Epidemiological studies have observed an association between pesticide exposure and the development of Parkinson’s disease, but have not established causality. The concept of an adverse outcome pathway (AOP) has been developed as a framework for the organization of available information linking the modulation of a molecular target [molecular initiating event (MIE)], via a sequence of essential biological key events (KEs), with an adverse outcome (AO). Here, we present an AOP covering the toxicological pathways that link the binding of an inhibitor to mitochondrial complex I (i.e., the MIE) with the onset of parkinsonian motor deficits (i.e., the AO). This AOP was developed according to the Organisation for Economic Co-operation and Development guidelines and uploaded to the AOP database. The KEs linking complex I inhibition to parkinsonian motor deficits are mitochondrial dysfunction, impaired proteostasis, neuroinflammation, and the degeneration of dopaminergic neurons of the substantia nigra. These KEs, by convention, were linearly organized. However, there was also evidence of additional feed-forward connections and shortcuts between the KEs, possibly depending on the intensity of the insult and the model system applied. The present AOP demonstrates mechanistic plausibility for epidemiological observations on a relationship between pesticide exposure and an elevated risk for Parkinson’s disease development.
Pesticides such as dichlorodiphenyltrichloroethane (DDT), dieldrin, paraquat, and rotenone have been considered as potential contributing factors to the development of Parkinson’s disease (PD) (Baltazar et al. 2014; Sandström et al. 2014). Concerns about the contribution of environmental agents to parkinsonian disorders have led to epidemiological studies examining an association between human exposure to pesticides and the development of PD. Meta-analyses of these epidemiological studies confirmed a significant association between pesticide exposure and PD. Such observations are difficult to integrate into regulatory risk assessments, as exposure is currently evaluated retrospectively and indirectly in the vast majority of epidemiological studies. Therefore, these studies do not allow the identification of causal relationships (Breckenridge et al. 2016; Hernández et al. 2016; Van Maele-Fabry et al. 2012). Moreover, epidemiological observations usually do not provide a plausible link to molecular processes known to be associated with PD pathogenesis.
The adverse outcome pathway (AOP) concept organizes heterogeneous biological and toxicological data to provide information on possible sequences of events across multiple levels of biological organization (Bal-Price et al. 2017; Villeneuve et al. 2014a, b; Leist et al. 2017). An AOP represents a linear sequence of key events (KEs) causally connected through key event relationships (KERs) that provide a plausible link between a molecular initiating event (MIE) and an adverse outcome (AO). The MIE is defined as the first specific modification of a biological target at the molecular level by a chemical that can trigger the subsequent events, leading to pathology (i.e., the AO). Notably, AOPs of regulatory significance describe a sequence of biological processes (biochemical, cellular, physiological) and not the effects (mode of action) of a specific compound (AOPs are “compound agnostic”). The most important implication of this characteristic is that they do not describe or take into account toxicokinetics. The latter cannot be described in a generic way but is inseparably coupled to the molecular identity of a toxicant. Essential criteria for the overall evaluation of AOPs is the presence of a solid basis for the (1) essentiality of KEs, (2) the biological plausibility and empirical support for KERs. The empirical support for the KER is indeed complementing the biological plausibility for the KER and the essentiality of the KEs. These principles are described in the Organisation for Economic Co-operation and Development (OECD) guidance document on the development of an AOP (http://www.oecd-ilibrary.org). The ratings represent a comparative measure of the degree of confidence in the supporting weight of evidence, based on acquired collective experience. In the future, the AOP framework could contribute to an integrated approach to testing and assessment (IATA) that includes absorption, distribution, metabolism, excretion (ADME) information as well as quantification of effective threshold concentrations.
In a meta-analysis by Tanner et al. (Tanner et al. 2011), pesticides were classified by their presumed mechanism and not only by their chemical class. This study allowed the identification of significant associations between the inhibition of mitochondrial complex I and a parkinsonian phenotype. A causal role of complex I inhibition in the development of a parkinsonian phenotype is supported by broad evidence, and over the past 30 years, the complex I inhibitors rotenone and MPP+ emerged as the most widely applied experimental toxicants in PD research (Schildknecht 2017). On this basis, we constructed an AOP that describes the link between the inhibition of mitochondrial complex I and the manifestation of parkinsonian motor deficits. The full AOP (Fig. 1) can be found in the AOP Wiki (https://aopwiki.org/aops/3), an AOP platform established as part of the 2012 OECD AOP development work plan.
