The pathophysiology of PD remains unknown. There are two hypotheses of cell death of DA neurons based on two histopathological hallmarks in PD, i.e., the α-synuclein hypothesis (an α-synucleinopathy) and the NM hypothesis (Figure 1). The α-synuclein hypothesis on the possible molecular mechanism of neuronal death of DA neurons in sPD may be summarized as follows: Mitochondrial oxidant stress by various exogeneous or endogenous factors may produce mitochondrial dysfunction, especially complex I deficiencies [82,83,84,85,86,87], oxidation of DA in cytoplasm [64,88,89], and formation of oxidized DA accumulation, especially toxic 3,4-dihydroxyphenylacetaldehyde (DOPAL), formation of toxic ROS, accumulation of cytotoxic fibrillar aggregates of α-synuclein oligomers, mitophagy/autophagy dysfunction, and neuroinflammation [90,91,92,93,94,95,96,97]. DOPAL is thought to accumulate in PD due to the low aldehyde dehydrogenase activity that oxidizes DOPAL to DOPAC in the SN in PD [98] and DOPAL generates potential reactive intermediates as causative agents for its neurotoxicity [99,100].

Since the 1990s, it was found that Lewy bodies mainly consisted of α-synuclein protein, and that the fibrillar oligomers produced by misfolding of the protein were neurotoxic and may be related to the cause of DA cell death [95,101,102]. In 1997, mutation of the α-synuclein gene (SNCA) was found to cause a familial PD (PARK1) [9,10]. Prion-like properties of α-synuclein were proposed by Braak (Braak hypothesis); α-synuclein produced in the intestine or olfactory bulb might spread via the vagus nerve or olfactory pathway to the midbrain and basal ganglia by cell-to-cell transfer [103,104,105]. α-Synuclein aggregates may spread from neuron to neuron, apparently transmitting the disease process through the brain. However, precisely how α-synuclein aggregates build-up and spread in this way is still unknown. Another question is that α-synuclein is not specific to PD, and also found in Lewy body disease (LBD) and multiple system atrophy (MSA) [106]. Aggregates of α-synuclein in distinct synucleinopathies, PD and MSA, have been proposed to represent different conformational strains of α-synuclein [107]. Even with these questions about the α-synuclein hypothesis, α-synuclein has been extensively examined in relation to DA neuron death in PD. The p62 protein normally assists in autophagy, a waste-management system that helps cells get rid of potentially harmful protein aggregates. In cell and animal models of PD, p62 is S-nitrosylated at abnormally high levels in affected neurons. This alteration of p62 inhibits autophagy, causing a build-up of α-synuclein aggregates, which in turn, leads to the secretion of segregates by affected neurons, and some of these aggregates are taken up by nearby neurons [108]. There are many references to support the cytotoxic effects of α-synuclein in vitro, especially in cell culture systems [10,20,95]. A downsized and optimized intracellular library-derived peptide prevents α-synuclein primary nucleation and toxicity without impacting upon lipid binding [109]. An animal model of PD with prodromal symptoms as in human PD has been reported [110]. The α-synuclein gene, SNCA, is a risk gene for sPD. A bacterial artificial chromosome transgenic mouse harboring SNCA and its gene expression regulating region in order to maintain the native expression pattern of α-synuclein showed prodromal symptoms in human PD such as RBD and anosmia without motor symptoms [110,111]. This mouse model is similar to human sPD and shows that α-synuclein alone can cause PD [110].

In the NM hypothesis, NM alone is related to DA neuron death. This review focuses on two independent ways of PD pathology, α-synuclein and NM. While both pathways may indeed lead to dopaminergic cell death, a decisive link between them is proposed “iron”, as pointed out recently by Riederer et al. [106,112,113]. The pathophysiology of NM decrease in the SN of DA neurons as a hallmark of PD remains unknown, especially in its relation to DA neuron death. In parallel with the elucidation of the chemistry and biosynthesis of NM in the DA neurons in the SN in PD, the physiological and pathological roles on NM have been studied since 2000s. NM in the SN increases gradually during aging in healthy subjects [21]. In contrast, NM decreases in PD. In PD, DA neurons containing NM in the human SN might preferentially degenerate, in parallel with the marked reduction in NM in the SN [114]. This fact, although controversial, suggests that NM is related to neurodegeneration and DA neuron death.

