Tracking extranigral degeneration in animal models of Parkinson’s disease: quest for effective therapeutic strategies


Address correspondence and reprint requests to Naren L. Banik, Division of Neurology, Department of Neurosciences, Medical University of South Carolina, 96 Jonathan Lucas Street, Suite 309, Charleston, SC 29425, USA. E-mail:


J. Neurochem. (2011) 118, 326–338.


Sporadic Parkinson’s disease (PD) is now interpreted as a complex nervous system disorder in which the projection neurons are predominantly damaged. Such an interpretation is based on mapping of Lewy body and Lewy neurite pathology. Symptoms of the human disease are much widespread, which span from pre-clinical non-motor symptoms and clinical motor symptoms to cognitive discrepancies often seen in advanced stages. Existing symptomatic treatments further complicate with overt drug-irresponsive symptoms. PD is better understood by assimilation of extranigral degenerative pathways with nigrostriatal degenerative mechanisms. The term ‘extranigral’ appeared first in the 1990s to more rigorously define the nigral pathology by process of elimination. However, as clinicians progressively identified PD symptoms unresponsive to the gold standard drug l-DOPA, definitions of PD symptoms were redefined. Non-motor symptoms prodromal to motor symptoms just as pre-clinical to clinical, and conjointly emerged the concept of nigral versus extranigral degeneration in PD. While nigrostriatal degeneration is responsible for the neurobiological substrates of extrapyramydal motor features, extranigral degeneration corroborates a vast majority of other changes in discrete central, peripheral, and enteric nervous system nuclei, which together account for global symptoms of the human disease. As an extranigral site, spinal cord degeneration has also been implicated in PD progression. Interconnected to the upper CNS structures with descending and ascending pathways, spinal neurons participate in movement and sensory circuits, controlling movement and reflexes. Several clinical and in vivo studies have demonstrated signs of parkinsonism-related degenerative processes in spinal cord, which led to recent consideration of spinal cord as an area of potential therapeutic target. In a nutshell, this review explores how the existing animal models can actually reflect the human disease in order to facilitate PD research. Evolution of extranigral degeneration studies has been succinctly revisited, followed by a survey on animal models in light of recent findings in clinical PD. Together, it may help to develop effective therapeutic strategies for PD.

Abbreviations used



Alzheimer’s disease






incidental Lewy body disorder


Lewy body (bodies)


Lewy neurite




Parkinson’s disease


substantia nigra


substantia nigra pars reticulata


tyrosine hydroxylase

Parkinson’s disease (PD) is a complex disorder with slowly progressing, unremitting neuronal death. The disease represents multiple stages with progressive disability and has no known cure (Dickson et al. 2009). Multitude of animal models has been developed over several decades for PD research. Primary objective of generating any animal model of a human disease is to better understand the underlying pathology and to facilitate tests of potential therapeutics leading to discovery of drugs for the disease. However, with the evolution of newer concepts and refinements in the field from the clinical realm, experimental models need revisiting. In the present review, the focus is on the recent advances in PD research and its experimental animal models. Neurotoxin based model with MPTP in rodents and primates, environmental toxin based model with rotenone, pesticides paraquat and maneb based models, inflammatory compound bacterial lipopolysaccharide (LPS) induced model and a range of transgenic mice models, wherein single genes are considered as disease causing/modifying agents, have been included. While basic characterization of these models is beyond the scope of the present discourse, hence, each of these models is reconsidered to assess the extent of incorporation of the recent advances in PD from clinical perspective with special reference to extranigral degeneration. Such inclusion may culminate in disease modifying therapeutic strategies.

Current understanding of sporadic PD: nigral versus extranigral pathology

Clinical symptoms, which emerge during the course of sporadic PD, can neither be fully explained by lack of dopamine (DA) in midbrain nigrostriatal system, nor can they be completely addressed by dopaminergic (DAergic) replenishment therapy. It is a well-documented fact that the Lewy pathology (alpha-synuclein inclusion body pathology) associated with sporadic PD is distributed throughout the central, peripheral, and enteric nervous systems (Jellinger 1999, 2009). However, the subsequent neuronal dysfunction is not a random phenomenon, rather selectively confined to some neuroanatomical nuclei, and is not strictly neurotransmitter-dependent (Braak and Del Tredici 2008). Consequent neuronal loss in sporadic PD has been proposed to follow a topographic distribution pattern of the Lewy pathology (Jellinger 1999, 2009). Attempts for diagnostic staging in the brain of PD patients taking into account the nigral and extranigral degeneration have been robust (Braak et al. 2003; Przuntek et al. 2004), but had certain limitations as discussed recently (Dickson et al. 2009). A major omission in these attempts was exclusion of spinal cord in parkinsonian degeneration (Dickson et al. 2009; Lim et al. 2009) and hence, the current review includes existing evidences from clinical and experimental fields to address selective involvement of spinal cord degeneration in PD. Furthermore, the review converges on the central theme – revisiting the animal models of PD and verifying our current understanding of parkinsonian degeneration.