Key events, essential for the progression of the response evoked by inhibitor binding to complex I (MIE) towards the AO, were assessed on their essentiality, based on the present knowledge on how a KE works, on the availability of robust detection methods (Table 1) and on evidence in the literature, indicating that experimental inhibition of a KE reduces or abolishes downstream KE activation. Weight of evidence analysis for the rating of KE essentiality is summarized in Table 2.
Experimental studies illustrating a direct correlation between two adjacent KEs were also listed in the “experimental support” section of the respective KERs and in the respective figures. This organization of information is not fully in line with the recommendations of the AOP handbook. However, we opted for this solution for two reasons: (1) it allows a concise overview of the vast body of information included in the full version of the present AOP (AOP 3, AOP Wiki), as required for a journal article; (2) for many academic researchers, it is easier to understand the importance of a KER, if not only correlative data around the KER are present, but if this is combined with experimental evidence that modulation of the KE directly upstream of the KER leads to a modulation of the KE directly downstream of the respective KER.
Key event relationships (KERs)
KER 1: relationship between “binding of an inhibitor to mitochondrial complex I” (MIE) and “complex I inhibition” (KE 1) (Fig. 2)
Oxidation of nicotinamide adenine dinucleotide (NADH) is catalyzed by the flavine mononucleotide moiety of complex I (Vinogradov 1993; Degli and Ghelli 1994). In a sequential manner, the two electrons of NADH are transferred along a chain of eight Fe–S clusters to the ubiquinone-binding site where they reduce ubiquinone (Q), via ubisemiquinone (•Q) formation, to ubiquinol (QH2) (Kotlyar et al. 1990; Suzuki and King 1983; van Belzen et al. 1997). The majority of complex I inhibitors block the electron transfer onto ubiquinone (Palmer et al. 1968). Complex I inhibitors were categorized into three classes based on their potential binding site (Degli Esposti 1998; Friedrich et al. 1994). However, more recent research indicates the presence of a single inhibitor-binding pocket in the hydrophobic ubiquinone-binding region of complex I with several binding sites for structurally diverse inhibitors (Okun et al. 1999). The majority of currently described complex I inhibitors either prevent access of ubiquinone to its binding site, or the inhibitors act as electron acceptors interfering with the Fe–S cluster electron transport chain (Lümmen 1998; Ohnishi 1998). In all of these inhibitor-mediated cases, blockade causes electrons to back up, resulting in the full reduction of upstream Fe–S clusters (Brand 2010). These conditions promote an uncoordinated flux of electrons from reduced sites of complex I onto molecular oxygen to form the superoxide radical anion (•O2 −) (Grivennikova and Vinogradov 2006; Liu et al. 2002), and they all prevent reduction of ubiquinone and thus the transfer of electrons through complexes III and IV to molecular oxygen. The N2 cluster, as well as flavine in its fully reduced or semiquinone form, have been suggested as molecular sites of superoxide formation upon complex I inhibition. These observations, however, are all dependent on the experimental system and procedures applied. Hence, they allow no generally accepted conclusion on the precise molecular site responsible for superoxide formation upon complex I inhibition (Brand 2010; Galkin and Brandt 2005; Genova et al. 2001; Lambert and Brand 2004).
The experimental basis for a causal relationship between inhibitor binding and complex I inhibition is based on experiments performed with submitochondrial particles, isolated mitochondria, and neuronal cell cultures. Real-time displacement tests using fluorescent (e.g., aminoquinazoline) or radioactively labeled complex I inhibitors and their derivatives (e.g., 3H-dihydrorotenone, 3H-AE F119209) provide direct evidence for the binding of complex I inhibitors (Greenamyre et al. 1992; Higgins and Greenamyre 1996; Ino et al. 2003; Okun et al. 1999; Talpade et al. 2000). Complex I activity is assessed by detection of NADH oxidation (Gluck et al. 1994; Höllerhage et al. 2009; Shimomura et al. 1989). Time- and concentration-dependent inhibition of complex I in submitochondrial particles or isolated mitochondria was observed with rotenoids, piercidines, myxobacterial antibiotics, and vanilloids such as capsaicin (Cleeter et al. 1992; Degli Esposti et al. 1993, 1994; Friedrich et al. 1994; Greenamyre et al. 2001; Grivennikova et al. 1997; Höllerhage et al. 2009; Ichimaru et al. 2008; Lambert and Brand 2004; Miyoshi 1998; Okun et al. 1999). The inhibitory action of complex I inhibitors on electron transfer onto ubiquinone was independently confirmed by the expression of the inhibitor-insensitive oxidoreductase NDI-1 from Saccharomyces cerevisiae in cell models, which circumvents complex I and allows maintenance of the normal respiratory chain electron flux (Seo et al. 1998; Sherer et al. 2003).
