen consumption, but did not aid in electron transfer between mitochondria complexes I and III.
The addition of a side chain to the 10 nitrogen caused a significant loss of potency and efficacy as demonstrated by the differences between phenothiazine and chlorpromazine in the glutamate, IAA, and rotenone assays. Chlorpromazine and promethazine have previously been reported to have minor protective actions with micromolar potency, which corresponds to our results from the neurotoxicity assays 
. In addition, the neuroprotective effects of phenothiazine and chlorpromazine were previously compared in a rotenone neurotoxicity assay highlighting phenothiazine’s robust neuroprotection as compared to chlorpromazine’s lack of efficacy 
. This was later elaborated on in vivo
in a C elegans model of Parkinson’s disease, with phenothiazine again being highly efficacious 
The position 5 sulfur is as equally important as the availability of the free 10 nitrogen motif evidenced by the differences between TB and NR in the glutamate, IAA, and rotenone assays. The substitution of a nitrogen in place of the sulfur in the heterocyclic nucleus of the molecule (phenothiazine backbone replaced with phenazine backbone) significantly decreased both the potency and efficacy of NR as compared with MB.
MB’s neuroprotective effects have been demonstrated in models of Alzheimer’s disease, Parkinson’s disease, stroke, optic neuropathy, and hypoxia 
. In addition, phenothiazine has been demonstrated to be protective in models of Parkinson’s disease employing rotenone or MPP+
. However, previous studies have not compared the effects of MB and phenothiazine together. Our results indicate that MB and phenothiazine have very similar neuroprotective effects due to the availability of their heterocyclic nitrogen and the presence of the position 5 sulfur. The two structural analogs for phenothiazine and MB, 2-chlorophenothiazine and TB respectively, also exhibit nanomolar neuroprotective effects in our neurotoxicity assays. However, our cellular bioenergetics and mitochondria lysate results indicate an apparent difference between MB and phenothiazine.
The distinct neuroprotective action of MB was suggested by our mitochondrial lysate oxidation assay, where MB requires mitochondria and NADH to reduce oxidative stress. We predict that MB accepts electron(s) from NADH via mitochondria complex I and is reduced to leuco-MB, which can act as a direct free radical scavenger and recycle back to the oxidized form of MB. This unique action of MB makes it a mitochondria specific regenerative anti-oxidant. On the other hand, phenothiazine and 2-chlorophenothiazine can function as direct free radical scavengers independent of the presence of mitochondria and NADH. In addition, the enhancement of complex IV expression and activity associated with MB was not observed with 2-chlorophenothiazine indicating a distinct mechanism between these two compounds. With the addition of a side chain to the 10-nitrogen, chlorpromazine enhanced the oxidative reaction independent of the presence of mitochondria and NADH explaining its low neuroprotective potency.
MB has previously been shown to directly accept electrons from NADH, NADPH, and FADH2
. We predicted that MB derivatives derive their protective actions by acting in an electron donor/acceptor capacity between mitochondria complexes I-III similar to MB 
. Surprisingly, only two compounds, MB and TB, were capable of increasing the rate of cytochrome c reduction in our complex I-III assay. The identified action of MB on complex I-III is consistent with its action on mitochondrial oxidative phosphorylation as we published previously 
. Similar to MB, we also observed that TB and NR increased cellular oxygen consumption and decreased lactate production although the action and pattern of NR on ECAR was different from that of MB and TB. Interestingly, MB, TB, and NR all have similar negative redox potentials. MB, TB, and other phenazine and phenothiazine derivatives have previously been shown to enhance electron transfer in a microbial fuel cell system 
. MB, NR, and TB have similar structural characteristics distinct from the other derivatives suggesting that the amine side chains are likely the major factor for the negative redox potential of these compounds, thus, their action on oxygen consumption and lactate production. In addition, since NR, with the substitution of a nitrogen for sulfur in the heterocyclic ring, has a very weak neuroprotective effect and does not function as an alternative electron transfer carrier in mitochondria, we predict that the protective effect of MB is likely related to its action on electron transfer independent of its effect on oxygen consumption and lactate production.
In conclusion, our structure-activity relationship study of MB has demonstrated the distinct anti-oxidant properties of MB. MB acts on superoxide generated due to the blockage of the mitochondria electron transport chain by providing an alternative mitochondrial electron transfer carrier to bypass complexes I-III. In addition, reduced leuco-MB can directly scavenge superoxide and recycle back to the oxidized form MB (). As a more than one century old drug, MB has been used clinically for the treatment of multiple diseases with well known pharmacokinetics in humans for both acute intravenous and chronic oral administration 
. These make MB and some of its derivatives ideal candidates for future investigations for the treatment of neurodegenerative diseases.
Schematic illustration depicts the novel neuroprotective mechanism of MB.