Understanding the Pre-Clinical Stages of Parkinson’s Disease: Where Are We in Clinical and Research Settings?

Abstract

Parkinson’s disease (PD) is the second most common neurodegenerative disorder in the world. PD is characterized by motor and non-motor symptoms, but the diagnosis primarily relies on the clinical assessment of postural and movement abnormalities, supported by imaging and genetic testing. It is widely accepted that the disease process begins decades before the onset of overt symptoms. Emerging evidence suggests that neuroinflammation plays a central role in the pathogenesis of PD, particularly during the pre-clinical phase. Activated microglia, increased levels of pro-inflammatory cytokines, and persistent oxidative stress have all been associated with the gradual loss of dopaminergic neurons. Although earlier detection and diagnosis remain elusive, achieving these goals is crucial for advancing prevention and disease-modifying strategies. Clinical studies are ongoing. To fill the gap, research models that recapitulate the chronic disease progression of PD are crucial to test preventive and disease-modifying strategies. This review briefly summarizes clinical knowledge on PD as a starting point for improving research models. Furthermore, we will critically evaluate how the existing models have been utilized and highlight opportunities to overcome their limitations and enhance the translational relevance to clinical application.

1. Introduction

Despite the heterogeneity of motor and non-motor symptoms of Parkinson’s disease (PD), diagnosis is based primarily on the cardinal motor symptoms that include resting tremor, bradykinesia, rigidity, and postural instability. By the time a diagnosis is made, neurological damage, namely, dopamine neuron (DOPAn) loss in the substantia nigra (SN) and the consequent loss of dopamine (DA) in the putamen and caudate nucleus, have already advanced. The clinical symptoms observed at this stage are manifestations of pathological processes that began decades earlier (Figure 1) [1,2,3,4,5,6,7,8,9,10,11,12,13,14]. The progression of PD can be retrospectively categorized into the following four phases: (1) the clinical phase (post-diagnosis), (2) the prodromal phase, (3) the pre-clinical phase, and (4) the recently added risk phase (Figure 1) [1,12,15]. Among these phases, the clinical phase is the most extensively studied, both in human and research models. However, a significant limitation of focusing on the clinical stage is that observed molecular changes likely reflect the downstream consequences of the disease rather than the initial pathogenic mechanisms. As a result, such findings may not accurately capture the early biological processes involved in disease initiation. The other three phases encompass the pre-diagnostic stages of PD, which are critical for understanding its onset and for identifying potential preventive or disease modifying strategies. Due to the nonspecific nature of prodromal signs (see later) and lack of brain tissue from presymptomatic timepoints, human-based information on the molecular events underlying early disease progression in PD remains scarce. Thus, research models that recapitulate this time period are the key to solving this problem.

In this review we will first recap clinical knowledge and emergent hypotheses of PD onset and progression. We will then move on to how the PD models have complemented clinical research. We will conclude with suggestions for how basic science can fill knowledge gaps resulting from limitations of human studies.

2. What We Know About PD: The Clinical Readouts

A comprehensive review of PD is beyond the scope of this paper; thus, the following sections offer only a summary. Nevertheless, this background is necessary for modeling the disease and identifying potential readouts for research into the early stages of the disease.

2.1. Post-Diagnostic Phase

A PD diagnosis is based on the neurological exam focusing on hallmark motor symptoms [2,12,13,14], supported by imaging and family history, especially if genetic mutations are involved. Motor symptoms emerge with the loss of DA and DOPAn in the SN pars compacta and loss of DA in terminal regions in the caudate and putamen. The evolution of both motor symptoms and DA neuronal decline is described by the Hoehn & Yahr (H&Y) score (Figure 1) [16,17]. Autopsy data reveal rapid DA neuron loss in SN during the prodromal stage (see Section 2.1.1), with symptoms emerging with 50–70% of DOPAn loss in the SN, followed by a slower progression driven by multiple mechanisms [18,19,20,21,22,23].

Among the multiple pathologic mechanisms potentially responsible for DOPAn loss, protein aggregation (e.g., α-synuclein–αSyn, ubiquitin), proteasomal dysfunction, inflammation, mitochondrial dysfunction, reactive oxygen species (ROS) and nitrogen oxygen species (NOS), reduction in the anti-oxidant defense (GSH, reduced glutathione; SOD, superoxide dismutase), increased iron, altered calcium homeostasis, glutamate excitotoxicity, and reduced trophic factors, have all been described [2,23,24,25,26,27,28,29,30,31,32,33,34].

Based on the Braak hypotheses, linking αSyn accumulation and Lewy bodies (LBs)’ formation to disease progression, brain damage starts in non-DA structures of the lower brainstem or in the olfactory bulb (stages I–II), leading to the prodromal symptoms (see later); reaching the SN (stage III), resulting in the death of DOPAn and motor symptoms; and then spread caudally (stage IV–VI), contributing to more severe PD symptoms [19,35,36,37,38]. It has been suggested that αSyn may originate in the intestine and reach the brain via the vagus nerve (see risk phase). Based on this hypothesis, the clinical evaluation of αSyn in different peripheral sites (e.g., blood; cerebrospinal fluid, CSF; skin; and colon biopsies) is ongoing with the hope of achieving an early diagnosis of PD (see Figure 1) [12,39,40,41]. However, the pivotal role of αSyn has received increasing criticism [23,26,37,38,42,43,44]. Arguments against the synucleinopathy hypothesis include the following: (i) αSyn aggregates are not exclusive to PD [37,45]; (ii) Braak “pathology” (even staging 4 to 6) is reported in individuals without neurological impairment [46,47,48,49]; (iii) LBs are not always present in PD patients with motor symptoms [50]; (iv) the pattern of synucleinopathy described by Braak is inconsistent with observations, and there is no relationship between Braak’s stage and the clinical progression of PD described by the H&Y score [37,38,42,51]; (vi) Braak’s staging describes the accumulation of αSyn, not DOPAn loss, which has been reported to precede αSyn deposition in SN [23,52]; and finally, (vii) it is unclear whether αSyn-related pathological mechanisms correlate with neurologic dysfunction [2,26]. From a molecular point of view, αSyn is abundantly expressed in healthy brains, where it plays physiologic roles at the cell membrane and at the synapse [45,53], and it does not induce major histocompatibility complex II (MHC II) or persistent microglial activation [54,55]. αSyn misfolded fibrils can provoke an immunogenic response in the brain and aggregate inside the cells forming LBs [45,56]. PD has multiple manifestations, from rapidly progressing prominent non-motor signs that do not respond to treatments to mild progressing, predominant motor signs that do respond to drugs, and to mixed manifestations and prognoses [34,57]. Braak staging described progression in a specific subtype of PD with early onset and prolonged duration [38,43,58,59,60]. If synucleopathy is not the mechanism for DOPAn cell death and does not provide a reliable outcome measure, this has major translational implications for developing pre-clinical PD models.

The second most reported mechanism underlying neuronal loss is microgliosis and inflammation. Microgliosis, astrogliosis, and the presence of CD8⁺ and CD4⁺ adaptive immune T cells have been documented in the pons, basal ganglia, frontal and temporal cortical regions of PD human brains [33,61]. These may lead to the production of NOS, ROS, cyclooxygenase (COX) activation; release of the pro-inflammatory cytokine C-X-C motif chemokine ligand 12 (CXCL12); tumor necrosis factor alpha (TNFα); interferon-γ (IFNγ); and interleukin 6 (IL6) and interleukin 1β (IL1β) [62,63,64,65]. In line with this, increases in TNFα, IL1β, interleukin 2 (IL2), and interleukin 10 (IL10) are found in post mortem human PD brains and the serum of PD subjects [66,67,68]. These findings are likely more related to the late stages of the disease than to causative mechanisms [65]. However, glial cells also have homeostatic and protective roles during neurodegenerative conditions [69,70], clearing extracellular debris, and releasing trophic factors like the brain-derived neurotrophic factor (BDNF) and glial-derived neurotrophic factor (GDNF) [65]. Thus, glial activation is not synonymous with neuronal death per se. The dynamic role of glial cells during PD progression needs further study to distinguish cause from effect. Notably, serum TNFα levels in PD patients correlate with cognitive and mood decline [71,72], sleep disorders [71,72,73], and overall H&Y progression [24,72,73,74,75,76]. Moreover, infections may trigger PD symptoms, and PD-related genes overlap with inflammatory pathways, supported by reduced PD incidence or severity in patients undergoing anti-inflammatory treatment [22,61,69,70,77,78,79,80,81,82]. Thus, inflammation is hypothesized to work in two separate phases of PD, it likely precedes the clinical diagnostic phase, and may also drive neurodegeneration in the advanced stages of PD [83].

Mitochondrial dysfunction and redox imbalance are also key contributors to PD. Post mortem studies reveal mitochondrial complex-I activity and oxidative stress in post mortem PD brains [24,84,85,86,87]. A 4-fold higher expression of heme oxygenase 1 (HMOX1), a sensor for ongoing redox stress, has been detected in the SN but not in other brain regions of PD patients compared to age-matched controls [88,89,90]. HMOX1 hyperactivation may enhance ROS production by iron accumulation, a well-known contributor to neurodegenerative diseases [91,92,93,94].

Redox stress and inflammation are intimately linked [95,96], and neuro-inflammation and redox stress may lead to protein disruption and αSyn aggregation [62]. The time course of these processes is not fully understood.

While SN lesions (DA and DOPAn loss) are responsible for symptoms at diagnosis (Figure 1), non-motor symptoms worsen and become predominant and potentially lethal in the late stage of clinical PD (Figure 1) [97,98,99,100]. Although decreased DA, DA transporters, and DOPAn in non-SN regions have been reported (Figure 2) [99,101], they do not fully explain non-motor symptoms (Figure 1). Instead, the degeneration of noradrenergic (NA), serotonergic (SE), and cholinergic (Chol) systems in other brain regions contributes significantly, with 30–90% loss in these neurons by the late stages [26,57,98,99,100,102,103,104,105,106,107,108,109,110]. These systems are linked to sleep disorders, non-DA-responsive gait and balance impairments, dementia, anosmia, and cognitive decline [98,101,108,109,111,112,113,114,115,116,117,118,119,120]. These circuits interact extensively with DA circuits [98,108,109,117,118,119,120].

Many non-motor symptoms have also been documented in the prodromal phase (Figure 1). Determining the mechanisms that contribute to the involvement of these neural circuits, from the prodromal phase to clinical PD, should help to reveal the real triggers of the disease.

2.1.1. Prodromal Phase

Despite the age-related onset of clinical PD between 50 and 60 years of age, it is well known that the disease starts much earlier, possibly decades before the development of diagnostic symptoms [1,3,9,11,26,121]. At that stage, non-motor symptoms (Figure 1), such as olfactory impairment potentially strictly linked with cognitive problems [42], constipation, apathy, sleep disorders, anxiety, depression, REM sleep behavior disorders (RBD), hyposmia, fatigue, excessive daytime sleepiness, depression, and pain are usually already present, indicating that the neuropathologic mechanisms have begun [1,2,3,4,5,6,7,8,9,10,11].

Imaging studies have revealed early changes in neural circuits between patients with PD and healthy controls [10,122]. Decreased DA transporter binding has been measured in the nigrostriatal pathway in PD, with greater decrease in the terminal regions (caudate and putamen) than in the SN [123]. This suggests a retrograde degeneration of DA neurons in PD and opens a temporal window for improving the therapeutic efficacy. Bidirectional association between sleep disruption and dopaminergic loss is present. Aberrant αsyn deposition, neuroinflammation, and glymphatic system disruption are powered by sleep deprivation [124,125].