The question of how electron transfer is coupled to proton pumping has still not been answered in a conclusive manner (Hirst 2013; Sharma et al. 2009). Electron paramagnetic resonance (EPR) analyses indicated the presence of two ubisemiquinone species during electron transport (Vinogradov et al. 1995). However, it is not known whether these species represent two independent ubisemiquinone molecules or two forms of the same semiquinone (Albracht et al. 1997). Inhibitor binding studies are performed with submitochondrial particles, containing membranes. Due to the lipophilicity of most complex I inhibitors, these investigations suffer from high background values as a result of unselective membrane binding (Horgan and Casida 1968). Although complex I inhibitors prevent ubiquinol formation, the precise inhibitor binding site(s) have not been identified yet. Furthermore, it is not evident whether •O2 − generation upon complex I inhibition is mainly derived from F–S cluster or from semiquinone-dependent electron transfer onto molecular oxygen.
KER 2: relationship between “complex I inhibition” (KE 1) and “mitochondrial dysfunction” (KE 2) (Fig. 3)
Complex I represents the principal gateway for the entry of electrons into the mitochondrial respiratory chain (Friedrich et al. 1994; Wirth et al. 2016). A functional respiratory chain generates a proton gradient across the inner mitochondrial membrane, exploited in a subsequent step by mitochondrial ATPases to generate ATP (Brandt 1997; Mailloux 2015). Disturbances in the electron transfer through complex I lead to an impaired proton gradient and reduced ATP generation. As a consequence of limited ATP availability, mitochondrial Ca2+ homeostasis is disturbed, thus contributing to mitochondrial dysfunction (high energy demand of Ca2+ ATPases) (Sheehan et al. 1997; Willems et al. 2008). In parallel to the reduction in ATP generation, blockade of the electron flow along the respiratory chain results in an accidental reduction of molecular oxygen to form superoxide (•O2 −) (Mailloux 2015). Elevated reactive oxygen species (ROS) levels promote oxidative damage of mitochondrial DNA, proteins, and lipids, and trigger mitochondrial fragmentation (Koopman et al. 2007; Willems et al. 2009). The loss of the mitochondrial transmembrane potential, impaired mitochondrial ATP generation, disturbances in mitochondrial Ca2+ homeostasis, as well as the production of harmful ROS levels are features collectively referred to as mitochondrial dysfunction (Bose and Beal 2016). Deficiencies in complex I activity are regularly observed in association with mutations in mtDNA or nuclear DNA-encoded complex I genes. Mutations in nuclear-encoded complex I genes have been demonstrated for 12 structural subunits of complex I (Berger et al. 2008; Fernandez-Moreira et al. 2007; Hoefs et al. 2008) and for five complex I assembly factors (Dunning et al. 2007; Janssen et al. 2006; Lazarou et al. 2009; Ogilvie et al. 2005; Pagliarini et al. 2008; Saada et al. 2008). Fibroblasts of patients with such complex I mutations exhibit a decreased mitochondrial transmembrane potential and mitochondrial ATP generation, as well as elevated •O2 − formation by complex I and hence meet the definition of mitochondrial dysfunction (Koopman et al. 2007).
Experimental support for a causal relationship between complex I inhibition and mitochondrial dysfunction is largely based on observations made with the complex I inhibitors rotenone and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP). A rich experimental basis indicates the direct correlation between complex I inhibition and the emergence of features of mitochondrial dysfunction in cellular and in vivo models exposed to rotenone or MPTP/MPP+ (Bi et al. 2008; Dukes et al. 2016; Giordano et al. 2012; Han et al. 2016; Li et al. 2014; Nakai et al. 2003; Piao et al. 2012; Schildknecht et al. 2009; Scholz et al. 2011; Wang et al. 2011; Wu et al. 2009; Ye et al. 2015). Initial studies using proton magnetic resonance spectroscopy (1H-MRS) and positron emission tomography (PET) have illustrated the onset of mitochondrial dysfunction by live measurements in living animals exposed to MPTP (Brownell et al. 1998; Koga et al. 2006). Experimental interventions to prevent impaired mitochondrial ATP generation, e.g., by expression of the inhibitor-insensitive oxidoreductase NDI-1 from S. cerevisiae to circumvent impaired endogenous complex I, protect from mitochondrial dysfunction (Seo et al. 1998; Sherer et al. 2003). Application of alternative electron shuttles, such as methylene blue or coenzyme Q10, (Moon et al. 2005; Shults et al. 2002; Wen et al. 2011) or boosting of cellular ATP levels by supplementation of cells exposed to complex I inhibitors with creatine/phosphocreatine (Beal 2011; Matthews et al. 1999; Yang et al. 2009) also protect from mitochondrial dysfunction and neuronal demise.
