Role of rodent models in advancing precision medicine for Parkinson’s disease

Abstract

With a current lack of disease-modifying treatments, an initiative toward implementing a precision medicine approach for treating Parkinson’s disease (PD) has emerged. However, challenges remain in how to define and apply precision medicine in PD. To accomplish the goal of optimally targeted and timed treatment for each patient, preclinical research in a diverse population of rodent models will continue to be an essential part of the translational path to identify novel biomarkers for patient diagnosis and subgrouping, understand PD disease mechanisms, identify new therapeutic targets, and screen therapeutics prior to clinical testing. This review highlights the most common rodent models of PD and discusses how these models can contribute to defining and implementing precision medicine for the treatment of PD.

INTRODUCTION

As the second most common neurodegenerative disease, Parkinson’s disease (PD) is estimated to impact more than 12 million people worldwide by 2040 (GBD 2016 Parkinson’s Disease Collaborators, 2018; Dorsey et al., 2018). PD is clinically characterized by motor symptoms including resting tremor, bradykinesia, rigidity, and postural instability. The hallmark pathology associated with symptoms includes dysfunction and degeneration of dopaminergic neurons in the substantia nigra pars compacta (SNpc) and the development of Lewy bodies and Lewy neurites, cytoplasmic inclusions containing aggregated forms of the presynaptic protein α-synuclein (aSyn; Spillantini et al., 1997; Cheng et al., 2010). Currently, available therapeutics aim to relieve motor symptoms through dopamine replacement or deep brain stimulation; however, no disease-modifying treatments are available to prevent, slow, or stop the progression of degeneration.

Following continued failures in clinical trials and a growing appreciation for the heterogeneity in PD pathology, the notion of precision medicine has emerged as an innovative new approach for treating PD. Precision medicine is a term that has evolved over time and refers to choosing treatments that are optimally timed and targeted to the specific disease characteristics of the individual patient (Espay et al., 2017). While the concept of precision medicine is relatively agreed upon, the challenge remains in how this broad definition can be applied in the clinic. Due to the heterogeneous nature of PD, clinicians and researchers recommend stratifying PD patients into subgroups based on shared disease characteristics. However, there is a clear lack of consensus regarding which factors should be used to define these subgroups and how those factors will be measured in patients. In this review, a role for rodent models in the pursuit for a precision medicine approach in PD is discussed.

CHALLENGES TO SUBGROUPING IN PD

Several studies have attempted to subgroup PD patients based on clinical representation of motor and nonmotor symptoms (Lewis et al., 2005; Selikhova et al., 2009; Brennan et al., 2017; Fereshtehnejad et al., 2017; De Pablo-Fernández et al., 2019). Indeed, these classifications have merit in their ability to distinguish patients with motor only symptoms versus those with motor and cognitive symptoms, and to predict the rate of progression in some subgroups of patients (Lewis et al., 2005; Selikhova et al., 2009; Brennan et al., 2017; Fereshtehnejad et al., 2017; De Pablo-Fernández et al., 2019). These classifications have been useful for choosing the optimal symptom-targeted treatment regimen and beneficial in understanding diverse clinical phenotypes; however, symptomatic parameters do not necessarily reflect the underlying mechanisms of the disease and are not always represented by pathological biomarkers (Selikhova et al., 2009; De Pablo-Fernández et al., 2019). Additionally, symptom presentation typically occurs several years after the initiation of pathology and substantial irreversible damage has already occurred (Ma et al., 1997; Kilzheimer et al., 2019). Thus, biomarkers for earlier detection and patient subgrouping are crucial as disease-modifying therapies are developed and tested in clinical trials.