An overview of extranigral degeneration in PD

It is worthwhile to re-emphasize that establishment of degeneration in the midbrain nigrostriatal DAergic pathway was a major step in the field of PD research (Hornykiewicz 2008). It enormously enriched our knowledge of motor circuits in the brain. For decades nigrostriatal degeneration was the sole focus of PD research because of the strong quantitative correlation between motor phenotype of clinical PD and loss of A9 (substantia nigra pars compacta) tyrosine hydroxylase (TH)-positive DAergic neurons causing reduction in TH-positive striatal terminals and loss of DA transporters in striata. However, to address the global symptoms of the human disease, one has to account for degeneration in several extranigral sites. A substantial amount of recent literature has elaborated these aspects. Lim and colleagues have discerned three broad categories of early pre-motor, later non-motor and advanced l-DOPA-unresponsive motor symptoms (Kalaitzakis et al. 2009; Lim et al. 2009). We revisited the salient features of extranigral degeneration with an aim to discern what can amount to testable, quantifiable aspect in animal models of experimental parkinsonism.

Extranigral features like cortical atrophic brain lesions and ventricular enlargement that could be correlated with the parkinsonian symptomatology were reported almost three decades ago (Schneider et al. 1979). The presence of Lewy bodies (LBs) was confirmed in extranigral sites along with nigral nuclei (Forno 1987). However, LBs were found in the processes in extranigral sites, unlike the perikaryon location in nigra (Forno 1987); consequently, significance of LBs in diagnosis of idiopathic PD was explored (Gibb and Lees 1989). Furthermore, association between LB disease and Alzheimer’s disease (AD) led to pathological studies of dementia component in PD (Gibb et al. 1989a; Morris et al. 1989). Severity of the loss of neuromelanin-containing noradrenergic neurons in locus coeruleus was linked with occurrence of dementia in PD (Zweig et al. 1993). In 1995, Braak and colleagues have inferred that nigral and extranigral lesions co-exist consistently in PD brain, the neuronal cytoskeleton is impaired and the efferent structures are predominantly destroyed unlike the afferents in AD (Braak et al. 1995). Such an interpretation was gradually developed into a pattern of brain destruction in PD as well as in AD (Braak et al. 1996). Postmortem studies in PD patients were conducted to detect specific brain areas for symptoms based on the distribution of LBs, being frequent or rare, more intense or less, and the major neurotransmitter of the associated nuclei; this led to conclusion that clinical subtypes of PD could harbor morphological lesion patterns (Jellinger 1999). Pathoanatomical mapping of PD progression depicted that most important centers of both the limbic system and the motor system were impaired in PD brains, which called for broadening the definition of PD beyond midbrain DAergic nigrostriatal degeneration so as to delve into the root of illness (Braak and Braak 2000; Braak et al. 2000). In a separate study, it was confirmed that simple movement sequences better correlate with plasma l-DOPA levels, than complex ones, in which concomitant cognitive efforts were needed, with consecutive hypothetical involvement of extranigral non-DAergic system (Muhlack et al. 2004). Visual hallucination, yet another PD symptom emerging from the extranigral pathology, was reported to be associated with cortical and amygdalar LB pathology (Muhlack et al. 2004). Olfactory dysfunction, a prominent PD symptom found in 70–100% of PD patients, has extranigral origination and is reported to be a cardinal sign in early and differential diagnosis of PD (Herting et al. 2008). Recently, olfactory testing coupled to DAT SPECT (DA transporters single-photon emission computed tomography) imaging has been reported as useful for diagnosis of pre-motor PD (Berendse and Ponsen 2009). While extranigral non-DAergic degeneration is responsible for several PD symptoms, it may be noted that extranigral DAergic nuclei in A8 (retrorubral field and midbrain reticular formation) and A10 region (parabrachial pigmented nucleus, paranigral nucleus and ventral tegmental area) are rather preserved in PD, thus differentiating it from progressive supranuclear palsy and multiple system atrophy (Murphy et al. 2008). Indeed the focus on damage to non-DAergic and extranigral sites is gaining importance for clinical practice (Braak and Del Tredici 2008). Certainly, such an approach recalls for revisiting the existing animal models. There is an urgent need for biomarkers that can faithfully reflect progression of neurodegeneration and can be used in clinical trials to test putative disease modifying agents (Poewe 2009; Poewe and Mahlknecht 2009).

Although diagnosis and handling of PD during the pre-motor phase has been greatly encouraged (Truong and Wolters 2009), it is worthwhile to emphasize that several extranigral degeneration may be random and less specific; and can be frequently misdiagnosed or under-diagnosed (Wolters 2009). Strong evidences from experimental research are needed to causally associate them with PD progression or test a therapeutic regimen. Certain extranigral abnormalities such as color vision are reported not to correlate with nigrostriatal DAergic degeneration (Müller et al. 1998). Likewise, attempts to detect PD-linked extranigral metabolic abnormalities to assess neurodegenerative changes outside the substantia nigra (SN) in PD, if they exist, have been reported as difficult to detect with the current 1H MRS (magnetic resonance spectroscopy) methods (O’Neill et al. 2002).