Recent experimental data suggest inflammation as a critical factor in PD (see later, in the risk phase section, and Figure 1). In agreement, physical activity throughout life shows neuroprotective action, possibly by decreasing systemic inflammation, or by maintaining the high level of BDNF, known to be decreased in PD [126,127]. Uric acid and deglycase J1 (DJ1) are under consideration as potential inflammatory-related biomarkers [10,128,129,130] (Figure 1). HMOX1 upregulation in plasma and saliva has been suggested as an early marker of PD onset [131,132,133]. Following Braak’s hypotheses, the evaluation of αSyn quantification in blood, CSF, and peripheral tissues is ongoing [10,39,122,128,134,135]. Pathological αSyn inclusions can also be found throughout the gastrointestinal tract up to 20 years before PD diagnosis, supporting Braak’s model of PD pathophysiology and making constipation one of the earliest identifiable prodromal features [124,125,136,137]. Abnormal αsyn deposition localized in the olfactory bulb of PD subjects has 95% of sensitivity and 91% of specificity compared to healthy elderly controls. In an MRI study on the anatomical changes in brain structures of an aged human population, PD patients showed a significant reduction in olfactory bulb volume compared to heathy controls. Microstructural degradation of the olfactory tract and substantia nigra are linked to the progressive dysfunction of the putaminal dopaminergic area. Moreover the olfactory function has been reported to correlate with the integrity of other neurotransmitter systems in PD [124,125] (Table 1).

Unfortunately, the predictivity of these markers is currently weak and additional research is needed [1]. Future efforts using open access datasets, bioinformatics, and “omics”-based approaches will likely play a role in optimizing markers to increase their predictivity [15,138,139,140,141,142,143,144,145,146,147]. The International Parkinson and Movement Disorder Society has established research criteria for identifying prodromal PD. These criteria include markers such as RBD, scores on the Unified PD Rating Scale (UPDRS), olfactory loss, constipation, excessive daytime sleepiness, symptomatic hypotension, severe erectile dysfunction, urinary dysfunction, and depression [148]. Given the absence of a clear neuroprotective or disease-modifying therapy for PD, the potential ethical issues of disclosing disease risk in a nonmedical context, and the uncertainty inherent in this emerging field, it is important to approach their use with caution (Table 1).

Table 1. Pre-clinical markers of the risk of developing PD under evaluation and suggestions from the models.

Marker Under Evaluation	Rational	Sampling	Supportive Molecular and/or Pathologic Evidences	References
αSyn presence	Higher in early PD patients. Suggested negative correlation with cognitive decline (CSF). GI: based on Braakt hypothesis.	Blood, CSF, GI	Still debated if its plasma level in PD subject is higher than in healthy individuals. CSF: usually reported lower in PD subject vs. healthy controls. GI: detected.	[10,39,122,128,134,135,136,137,148]
Uric acid	Higher levels are associated with a significantly decreased risk of PD.	Blood and urine	Anti-oxidant action.	[128]
DJ1	Higer in early PD patients.	CSF	Plays a role in mitochondrial dysfunction, oxidative stress, and chaperons activity.	[128]
HMOX1	Higher in early-on-set PD patients.	Plasma and saliva	Sensor of redox stress. Tentative protection by bilirubin production. By the other side: potentially dangerous by producing iron.	[68,131,132,133,149]
Bilirubin	Higher serum bilirubin level in early-stage PD.	Serum, plasma	Elevated serum or plasma bilirubin levels in PD may result from the overexpression of HMOX1, which leads to anti-oxidant and anti-inflammatory effects and may contribute to neuroprotection.	[150,151,152]
DA	Detecting DOPAn initial suffering that preceed DOPAn loss and the appearance of cardinal motor signs.	DAT-scan	Decrease in DA transporter binding.	[10,122,123]
Additional information from PD models
DA	Retrograde degeneration of DA in the nigrostriatal system.	In models	Synaptic spine number and neurite length decrease after MPTP/rotenone treatment. VMAT2 blockage caused by toxins.	[153,154,155]
NE, SE, Chol reduced levels in the extra nigrostriatal areas	Hyposmia, REM sleep deficits and disturbances.	In models	Suggested dysfunction of the neurological circuits. Correlation among symptoms and neurotransmitter’s level.	[156,157,158,159,160,161,162,163]
Occurrence of gastrointestinal symptoms in prodromal and frank PD	Constipation, reduced intestinal motility.	In models	Alterations in the resident neuronal populations of the GI tract. Increased αSyn detected.	[164,165,166]

Legend: Clinical trials aim mainly in identifying and validating novel and precocious biomarkers for predicting the risk of developing PD. Models might add information on the correlation among prodromal molecular changes and prodromal symptoms, suggesting new markers to follow at clinical level with even minimally invasive approaches (imaging, GI biopsy). Abbreviations: αSyn: αSynuclein; PD: Parkinson’s Disease; CSF: cerebrospinal fluid; GI: Gastrointestinal; DJ1: deglycase J1; HMOX1: heme oxygenase 1; DA: dopamine; DOPAn: dopaminergic neurons; DAT-scan: dopamine transporter single-photon emission computed tomography; MPTP: 1-Methyl-4-phenyl-1,2178,3,6-tetrahydropyridine; VMAT2: vesicular monoamine transporter 2; NE: norepinephrine; SE: serotonin; Chol: cholinergic; and REM: rapid eye movement.

Table 1. Pre-clinical markers of the risk of developing PD under evaluation and suggestions from the models.

Marker Under Evaluation	Rational	Sampling	Supportive Molecular and/or Pathologic Evidences	References
αSyn presence	Higher in early PD patients. Suggested negative correlation with cognitive decline (CSF). GI: based on Braakt hypothesis.	Blood, CSF, GI	Still debated if its plasma level in PD subject is higher than in healthy individuals. CSF: usually reported lower in PD subject vs. healthy controls. GI: detected.	[10,39,122,128,134,135,136,137,148]
Uric acid	Higher levels are associated with a significantly decreased risk of PD.	Blood and urine	Anti-oxidant action.	[128]
DJ1	Higer in early PD patients.	CSF	Plays a role in mitochondrial dysfunction, oxidative stress, and chaperons activity.	[128]
HMOX1	Higher in early-on-set PD patients.	Plasma and saliva	Sensor of redox stress. Tentative protection by bilirubin production. By the other side: potentially dangerous by producing iron.	[68,131,132,133,149]
Bilirubin	Higher serum bilirubin level in early-stage PD.	Serum, plasma	Elevated serum or plasma bilirubin levels in PD may result from the overexpression of HMOX1, which leads to anti-oxidant and anti-inflammatory effects and may contribute to neuroprotection.	[150,151,152]
DA	Detecting DOPAn initial suffering that preceed DOPAn loss and the appearance of cardinal motor signs.	DAT-scan	Decrease in DA transporter binding.	[10,122,123]
Additional information from PD models
DA	Retrograde degeneration of DA in the nigrostriatal system.	In models	Synaptic spine number and neurite length decrease after MPTP/rotenone treatment. VMAT2 blockage caused by toxins.	[153,154,155]
NE, SE, Chol reduced levels in the extra nigrostriatal areas	Hyposmia, REM sleep deficits and disturbances.	In models	Suggested dysfunction of the neurological circuits. Correlation among symptoms and neurotransmitter’s level.	[156,157,158,159,160,161,162,163]
Occurrence of gastrointestinal symptoms in prodromal and frank PD	Constipation, reduced intestinal motility.	In models	Alterations in the resident neuronal populations of the GI tract. Increased αSyn detected.	[164,165,166]

Legend: Clinical trials aim mainly in identifying and validating novel and precocious biomarkers for predicting the risk of developing PD. Models might add information on the correlation among prodromal molecular changes and prodromal symptoms, suggesting new markers to follow at clinical level with even minimally invasive approaches (imaging, GI biopsy). Abbreviations: αSyn: αSynuclein; PD: Parkinson’s Disease; CSF: cerebrospinal fluid; GI: Gastrointestinal; DJ1: deglycase J1; HMOX1: heme oxygenase 1; DA: dopamine; DOPAn: dopaminergic neurons; DAT-scan: dopamine transporter single-photon emission computed tomography; MPTP: 1-Methyl-4-phenyl-1,2178,3,6-tetrahydropyridine; VMAT2: vesicular monoamine transporter 2; NE: norepinephrine; SE: serotonin; Chol: cholinergic; and REM: rapid eye movement.

2.1.2. Pre-Clinical Phase

Despite the initiation of pathophysiological processes, overt PD manifestations are not present in the pre-clinical phase. This gap suggests the possibility of identifying PD biomarkers and represents the primary target for modeling the disease in the laboratory (Figure 1). Due to the absence of any pathological sign, retrospective re-analyses of previous studies are the only approaches that might be applied to improve the specificity of biomarkers in clinics. Ethical issues accompany this strategy [167]; moreover, the absence of symptoms does not justify performing expensive analyses (e.g., 18-F DAT SPECT), which are seldom expensive. Even minimally invasive approaches (e.g., blood and saliva) are accompanied by ethical issues. Progress is being made by system biology tools in the identification of new genomic, transcriptomic, proteomic, lipidomic, and metabolomic molecules and new signaling pathways that may be quantified by minimally invasive approaches and that may be relevant to the pathogenesis [122]. Models may help by avoiding (in vitro models) or reducing (in vivo models) the ethical limitations, reducing the costs of the studies, but even more relevant giving access to central nervous system (CNS) biomolecular analysis in a dynamic form from the pre-symptomatic to the full-blown phases.

2.1.3. Risk Phase

Determining the etiology of PD has been elusive, but many risk factors for PD have been reported. It is likely that PD (as the other age-related diseases) may develop as consequence of the lifelong accumulation of environmental and lifestyle “stressors” and their interactions with genetic factors [2,10,68,168] (Figure 1).

The list of the PD-related stressors is quite extensive, and it includes environmental factors (exposure to metals, pesticides and herbicides) [169,170,171]; the high consumption of milk and dairy products [172,173]; gut dysbiosis [5,79]; substance abuse; insulin resistance in metabolic syndrome; brain trauma; and the physiologic neuroinflammation accompanying aging [79,172,174,175,176]. Physical activity is also associated with a 34% decreased risk of PD, acting by increasing BDNF, and/or DA [79] or by decreasing the systemic inflammation [172,177], with systemic reduction in DA suggested preceding CNS DOPAn loss [122]. Notably, systemic DA cannot enter the brain but may be immunomodulatory [62].

PD incidence differs depending on ethnic origin, being higher in Hispanic people, then non-Hispanic White people, Asian people, and lower in Black people [178]. Genome-wide association studies identified 24 loci with clinically significant association with increased risk, and 11 variants showing a decrease in the risk of developing PD [179].

Some functional correlations between gene variants and environmental/lifestyle factors have been individuated [2]. Key molecular pathways presumed to be important in both familial and sporadic PD have been identified by fitting PD-related genes into common intracellular networks [180]. Recent perspectives on the etiology of PD suggest that the genes involved in PD may share biochemical mechanisms or signaling pathways with inflammatory pathways [69,70,74,77,78,81,121,181]. These include AMP/PKA/MAPK (cyclic adenosine monophosphate/ protein kinase A/ mitogen-activated protein kinases), NRF2 (nuclear factor erythroid 2–related factor 2), PPARs (peroxisome proliferator-activated receptors), NFκB (nuclear factor kappa-light-chain-enhancer of activated B cells), TLR4 (toll-like receptor 4), and inflammasome-NLRP3 (NOD-, LRR- and pyrin domain-containing protein 3) [69,174,175,182,183,184,185,186,187,188].

3. Where We Are in Understanding the Early Stages of PD Through Research Models

Research models for studying PD span different levels of complexity, from cell lines to non-human primates (NHPs). These models include the administration of toxins or proteins that result in various PD-like outcomes (see Table 2). Several reviews summarize the protocols used to mimic the disease, discussing the animal model [189]; route of administration [190,191,192]; dosages of the inducing agent [190,191,192]; the timing for developing motor signs and DA loss [193] and non-motor features [160,161,194]; and the effect of the animal age on the resulting phenotype [195].