More recently, researchers and clinicians have been subgrouping PD patients based on genetic variations which are linked to an increased risk of PD. These studies have provided useful insight into the clinical phenotypes and rate of disease progression associated with specific genetic mutations. Furthermore, pathological features of genetic forms of PD, such as mitochondrial dysfunction or accumulation of aSyn, have been observed in patients with sporadic PD suggesting that they may provide valuable information regarding mechanistic pathways of degeneration in both familial and idiopathic PD. However, stratification of PD patients based on genetics alone, while a useful starting point, is not directly applicable to the majority of PD cases as only 1% to 2% of PD cases are caused by genetic mutations and 5% to 10% are strongly associated with risk genes (Pankratz & Foroud, 2007; Bandres-Ciga et al., 2020). Therefore, defining subgroups based on a combination of factors including both genetics and mechanistic similarities such as dysfunctional organelles (i.e., mitochondria or lysosomes), accumulation of pathological aSyn, and/or inflammatory markers may be a promising approach for stratification of idiopathic cases of PD. For mechanistic stratification to become possible, it will require a better understanding of the cellular and molecular features underlying PD to define subgroup parameters, determination of appropriate biomarkers to identify patients belonging to each subgroup, and identification of the optimal treatments to target specific mechanisms of pathology. Studies in animal models will be an important component to achieving this goal.

RATIONALE FOR MODELING PD IN RODENTS

Establishing stratification of PD patients in clinical studies is challenging in that the trials can be costly and lengthy in duration; it is difficult to obtain enough volunteers to subdivide patients while maintaining statistical power; environmental factors and lifestyle habits are difficult to control and may be confounding factors. For these reasons, mechanistic studies in animal models of PD will continue to be an essential component to better understand the heterogeneous etiologies of PD, identify subgroup biomarkers, and, ultimately, to advance the use of precision medicine in PD patients. In fine with the heterogeneity of PD in humans, a diverse array of animal models representing genetic mutations, environmental toxins, and common pathological features seen in human PD patients are available to investigate mechanisms and biomarkers of disease and to screen novel therapies prior to clinical testing. Highlighted below are a selection of the most common rodent models of PD and their value in helping to define and advance a precision medicine approach for PD.

Rodents have been used for decades to model disease and test potential therapies. They are part of the translational spectrum that includes different in vitro cell systems and patient-derived induced pluripotent stem cells (IPSCs), three dimensional in vitro systems such as stem cell-derived organoids, and drosophila and zebrafish for gene x environment studies and screening (De Souza, 2018; Valadez-Barba et al., 2020). In PD, rodent models were instrumental in the discovery of dopamine, its association with movement dysfunction in PD, and for L-DOPA administration as symptomatic treatment for PD (Carlsson, 1993; Lees et al., 2015). Rodents are good models for studying PD as they are a mammalian species with strong overlap in anatomy, physiology, and genetics to humans. Rodents also have a basic repertoire of behaviors that are similar to humans (e.g., voluntary movement, fear response, and cognitive functions that include attention, strategy switching, and different forms of memory). Their breeding, short gestation period, litter size, and physical size are amenable to genetic manipulation and studying mechanisms underlying cellular function and dysfunction. However, since there are no reported cases of PD developing spontaneously in animals, only humans are the perfect model of the disease. Some of the major limitations of using rodents to model PD include the relatively short lifespan (~2 years), limited higher order behaviors, and limited fine motor skills. In addition, while there is good anatomical overlap between rodents and humans they are obviously not the same. In the rodent brain, the caudate and putamen are not separated by the internal capsule as they are in nonhuman and human primate brains and dopamine neurons in the substantia nigra pars compacta lack melanin in rodents but not in nonhuman and human primates. Furthermore, rodents are quadrupedal animals, using all four limbs to bear weight during locomotion rather than bipedal as is the case with humans. While rodents provide a strong platform for understanding the basis of dysfunction in PD, the direct translation of results in rodents to humans needs to take into consideration the limits of the animal system.