The schematic diagram (Fig. 1) postulates the current understanding on progression of disability in PD, considering both, extranigral and nigral degenerative pathways, and highlights the avenues for further research in PD.

Figure 1.

 Current paradigm on the progression of disability in PD, and prospective interventions that can improve quality of life in PD patients. Black boxes indicate the commonly accepted postulates of nigrostriatal degeneration-based PD progression, derived from clinical observations and experimentally confirmed in existing animal models. Gray boxes represent pathways of extranigral degeneration proposed by clinicians that need to be explored in animal models. White boxes reflect the spots of prospective interventions in order to attain the ultimate goal of diminishing progression of PD.

Spinal cord involvement in PD: clinical and experimental evidences

William Langston had once exclaimed in a review that both, clinicians and researchers typically view PD as confined to the tip of an iceberg (Langston 2006). However, the disease is much widespread throughout the central and peripheral nervous system – the body of the iceberg (Langston 2006). Originations of certain symptoms of PD are now selectively linked to discrete nuclei in spinal cord (see for review Vivacqua et al. 2011b). Earlier reports on involvement of intermediolateral nucleus in PD were further confirmed by patient studies that showed significant loss of neurons in these nuclei (Wakabayashi and Takahashi 1997). l-DOPA associated pain in PD patients either originates in or is mediated by spinal DAergic system, as suggested by a case report (Sage et al. 1990). Furthermore, incidental LB disorder (ILBD) in neurologically unimpaired elderly subjects, which is considered to be equivalent to pre-symptomatic PD, has served as relevant tool for studying the progression of alpha-synuclein pathology in PD. Klos et al. (2006) reported that ILBD is not only confined to the brain, but also involves spinal cord. Alpha-synuclein pathology in spinal cord was documented in 69% of ILBD subjects and 8% of overall group out of 106 neurologically normal individuals. It has been suggested that spinal cord pathology occurs early in the Braak staging scheme, but probably after involvement of lower brainstem, consistent with at least PD diagnosing-stage 2. The study confined to examination of thoracic spinal cord revealed alpha-synuclein pathology mostly in intermediolateral cell column; whether alpha-synuclein pathology is found in other regions of spinal cord (probably in sacral parasympathetic nuclei), should be clarified. The neurons in thoracic region include sympathetic pre-ganglionic autonomic neurons; their involvement has been implicated in autonomic dysfunction in PD, where there is consistent neuronal pathology and neuronal loss (Klos et al. 2006). The same group had earlier reported pathology of PD with overlapping motor neuron disease (Klos et al. 2005). Employing ILBD cases, another group of investigators suggested that alpha-synuclein pathology of PD starts in lower brainstem nuclei and olfactory bulb. Like Klos and colleagues, they indicated that spinal autonomic nuclei and peripheral autonomic nervous system belong to the most consistently and earliest affected-regions next to medullary structures and olfactory nerves (Bloch et al. 2006). In a recent study, high incidence of alpha-synuclein was observed in spinal cord of PD patients (Kalaitzakis et al. 2008).

Pyramidal, basal ganglia and cerebellar pathways converge to form the basis for integration of fine motor control and gait, spinal cord being the final common pathway for outflow of central nervous motor systems. Control of walking and autonomic function requires systems that are integrated into the spinal cord itself. Gait failure (freezing) and dysautonomia (orthostatic hypotension, bladder bowel and sexual dysfunction) are ascribed to pathology in the descending spinal pathways of the pedunculopontine nucleus (Pahapill and Lozano 2000) and intermediolateral cell columns, respectively. Somatic pain, an under-recognized, but common early symptom of PD likely results from deregulation of spinal systems as well (Defazio et al. 2008). Pathologically, the autonomic nuclei of the spinal cord and the peripheral autonomic nervous system belong to the most consistently and earliest affected regions (Probst et al. 2008). Synuclein immunoreactivity has been found in Lamina I dorsal horn spinal cord neurons (Braak et al. 2007). Lewy like inclusions staining for synuclein and ubiquitin have been found in spinal cord in a subgroup of genetic PD (Sasaki et al. 2008). Stimulation of the pedunculopontine nucleus is an experimental therapy for freezing gait in PD patients and has been shown to alter spinal cord reflex physiology (Pierantozzi et al. 2008). Likewise, spinal cord has also been proposed as a prospective site of stimulation to facilitate motor symptoms in PD, as less invasive than deep-brain stimulation, based on a recent animal study. It was demonstrated that epidural electrical stimulation of the dorsal columns in combination with significantly low doses of l-DOPA in the spinal cord restored locomotion in DA transporter knock-out mice and 6-hydroxydopamine (6-OHDA) lesioned rat models of PD (Fuentes et al. 2009). The extensive electrophysiological study explains how activation of afferent fibers terminating in the dorsal column nuclei and ascending through the lemniscal pathway to cortical areas unlocks the basal ganglia-cortical circuits and thus, permits initiation of locomotion. The data on combined DA replacement therapy and spinal electrical stimulation is suggested to be helpful in clinical trials to alleviate motor symptoms in PD (Fuentes et al. 2010).