In order to complement the existing literature, we will review the extent to which published studies address different stages of PD, with maximal interest in the accessibility to the pre-clinical stages, where research is available. The intention is to critically review whether the actual modeling of PD attempts to explore the sequence of events involved in the progression of the disease. This is a prerequisite to advance prevention and disease-modifying strategies.

Table 2. Common PD models: mechanisms and limitations.

Reserpine	Inhibition of VMAT2, which is expressed on synaptic vesicles of DOPAn, NE, and SE neurons and regulates the release of neurotransmitters [196]. Accordingly, dysfunction of these circuits in PD has been reported [161,189,197]. This model reproduces motor signs [161,198,199,200]; DA loss [189,199,200]; non-motor signs [161]; αSyn presence [199]; mitochondrial dysfunction and redox stress [155,161,199,201]; autophagy [155,161,199,202]; and inflammation [161,199].
6-OHDA	Production of ROS and inhibition of mitochondrial respiratory chain complexes I and IV [203]. Reduces GSH and SOD reduction. Increases glutamate (Glut); astrogliosis, autophagy, and proteasomal dysfunction induction [161,204]. Does not cross the blood–brain barrier (BBB), so intracranial injections are necessary. High mortality rate when administered bilaterally. Endogenous production of 6-OHDA has been reported in the brains of PD patients [55,189,203,205]. This model reproduces motor signs [55,156,161,189,198,199,205,206,207,208,209,210,211,212,213,214,215,216,217,218,219]; DA loss [55,156,189,198,199,205,206,207,208,209,210,211,212,213,214,215,216,217,218,219,220,221,222]; non-motor signs [156,160,161,207,211,212,213,218,223,224]; non-DA loss [156,160,189,212,213,214,216]; αSyn presence [208] and non-detected [55,189,199,205,206]; mitochondrial dysfunction and redox stress [161,199,209,210,221,223,225,226,227,228,229,230,231,232,233]; autophagy [161,199,234]; and inflammation [161,199,207,217,219,220,235].
MPTP	MPTP crosses the BBB due to its lipophilic nature. In the brain MPTP is converted to MPTP+ by the glial MAO-B, spontaneously oxidized to MPP+, then taken up into DA neurons by DAT [198,236,237]. MPTP may also be administered, but the positive charge reduces its brain bioavailability. Stereotaxic injection provide better results [193]. In cells MPP+ acts by accumulating in VMAT2 vesicles, leading to DA release and toxic auto-oxidation. MPP+ also binds to mitochondrial NADH, decreasing ATP production and inhibiting complex I with ROS production [208]. MPTP increases Glut, induces astrogliosis, microgliosis and cytokine release [238,239]. Rats are resistant to MPTP, thus requiring higher doses, and presenting high mortality rates secondary to massive loss of neurons [193,208,240]. Acute doses in MPTP models like mice, rats, zebrafish, and monkeys lead to a quick decrease in ATP, potentially resulting in rapid DOPAn death by apoptosis and necrosis [189,205,208,237], possibly being too rapid to allow proper access to each stage. Chronic administration of low doses in PD features in a small percentage of animals [189,193,197]. This model reproduces motor signs [156,161,189,199,205,208,214,237,241,242,243,244,245,246,247,248,249,250,251,252,253,254,255,256]; DA loss [55,156,189,198,199,205,206,208,214,238,241,242,243,244,245,246,248,249,250,251,252,254,255,256,257,258]; non-motor signs [156,160,161,241,242,250,255,256,259,260,261,262] and non-detected [249]; non-DA loss: [156,160,189,214,242,247,248,252,255,261,262] and non-detected [249,250]; αSyn presence [55,189,198,199,206,208,241,242,261,262] and non-detected [205,249]; absence of LBs [205,249,261,262]; mitochondrial dysfunction/redox stress [161,189,199,242,244,245,257,263,264,265,266,267,268,269,270]; autophagy [161,199,234]; inflammation [161,199,238,241,242,243,257,258,261,262,267,269]; and GI [164].
Paraquat	Agricultural herbicide linked to PD in farmers; not DA-specific. Paraquat induces mitochondrial dysfunction and ROS production [84], reduces GSH and thioredoxin, leads to lipid, protein, and DNA damage, inflammation, autophagy, and proteasomal dysfunction [189,197,205,237]. Potentially toxic to the liver, kidney, and lungs; high mortality rate [208]. This model reproduces motor signs [55,156,161,198,199,205,206,208,271,272,273,274]; DA loss [55,156,198,199,205,206,208,272,273,274]; non-motor signs [161,272]; non-DA loss [156,271,274]; αSyn presence [55,199,206] and non-detection [205,208]; but with the absence of LBs [55,189]; mitochondrial dysfunction and redox stress [161,189,199,263,271,274,275,276,277]; autophagy [161,199]; and inflammation [161,199].
Rotenone	Agricultural herbicide and insecticide linked to PD. Acts similarly to paraquat on mitochondria, redox state [79,84,161,189,193,205,237,278,279], and inhibiting proteasome activity [208], but it does not activate the caspase pathway [280]. Recently, low doses have been shown to have a pro-inflammatory effect [281,282,283]. It is water and alcohol insoluble, rapidly degraded by light, and metabolized in the liver and gastric mucosa. Thus, its usage in vivo is somehow difficult [192]. Acute doses may lead to systemic toxicity and necrosis in the brain [284], with potential lethality [192]. Heterogeneity in the time scale of symptom development and low percentage of animals developing the disease are scientific and ethic problems, overstepped recently by the use of older animals and adjuvants [192,195,285]. Rotenone reproduces motor signs [55,156,161,189,198,199,205,206,208,282,283,285,286,287,288,289]; DA loss: [55,156,189,198,199,205,206,208,282,283,285,286,287,288,289]; non-motor signs [156,161,208]; non-DA loss [156,206]; αSyn presence [55,189,199,206,282,285,288] and not detected [205,208]; LBs [189,288]; mitochondrial dysfunction and redox stress [161,199,263,266,286,289,290,291]; autophagy [161,199,292,293,294,295,296]; inflammation [154,161,199,282,283,288,289,290,294,297]; and gastro intestinal manifestations (GI) [156,165,166].
LPS	Pro-inflammatory [193], but also induces mitochondrial dysfunction and redox stress [161,194,199]. Largely used in recent years to explore the most recent hypotheses that inflammation is important in PD outcome. LPS reproduces motor signs [67,161,194,199,298,299,300,301,302,303,304,305,306,307]; DA loss [194,199,298,299,302,303,304,305,306,307,308,309]; non-motor signs [161,194,300,301]; non-DA loss [194,298,299,300,301,302,305]; αSyn presence [194,199,302,305,309] and non-detected [194]; mitochondrial dysfunction and redox stress [161,199,298,300,301,305,306,308]; autophagy [161,199]; inflammation: [67,161,199,298,299,302,303,304,305,306,307,308]; and GI [307]
αSyn	Protein physiologically present. It can be oxidated, modifying its folding. Misfolding provokes the formation of amyloid fibrils that will result in forming the LBs [45,53,54,55,56], motor signs [310,311,312]; DA loss [54,310,311,312]; αSyn presence [54,310,311,312]; LBs [310]; inflammation [54,310,312].
Genetic models	Multiple models based on multiple genetic variants for each PD gene exist. Large variability of the PD-like features is reached in all genetic animal models. The variability is due both to the genetic variant reproduced and the promoter that controls the gene of interest expression [199,313,314,315]. PARK1/4(SNCA) * has been reported to reproduce motor signs [55,161,189,198,199,205,206,313,314]; DA loss [55,160,198,314] and non-detected [189,199]; non-motor signs [160,161,314]; non-DA loss [160,313]; αSyn presence [55,199,206,313,314] and non-detected [206,313,314]; mitochondrial dysfunction [161,199]; redox stress [161]; autophagy [161,199]; GI [314,316]; inflammation [161,199]. PARK8 (LRRK2) ** has been reported to reproduce motor signs [55,189,198,206,313] and non-detected [161,199,205,314]; DA loss [198,314] and non-detected [199,205]; non-motor signs [161]; non-DA loss [313]; αSyn presence [198] and non-detected [55,189,205,313]; absence of LBs [206,313]; mitochondrial dysfunction [161,199]; redox stress [161]; autophagy [161,199]; GI [314,317]; and inflammation [161,199]. PARK2 (PRKN) *** has been reported to reproduce motor signs [199,313,314]; absence of DA loss [314]; non-motor signs; non-DA loss [189,199,206,314]; αSyn presence [314] and non-detected [189,206,313]; redox stress [199]; GI [199]; and inflammation [189]. PARK7 (DJ1) has been reported to reproduce motor signs [313,314] and non-detected [199,205,206]; DA loss [314] and non-detected [189,199,205,206]; non-DA loss [313]; αSyn presence [314] and non-detected [199,205,206]; absence of LBs [313]; mitochondrial dysfunction [199]; inflammation [199]. PARK5 (UCHL1) has been reported to reproduce motor signs [198,199,206]; DA loss [198]; non-motor signs [314]; absence of αSyn [199]; absence of LB [206]. PARK6 (PINK1) has been reported to reproduce motor signs [313,314] and non-detected [205]; DA loss [189,314] and non-detected [205]; non-motor signs [314]; non-DA loss (only in mice) [313]; αSyn presence [314] and non-detected [189,205,313]; absence of LBs [313]. VMAT2 has been reported to reproduce motor signs [160]; DA loss [160]; non-motor signs [160]; non-DA loss [160].

This table recaps the major outcomes and weaknesses in modeling PD based on the disease inducer. Abbreviations: PD: Parkinson’s Disease; VMAT2: vesicular monoamine transporter type 2; DA: dopamine; NE: norepinephrine; SE: serotonin; 6-OHDA: 6-Hydroxydopamine; ROS: reactive oxygen species; GSH: reduced glutathione; SOD: super oxide dismutase; Glut: glutamate; MAO-B: glial monoamine oxidase B; MPP⁺: 1-methyl-4-phenylpyridinium; MPTP: 1-Methyl-4-phenyl-1,2,3,6-tetrahydropyridine; DAT: dopamine transporter; NADH: nicotinamide adenine dinucleotide-coenzyme 1; ATP: Adenosine triphosphate; DOPAn: dopaminergic neurons; BBB: blood–brain barrier; SNCA: synuclein alpha gene; PRKN: parkin RBR E3 ubiquitin protein ligase; DJ1: deglycase J1; UCHL1: ubiquitin carboxyl-terminal hydrolase isozyme L1; PINK1: PTEN-induced kinase 1; LRRK2: leucine-rich repeat kinase 2; αSyn: α-synuclein; LB: Lewy body; GI: gastrointestinal signs; LPS: lipopolysaccharides. * Depending on the promoter. Synucleinopathy, neuronal degeneration and motor and/or non-motor signs are present only with mPDGFb/Pitx3/mPrp (hA53T)/mThy1 (hwt) promoters. Synucleinopathy and motor/non-motor signs (but not neurodegeneration) are present under PAC (hA53T) control. Synucleinopathy and motor signs (but not neurodegeneration and non-motor signs) in mPrp (E46k)/rTH (wt1–120)/mTHy1 (hsyn-A30P). ** Synucleinopathy and non-motor signs under Cam-TA /PAC (hA30P) in rodents. Synucleinopathy, neuronal degeneration and motor signs in mTh1 (A53T) in rodents [314]. *** Only the bacterial artificial chromosome (BAC) model shows synucleinopathy, neuronal degeneration and motor signs in rodents [314]. NB: In zebrafish neurodegeneration and motor deficits occur in the absence of synucleinopathy for SNCA, LRKK2, DJ1, PARKIN and PINK mutational variants [314].