GENETIC MODELS OF PARKINSON’S DISEASE

Numerous transgenic and knockout models have been developed to reflect genetic variations known to cause or increase the risk of developing PD (Table 1.1.). As many transgenic models of PD are generated using the same human gene mutations seen in PD patients, they can provide valuable information on the sequence of events leading to cellular and behavioral dysfunction. However, many genetic models of PD do not develop extensive and progressive degeneration of nigrostriatal dopaminergic neurons. This is likely due in part to the short lifespan of rodents and the slow progression of pathology seen in PD. While this is certainly a limitation of many genetic models, it may also present an opportunity to investigate dysfunction earlier in the disorder to facilitate identification of biomarkers of disease onset and novel therapeutic targets prior to irreversible neuron loss.

Table 1.1.

Pathological mechanisms in genetic models of Parkinson’s disease

Rodent model	Gene	Primary PD disease mechanism(s)	Secondary disease mechanism(s)	Dopaminergic neuron loss	Motor deficits	Non-motor deficits
A53T-aSyn	SNCA	Mutated aSyn-mediated pathology	Mitochondrial dysfunction Inflammation	No	Yes	Cognitive Sensorimotor
Thy1-aSyn “Line 61”	SNCA	Human WT aSyn-mediated pathology	Mitochondrial dysfunction Inflammation	No	Yes	Cognitive Sensorimotor Olfactory Disrupted sleep patterns
GBA KO	GBA	Lysosomal dysfunction	None	No	No	None
GBA L444P heterozygous mutation	GBA	Lysosomal dysfunction	Disrupted mitophagy & Autophagy aSyn pathology	No	Yes
LRRK2 KO	LRRK2	None observed	Abnormal peripheral organ function (kidney, liver, lung, & spleen)	No	No	None observed
LRRK2 G2019S mutation	LRRK2	Vesicle trafficking, autophagy	Mitochondrial dysfunction Neuroinflammation aSyn pathology	Yes- in some mouse models	Yes	Anxiety & depressive-like behavior
PINK1 KO	PINK1	Mitochondrial dysfunction	aSyn pathology	Yes	Yes	Swallowing Vocalizations Olfactory

Furthermore, disruptions in biochemical pathways triggered by PD-associated genetic mutations are the same pathways implicated in many cases of sporadic PD, allowing for reasonable translation to sporadic forms of PD. As PD is considered to be triggered by a combination of genetic susceptibility and environmental factors, genetic models can also be combined with neurotoxin exposure to assess the vulnerability of genetic mutations to specific environmental factors. Currently, with the advancement of gene editing techniques, genetic models will continue to be invaluable for understanding PD pathogenesis, mechanistic stratification of patients, identifying novel therapeutic targets, and screening new therapies.

A53T aSyn mice

The first familial form of PD was identified as having a missense mutation with an amino acid substitution of alanine-to-threonine at codon 53 in the SNCA gene that encodes for aSyn (Polymeropoulos et al., 1997). Mice overexpressing the human A53T mutation were one of the earliest transgenic models generated and are still commonly used to study aSyn-related pathological mechanisms in PD (Giasson et al., 2002; Lee et al., 2002). The frequently used strains overexpress the human A53T mutation via the prion promoter and results in aSyn pathology in the midbrain, cerebellum, brainstem, and spinal cord in association with astrogliosis, microgliosis, and mitochondrial dysfunction (Giasson et al., 2002; Lee et al., 2002; Ordonez et al., 2018; Hu et al., 2019; Li et al., 2019). Inclusions of aSyn develop in this model and resemble filamentous aSyn inclusions seen in human PD (Giasson et al., 2002; Lee et al., 2002). However, A53T mice develop degeneration within the spinal cord and show spinal cord-related motor decline (Giasson et al., 2002; Lee et al., 2002). Furthermore, they do not develop dopaminergic neuron loss. Most studies using this model focus on aSyn-related dysfunction prior to spinal cord degeneration. Indeed, substantial molecular changes such as alterations in long-term depression and synaptic transmission and behavioral deficits such as impairments in rotarod performance and Y-maze performance develop prior to severe motor decline (Paumier et al., 2013). Although this is a commonly used model for mechanistic and target validation studies, it is important to recognize that A53T mutations in PD are very rare and findings may not be representative of what occurs in sporadic cases of PD.