Alternatively, NM in DA neurons has generally been regarded as acting for neuroprotection, since NM inactivates toxic free radical species via its ability to chelate transition metals, especially iron. Iron also accumulates in DA neurons [112,115,116,117]. Iron is bound to NM in the ferrous (II) iron form, a redox-active form that is involved in a Fenton-like reaction to produce toxic free radical species. NM also eliminates various toxic substances in cytoplasm. Thus, NM may act for neuroprotection also in vivo. However, during the progress of PD, the release of toxic substances bound to NM owing to intracellular NM degradation may result in activation of microglia to release cytotoxic cytokines that produce neuroinflammation and neurodegeneration [78,118]. PD occurs spontaneously only in humans. To produce the PD phenotype in various models of PD in mammals such as in mice and rats that nearly lack NM in the brain, it is necessary to trigger the DA neurodegeneration by some toxic chemicals such as 1-phenyl-4-methyl-1,2,3,6-tetrahydropyridine (MPTP) that inhibits the mitochondrial complex I [97]. Vila’s group reported that NM accumulation in DA neurons during aging over a threshold causes DA neuron death and PD phenotype [119,120,121]. They created a rat model of human PD by overexpression of human NM in the right SNpc by stereotaxic injection of an adeno-associated viral (AAV) vector expressing human tyrosinase [119]. The rats showed age-dependent production of human-like NM within nigral DA neurons, up to levels in elderly humans. Intracellular NM aggregation above a specific threshold is associated with an age-dependent PD phenotype, including hypokinesia. Enhancing lysosomal proteostasis reduces intracellular NM and prevents neurodegeneration in tyrosinase-overexpressing rats. Intracellular NM levels may set the threshold for the initiation of PD. Furthermore, extracellular NM leaked from dead NM-containing DA neurons may activate microglia to produce neuroinflammation and to further promote DA cell death [122].

NM is a hot candidate to trigger PD and/or to lead to progressed degeneration of PD, however, there are unsolved problems arising from the treatment of PD with levodopa: (1) Post-mortem data have not shown an increase of NM in surviving NM containing dopaminergic neurons of the SN after levodopa long term treatment and (2) long term levodopa treatment has not demonstrated a significant increase in the progression of PD. The NM theory well fit to the phenotypes of human sPD. In addition, there is much evidence on the cytotoxicity of α-synuclein [10,19,123].

Recently, the role of NM in inducing α-synuclein expression and aggregation has been suggested as a mechanism for this pigment to modulate neuronal vulnerability in PD [124]. α-Synuclein reacts with tyrosinase, and the chemical modifications on the tyrosinase-treated α-synuclein strongly influence its aggregation properties and increase the toxicity, and α-synuclein may influence synthesis of NM [125,126]. Iron redox chemistry promotes the aggregation of α-synuclein, and protein-metal complex aggregates are directly involved in ROS production, exacerbating the oxidative damage [127]. Furthermore, DA neurons easily express MHC-I, and induction of MHC-I is promoted by activation of microglia either by α-synuclein or by NM, as well as by interferon gamma or high cytosolic DA and oxidative stress [128]. The activated microglia in PD brains express major histocompatibility complex class II (MHC-II) molecules. The number of MHC-II positive microglia in the SN and putamen increase as the neuronal degeneration of the SN proceeds [129].

An evolution theory has been proposed to explain human-specific PD based on the greater development of human cerebral cortex than that of basal ganglia [130,131,132]. Clinically, PD is a systemic disease, and it is difficult to explain the degenerative processes, especially in the autonomic nervous system, exclusively by NM theory, although there is accumulating evidence that the pathogenesis of PD is complex and involves energy metabolism disorders, oxidative stress, proteasomal abnormalities, α-synuclein accumulation, alterations of gut microbiota metabolites, and neuroinflammation [133,134]. In this context, the evolutional point of view on the NM system and α-synuclein system is also of interest.