Table 2. Common PD models: mechanisms and limitations.

Reserpine	Inhibition of VMAT2, which is expressed on synaptic vesicles of DOPAn, NE, and SE neurons and regulates the release of neurotransmitters [196]. Accordingly, dysfunction of these circuits in PD has been reported [161,189,197]. This model reproduces motor signs [161,198,199,200]; DA loss [189,199,200]; non-motor signs [161]; αSyn presence [199]; mitochondrial dysfunction and redox stress [155,161,199,201]; autophagy [155,161,199,202]; and inflammation [161,199].
6-OHDA	Production of ROS and inhibition of mitochondrial respiratory chain complexes I and IV [203]. Reduces GSH and SOD reduction. Increases glutamate (Glut); astrogliosis, autophagy, and proteasomal dysfunction induction [161,204]. Does not cross the blood–brain barrier (BBB), so intracranial injections are necessary. High mortality rate when administered bilaterally. Endogenous production of 6-OHDA has been reported in the brains of PD patients [55,189,203,205]. This model reproduces motor signs [55,156,161,189,198,199,205,206,207,208,209,210,211,212,213,214,215,216,217,218,219]; DA loss [55,156,189,198,199,205,206,207,208,209,210,211,212,213,214,215,216,217,218,219,220,221,222]; non-motor signs [156,160,161,207,211,212,213,218,223,224]; non-DA loss [156,160,189,212,213,214,216]; αSyn presence [208] and non-detected [55,189,199,205,206]; mitochondrial dysfunction and redox stress [161,199,209,210,221,223,225,226,227,228,229,230,231,232,233]; autophagy [161,199,234]; and inflammation [161,199,207,217,219,220,235].
MPTP	MPTP crosses the BBB due to its lipophilic nature. In the brain MPTP is converted to MPTP+ by the glial MAO-B, spontaneously oxidized to MPP+, then taken up into DA neurons by DAT [198,236,237]. MPTP may also be administered, but the positive charge reduces its brain bioavailability. Stereotaxic injection provide better results [193]. In cells MPP+ acts by accumulating in VMAT2 vesicles, leading to DA release and toxic auto-oxidation. MPP+ also binds to mitochondrial NADH, decreasing ATP production and inhibiting complex I with ROS production [208]. MPTP increases Glut, induces astrogliosis, microgliosis and cytokine release [238,239]. Rats are resistant to MPTP, thus requiring higher doses, and presenting high mortality rates secondary to massive loss of neurons [193,208,240]. Acute doses in MPTP models like mice, rats, zebrafish, and monkeys lead to a quick decrease in ATP, potentially resulting in rapid DOPAn death by apoptosis and necrosis [189,205,208,237], possibly being too rapid to allow proper access to each stage. Chronic administration of low doses in PD features in a small percentage of animals [189,193,197]. This model reproduces motor signs [156,161,189,199,205,208,214,237,241,242,243,244,245,246,247,248,249,250,251,252,253,254,255,256]; DA loss [55,156,189,198,199,205,206,208,214,238,241,242,243,244,245,246,248,249,250,251,252,254,255,256,257,258]; non-motor signs [156,160,161,241,242,250,255,256,259,260,261,262] and non-detected [249]; non-DA loss: [156,160,189,214,242,247,248,252,255,261,262] and non-detected [249,250]; αSyn presence [55,189,198,199,206,208,241,242,261,262] and non-detected [205,249]; absence of LBs [205,249,261,262]; mitochondrial dysfunction/redox stress [161,189,199,242,244,245,257,263,264,265,266,267,268,269,270]; autophagy [161,199,234]; inflammation [161,199,238,241,242,243,257,258,261,262,267,269]; and GI [164].
Paraquat	Agricultural herbicide linked to PD in farmers; not DA-specific. Paraquat induces mitochondrial dysfunction and ROS production [84], reduces GSH and thioredoxin, leads to lipid, protein, and DNA damage, inflammation, autophagy, and proteasomal dysfunction [189,197,205,237]. Potentially toxic to the liver, kidney, and lungs; high mortality rate [208]. This model reproduces motor signs [55,156,161,198,199,205,206,208,271,272,273,274]; DA loss [55,156,198,199,205,206,208,272,273,274]; non-motor signs [161,272]; non-DA loss [156,271,274]; αSyn presence [55,199,206] and non-detection [205,208]; but with the absence of LBs [55,189]; mitochondrial dysfunction and redox stress [161,189,199,263,271,274,275,276,277]; autophagy [161,199]; and inflammation [161,199].
Rotenone	Agricultural herbicide and insecticide linked to PD. Acts similarly to paraquat on mitochondria, redox state [79,84,161,189,193,205,237,278,279], and inhibiting proteasome activity [208], but it does not activate the caspase pathway [280]. Recently, low doses have been shown to have a pro-inflammatory effect [281,282,283]. It is water and alcohol insoluble, rapidly degraded by light, and metabolized in the liver and gastric mucosa. Thus, its usage in vivo is somehow difficult [192]. Acute doses may lead to systemic toxicity and necrosis in the brain [284], with potential lethality [192]. Heterogeneity in the time scale of symptom development and low percentage of animals developing the disease are scientific and ethic problems, overstepped recently by the use of older animals and adjuvants [192,195,285]. Rotenone reproduces motor signs [55,156,161,189,198,199,205,206,208,282,283,285,286,287,288,289]; DA loss: [55,156,189,198,199,205,206,208,282,283,285,286,287,288,289]; non-motor signs [156,161,208]; non-DA loss [156,206]; αSyn presence [55,189,199,206,282,285,288] and not detected [205,208]; LBs [189,288]; mitochondrial dysfunction and redox stress [161,199,263,266,286,289,290,291]; autophagy [161,199,292,293,294,295,296]; inflammation [154,161,199,282,283,288,289,290,294,297]; and gastro intestinal manifestations (GI) [156,165,166].
LPS	Pro-inflammatory [193], but also induces mitochondrial dysfunction and redox stress [161,194,199]. Largely used in recent years to explore the most recent hypotheses that inflammation is important in PD outcome. LPS reproduces motor signs [67,161,194,199,298,299,300,301,302,303,304,305,306,307]; DA loss [194,199,298,299,302,303,304,305,306,307,308,309]; non-motor signs [161,194,300,301]; non-DA loss [194,298,299,300,301,302,305]; αSyn presence [194,199,302,305,309] and non-detected [194]; mitochondrial dysfunction and redox stress [161,199,298,300,301,305,306,308]; autophagy [161,199]; inflammation: [67,161,199,298,299,302,303,304,305,306,307,308]; and GI [307]
αSyn	Protein physiologically present. It can be oxidated, modifying its folding. Misfolding provokes the formation of amyloid fibrils that will result in forming the LBs [45,53,54,55,56], motor signs [310,311,312]; DA loss [54,310,311,312]; αSyn presence [54,310,311,312]; LBs [310]; inflammation [54,310,312].
Genetic models	Multiple models based on multiple genetic variants for each PD gene exist. Large variability of the PD-like features is reached in all genetic animal models. The variability is due both to the genetic variant reproduced and the promoter that controls the gene of interest expression [199,313,314,315]. PARK1/4(SNCA) * has been reported to reproduce motor signs [55,161,189,198,199,205,206,313,314]; DA loss [55,160,198,314] and non-detected [189,199]; non-motor signs [160,161,314]; non-DA loss [160,313]; αSyn presence [55,199,206,313,314] and non-detected [206,313,314]; mitochondrial dysfunction [161,199]; redox stress [161]; autophagy [161,199]; GI [314,316]; inflammation [161,199]. PARK8 (LRRK2) ** has been reported to reproduce motor signs [55,189,198,206,313] and non-detected [161,199,205,314]; DA loss [198,314] and non-detected [199,205]; non-motor signs [161]; non-DA loss [313]; αSyn presence [198] and non-detected [55,189,205,313]; absence of LBs [206,313]; mitochondrial dysfunction [161,199]; redox stress [161]; autophagy [161,199]; GI [314,317]; and inflammation [161,199]. PARK2 (PRKN) *** has been reported to reproduce motor signs [199,313,314]; absence of DA loss [314]; non-motor signs; non-DA loss [189,199,206,314]; αSyn presence [314] and non-detected [189,206,313]; redox stress [199]; GI [199]; and inflammation [189]. PARK7 (DJ1) has been reported to reproduce motor signs [313,314] and non-detected [199,205,206]; DA loss [314] and non-detected [189,199,205,206]; non-DA loss [313]; αSyn presence [314] and non-detected [199,205,206]; absence of LBs [313]; mitochondrial dysfunction [199]; inflammation [199]. PARK5 (UCHL1) has been reported to reproduce motor signs [198,199,206]; DA loss [198]; non-motor signs [314]; absence of αSyn [199]; absence of LB [206]. PARK6 (PINK1) has been reported to reproduce motor signs [313,314] and non-detected [205]; DA loss [189,314] and non-detected [205]; non-motor signs [314]; non-DA loss (only in mice) [313]; αSyn presence [314] and non-detected [189,205,313]; absence of LBs [313]. VMAT2 has been reported to reproduce motor signs [160]; DA loss [160]; non-motor signs [160]; non-DA loss [160].

This table recaps the major outcomes and weaknesses in modeling PD based on the disease inducer. Abbreviations: PD: Parkinson’s Disease; VMAT2: vesicular monoamine transporter type 2; DA: dopamine; NE: norepinephrine; SE: serotonin; 6-OHDA: 6-Hydroxydopamine; ROS: reactive oxygen species; GSH: reduced glutathione; SOD: super oxide dismutase; Glut: glutamate; MAO-B: glial monoamine oxidase B; MPP⁺: 1-methyl-4-phenylpyridinium; MPTP: 1-Methyl-4-phenyl-1,2,3,6-tetrahydropyridine; DAT: dopamine transporter; NADH: nicotinamide adenine dinucleotide-coenzyme 1; ATP: Adenosine triphosphate; DOPAn: dopaminergic neurons; BBB: blood–brain barrier; SNCA: synuclein alpha gene; PRKN: parkin RBR E3 ubiquitin protein ligase; DJ1: deglycase J1; UCHL1: ubiquitin carboxyl-terminal hydrolase isozyme L1; PINK1: PTEN-induced kinase 1; LRRK2: leucine-rich repeat kinase 2; αSyn: α-synuclein; LB: Lewy body; GI: gastrointestinal signs; LPS: lipopolysaccharides. * Depending on the promoter. Synucleinopathy, neuronal degeneration and motor and/or non-motor signs are present only with mPDGFb/Pitx3/mPrp (hA53T)/mThy1 (hwt) promoters. Synucleinopathy and motor/non-motor signs (but not neurodegeneration) are present under PAC (hA53T) control. Synucleinopathy and motor signs (but not neurodegeneration and non-motor signs) in mPrp (E46k)/rTH (wt1–120)/mTHy1 (hsyn-A30P). ** Synucleinopathy and non-motor signs under Cam-TA /PAC (hA30P) in rodents. Synucleinopathy, neuronal degeneration and motor signs in mTh1 (A53T) in rodents [314]. *** Only the bacterial artificial chromosome (BAC) model shows synucleinopathy, neuronal degeneration and motor signs in rodents [314]. NB: In zebrafish neurodegeneration and motor deficits occur in the absence of synucleinopathy for SNCA, LRKK2, DJ1, PARKIN and PINK mutational variants [314].

3.1. Modeling PD

Based on clinical and pathological features, a successful animal model of PD should include DA/DOPAn loss, motor deficits, αSyn, and/or LBs formation. Protein misfolding, inflammation, and redox impairment are the major pathologic mechanisms suspected in clinic and frequently investigated in PD models (in vitro to in vivo) (Table 2, Table 3 and Table 4).