Thy1-aSyn mice (a.k.a. Line 61)

In addition to missense mutations in aSyn, triplication and duplication of the aSyn locus have also been identified in familial PD (Singleton et al., 2003; Chartier-Harlin et al., 2004). These cases indicate that elevated levels of aSyn can lead to the development of PD. Multiple lines of mice that overexpress human wild-type aSyn have been generated to study the pathological effects of elevated aSyn (Rockenstein et al., 2002). Mice that overexpress wild-type human aSyn under the Thy 1 promoter are one of the most well characterized aSyn transgenic lines that continues to be utilized today (Rockenstein et al., 2002). They have a broad distribution of aSyn accumulation and age-related reduction in striatal dopamine (Rockenstein et al., 2002; Lam et al., 2011). Importantly, phosphorylated (Ser 129) and proteinase K-resistant aSyn is observed in this model (Chesselet et al., 2012; Lam et al., 2011). These aSyn aggregates are present in several brain regions including the striatum, substantia nigra, locus coeruleus, cortex, and olfactory bulb (Chesselet et al., 2012). These mice display PD-like motor symptoms as well as early non-motor symptoms often found in human PD such as cognitive impairments, olfactory deficits, sleep dysregulation, gastrointestinal dysfunction, and autonomic cardiovascular dysfunction (Fleming et al., 2004, 2008, 2013; Wang et al., 2008, 2012; Magen et al., 2012; McDowell et al., 2014). Mechanistically, early and prolonged inflammation including activated microglia, reactive astrocytes, and upregulation of the pro-inflammatory cytokine TNF-α have been reported (Watson et al., 2012). Mitochondrial alterations are also prevalent in this model with mitochondrial respiration defects shown in the ventral midbrain and striatum, and specifically inhibition of complex 1 activity and increased oxidative stress in the ventral midbrain (Subramaniam et al., 2014). While this model displays several PD-associated biochemical and behavioral features, Thy1-aSyn transgenic mice do not develop dopaminergic neuron degeneration. Despite this limitation, the myriad of PD-associated pathological features in this model make it valuable for the investigation of early biomarkers and therapeutic targets prior to the irreversible loss of dopamine neurons. Additionally, this model has been shown to display an increased vulnerability to neurotoxins such as 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine and paraquat, which can exacerbate pathology resulting in mitochondrial dysfunction and neuronal axon degeneration (Song et al., 2004; Fernagut et al., 2007). As PD is often considered to result from a combination of genetic vulnerability and environmental exposures, this model can be combined with neurotoxins to induce specific pathological features seen in individual patients. While this model is still commonly utilized, elevated levels of aSyn are not consistently reported in sporadic PD (Duffy et al., 2018a,b).

GBA1

Gaucher’s disease is the most common lysosomal storage disorder and is associated with mutations in GBA1, the gene encoding the lysosomal enzyme glucocerebrosidase (GCase). Mutations in GBA1 are associated with an increased risk of developing Lewy body disorders including PD (Goker-Alpan et al., 2004; Satake et al., 2009). Furthermore, GBA1 variants are more common in PD patients than in age-matched controls (Sidransky et al., 2009). Additionally, patients with sporadic PD without GBA1 mutations also demonstrate reduced GCase in the brain, supporting the idea of shared mechanisms of pathology between familial and sporadic PD and suggesting that sporadic PD patients may also benefit from therapeutics developed and screened in gba genetic mouse models (Murphy et al., 2014; Chiasserini et al., 2015).