The second strategy to protect from impaired complex I-dependent mitochondrial dysfunction targets complex I-mediated •O2 − formation. Overexpression of superoxide dismutase (SOD) protects from the toxic influence of MPTP (Przedborski et al. 1992) whereas knockdown of endogenous SOD elevates the sensitivity of mice towards MPTP-dependent mitochondrial dysfunction and nigrostriatal cell loss (Zhang et al. 2000). In cellular and in vivo models exposed to rotenone or MPTP, antioxidants protect from complex I inhibition-dependent mitochondrial dysfunction (Filomeni et al. 2012; Lee et al. 2011; Liu et al. 2015; Nataraj et al. 2016; Thomas et al. 2012; Tseng et al. 2014; Sherer et al. 2003; Wang et al. 2015).
A more detailed analysis indicates a mutual interaction between ROS and complex I. While complex I acts as a potent source of •O2 − following its inhibition, an experimental decline of cellular glutathione levels (e.g., γ-glutamylcysteine synthetase knockdown, treatment with buthionine sulfoximine) (Jha et al. 2000; Chinta and Andersen 2006) correlates with a reduction of complex I activity, the onset of mitochondrial dysfunction, and ultimately with the demise of dopaminergic (DA) neurons. All of these KEs were prevented by the application of thiol antioxidants such as dithiothreitol (DTT) or N-acetylcysteine (NAC) (Chinta and Andersen 2006; Jha et al. 2000). Mechanistic investigations unraveled reversible S-nitrosation of complex I, leading to its inhibition (Dahm et al. 2006; Burwell et al. 2006). Thiol antioxidants evoke a de-nitrosation and are associated with a re-activation of complex I, preventing mitochondrial dysfunction and protecting from neurodegeneration (Dahm et al. 2006; Borutaite et al. 2000).
These observations indicate that targeting either impaired ATP generation or elevated •O2 − formation as the two direct consequences of complex I inhibition, represents an effective intervention strategy, capable of preventing the activation of KEs downstream of complex I inhibition (KE 1).
Complex I inhibition results in a reduction in mitochondrial ATP generation and an elevation of •O2 − formation. To date, the respective contribution of these two factors to mitochondrial dysfunction has not been quantified. A cell model devoid of classical complex I activity (Choi 2008) still shows effects of rotenone or MPP+. This may be due to a contribution by off-target effects, e.g., on microtubules (Brinkley et al. 1974; Marshall and Himes 1978), but the data have not been confirmed by others. The vast majority of experimental evidence on the relationship between complex I inhibition and the onset of parkinsonian motor deficits is based on the use of the complex I inhibitors rotenone and MPTP/MPP+. A relatively wide spectrum of structurally different complex I inhibitors have been described over the course of recent decades. Prominent examples are acetogenins (Bermejo et al. 2005), tetrahydroisoquinolines (Morikawa et al. 1996), antibiotics such as piericidin A (Degli Esposti 1998; Friedrich et al. 1994; Kubota et al. 2003; Horgan et al. 1968; Singer 1979), insecticides such as quinazolines or acetogenins (Ahammadsahib et al. 1993; Hollingworth et al. 1994), quinones (Kean et al. 1971), and vanilloids (Shimomura et al. 1989). All of these structurally different complex I inhibitors have been characterized with isolated mitochondria or with submitochondrial particles. Robust K i values and functional studies involving neuronal cell cultures or in vivo models are rather rare. A systematic comparison of the half maximal inhibitory concentration (IC50) values for complex I inhibition and half maximal binding concentration (EC50) values for the reduction of ATP levels was performed with rat fetal striatal neurons (Höllerhage et al. 2009). Due to the lipophilicity of most of the complex I inhibitors tested, the detected EC50 values were in most cases lower than the IC50 values detected for complex I inhibition.