3.1.1. Modeling the Early Stages of Clinical PD

Motor symptoms (rigidity and akinesia) together with a DOPAn loss up to 75% in SN, and they may be presented in rodents and zebrafish exposed to 6-OHDA [190,193,217,252,318,319,320,321,322]. While rarely, if ever, used in NHP, rotenone reproduces key features of early and late clinical PD in rodents and zebrafish, with the presence of motor defects, DA loss, and αSyn inclusions in nigral neurons, and LBs [162,189,191,192,197,205,282,283,285,288,323,324,325,326]. Bradykinesia, rigidity, stooped posture, and mild postural instability are reached with the acute dosing (high dosage, short duration) of MPTP in mice and NHP, accompanied by a rapid degeneration of the DOPAn [198,205,284]. Chronic challenge with lower doses of MPTP results in clinical motor signs and decreased DOPAn numbers in middle-aged NHP [327]. Notably, in acute 6-OHDA and MPTP, and in reserpine models, the initial tyrosine hydroxylase (TH⁺, DA precursor) decrease may revert, suggesting reduced DA production rather that a neuron loss [189,193,197,200,328]. Moreover, acute dosing regimens are seldom accompanied by toxicity potentially leading to death (see Table 2). This strongly suggest that acute dosing regimen do not model well the etiopathogenesis of human PD. This may limit not only the translational relevance of understanding disease mechanisms but also using this knowledge to develop and test therapeutic approaches.

Mechanistically speaking, in addition to their well-known pro-oxidant effect, MPTP and 6-OHDA also produce proteasome inhibition [234], leading to endoplasmic reticulum stress [329] and the upregulation of chaperones involved in cell death [330]. Proteasome inhibition is strongly connected with the concept of protein misfolding in PD, and proteasome inhibitors lead to DOPAn loss, motor abnormalities, and αSyn accumulation in vivo [331]. Small inclusion bodies were also reported in c. elegans, rodents, minipigs, and NHPs exposed to, e.g., 6-OHDA and proteosome inhibitor, but the pathogenic role of αSyn in these models has not been determined. In vitro, proteasome inhibition leads to intracellular accumulation of p52, a pro-apoptotic mediator [332,333]. In line with this, proteasome inhibitors can induce the mitochondrial apoptotic signaling cascade, with the increased release of cytochrome c, ROS, and Bcl-2-associated X (BAX) and activation of caspase 3 and caspase 9 [334,335]. Notably, apoptosis is induced by rotenone and 6-OHDA but not by MPTP in the MN9D cell line and primary cultures [232,233,280], highlighting how the chemical used to induce PD may affect the results and their translation to the clinic. Both proteasome inhibitors (specifically actacystin) and rotenone impair mitochondrial function and reduce intracellular ATP [280,336], which might impair vesicular trafficking and mis/unfolded protein clearance, leading to αSyn accumulation [337]. In a cellular model of rotenone-induced PD, transcriptomic sequencing revealed PTEN-induced putative kinase 1 (PINK1) as one of the most altered pathways in mitophagy, stressing its potential contribution in this specific disease mechanism [291].

Notably, in animal models of early PD phase that recapitulates the presence of αSyn, reserpine, 6-OHDA (independently from the dosage and site of injection) and MPTP (at least under acute schemes) do not result in LBs despite the presence of motor abnormalities [114,189,193,197,198,200,205,208,284,328]. Multiple injections (across 8 weeks) of paraquat lead to 50% reduction in TH⁺ cells and αSyn presence but with the absence of LBs in rodents [53,55,156]. Even genetic models largely fail in this aim (see Table 2), except for the SNCA gene coding for αSyn. Altogether, these results may be interpreted in two ways as follows: (1) as a failure of the model, or (2) as an indication that αSyn is not causative for PD symptoms, as discussed in Section 2.1. Supporting criticisms of a mechanistic role of αSyn and LBs in clinical PD is the fact that at moderate/chronic doses of MPTP leads to αSyn aggregation in mice but not DOPAn loss or motor symptoms [198]. Unilateral injection of αSyn preformed fibrils into mouse striatum leads to neuroinflammation preceding accumulation of pathologic form of αSyn, and thereafter neurodegeneration and motor deficits [310] (Table 3). This highlights the importance of the experimental setting and analysis (e.g., αSyn gene expression, physiologic protein, and fibrils) in elucidating translational pathologic mechanisms [54,55]. Interestingly, the uninjected contralateral striatum displays similar pathophysiological effects to the side injected with αSyn preformed fibrils [310]. This has been interpreted as reflecting a pivotal role for inflammatory mediators in initiating the disease (see also the risk phase; 2.1.3: clinic; and 3.3: models) [53,189,193,310,338,339]. Of relevance for the review, there is a general lack of models exploring inflammatory pathways during the early clinical PD stage (Table 3 and Table 4). Some in vitro–ex vivo models (e.g., cell lines, induced pluripotent stem cells—iPSCs, organoids, and organotypic brain cultures—OBCs) have been used to address specific mechanisms of PD. In SH-SY5Y cells exposed to 6-OHDA, the upregulation of 12 necroptosis-related genes belonging to the inflammatory cascade previously identified in patients with PD has been reproduced [235]. In agreement with the inflammatory hypotheses, in PC12 cells, 6-OHDA initiates a physical interaction among the NEDD4 E3 ubiquitin ligase and the pro-apoptotic regulator RTP801 [340]. A decrease in NEDD4 is responsible for increased RTP801 gene expression, which contributes to neuroinflammation, memory impairment, and mortality in Alzheimer’s disease (AD) [341]. The NEDD4-RTP801 complex has been localized at the synapse level in primary cortical neurons and in TH⁺ neurons from PD patients [340]. Moreover, the lncRNA (long non-coding RNA) Metastasis Associated Lung Adenocarcinoma Transcript 1 (MALAT1) induces inflammatory response in microglia BV2 cells, promoting down regulation of miR-23b-3p (micro RNA-23b-3p), reduced autophagy, increased αSyn formation, and apoptosis in dopaminergic neuronal MN9D cells exposed to rotenone [290,294]. This is not a unique example of epigenetic regulation in PD. A reduction in αSyn accumulation in vitro may be obtained by silencing the lncRNA small nucleolar RNA host gene 14 (SNHG14)/miR-133b [293], the circular RNA circ-Pank1 [295], and NADPH oxidase 4 (NOX4) gene [268], suggesting potential translational alternative therapeutic approaches to PD. Similarly, modulating histone deacetylase 4 (HDAC4), which is upregulated at both a transcript and protein level in PD, may increase cell survival and reduce αSyn levels in cell lines exposed to rotenone [292].

Supportive of the interconnection between epigenetic, redox status, and αSyn accumulation [268,293,295], the lncRNA HOXA transcript antisense RNA, myeloid-specific 1 (HOTAIRM1) is upregulated in PD patients. Its silencing induces HMOX1, a redox sensor and activator of different anti-oxidant genes, preventing apoptosis [264]. In line with this, sorting nexin 5 (SNX5), endosomal trafficking regulator, and ferroptosis are also up regulated, and its knockout reverses GSH/GSSG depletion, ROS increase, and cell death in PC-12 cell and rat model induced by 6-OHDA [229]. Ferroptosis-linked cell death has also been reported in SH-SY5Y exposed to MPTP and paraquat, and PC12 cells challenged MPTP and 6-OHDA [230,270], with cell viability completely restored by deferoxamine (DFO), a ferroptosis inhibitor, and partly restored by the overexpression of the ferritin heavy chain 1 (FTH1) [230]. Ferroptosis, redox stress, protein misfolding [342], and αSyn aggregation are connected [230,270,277,342], but the timeline of the events requires experimental confirmation. Of relevance, iron is one of the anabolites of HMOX1 activity, and hyperactivation of HMOX1 is believed to enhance neurodegeneration by increasing oxidative stress [88,91,92,94,149]. As a matter of notice, bilirubin is another anabolic product of HMOX1 activity, with a speculated opposite (protective) action in DOPAn survival [68,297]. Thus, the use of HMOX1 as a marker of PD is supported by the literature, even if the exact role must be elucidated. Also, the excessive release of DA into the cytoplasm or extracellular space may enhance the pro-oxidant status of the milieu. DA can undergo oxidation, enhancing redox stress, as demonstrated in vitro by Li et al. through the blocking of vesicular monoamine transporter 2 (VMAT2) with reserpine in SH-SY5Y cells [155]. Notably this is one of the few studies performed in a time course approach, thus overlapping prodromal (10% cell death at 12 h) and early clinical PD (50% cell death at 24 h—Table 4). The majority of analyses have been performed only at 24 h, when autophagy and αSyn levels were already significantly increased [155]. This highlights the limited number of studies that focus on understanding the early molecular mechanisms of DOPAn loss. As a basis for this review, we have a solid technical background to mimic the prediagnostic stage both in vitro and in vivo models thus the gap can be filled. Iron-induced redox stress is responsible for lipid peroxidation [277], and it is potentially related to the changes in the level of saturation and length of the phospholipid chains of fatty acids present in PD brain tissue that presumably contribute to αSyn accumulation [343]. Of relevance in modeling PD, a mitochondrial oxidative phosphorylation metabolism prevails at low MPTP dosage (30 μM, 90% cell viability), and a glycolytic metabolism prevails at high MPTP challenge in SH-SY5Y cells (300 μM, 70% of cell viability) [344,345].

Table 3. Mechanistic findings in pre-clinical models.