Investigations have provided further evidence of selective degeneration of spinal cord in experimental PD. In autosomal dominant familial PD, two mutations in the alpha-synuclein gene (Ala53-> Thr and Ala30-> Pro) occur. Alpha-synuclein is found mostly in pre-synaptic nerve terminals. Mice, over-expressing the human A53T mutant alpha-synuclein, develop a severe movement disorder, paralysis, and synucleinopathy, the mechanisms of which have been explored in transgenic mice, expressing human wild-type or familial PD-linked A53T or A30P mutant alpha-synuclein that develop neuronal degeneration and cell death. Lee et al. (2002) examined the mutant mice at early-to-mid stage disease and at near end-stage disease, compared to age-matched non-transgenic littermates as controls. In A53T mice, neurons in brainstem and spinal cord exhibited large axonal swellings, somal chromatolytic changes, and nuclear condensation. Spheroid eosinophilic LB-like inclusions were present in the cytoplasm of cortical neurons and spinal motor neurons. These inclusions contained human alpha- and nitrated synuclein. Motor neurons were depleted (∼ 75%) in A53T mice, but were affected less in A30P mice. Axonal degeneration was present in many regions. Electron microscopy confirmed the cell and axonal degeneration and revealed cytoplasmic inclusions in dendrites and axons. Some inclusions presented degenerating mitochondria and were immunopositive for human alpha-synuclein. Mitochondrial complex IV and V proteins were at control levels, but complex IV activity was reduced significantly in spinal cord. Subsets of neurons in neocortex, brainstem, and spinal cord ventral horn were positive for TUNEL (terminal deoxynucleotidyl transferase-mediated biotinylated UTP nick end labeling), and cleaved caspase-3 and p53; mitochondria in neurons had TUNEL-positive matrices and p53 at the outer membrane. Thus, A53T mutant mice develop intraneuronal inclusions, mitochondrial DNA damage and degeneration, and apoptotic-like death of neocortical, brainstem, and spinal cord motor neuron (Lee et al. 2002; Martin et al. 2006).

Recently, applying selective monoclonal antibodies for pre-synaptic terminals and fibers as well as neuronal nuclei respectively, alpha-synuclein was detected throughout the rat (Vivacqua et al. 2009) and mice (Vivacqua et al. 2011a) spinal cord neuronal fibers and synapses of Lamina I, II, VII and X, with extensive localization in neuronal nuclei of Lamina VII and X. Authors suggested that these findings could elucidate the genesis of some early clinical symptoms of PD, and indicate spinal cord as the probable starting point, as proposed earlier by other group of investigators also. Although pathological increase of alpha-synuclein was observed in A30P transgenic mice (Schell et al. 2009), yet, how the pathological alpha-synuclein accumulation correlates with neurotoxin-induced animal models of experimental PD needs to be investigated.

Transgenic mice over-expressing alpha-synuclein showed intense astrocytic cell death in spinal cord associated with extensive gliosis and activation of microglia, indicating augmentation of inflammatory reactions in spinal cord of these mice (Mendritzki et al. 2010). In addition, more severe mitochondrial impairments were found in astrocytes and oligodendrocytes than in spinal neurons. Notably, pathological alterations in spinal glial cells exceeded those in mesencephalon, suggesting caudal to rostral progression of the disease (Mendritzki et al. 2010). Importantly, astrocytic expression of A53T alpha-synuclein induced the initiation of non-cell autonomous killing of neurons and resulted in severe spinal cord and brainstem pathology along with concomitant midbrain degeneration, thus, suggesting the potential role of reactive astrocytes and microglia as therapeutic targets for PD (Gu et al. 2011). A substantial reduction in norepinephrine levels in spinal cord in transgenic A53T mutant mice under the Prion promoter was reported, which closely, recapitulated the major damage to the noradrenergic system that occured in patients with PD (Sotiriou et al. 2010).

Research in our laboratory in recent years has provided evidence in support of the hypothesis that spinal cord, the final coordinator of movement, is also involved during parkinsonian degeneration (Ray et al. 2000; Chera et al. 2002, 2004; Samantaray et al. 2008a). This was further confirmed using two distinct experimental parkinsonism models induced by the neurotoxin MPTP (Samantaray et al. 2008b) and the environmental toxin rotenone (Samantaray et al. 2007). A key focus of our studies is the role that calpain, a Ca2+-activated neutral protease, plays in disrupting the structural–functional integrity of the spinal cord in the context of spinal cord degeneration in experimental parkinsonism (Samantaray et al. 2008c). We examined the mechanisms of calpain-mediated neuronal death in differentiated spinal cord motoneuron cultures following exposure to active parkinsonian toxins MPP+ and rotenone, and also tested the neuroprotective efficacy of calpeptin, a calpain inhibitor, in these cell culture models of experimental parkinsonism. Our results implied that spinal cord motoneurons could be potential extranigral area of neurodegeneration during pathogenesis of PD in the CNS and that calpain inhibition could provide neuroprotection (Samantaray et al. 2006, 2008a). An outline of major findings corroborating spinal cord degeneration in PD is enlisted in Table 1.