Multiple mutations in gba have been modeled in mice including L444P, D409H, and V394L in addition to knocking out gba (Migdalska-Richards et al., 2017; Yun et al., 2018; Li et al., 2019). Some models can develop aSyn accumulation, for example, at 24 months, heterozygous gba null mice show aSyn aggregation in the hippocampus and striatum (Xu et al., 2011; Migdalska-Richards et al., 2017). However, in several studies the Gaucher models are crossbred with aSyn transgenic mice or aSyn is overexpressed in specific brain regions using adeno-associated viral vectors to determine the effect of impaired GCase function on aSyn toxicity (Fishbein et al., 2014; Migdalska-Richards et al., 2017; Kim et al., 2018; Henderson et al., 2020). Indeed, combining the L444P heterozygous GBA mutation or heterozygous null mutation with A53T-aSyn mice or aSyn overexpressing mice via adeno-associated vims injection exacerbated aSyn pathology and enhanced dopaminergic neuronal degeneration (Fishbein et al., 2014; Migdalska-Richards et al., 2017). Potential therapies targeting GCase function have also been tested in PD models (Richter et al., 2014; Burbulla et al., 2021).

LRRK2

Mutations in the leucine-rich repeat kinase 2 (LRRK2) gene represent the most common genetic cause of PD accounting for roughly 4% to 5% of familial cases and 1% to 3% of sporadic cases (Gilks et al., 2005; Khan et al., 2005; Lesage et al., 2007; Paisán-Ruíz et al., 2008). Importantly, LRRK2-associated PD is phenotypically and pathologically indistinguishable from sporadic PD suggesting that information gained from LRRK2 models of PD may have strong translational relevance to several other subgroups of sporadic PD patients (Paisán-Ruíz et al., 2004; Zimprich et al., 2004). Given its prominence in both familial and sporadic PD, mutant LRRK2 animal models have been important in elucidating the mechanisms underlying pathology and identifying novel potential therapeutics.

The LRRK2 gene encodes for an intracellular enzymatic protein involved in a diverse range of cellular processes, including vesicle trafficking, autophagy, and protein translation (Shin et al., 2008; Piccoli et al., 2011; Nikonova et al., 2012; Pellegrini et al., 2018; Sheehan & Yue, 2019). Mutations in LRRK2 can affect kinase and GTPase activity leading to elevated kinase activity (Healy et al., 2008; Islam & Moore, 2017; Steger et al., 2016). Mice and rats share a LRRK2 homolog with 86% to 88% conserved sequence identity to the human LRRK2 gene, including all residues impacted by pathogenic mutations identified in human PD (West et al., 2014). Multiple LRRK2 knockout, knockin, and overexpression rodent models have been developed to study the function of LRRK2 in health and disease. In general, the LRRK2 rodent models on their own show modest to no effects on the nigrostriatal system and aSyn pathology (Lin et al., 2009; Tong et al., 2010, 2012; Ramonet et al., 2011; Zhou et al., 2011; Chen et al., 2012; Hinkle et al., 2012; Baptista et al., 2013; Daher et al., 2014; West et al., 2014; Sloan et al., 2016; Weng et al., 2016; Xiong et al., 2018). However, when combined with aSyn overexpression or intracranial injection of aSyn preformed fibrils, the LRRK2 G2019S mutation can cause an increase in aSyn burden and enhanced cell loss in the substantia nigra (Lin et al., 2009; Volpicelli-Daley et al., 2016; Henderson et al., 2019; Bieri et al., 2019). Inhibition of LRRK2 activity with small-molecule inhibitors or anti-sense oligonucleotides has been shown to be beneficial in animal models (Lee et al., 2010; Daher et al., 2015; Zhao et al., 2017). LRRK2 is being pursued in biomarker development and LRRK2 inhibitors are currently in clinical trials (Kelly and West, 2020).