	DA Loss/Involvement	Non-DA Loss/Involvement	Motor Deficits	Non-Motor Deficits	αSyn	LBs	Inflammation	Mitochondria and Redox	Ref. and Short Description
Post-diagnosis phase	Y	Y	Y	Y	Y	°	Y	°	[241] Mice, MPTP
	Y	Y	Y	Y	Y	°	Y	Y	[242] Mice, MPTP
	Y	Y	Y	°	°	°	Y	Y	[298] Rat, LPS
	Y	Y	Y	°	°	°	Y	°	[299] Rat, LPS
	°	Y	Y	Y	°	°	°	Y	[300,301] Rat, LPS
	Y	Y	Y	°	Y	°	Y	°	[302] Mice, LPS
	°	Y	Y	°	Y	No	Y	°	[261,262] NHP, MPTP
	Y	°	Y	Y	°	°	Y	°	[207] Mice, 6-OHDA
	Y	°	Y	°	°	°	°	Y	[209] Rat, 6-OHDA
	Y	°	Y	°	°	°	Y	°	[219] Mice, 6-OHDA *
	Y	°	Y	°	°	°	Y	°	[243] Mice, MPTP
	Y	°	Y	°	°	°	°	Y	[210] Mice, 6-OHDA
	Y	°	°	°	°	°	Y	°	[220] Rat, 6-OHDA *
	Y	°	°	°	°	°	°	Y	[221] Zebrafish, 6-OHDA
	Y	°	°	°	°	°	Y	Y	[257] Mice, MPTP *
	Y	°	°	°	°	°	Y	°	[258] Mice, MPTP
	Y	°	°	°	°	°	Y	°	[238] Mice, MPTP
	Y	°	Y	°	°	°	°	Y	[244] Mice, MPTP *
	Y	°	Y	°	°	°	°	Y	[245] Zebrafish, MPTP *
	Y	°	Y	°	°	°	°	Y	[286] Rat, rotenone
	Y	°	Y	°	°	°	Y	°	[283] Mice, rotenone
	Y	°	Y	°	°	°	°	Y	[271] Rat, paraquat
	Y	°	Y	°	°	°	Y	°	[303] Rat, LPS
	Y	°	Y	GI	°	°	Y	°	[307] Rat, LPS
	Y	°	°	°	Y	°	Y	°	[54] Rat, αSyn *
	Y	Y	Y	Y	°	°	°	°	[211] Rat, 6-OHDA *
	Y	Y	Y	Y	°	°	°	°	[212] Rat, 6-OHDA *
	Y	Y	Y	Y	°	°	°	°	[213] Rat, 6-OHDA *
	Y	Y	Y	°	°	°	°	°	[214] Zebrafish, 6-OHDA or MPTP *
	Y	°	Y	°	°	°	°	°	[246] NHP, MPTP *
	Y	Y	Y	°	°	°	°	°	[247] NHP, MPTP
	Y	Y	Y	°	°	°	°	°	[248] NHP, MPTP
	Y	No	Y	No	No	No	°	°	[249] NHP, MPTP *
	Y	Y	Y	Y	°	°	°	°	[250] Mice, MPTP *
	Y	°	Y	°	°	°	°	°	[251] Zebrafish, MPTP
	Y	Y	Y	°	°	°	°	°	[252] Zebrafish, MPTP
	°	°	Y	°	°	°	°	°	[253] Zebrafish, MPTP
	Y	°	Y	°	Y	°	°	°	[285] Rat, rotenone
	Y	°	Y	°	°	°	°	°	[287] Rat, rotenone
	Y	°	Y	Y	°	°	°	°	[272] Zebrafish, paraquat
	Y	°	Y	°	°	°	°	°	[273] Rat, paraquat *
Prodromal to post-diagnosis	Y	°	Y	°	°	°	Y	°	[346] DJ1 mice genetic model *
	Y	°	Y	°	°	°	°	°	[246] NHP, MPTP *
	Y	°	Y	°	°	°	°	°	[200] Mice, reserpine *
	Y	°	Y	°	°	°	Y	°	[346] Park2 mice genetic model
	Y	°	Y	°	No	No	Y	°	[347] Parkin rat genetic model *
	Y	°	Y	°	Y	Y	Y	°	[310] Mice, αSyn *
	Y	°	°	°	°	°	°	°	[222] Rat, 6-OHDA *
	Y	°	Y	°	Y	°	°	°	[311] Mice, LPS and αSyn *
	°	°	Y	°	°	°	Y	°	[67] Rat, LPS
	Y	°	Y	°	°	°	Y	°	[304] Mice, LPS *
	Y	Y	Y	°	°	°	°	Y	[274] Zebrafish, paraquat *
	Y	Y	Y	°	Y	°	Y	Y	[305] Rat, LPS *
	Y	°	Y	°	Y	Y	Y	°	[288] Mice, rotenone *
	Y	°	Y	°	°	°	Y	Y	[289] Mice, rotenone *
	Y	°	Y	°	Y	°	Y	°	[282] Rat, rotenone *
	Y	°	Y	°	°	°	Y	°	[254] NHP, MPTP
	Y	°	Y	°	°	°	°	°	[215] Rat, 6-OHDA
	Y	Y	Y	°	°	°	°	°	[216] Rat, 6-OHDA
	Y	°	Y	Y	°	°	°	°	[255] Rat, 6-OHDA * [255] Rat, MPTP *
	Y	Y	Y	Y	°	°	°	°
	Y	°	Y	°	weak	°	Y	°	[312] Mice, αSyn
	Y	°	Y	°	°	°	Y	°	[217] Zebrafish, 6-OHDA *
Prodromal phase	Y	No	Y	Y	°	°	°	°	[218] Rat, 6-OHDA
	°	°	°	Y	°	°	°	Y	[223] Rat, 6-OHDA
	Y	°	Y	Y	°	°	°	°	[256] NHP, MPTP
Risk phase	Y	°	°	°	°	°	Y	Y	[306] Rat, LPS *
Out of classification based on the clinical cardinal symptoms	Y	°	°	°	Y	Y	Y	°	[348] αSyn mice genetic model, LPS
	°	°	°	°	°	°	Y	Y	[269] Mice, MPTP *
	Y	°	°	°	Y	°	°	°	[309] Mice, LPS *
	Y	°	°	°	°	°	Y	Y	[308] Mice, LPS

°: not assessed; Y: yes present; No: not present. Abbreviations: LBs: Lewy bodies; DA: dopamine; αSyn: α-synuclein; 6-OHDA: 6-hydroxydopamine; MPTP: 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine; LPS: lipopolysaccharide; NHP: non-human primates; GI: gastro intestinal. * Models investigating changes along a time course experimental scheme.

Table 3. Mechanistic findings in pre-clinical models.

	DA Loss/Involvement	Non-DA Loss/Involvement	Motor Deficits	Non-Motor Deficits	αSyn	LBs	Inflammation	Mitochondria and Redox	Ref. and Short Description
Post-diagnosis phase	Y	Y	Y	Y	Y	°	Y	°	[241] Mice, MPTP
	Y	Y	Y	Y	Y	°	Y	Y	[242] Mice, MPTP
	Y	Y	Y	°	°	°	Y	Y	[298] Rat, LPS
	Y	Y	Y	°	°	°	Y	°	[299] Rat, LPS
	°	Y	Y	Y	°	°	°	Y	[300,301] Rat, LPS
	Y	Y	Y	°	Y	°	Y	°	[302] Mice, LPS
	°	Y	Y	°	Y	No	Y	°	[261,262] NHP, MPTP
	Y	°	Y	Y	°	°	Y	°	[207] Mice, 6-OHDA
	Y	°	Y	°	°	°	°	Y	[209] Rat, 6-OHDA
	Y	°	Y	°	°	°	Y	°	[219] Mice, 6-OHDA *
	Y	°	Y	°	°	°	Y	°	[243] Mice, MPTP
	Y	°	Y	°	°	°	°	Y	[210] Mice, 6-OHDA
	Y	°	°	°	°	°	Y	°	[220] Rat, 6-OHDA *
	Y	°	°	°	°	°	°	Y	[221] Zebrafish, 6-OHDA
	Y	°	°	°	°	°	Y	Y	[257] Mice, MPTP *
	Y	°	°	°	°	°	Y	°	[258] Mice, MPTP
	Y	°	°	°	°	°	Y	°	[238] Mice, MPTP
	Y	°	Y	°	°	°	°	Y	[244] Mice, MPTP *
	Y	°	Y	°	°	°	°	Y	[245] Zebrafish, MPTP *
	Y	°	Y	°	°	°	°	Y	[286] Rat, rotenone
	Y	°	Y	°	°	°	Y	°	[283] Mice, rotenone
	Y	°	Y	°	°	°	°	Y	[271] Rat, paraquat
	Y	°	Y	°	°	°	Y	°	[303] Rat, LPS
	Y	°	Y	GI	°	°	Y	°	[307] Rat, LPS
	Y	°	°	°	Y	°	Y	°	[54] Rat, αSyn *
	Y	Y	Y	Y	°	°	°	°	[211] Rat, 6-OHDA *
	Y	Y	Y	Y	°	°	°	°	[212] Rat, 6-OHDA *
	Y	Y	Y	Y	°	°	°	°	[213] Rat, 6-OHDA *
	Y	Y	Y	°	°	°	°	°	[214] Zebrafish, 6-OHDA or MPTP *
	Y	°	Y	°	°	°	°	°	[246] NHP, MPTP *
	Y	Y	Y	°	°	°	°	°	[247] NHP, MPTP
	Y	Y	Y	°	°	°	°	°	[248] NHP, MPTP
	Y	No	Y	No	No	No	°	°	[249] NHP, MPTP *
	Y	Y	Y	Y	°	°	°	°	[250] Mice, MPTP *
	Y	°	Y	°	°	°	°	°	[251] Zebrafish, MPTP
	Y	Y	Y	°	°	°	°	°	[252] Zebrafish, MPTP
	°	°	Y	°	°	°	°	°	[253] Zebrafish, MPTP
	Y	°	Y	°	Y	°	°	°	[285] Rat, rotenone
	Y	°	Y	°	°	°	°	°	[287] Rat, rotenone
	Y	°	Y	Y	°	°	°	°	[272] Zebrafish, paraquat
	Y	°	Y	°	°	°	°	°	[273] Rat, paraquat *
Prodromal to post-diagnosis	Y	°	Y	°	°	°	Y	°	[346] DJ1 mice genetic model *
	Y	°	Y	°	°	°	°	°	[246] NHP, MPTP *
	Y	°	Y	°	°	°	°	°	[200] Mice, reserpine *
	Y	°	Y	°	°	°	Y	°	[346] Park2 mice genetic model
	Y	°	Y	°	No	No	Y	°	[347] Parkin rat genetic model *
	Y	°	Y	°	Y	Y	Y	°	[310] Mice, αSyn *
	Y	°	°	°	°	°	°	°	[222] Rat, 6-OHDA *
	Y	°	Y	°	Y	°	°	°	[311] Mice, LPS and αSyn *
	°	°	Y	°	°	°	Y	°	[67] Rat, LPS
	Y	°	Y	°	°	°	Y	°	[304] Mice, LPS *
	Y	Y	Y	°	°	°	°	Y	[274] Zebrafish, paraquat *
	Y	Y	Y	°	Y	°	Y	Y	[305] Rat, LPS *
	Y	°	Y	°	Y	Y	Y	°	[288] Mice, rotenone *
	Y	°	Y	°	°	°	Y	Y	[289] Mice, rotenone *
	Y	°	Y	°	Y	°	Y	°	[282] Rat, rotenone *
	Y	°	Y	°	°	°	Y	°	[254] NHP, MPTP
	Y	°	Y	°	°	°	°	°	[215] Rat, 6-OHDA
	Y	Y	Y	°	°	°	°	°	[216] Rat, 6-OHDA
	Y	°	Y	Y	°	°	°	°	[255] Rat, 6-OHDA * [255] Rat, MPTP *
	Y	Y	Y	Y	°	°	°	°
	Y	°	Y	°	weak	°	Y	°	[312] Mice, αSyn
	Y	°	Y	°	°	°	Y	°	[217] Zebrafish, 6-OHDA *
Prodromal phase	Y	No	Y	Y	°	°	°	°	[218] Rat, 6-OHDA
	°	°	°	Y	°	°	°	Y	[223] Rat, 6-OHDA
	Y	°	Y	Y	°	°	°	°	[256] NHP, MPTP
Risk phase	Y	°	°	°	°	°	Y	Y	[306] Rat, LPS *
Out of classification based on the clinical cardinal symptoms	Y	°	°	°	Y	Y	Y	°	[348] αSyn mice genetic model, LPS
	°	°	°	°	°	°	Y	Y	[269] Mice, MPTP *
	Y	°	°	°	Y	°	°	°	[309] Mice, LPS *
	Y	°	°	°	°	°	Y	Y	[308] Mice, LPS

°: not assessed; Y: yes present; No: not present. Abbreviations: LBs: Lewy bodies; DA: dopamine; αSyn: α-synuclein; 6-OHDA: 6-hydroxydopamine; MPTP: 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine; LPS: lipopolysaccharide; NHP: non-human primates; GI: gastro intestinal. * Models investigating changes along a time course experimental scheme.

3.1.2. Modeling the Late Stages of Clinical PD

As shown in Figure 1 in the late PD stage both motor and non-motor symptoms are present. Behaviors modeling anxiety, depression, and apathy, as well as aphagia, adipsia, and death can result from 6-OHDA and MPTP administration, with a consistent neuron loss (NE loss up to 75%; DA 83–98%) in rodents with intracisternal injection [156,160,189,193,318]. Parallel effects are observed in zebrafish [190,318,324,349,350]. Non-motor PD signs [351], NMDA-linked parkinsonian symptoms [248,352], and NE depletion [247,351] result from chronic treatments with MPTP in NHPs. The degeneration of nigrostriatal DOPAn results in increased GABAergic activity in the striatum [353]. In OBCs from the nucleus of Meynert, cortex, ventral mesencephalon, and dorsal striatum, rotenone treatment results in cholinergic cell death [354]. Loss of SE, Chol, and NE neurons [55,254], with consequent non-motor signs of PD [160,191,324,355], including gastrointestinal dysfunctions, have been documented [53,116,162,191,192,198,283,288,323]. A recent study revealed that MPP⁺-induced effects on DA release are mediated in part by mechanisms related to acetylcholine regulation [356]. The use of rotenone into rodents has been limited by variability in the percentage of animals affected, the long time required for chronic dosing, and high mortality with acute dosing [192]. These limitations have been overcome by using middle-aged rats and medium-chain triglyceride Miglyol 812 N as adjuvant. This regimen results in PD signs in 100% of animals within 3 weeks and minimal inter-individual variability of disease severity [195,285]. Given the completeness of phenotypic signs and the homogeneity in their development, this is likely one of the most effective models currently available for performing time course evaluations of biomolecular events ongoing in the interconnected CNS circuits. Moreover, it also provides valuable insights into disease onset and progression, with minimal ethical and economic investments.