Table 1.   Spinal cord studies in clinical and experimental PD
Clinical studies
 In ILBD subject spinal cord
  Alpha-synuclein pathology in 69% of subjects (Klos et al. 2006)
 In PD patient spinal cord
  Loss of neurons in intermediolateral nucleus (Wakabayashi and Takahashi 1997)
  Synuclein immunoreactivity in lamina I dorsal neurons (Braak et al. 2007)
  Lewy like inclusions containing synuclein and ubiquitin (Sasaki et al. 2008)
  High incidence of alpha-synuclein (Kalaitzakis et al. 2008)
Experimental studies
 Neurotoxin-induced animal models
   Epidural electrical stimulation of dorsal column in spinal cord restored locomotion (Fuentes et al. 2009, 2010)
   Accumulation of 3H-MPTP in spinal cord (Lyden et al. 1985)
   Altered glucose metabolism in spinal cord (Schwartzman and Alexander 1985)
   Presence of MPTP metabolizing system in spinal cord (Samantaray et al. 2008b)
   Neurodegeneration in cervical and lumbar spinal cord neurons (Samantaray et al. 2008a,b)
   Upregulation of calpain expression and activity in spinal neural cells (Ray et al. 2000; Chera et al. 2002, 2004)
   Acupuncture at GB34 and LR3 acupoints altered gene expression in cervical spinal cord (Choi et al. 2011) further supporting earlier reports showing benefits of this procedure against nigrostriatal degeneration (Kang et al. 2007; Choi et al. 2009)
   Proinflammation, proteolytic events, and neurodegeneration in cervical and lumbar neurons (Samantaray et al. 2007)
 Transgenic animal models
   Large axonal swellings, somal chromatolysis, and nuclear condensation in spinal cord (Lee et al. 2002)
   Presence of LB-like inclusions in cytoplasm of spinal cord motoneurons, reduced complex IV activity, positive TUNEL immunoreactivity, and cleaved caspase-3 and p53 (Martin et al. 2006)
   Sequestration of cell signalling proteins 14-3-3 and synphilin-1 by alpha-synuclein causing neuronal death subsequently (Shirakashi et al. 2006)
   Selective noradrenergic vulnerability (Sotiriou et al. 2010)
   Significant loss of spinal cord motoneurons and microglia activation (Gu et al. 2011)
   Alpha-synuclein pathology and astrogliosis with progressive loss of locomotor function (Neumann et al. 2002)
   Extensive immunoreactivity for insoluble phosphorylated on serine 129 isoform of alpha-synuclein and ubiquitin, associated with caspase 9 activation in spinal cord of end-stage symptomatic A30P mice. Parkin deficiency mitigated neuritic pathology in end-stage symptomatic A30P mice (Fournier et al. 2009)
  Dopamine transporter knock-out
   Epidural electrical stimulation of dorsal column in spinal cord restored locomotion (Fuentes et al. 2009, 2010)
  Alpha-synuclein over-expression
   Intense astrocytic cell death, extensive gliosis, and activation of microglia (Mendritzki et al. 2010)

Existing animal models and understanding of PD

Selective degeneration in extranigral nuclei has been addressed in several animal models to interpret parkinsonian mechanisms. Data presented in the Table 2 outline diverse animal models, which have addressed such investigations. The following account on studies in animal models and extranigral degeneration may not be an exhaustive list of all available literature, but a coherent collection that may direct towards the central theme of this review. List of extranigral sites is based on the earlier reviews, which enlist such nuclei that are implicated in PD (Jellinger 1999, 2009). The order in which the extranigral sites have been addressed in the current review is not topographical, rather alphabetic, as there is still more evidence needed to affirm the caudo-rostral progression of alpha-synuclein pathology in PD (Kalaitzakis et al. 2008). A recent study even suggests existence of multicenter process from the earliest stages in PD (Dickson et al. 2009). Moreover, mapping LB has to be substantiated with actual evidence of neuronal death (Jellinger 2009), and the exact role of LB in neuronal death has to be investigated in greater detail. Along with well-established 6-OHDA, MPTP and more recent rotenone-induced PD models, prenatal exposure of LPS to animals was validated for PD studies. Notably, lentiviral vectors encoding wild type and A30P mutant human alpha-synuclein, which were microinfused in striatum, SN and amygdala, were useful to create transgenic models of PD. This promoted over-expression of alpha-synuclein in different brain areas, induced Lewy-like pathology, alpha-synuclein-positive neuritic varicosities and cytoplasmic inclusions (Lauwers et al. 2003). The extranigral sites were examined, often concurrently, because of their morphological and neurochemical involvement in motor circuits, projecting neurons, and neuromodulatory effects (Table 2).