PINK1

PINK1 mutations are found in 4% to 9% of PD patients in the Asian population and 2% to 4% of Caucasian PD patients, making them the second most common cause of autosomal recessive PD (Schulte & Gasser, 2011). PINK1 is a mitochondrial serine/threonine protein kinase involved in mitochondrial control and dynamics. Studies show PINK1 acts as a sensor which promotes Parkin activation for the degradation of defective mitochondria. Indeed, PINK1 knockout rodents demonstrate mitochondrial dysfunction and oxidative stress similar to that seen in human PD (Gautier et al., 2008; Stauch et al., 2016; Villeneuve et al., 2016). PINK1 knock-out mice and rats are robust models for investigating early biomarkers in PINK1 related human PD cases as they display similar prodromal symptoms to humans such as olfactory, swallowing, vocalization, and gait disturbances (Dave et al., 2014; Grant et al., 2015; Cullen et al., 2018; Kelm-Nelson & Gammie, 2020). Although inconsistent, there are reports of nigrostriatal dopaminergic cell loss in the rat PINK1 knockout model (Dave et al., 2014; De Haas et al., 2019). Hallmark aSyn pathology has also been detected in the substantia nigra, nucleus ambiguous, and locus coeruleus (Grant et al., 2015). Importantly, a recent large-scale gene sequencing study matched the gene databases of PINK1 null rat brainstem tissue with human idiopathic PD and PARK6/2 tissue, demonstrating valid translational relevance of this model with human PD (Kelm-Nelson & Gammie, 2020). This model has been particularly important for studying cranial sensorimotor impairments related to PD such as vocalization and swallowing (Grant et al., 2015; Hoffmeister et al., 2021). These are significant PD symptoms that can contribute to reduced quality of life and mortality and are understudied in the disease. Exciting new studies show a relationship between brainstem norepinephrine and impairments in vocalization and anxiety in this model (Hoffmeister et al., 2021).

INDUCIBLE α-SYNUCLEIN OVEREXPRESSING MODELS

Alpha-synuclein accumulation is a hallmark feature of human pathology in both familial and sporadic cases of PD. It is clear that aSyn accumulation is an important factor in PD and, thus, aSyn models currently dominate the field (Table 1.2). In addition to transgenic models of aSyn, other models of aSyn expression include recombinant adeno-associated viral vector (AAV) mediated aSyn and the aSyn preformed fibril (PFF) models (Kirik et al., 2002; Luk et al., 2012). AAV-aSyn models have provided important information regarding the biochemical properties of aSyn aggregates, and how aSyn overexpression contributes to nigrostriatal degeneration while the aSyn PFF model is a powerful tool to study aSyn seeding.

Table 1.2.

Inducible models of aSyn pathology

α-Syn overexpression model	α-Syn pathology	Dopaminergic neuron loss
Viral vector-mediated aSyn overexpression	Substantia nigra	Yes
Preformed Fibril Injection (striatum)	Striatum Frontal cortex Olfactory bulbs Amygdala	Yes
Preformed Fibril Injection (gut)	Medulla Locus Coeruleus Striatum Prefrontal cortex Amygdala Hippocampus Olfactory bulbs	One study showed DA neuron loss (Kim et al., 2019

Viral Vector-Induced Alpha-Synuclein Overexpression

Viral vector mediated overexpression of wild-type or mutant aSyn is a useful tool for inducing PD-like aSyn pathology. Mutated or wild-type aSyn genes are overexpressed via a recombinant adeno-associated viral (AAV) or lentiviral vector allowing for targeted delivery into neuronal cells of specific tissues such as dopaminergic neurons in the substantia nigra. One of the advantages of targeted viral vector-mediated aSyn overexpression is the reproducibility of dopaminergic cell loss in the substantia nigra which is generally lacking in the transgenic models. In rodents, targeted AAV-aSyn overexpression results in progressive loss of dopaminergic neurons in the substantia nigra and motor deficits (Kirik et al., 2002). Pathology initiates rapidly in the AAV-aSyn model with significantly reduced dopamine reuptake as early as 10 days postinjection, followed by a slow and progressive pattern of degeneration with synaptic and axonal deficits occurring prior to dopaminergic cell death (Butler et al., 2015; Kordower et al., 2013; Lundblad et al., 2012). AAV-aSyn models also develop a robust neuroinflammatory response that includes changes in early microglial activation, upregulation of pro-inflammatory cytokines, and infiltration of lymphocytes prior to dopaminergic neuronal loss (Chung et al., 2009; Cao et al., 2010; Daher et al., 2014; Reish & Standaert, 2015; Harms et al., 2018). Furthermore, coinjections of viral-induced aSyn overexpression and GBA or Parkin genes, or AAV-aSyn overexpression in transgenic models of PD allow for investigation of the interaction between aSyn and PD-associated genes in the progression of pathology (Lo Bianco et al., 2004; Rocha et al., 2015). Indeed, AAV-aSyn overexpression in mice expressing the G2019S-LRRK2 mutation showed enhanced inflammation and dopaminergic neuron cell loss which was ameliorated with administration of a LRRK2 inhibitor (Daher et al., 2014). AAV-aSyn models alone or in combination with other genetic or environmental factors are robust models of aSyn pathology in PD and may help elucidate aSyn mediated degeneration to identify aSyn-related targets for earlier diagnosis and therapeutic intervention. However, like the transgenic aSyn overexpression models, increased levels of aSyn is not consistently reported in sporadic PD (Duffy et al., 2018a,b).