Table 4. Mechanistic findings in in vitro and ex vivo models.

	Inflammation	Redox Stress	Autophagy, lysosome, αSyn	Apoptosis	Ref. and Short Description
Post-diagnosis phase (cell death of at least 50%)	°	°	°	°	[357] OBCs, reserpine
	°	Y	Y	Y	[155] SH-SY5Y, reserpine
	°	°	°	Y	[358] OBCs, 6-OHDA
	°	°	Y §	°	[343] SH-SY5Y, 6-OHDA §
	°	°	°	Y	[359] SH-SY5Y, 6-OHDA
	°	Y	°	Y	[225] SH-SY5Y, 6-OHDA
	°	°	°	Y	[359] PC12, 6-OHDA
	°	Y	°	Y	[229] PC12, 6-OHDA
	°	Y	°	°	[230] PC12, 6-OHDA
	°	°	°	Y	[360] LUHMES, 6-OHDA
	°	Y	°	Y	[276] SH-SY5Y, paraquat and physical stretch
	°	Y	°	Y	[277] SH-SY5Y, paraquat
	°	°	°	°	[361] SH-SY5Y, paraquat
	°	Y	°	Y	[263] SH-SY5Y, paraquat *
	°	°	°	Y	[354] OBCs, rotenone
	°	Y	Y	Y	[293] MN9D, rotenone
	Y	°	Y	Y	[294] MN9D, rotenone
	°	°	Y	Y	[295] MN9D, rotenone
	Y	°	Y	Y	[290] LUHMES, rotenone
	°	Y	°	Y	[266] PC12, rotenone
	°	°	Y	°	[330] SH-SY5Y, MPTP
	°	Y	°	Y	[264] SH-SY5Y, MPTP
	°	°	°	Y	[360] SH-SY5Y, MPTP
	°	Y	Y	Y	[362] SH-SY5Y, MPTP
	°	Y	°	Y	[265] SH-SY5Y, MPTP *
	°	Y	°	Y	[266] PC12, MPTP
	Y	Y	°	Y	[363] PC12, MPTP
	°	Y	°	Y	[270] PC12, MPTP
	Y	Y	°	Y	[267] MN9D, MPTP
	°	°	°	Y	[364] MN9D, MPTP
	°	Y	°	Y	[268] MN9D, MPTP
Prodromal to post-diagnosis phase	Y	Y	°	Y	[297] OBCs, rotenone *
	Y	Y	°	Y	[154] OBCs, rotenone *
	°	Y	°	Y	[263] SH-SY5Y, rotenone *,#
	°	Y	°	Y	[263] SH-SY5Y, MPTP *,#
	°	°	°	Y	[280] MN9D, rotenone #
	°	°	°	°	[345] SH-SY5Y, MPTP
Out of classification based on the clinical cardinal symptoms (no data on cell viability, studied only the molecular mechanisms)	°	°	Y	°	[202] PC12, reserpine
	°	Y	°	°	[201] PC12, reserpine
	°	Y	°	°	[227] SH-SY5Y, 6-OHDA *
	°	Y	°	Y	[228] SH-SY5Y, 6-OHDA
	Y	°	°	°	[235] SH-SY5Y, 6-OHDA
	°	°	°	Y	[340] PC12, 6-OHDA
	°	Y	°	Y	[226] MN9D, 6-OHDA
	°	°	°	Y	[232] MN9D, 6-OHDA
	°	°	Y	°	[234] MN9D, 6-OHDA
	°	Y	°	Y	[233] MN9D, 6-OHDA *
	°	Y	°	Y	[231] MN9D, 6-OHDA
	°	°	°	°	[365] LUHMES, 6-OHDA
	°	Y	°	Y	[275] OBCs, paraquat
	°	°	Y	Y	[292] SH-SY5Y, rotenone
	Y	Y	°	Y	[366] SH-SY5Y, rotenone
	°	°	Y	Y	[296] PC12, rotenone
	°	°	°	°	[329] PC12, rotenone
	°	°	°	°	[153] OBCs, MPTP
	°	°	°	°	[356] OBCs, MPTP
	°	Y	°	Y	[367] SH-SY5Y, MPTP
	°	°	°	°	[329] PC12, MPTP
	°	Y	Y	Y	[368] PC12, MPTP
	°	°	Y	°	[234] MN9D, MPTP

^§ supposingly linked to αSyn quantity; ^# depend on time and concentration of the treatment; * models investigating changes along a time course experimental scheme. Abbreviations: Y: assessed; °: not assessed; αSyn: α-synuclein; 6-OHDA: 6-hydroxydopamine; MPTP: 1-Methyl-4-phenylpyridinium; OBCs: organotypic brain culture; LUHMES: Lund human mesencephalic cell line; MN9D: mouse dopaminergic neuronal cell line.

Table 4. Mechanistic findings in in vitro and ex vivo models.

	Inflammation	Redox Stress	Autophagy, lysosome, αSyn	Apoptosis	Ref. and Short Description
Post-diagnosis phase (cell death of at least 50%)	°	°	°	°	[357] OBCs, reserpine
	°	Y	Y	Y	[155] SH-SY5Y, reserpine
	°	°	°	Y	[358] OBCs, 6-OHDA
	°	°	Y §	°	[343] SH-SY5Y, 6-OHDA §
	°	°	°	Y	[359] SH-SY5Y, 6-OHDA
	°	Y	°	Y	[225] SH-SY5Y, 6-OHDA
	°	°	°	Y	[359] PC12, 6-OHDA
	°	Y	°	Y	[229] PC12, 6-OHDA
	°	Y	°	°	[230] PC12, 6-OHDA
	°	°	°	Y	[360] LUHMES, 6-OHDA
	°	Y	°	Y	[276] SH-SY5Y, paraquat and physical stretch
	°	Y	°	Y	[277] SH-SY5Y, paraquat
	°	°	°	°	[361] SH-SY5Y, paraquat
	°	Y	°	Y	[263] SH-SY5Y, paraquat *
	°	°	°	Y	[354] OBCs, rotenone
	°	Y	Y	Y	[293] MN9D, rotenone
	Y	°	Y	Y	[294] MN9D, rotenone
	°	°	Y	Y	[295] MN9D, rotenone
	Y	°	Y	Y	[290] LUHMES, rotenone
	°	Y	°	Y	[266] PC12, rotenone
	°	°	Y	°	[330] SH-SY5Y, MPTP
	°	Y	°	Y	[264] SH-SY5Y, MPTP
	°	°	°	Y	[360] SH-SY5Y, MPTP
	°	Y	Y	Y	[362] SH-SY5Y, MPTP
	°	Y	°	Y	[265] SH-SY5Y, MPTP *
	°	Y	°	Y	[266] PC12, MPTP
	Y	Y	°	Y	[363] PC12, MPTP
	°	Y	°	Y	[270] PC12, MPTP
	Y	Y	°	Y	[267] MN9D, MPTP
	°	°	°	Y	[364] MN9D, MPTP
	°	Y	°	Y	[268] MN9D, MPTP
Prodromal to post-diagnosis phase	Y	Y	°	Y	[297] OBCs, rotenone *
	Y	Y	°	Y	[154] OBCs, rotenone *
	°	Y	°	Y	[263] SH-SY5Y, rotenone *,#
	°	Y	°	Y	[263] SH-SY5Y, MPTP *,#
	°	°	°	Y	[280] MN9D, rotenone #
	°	°	°	°	[345] SH-SY5Y, MPTP
Out of classification based on the clinical cardinal symptoms (no data on cell viability, studied only the molecular mechanisms)	°	°	Y	°	[202] PC12, reserpine
	°	Y	°	°	[201] PC12, reserpine
	°	Y	°	°	[227] SH-SY5Y, 6-OHDA *
	°	Y	°	Y	[228] SH-SY5Y, 6-OHDA
	Y	°	°	°	[235] SH-SY5Y, 6-OHDA
	°	°	°	Y	[340] PC12, 6-OHDA
	°	Y	°	Y	[226] MN9D, 6-OHDA
	°	°	°	Y	[232] MN9D, 6-OHDA
	°	°	Y	°	[234] MN9D, 6-OHDA
	°	Y	°	Y	[233] MN9D, 6-OHDA *
	°	Y	°	Y	[231] MN9D, 6-OHDA
	°	°	°	°	[365] LUHMES, 6-OHDA
	°	Y	°	Y	[275] OBCs, paraquat
	°	°	Y	Y	[292] SH-SY5Y, rotenone
	Y	Y	°	Y	[366] SH-SY5Y, rotenone
	°	°	Y	Y	[296] PC12, rotenone
	°	°	°	°	[329] PC12, rotenone
	°	°	°	°	[153] OBCs, MPTP
	°	°	°	°	[356] OBCs, MPTP
	°	Y	°	Y	[367] SH-SY5Y, MPTP
	°	°	°	°	[329] PC12, MPTP
	°	Y	Y	Y	[368] PC12, MPTP
	°	°	Y	°	[234] MN9D, MPTP

^§ supposingly linked to αSyn quantity; ^# depend on time and concentration of the treatment; * models investigating changes along a time course experimental scheme. Abbreviations: Y: assessed; °: not assessed; αSyn: α-synuclein; 6-OHDA: 6-hydroxydopamine; MPTP: 1-Methyl-4-phenylpyridinium; OBCs: organotypic brain culture; LUHMES: Lund human mesencephalic cell line; MN9D: mouse dopaminergic neuronal cell line.

3.2. Modeling the Prodromal Phase

The constellation of non-motor symptoms that are present during the prodromal phase of PD (Figure 1) provide opportunities to establish an early therapeutic window. These symptoms result from the pathology in DA and non-DA pathways in brain regions outside the nigrostriatal pathway (Figure 2). Both the pre-diagnosis symptoms, as well as NA, SE, and Chol dysfunction in extra nigrostriatal areas have been reproduced by in vivo models (6-OHDA; MPTP; paraquat; rotenone; reserpine; and VMAT2, LRRK2, and hA53T αSyn models, see Table 2) [156,157,158,159,160,161,162,163]. For example, hyposmia results from the intranasal infusion of a low MPTP dose in rats [369]. Accompanying hyposmia were TH⁺ decreases in the olfactory bulb, SN, striatum, and prefrontal cortex, followed by cognitive and motor deficits. This model is interesting because these effects recapitulate the prodromal stage, the symptomatic PD stage, and the risk phase as they result from an environmental toxin entering the brain through the olfactory tract. Similar results have been reported following intranigral stereotaxic injection of rotenone in rats. In the model, deceased DA in the SN and hyposmia are correlated with rapid eye movement (REM) sleep deficits [370]. Sleep disturbances are also produced also by 6-OHDA in rodents [224] and NHP [224,259,260]. Ex vivo (OBCs) VMAT2 blockage with reserpine inhibits DA and glutamatergic synaptic activity in the prefrontal cortex, hippocampus, and VTA, key cognitive and affective nuclei involved in prodromal and late PD symptoms. Applying DA restored synaptic activity early during the reserpine challenge but not later on [357], in agreement with L-DOPA therapy in the clinical setting. By extensively reproducing clinical data in prodromal PD, OBCs have shown a decreased synaptic spine number when treated with MPTP [153] and decreased neurite length and number when treated with rotenone [154]. This mimics the decreased synaptic DA preceding DOPAn loss noticed in clinic by 18-F Dopa scan/PET scan, thus supporting the utility of these ex vivo slow degenerating models in studying prodromal mechanisms of PD. In PC12 cells exposed to reserpine, the depletion of intracellular DA results in increased amounts of intracellular GSH, while GSH depletion results in DA demise, supporting the protective anti-oxidant role of GSH on DA, not only in the extracellular matrix, but also intracellularly [201].