Table 2.   Studies on specific CNS nuclei in animal models of PD
CNS nuclei frequently affected with LB in PDAnimal models of PD
Neurotoxin-inducedEnvironmental toxin-inducedInflammatory 
  1. aTransgenic (Tg) mice include: D2 receptor deficient mice; D2 receptor knock-out mice; leucine-rich repeat kinase 2 (LRRK2, Lrrk2) DA beta-hydroxylase knock-out mice that lack noradrenaline altogether; parkin knock-out mice; and noradrenaline transporter knock-out mice.

  2. N/F, not found.

Anterior cingulate gyrusRodentsPrimatesN/FN/FN/FN/F
AmygdalaRodentsRodents, primatesN/FN/FRodentsTg mice
Nucleus basalis of MeynertN/FPrimatesCo-culture brain slice modelN/FN/FN/F
ThalamusRodents, primatesRodents, primatesRodentsN/FN/FTg mice
HypothalamusRodentsRodents, dogs, primatesRodentsN/FRodentsTg mice
Locus coeruleusRodentsRodents, primatesRodentsN/FN/FTg mice

Anterior cingulate gyrus of cerebral cortex, which contains pyramidal neurons projecting to the subthalamic nucleus, was examined in primates (Jenkins et al. 2004) and rodents (Parr-Brownlie et al. 2007) to assess firing patterns of subthalamic nucleus and dysfunction in DA circuitry. In rodents, lesion in DAergic neurons was induced by infusion of 6-OHDA into medial forebrain bundle, which caused severe DA depletion in striatum and partial DA depletion in anterior cingulate cortex. However, no association was found between changes in activity of subthalamic nucleus and anterior cingulate cortex (Parr-Brownlie et al. 2007).

Amygdala, a catecholaminergic limbic structure, responsible for emotional processing, was variously tested in animal models of PD. MPTP prominently reduced TH-positive fiber densities in basolateral, lateral and central nucleus of amygdala, thus, explaining morphological basis of emotional responses in animals (von Bohlen und Halbach et al. 2005). Amygdala was also tested in 6-OHDA infused rat model for catecholamine depletion and its effects on olfactory deficit (Fernandez-Ruiz et al. 1993). Comparative profiles of mRNA encoding alpha1 subunit of GABA receptor were studied to account for changes in GABAergic activity, occurring downstream of striatal DA loss; there were smaller changes in lateral amygdala and striatum compared to SN pars reticulata (SNpr), globus pallidus, and parafasicular nucleus (Chadha et al. 2000). In central amygdaloid nucleus, estradiol stimulated DA release, thus inferring beneficial role for reducing the DAergic dysfunction in mesolimbic structures (Liu and Xie 2004). Amygdala was one of the regions in DAergic pathway, which was reported to have increased levels of manganese, iron, copper, and zinc after 6-OHDA infusion (Tarohda et al. 2005). Amygdala was further implicated in the interaction between basal ganglia (ventral pallidum, in particular) and limbic system via Fos-signaling (Turner et al. 2008). Unexpected Fos gene activation in amygdala may contribute in part to the complications associated with l-DOPA therapy (Cole et al. 1993). Increased sensitivity of amygdala to l-DOPA was also indicated in hemiparkinsonian rhesus monkeys (Chen et al. 1999). Opiate receptors in amygdala of MPTP-lesioned rhesus monkeys showed reduced avidity, which indicated adaptable contribution of opiate pathway to changes in basal ganglia circuits that forestall the initial clinical manifestations of PD (Cohen et al. 1999). Additionally, decreased concentration of corticotropin-releasing factor, found in PD patients, was also recorded in amygdala and paraventricular nucleus of hypothalamus in MPTP-treated rats (Huang and Lee 1995). Significant loss of DA and 5-hydroxytryptamine (serotonin) was recorded in amygdala along with striatum, frontal cortex, nucleus accumbens, hippocampus, and hypothalamus in young adult rats prenatally exposed to bacterial LPS (Wang et al. 2009a). Further, amygdala was examined in A30P alpha-synuclein transgenic mice for age-dependent fibrillization of alpha-synuclein in specific cortical regions, including central nucleus of amygdala; and involvement of amygdala in cognitive behavior of mice was demonstrated (Freichel et al. 2007). Pathologic phosphorylation of alpha-synuclein in lateral/basolateral amygdalar nuclei along with hippocampal and cortical areas was shown; the damage was extended with age to central amygdalar nucleus and one of its projection areas, the lateral hypothalamus (Schell et al. 2009).

Basal nucleus of Meynert shows intense alpha-synuclein pathology in human PD patients (Kalaitzakis et al. 2008) and may be associated with impairment of intellectual capacity in PD as cholinergic hypofunction. Although a very early study failed to find any loss in cholinergic cells in this nucleus after MPTP administration to marmosets (Garvey et al. 1986), it was found recently that exposure of organotypic brain slices to rotenone significantly decreased number of cholinergic neurons in basal nucleus of Meynert (Ullrich and Humpel 2009).