Alpha-Synuclein Preformed Fibril

The development of synthetic aSyn preformed fibrils (PFF) has been a significant advancement in modeling PD. Exogenous aSyn fibrils injected into the striatum act as seeds which trigger phosphorylation and accumulation of endogenous aSyn into neurotoxic oligomers and fibrils (Luk et al., 2012; Paumier et al., 2015; Duffy et al., 2018a,b; Patterson et al., 2019). Injection of PFFs closely mimic several pathological characteristics of human PD including aSyn accumulation, mitochondrial impairment, lysosomal dysfunction, synaptic alterations, progressive degeneration of dopaminergic neurons, and seeding of aSyn pathology through cell-to-cell transmission into regions of the brain directly connected to the site of injection be it striatum or substantia nigra (Volpicelli-Daley et al., 2011; Luk et al., 2012; Luk & Lee, 2014; Paumier et al., 2015; Duffy et al., 2018a,b; Patterson et al., 2019). This model is now widely used for target validation, mechanistic studies, and testing novel therapeutics.

Lewy body pathology is not confined to the central nervous system in PD and has been observed in peripheral tissues including the enteric nervous system (Wakabayashi et al., 1988, 1990; Braak et al., 2006; Beach et al., 2010, 2016; Barrenschee et al., 2017). Gastrointestinal (GI) dysfunction and constipation are established nonmotor symptoms in PD, often occurring prior to the onset of motor symptoms (Chen et al., 2015; Gao et al., 2011; Park et al., 2015; Sung et al., 2014). In addition, GI dysfunction is also reported in animal models overexpressing aSyn (Wang et al., 2008, 2012; Kuo et al., 2010; Hallett et al., 2012; Farrell et al., 2014). These observations combined with pathological staging studies by Braak et al. (2003a,b) led to the hypothesis that the gut may be a site of PD initiation and progress via the vagus to the central nervous system. While this hypothesis has received some pushback from studies demonstrating conflicting patterns of Lewy pathology (Jellinger, 2019), this progression of PD pathology from gut to the central nervous system may help to define subpopulations of PD patients and help to advance biomarker development and precision medicine approaches. To date, several groups have tested this hypothesis in rodents by injecting aSyn PFFs into the gut and tracking aSyn progressive pathology (Manfredsson et al., 2018; Uemura et al., 2018, 2020; Kim et al., 2019; Challis et al., 2020). Injection of aSyn PFFs in the gut have met with mixed results including development of pathological aSyn expression in enteric neurons, GI dysfunction, intestinal inflammation, and in some cases spread of aSyn pathology to CNS structures in hindbrain and one report showing aSyn pathology in the nigrostriatal system (Manfredsson et al., 2018; Uemura et al., 2018; Kim et al., 2019; Challis et al., 2020). Discrepancies in reported outcomes may be due to differences in aSyn PFF concentrations, site of injection, and timing. While it will be important to define the precise parameters for aSyn PFF injection in the gut, this model would be vital for identifying early peripheral biomarkers for a subgroup of PD patients prior to significant damage in the central nervous system and onset of motor symptoms. Furthermore, animal models of gut-brain aSyn propagation will be essential to develop novel therapeutics to prevent the spread of pathology from the enteric nervous system to central nervous system structures.