Although rarely studied, the blood–brain barrier (BBB) may play a role in PD. As reported in the Lund Human Mesencephalic (LUHMES) cell model with a mild lesion induced by 6-OHDA and MPTP, damage may be reversed by exposing the cells to a pericyte-conditioned medium, emphasizing the importance of cells surrounding the BBB in the initial stages of the disease [360].

Gastrointestinal disorders (GI in Table 2), such as constipation and reduced intestinal motility, both of which are thought to precede classical signs of PD in humans, have been reported in rodents and primates exposed to chronic dosages of MPTP and rotenone [164,165,166]. Notably, the enteric DA level is also altered in these models. Enteric DA plays an immunomodulatory role [68], so the PD-induced DA modulation supports an additional link between PD and inflammation. Considering the putative role of the gastroenteric system in the risk-phase of PD, the causative dynamic among gastrointestinal disorders, DA loss, inflammation, and signs of PD should receive more attention.

3.3. Modeling the Risk Phase

Intragastric rotenone has been used to test (with positive results) the Braak hypotheses that PD synucleinopathy starts in the enteric system [371] and reaches the brain through the vagus nerve where it propagates like a prion [35,38,372]. Olfactory dysfunction is a well-known symptom of PD, occurring in 90%–96% of patients [373,374]. The absence of suitable models limits the study of olfactory pathology to post mortem samples from PD patients. Nevertheless, in organotypic olfactory bulb cultures, monoamine-induced DA loss occurs from the prediagnostic to prodromal phases of PD. This is mainly due to oxidative stress, mitochondrial dysfunction, and transcriptomic alterations [375]. The alterations include the upregulation of the amyloid-beta precursor protein (APP), hyperexpression of BDNF, and dysregulation of PARK genes such as PINK1 and DJ-1/PARK7, key factors involved in mitochondrial functioning and oxidative stress regulation [375].

Multiple genetic models of PD exist, frequently failing to reproduce all the classic clinical signs (Table 2) [314,376]. This supports the idea that only a minority of PD cases have a purely genetic origin. In short, mutations in DJ1/PARK7 belonging to the anti-oxidant defense system, lead to a weak PD phenotype without DA loss [193]. Supportive is the fact that the DJ1 genetic model in zebrafish has behavioral deficits, but only in the mitochondrial dysfunction among the cardinal molecular signs of PD [318,377]. In line with the involvement of DJ1 in redox-mediated disease processes, rodents with DJ1 KO challenged with 6-OHDA exhibited greater DA depletion, loss of TH⁺ neurons, and worse performance in behavioral tests compared to WT mice treated with 6-OHDA [378]. Considering the interplay among genes and environment, these findings support this mutation as a risk factor associated with redox stress [379], a condition that is prevalent in many neurodegenerative diseases. As a part of the PINK1–PARKIN protein complex that is required for the removal of damaged mitochondria, their malfunctioning leads to secondary absence of autophagy and accumulation of αSyn [379]. Specifically, the loss of the E3 ubiquitin ligase PARKIN leads to proteasomal dysfunctions and to the accumulation of parkin substrates (e.g., synphilin-1, O-glycosylated alpha-synuclein, Pael-R, CHIP, cdc-Rel1A, cyclin E, and synaptotagmin X1) [380]. Despite this, the genetic models do not reproduce LB formation and neurodegeneration [189,376]. Again, only the PARKIN-Q311X-DAT-BAC mice manifest progressive motor deficits and loss of nigrostriatal DA neurons. This mutation also results in dose-dependent neurodegeneration in rats [347]. Effects of PINK1 mutation in zebrafish is limited to behavioral deficits and mitochondrial dysfunction [318]. These models exhibit a lack of DOPAn loss in SN, mild locomotor disabilities [189,376], and mild anxiety behaviors [160,197]. PARKIN deficiency may also enhance the vulnerability of SN neurons to inflammation [381]. Supporting this was a study reporting that when PARKIN^+/− were injected with LPS, an inflammasome inhibitor reduced inflammation, neurodegenerative damage, and increased TH⁺ rescue compared to WT animals undergoing the same treatment. This highlights the role of PARKIN in modulating the inflammasome [382]. αSyn (SNCA) models often do not present DA loss, and most studies do not include behavioral testing, instead focusing on the measures of inflammation and microglia activation [189,193,197,205]. The zebrafish SNCA model exhibits reduced DA, mitochondrial activity, and motor deficits [318,383,384,385]. As discussed in the clinical section, the interplay among genetic, environmental, and lifestyle factors is now considered the most likely responsible for PD onset. In line, in a rodent αSyn transgenic model challenged with LPS, LPS triggered a progressive degeneration of the nigrostriatal dopamine pathway, αSyn aggregation and LBs formation, not present in the pure genetic model [348]. Interestingly, rotenone administration in genetic models of αSyn recapitulate the constellation of symptoms present in human PD. This is not possible using a single inducer, suggesting that αSyn overexpression enhances DA sensitivity to mitochondrial impairment and oxidative stress [191]. Nevertheless, as for the other PD-linked mutations, iPSCs derived from patients with SNCA variants are powerful tools for developing personalized medicine approaches [316,386]. Generally, mutations in LRRK-2 lead to mitochondrial dysfunction, oxidative stress, inflammation, autophagy, and proteasomal dysfunction [161], but not αSyn and LB formation, neurodegeneration, or frank motor deficits in animal models [55,161,189,197]. DOPAn loss in SN is present only with the G2019S LRRK2 mutation in rodents [387]. This mutation in iPSCs increases cell death and reduces branching complexity in DOPAn [388]. Like αSyn, zebrafish with LRRK2 mutation exhibit a more complete classic PD phenotype, with neuronal loss, αSyn aggregation, and motor deficits [318]. In iPSCs cells from a patient with LRRK2 variants and early on-set PD, increased expression of different oxidative stress-related genes and αSyn protein were detected [317].

Inflammation is gaining recognition as a potential trigger of clinical PD, and emerging studies are beginning to explore this hypothesis (Table 3). Choi et al. described in rats a progressive reduction when injected with LPS in the number of TH⁺ cells in SN (21%, 38%, and 41% at 1, 2, and 4 weeks, respectively) compared to the non-injected side [194,305]. These resulting data have been reported in rats [305,389,390] and mice [391], supporting the reproducibility of the model. Of relevance, these studies belong to the minority of experimental studies that measure PD readouts over time (Table 3 and Table 4), providing valuable temporal mechanisms that may be related to PD progression. Even more relevant to the risk phase, intrauterine injection of LPS in pregnant rodents leads not only to decreased DA in the brains of offspring, but also to enhanced damage from pro-inflammatory stimuli administered in the post-natal age [306,392]. In line with this, in OBCs challenged with low doses of rotenone to recreate a slow degenerative model of PD, inflammation, and specifically TNFα, has been demonstrated to be the determinant of DOPAn demise [154,297]. Notably, TNFα overexpression in SN leads to progressive neurodegeneration, with motor signs and microgliosis, demonstrating the pivotal role of this cytokine in DA loss [67]. Along risk factors, low systemic inflammation is present in diabetes, and in PC12 cells exposed to 6-OHDA cell death is almost completely reversed by the silencing of insulin growth factor binding protein 5 (IGFBP5) [359]. These results have been confirmed in vivo, where wild-type (WT) rats exposed to 6-OHDA presented motor abnormalities and TH⁺ decrease, while not observed in IGFBP5^−/− animals [359]. In line with this, in aging studies, IGFBP5 is correlated with cognitive dysfunction, also present in PD [393]. Aging, obesity, pre-diabetes, type 2 diabetes, and insulin resistance are debated risk factors in human PD, as discussed by others [33,172,177,394,395,396,397,398]. Conversely, low-dose LPS preconditioning to midbrain OBCs protected DOPAn against inflammatory damage by decreasing microglial activation and releasing pro-inflammatory factors [399]. A complementary in vitro approach (co-culture of SH-SY5Y cells, BV2 cells, and primary microglia) suggested that the adaptative anti-inflammatory behavior results from suppressing the TLR4-mediated NF-κB and MAPK p38 signaling pathways [399]. This suggests that the level of inflammation (and possibly TNFα) may determine the protective or damaging sequence in PD.

4. Conclusions

Due to the complexity of Parkinson’s disease (PD), current clinical management is largely limited to symptom control. There is a significant lack of prevention campaigns and disease-modifying strategies. To address these gaps, existing PD models should be more extensively and strategically utilized (Box 1). Clinical knowledge should serve as a validation tool for assessing the reproducibility of newly developed models, and it should not be the final goal. A critical need exists to accurately recapitulate the chronic progression of PD, especially to investigate the pre-diagnostic stages to identify biomolecular changes in the early phases. Our review highlights a striking lack of studies from risk to pre-clinical stages.

Box 1. Key-notes regarding the main failures/limits of the current models.

• Acute schemes approach does not model consistently the etiopathogenesis of human PD. This limits the translational relevance in the understanding of the disease and therapies screening.
• Focus on synucleopathy. The role of synucleopathy is debated, possibly relevant only to a subtype of PD. Possibly a late event.
• Up to now, large attention to the late (after diagnosis) molecular events. Need for discovery approaches, necessity in exploring the early stages.
• Disease complexity. Multiple mechanisms and neurotransmitters involved, in a different time scale. Need for experimental schemes exploring the different phases in a single study.
• Need for complex models. The temporal scale of human PD is too long for experimental studies. Models are necessary and fundamental.
• Need for solid, reproducible and shared models. The past studies build the basis on knowledge in how to properly mimic human PD. Using this background to pursuit the objectives suggested in the previous points by creating slow degenerative models, by using shared, uniform protocols, making data comparable and potentially complementary might be a plus.

Researchers often struggle to select the most appropriate model, as the predominant focus remains on the clinical (symptomatic) phase of PD. Research often isolates individual mechanisms without integrating them into the broader, interconnected reality of PD. Current knowledge demonstrates the extreme complexity of PD. Multiple and interconnected neurotransmitter systems are involved. Many non-motor symptoms described are present both in the prodromal and late phase. The extra CNS origin is more and more believed, not only the Braak hypotheses, but the interplay of genes with inflammation, as an example. The most promising approach to fill the gap looks to be the use of the diverse range of PD models with milder disease triggers replicating the early stages, both in vitro and in vivo. The study of multiple experimental times, in slow degenerative models, by a discovery approach will be ideal. In vitro models remain crucial for dissecting intrinsic cellular signaling pathways. Genetic models, although not fully representative of all PD symptoms or outcomes, are invaluable for studying the pathological consequences of specific mutations and the interaction between genomic variants, the environment, and lifestyle. Continuing developing and using complex in vivo models, despite the ethical concerns they raise, is fundamental to overstep the limitation of specimens, excessively long timing, and the ethical concerns of clinical studies on PD patients.

A comprehensive view that balances complexity with clinical applicability and reproducibility is essential. Despite its high public profile, PD biologically remains largely enigmatic. Clinical and, above all, basic research must continue to delve deeper into its mechanisms, exploiting new omics technologies to perform discovery of the disease.