Thalamic neurons were suggested to be involved in development of parkinsonian symptoms. These neurons function across the striato-pallidal-thalamic-cortical motor circuit, transferring motor information from basal ganglia output nuclei to subthalamic nucleus, striatum, and cortex. Thalamic degeneration is suggested as an early, rather than a late feature of PD. In 6-OHDA-lesioned rats, it was shown that pedunculopontine nucleus and parafascicular nucleus of thalamus that project to subthalamic nucleus became hyperactive after nigrostriatal DAergic denervation (Orieux et al. 2000). As opposed to this, metabolic profiles in selective thalamic neurons that are innervated by SN, pallidium and cerebellum were recently suggested to be hypoactive in two distinct models of nigrostriatal denervation by 6-OHDA and MPTP in rats and monkeys, respectively (Rolland et al. 2007). Moreover, akinetic deficits produced in Wistar rats by discrete bilateral 6-OHDA striatal infusions, which mimics striatal denervation in early symptomatic stage of PD, have been associated with metabolic changes in the cortico-basal ganglia-cortical loop downstream of striatum and pallidal complex (Oueslati et al. 2005). The centromedian-parafascicular thalamic complex was implicated in motivational responses and autonomic dysfunction in 6-OHDA model of PD (Henderson et al. 2005; Truong et al. 2009). The centromedian-parafascicular complex of thalamus was considered as a potential therapeutic target for neurosurgical treatment (Kerkerian-Le Goff et al. 2009; Jouve et al. 2010) and high-frequency stimulation (Lin et al. 2007). However, the robust oscillatory activities in basal ganglia output nuclei after DA cell lesion could not be directly correlated with oscillatory activity in thalamic parafascicular nucleus as revealed by electrophysiological studies in 6-OHDA-lesioned rats (Parr-Brownlie et al. 2009). Ongoing studies on thalamus in transgenic models, wherein pathogenic substitutions in leucine-rich repeat kinase 2 (LRRK2), linked to familial, late-onset parkinsonism, has been reported to be ubiquitously expressed and relatively abundant in most brain regions, including thalamus (Melrose et al. 2007).

Hypothalamus was implicated in MPTP-induced parkinsonism (Sandyk et al. 1990). Caudal hypothalamus contains histaminergic neurons, which is one of the nuclei responsible for conversion of MPTP to MPP+ prior to transfer to nigrostriatal system (Nakamura and Vincent 1986). Hypothalamic nucleus was also affected in marmosets with MPTP (Gibb et al. 1989b) and hypothalamic pathology was strongly suggested in behavioral disturbances caused by DA deficit (Willis and Sandyk 1992). Increased levels of histamine in hypothalamus along with other extranigral sites, such as hippocampus and medulla oblongata, where recorded following postnatal administration of 6-OHDA in rats; these animals developed parkinsonian behavioral disorders (Nowak et al. 2009). Modulatory role of neuroendocrine hypothalamic-pituitary-adrenal axis in PD was suggested in MPTP-treated dogs (Mizobuchi et al. 1993). The function of the tuberomamillar nucleus of hypothalamus in PD, which forms histamine input in SNpr, remains poorly understood and often controversial. It can be assumed that role in regulation of serotonin neurotransmission via histamine receptors may be important in mental functions (Threlfell et al. 2008).

Locus coeruleus is associated with severe neuronal loss in PD patients (Zarow et al. 2003). First, convincing signs of structural damage to locus coeruleus noradrenergic neurons were obtained using protracted regimen of MPTP in squirrel monkeys reporting lesions and eosinophilic inclusion in locus coeruleus (Forno et al. 1986). Damage to noradrenergic neurons in locus coeruleus substantially reduced the rate of recovery in MPTP-treated monkeys, which correlated with severe loss of nigrostriatal DAergic neurons (Mavridis et al. 1991). Correlation between denervation of locus coeruleus noradrenergic terminals and DAergic neurodegeneration was confirmed independently in different models of PD (Bing et al. 1994; Srinivasan and Schmidt 2003). In 6-OHDA model, nigrostriatal DA depletion was accompanied with modification of electrophysiological profile related to the locus coeruleus neuron basal activity (Wang et al. 2009b), which attributes to mechanisms of depression in PD (Miguelez et al. 2011). Besides, functional interconnection between lesion of locus coeruleus, and consequent hyperactivity of pyramidal neurons of medial prefrontal cortex and SNpr was demonstrated in rat model of PD (Wang et al. 2010a,b). Selective pre-treatment with α2-adrenoceptor antagonist protected DAergic neurodegeneration in PD (Srinivasan and Schmidt 2004). Contribution of locus coeruleus to motor dysfunction in PD was corroborated in DA beta-hydroxylase knock-out (Dbh−/−) mice (Rommelfanger et al. 2007). On the other hand, participation of locus coeruleus in l-DOPA-induced motor fluctuations was unrelated (Marin et al. 2008). Significantly reduced noradrenergic neurons in locus coeruleus has also been reported after rotenone infusions in rats (Hoglinger et al. 2003; Lin et al. 2008).