NEUROTOXIN MODELS OF PARKINSON’S DISEASE

Animal models used to study the pathological mechanisms underlying sensorimotor function in PD often involve the administration of toxins that selectively destroy nigrostriatal dopaminergic neurons (Table 1.3). These models include rodents treated with 6-hydroxydopamine (6-OHDA) and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), both have and continue to contribute to our understanding of basal ganglia circuitry and voluntary movement (Ungerstedt, 1968; Burns et al., 1983; Carlsson, 1993). More current toxin models include paraquat and rotenone-treated rodents and these are particularly useful for studying gene–environment interactions in PD (Brooks et al., 1999; Betarbet et al., 2000: Fernagut et al., 2007). Overall, the toxin models tend to have good face validity for sensorimotor signs in that they display impairments in movement initiation, weight shifting, and postural stability that are reversible with levodopa similar to patients with PD (Tillerson et al., 2002; Alam & Schmidt, 2004; Fleming et al., 2005).

Table 1.3.

Neurotoxic models of Parkinson’s disease

Toxin	Toxic mechanism(s)	Dopaminergic neuron loss	aSyn pathology
6-OHDA	Production of ROS Mitochondrial dysfunction Neuroinflammation	Yes	No
MPTP	Mitochondrial complex I inhibitor Oxidative stress Neuroinflammation	Yes	aSyn nitration
Paraquat	Oxidative stress	Yes	aSyn upregulation
Rotenone	Mitochondrial complex I inhibitor	Yes	aSyn aggregates

CONCLUSION

Parkinson’s disease is a multisystem heterogeneous neurodegenerative disorder with a wide range of genetic and environmental risk factors leading to diverse clinical and pathological phenotypes. In order to move toward precision medicine in PD, rodent models will continue to be an important component of the translational pathway (Table 1.4). Choosing the appropriate model or combination of models to represent potential subgroups of PD patients will help in elucidating the underlying mechanism of disease, identifying biomarkers to diagnose and stratify PD patients, and screening new therapeutics prior to clinical trials. Furthermore, as new therapeutics are developed, preclinical trials in animal models will continue to be a critical step in understanding the pharmacokinetics, ability to cross the blood–brain barrier, target specificity, potential toxicity, and drug interactions of each therapeutic prior to testing in humans.

Table 1.4.

The advantages of animal models in defining and implementing precision medicine for Parkinson’s disease

Model	Biomarker development	Elucidation of disease mechanisms & identification of therapeutic targets	Preclinical drug screening
Transgenic/Knockout	Display PD-associated prodromal symptoms and pathology	Directly related to genetic mutations in human PD patients Disease mechanisms overlap with sporadic PD Can be combined with aSyn overexpression or neurotoxic models to study gene-aSyn or gene-toxin interactions	Can represent genetic-linked subgroups of PD patients to predict safety and efficacy of therapeutics
aSyn Overexpression	Demonstrate early stages of PD pathology Display PD-associated prodromal symptoms	Representative of a major pathological mechanism in human PD Display PD-associated disease pathology and symptoms Can be combined with transgenic or neurotoxic models to study interaction of aSyn with genetic or environmental factors	Can predict impact of therapeutics on preventing spread of aSyn pathology Can demonstrate drug interaction with aSyn and related pathology
aSyn Preformed Fibril	Site of injection can mimic earliest region of aSyn pathology identified in human PD	Can study seeding and mechanisms of spreading pathology Display disease pathology seen in familial and sporadic human PD	Can predict impact of therapeutics on preventing spread of aSyn pathology
Neurotoxin	Mimic symptoms of human PD Can be used to identify early vulnerabilities or triggers for neurotoxin induced PD	Disease mechanisms overlap with human PD Some are directly related to environmental exposures impacting human PD	Can predict response of neurotoxin induced PD to drugs Can predict symptomatic response to drugs