Olfactory bulb deficits were observed in 6-OHDA-induced catecholamine depletion in amygdala; adrenal medullary transplants restored olfactory aversion and catecholamine levels (Fernandez-Ruiz et al. 1993). Olfactory bulb was profoundly affected at an early time-point after a single intranasal administration of MPTP as indicated by lipid peroxidation and oxidative stress, reinforcing the notion that the olfactory system could represent a particularly sensitive route for the transport of neurotoxins into the CNS that might be related to etiology of PD (Franco et al. 2007). In chronic MPTP mouse model, olfactory dysfunction appeared after the 1st MPTP injection, whereas motor dysfunction appeared after the 3rd and worsened upon subsequent administrations (Schintu et al. 2009). Systemic rotenone infusion to Lewis rats induced oxidative damage in olfactory bulb along with midbrain (Sherer et al. 2003). Transducing of DAergic neurons in alpha-synuclein knock-out mouse olfactory bulb using wild type human alpha-synuclein lentivirus, impact of alpha-synuclein on TH was studied (Alerte et al. 2008).

Autologous sympathetic ganglionic neurons were tested for efficacy as donors for cell therapy in PD patients (Nakao et al. 2004). Such investigations were initiated in 1990 with MPTP-treated monkeys, and animal behavior was quantified (Nakai et al. 1990). The superior cervical ganglion has been grafted into the brain of adult rats with an attempt to reverse the parkinsonian syndrome that follows destruction of central DA systems (Nakao et al. 1995); however, massive cell death largely limited the process (Nakao et al. 1995; Pallini et al. 1996). Investigations with autonomic ganglia were mainly focused on autotransplantation of cervical sympathetic ganglion grafts into DA-depleted caudate nuclei of MPTP-induced hemiparkinsonian monkey’s brain, with further assessment of morphological, biochemical, and behavioral parameters (Gao et al. 1998). Autotransplantation of superior cervical ganglion grafts into parietal cortex of rats and caudate nucleus of MPTP-treated monkeys was also conducted (Itakura et al. 1988), suggesting this approach as a new therapy for PD. Moreover, transplantation of cultured fetal human sympathetic neurons reversed amphetamine-induced turning behavior in 6-OHDA rats (Kamo et al. 1986).

While most animal models of PD are based on selective midbrain nigrostriatal degeneration, a few recent models address nigral and extranigral features simultaneously like human PD. In one such model with 95% reduction in vesicular monoamine transporter expression, symptoms like progressive deficits in olfactory discrimination, delayed gastric emptying, altered sleep latency, anxiety-like, and age-dependent depressive behavior, progressive loss of striatal DA, l-DOPA-responsive motor deficits, alpha-synuclein accumulation, and nigral DAergic cell loss were reported (Taylor et al. 2009). Due to deficits in other monoamine systems (norepinephrine, serotonin) this model reflects non-motor symptoms of PD well (Taylor et al. 2011). In proteasome inhibitor lactacystin induced rat model evolution of extranigral damage could predict behavioral deficits, resulting from nigrostriatal degeneration (Vernon et al. 2011).


Different animal models of PD have been extensively reviewed, and advantages as well as caveats of each model were discussed (see for review Melrose et al. 2006; Lane and Dunnett 2008; Meredith et al. 2008). The ability of existing models to imitate pathological aspects of the disease based on nigrostriatal DA loss was evaluated. The existing models, which correlate the motor symptoms of PD with nigrostriatal degeneration, have been validated for assessing symptomatic therapies. In the current review, we showed possible extranigral sites, which were concomitantly investigated in existing models, and were also involved in or contributed to parkinsonian degenerative processes. We have found that extranigral degeneration has been explored more frequently in MPTP and 6-OHDA models of PD, whereas, there is paucity of reports from the environmental toxin based models. As epidemiological surveys strongly support environmental toxins and occurrence of PD, hence, validation of extranigral degeneration in rotenone, paraquat, and maneb models may be immensely useful. Addressing extranigral issues in transgenic mice needs caution and should be critically evaluated. Amongst the various extranigral sites discussed in the current review, it may be corroborated that the existing animal models can accommodate testing of non-motor symptoms in addition to motor dysregulation in PD. Valuable resources are available to interpret the involvement of extranigral sites like amygdala, basal nucleus of Meynert, locus coeruleus, hypothalamus, and olfactory bulb; these may be beneficial for validating drugs for neuropsychiatric and dementia components, behavioral modulation, thus, attaining the quality of life in PD patients. Olfactory bulb can be more explored for early diagnosis and thalamus for the neurosurgical interventions and stimulation. Concomitant of nigrostriatal degeneration with extranigral DAergic sites, including mesolimbic, mesocortical and thalamic, and extranigral non-DAergic sites (noradrenergic, cholinergic, serotoninergic) may result in preventing complication, such as l-DOPA-induced dyskinesia. It would be advantageous to explore the feasibility of recapitulation of Braak’s stage in animal models and track early biomarkers. Spinal cord may be a region of interest in this regard.


This work has been supported in part by a grant (RO1 NS62327) from NIH-NINDS.