KER 3: relationship between “mitochondrial dysfunction” (KE 2) and “impaired proteostasis” (KE 3) (Fig. 4)
The two main systems for the removal of misfolded proteins are: (1) the autophagy–lysosomal pathway (ALP), which removes dysfunctional proteins, aggregates, and even subcellular organelles; and (2) the ubiquitin proteasomal system (UPS), which catalyzes the tagging of target proteins by ubiquitination, followed by their degradation via the 26S proteasome (Ding et al. 2003; Goldberg 2003; Komatsu et al. 2006; Martini-Stoica et al. 2016; Menzies et al. 2015; Zheng et al. 2016). The correlation between mitochondrial dysfunction and impaired proteostasis is based on the considerations that (1) proteostasis is an energy-consuming process requiring ATP from mitochondria, and that (2) components of the proteasomal system are subject to inhibition by ROS, generated from dysfunctional mitochondria (Finley 2009; Pickart and Cohen 2004). The 26S proteasome catalyzes ATP-dependent protein degradation and consists of a 20S core, associated with a regulatory 19S particle (Kim et al. 2011; Murata et al. 2009; Voges et al. 1999). Oxidative stress causes the dissociation of the 20S core from the regulatory 19S particle, leading to the loss of 26S proteasome activity (Wang et al. 2010). Under conditions of impaired ATP synthesis and elevated ROS levels, the interaction of the 20S core with alternative activation proteins yields a 20S proteasome without de-ubiquitination and ATPase activities (Schmidt et al. 2005; Ma et al. 1992). The ATP-independent 20S proteasome is also subject to posttranslational modifications such as hydroxynonenal modifications, carbonylation, or S-glutathionylation (Demasi et al. 2001, 2003; Farout et al. 2006; Ishii et al. 2005), but displays a higher resistance to oxidative stress in comparison with the 26S proteasome (Reinheckel et al. 1998, 2000). In parallel to the direct inhibition of cellular protein degeneration systems, oxidative stress increases the load of modified and misfolded proteins as substrates of the degradation machinery (Butterfield and Kanski 2001; Sayre et al. 2001), similar to what occurs in mitochondrial dysfunction. Analysis of ALP activity upon inhibition of complex I provides a heterogeneous picture, with several reports illustrating an impairment of ALP activity (Lim et al. 2011; Mader et al. 2012; Pan et al. 2009; Sarkar et al. 2014), while others describe an activation (Chen et al. 2007; Chu et al. 2013; Zhu et al. 2007a). Autophagy has been suggested as a component of the cellular antioxidant system, based on its removal of oxidatively modified proteins (Giordano et al. 2013). It is hence speculated that activation of ALP represents a countermeasure of the cell in early stages of mitochondrial dysfunction, while later stages are characterized by a decline in autophagy activity and an associated decline in cell viability. The PD-associated protein alpha synuclein (ASYN) emerged as a key element connecting mitochondrial dysfunction and impaired proteostasis. Knockdown of ASYN protects from complex I inhibition-mediated neurodegeneration (Zharikov et al. 2015), while ASYN overexpression sensitizes neurons towards secondary stressors (Chartier-Harlin et al. 2004; Singleton et al. 2003). Mitochondrial dysfunction leads to an accumulation and an aggregation of oxidative modified ASYN (Betarbet et al. 2006; Cannon et al. 2009). Vice versa, elevated ASYN levels evoke mitochondrial dysfunction (Hsu et al. 2000). In conclusion, mitochondrial dysfunction is characterized by an impaired ATP generation and elevated levels of ROS. Oxidative stress not only increases the load of misfolded proteins, it also leads to an impairment in the cellular protein degradation machineries. These energy-consuming processes are further hampered by the limitations in ATP supply under these conditions, hence resulting in an inadequate removal of misfolded proteins.
Most empirical support comes from cellular models exposed to rotenone and MPTP/MPP+. Moreover, an impairment of the UPS activity in the nigrostriatal system parallels mitochondrial dysfunction in PD patients (Ambrosi et al. 2014; Betarbet et al. 2005; McNaught and Jenner 2001a, b, 2003; Yu et al. 2009). Experimental induction of mitochondrial dysfunction in mice and rats by rotenone or MPTP is associated with a decline in nigrostriatal UPS (Fornai et al. 2005; Liu et al. 2013; Wu et al. 2015). In vitro models revealed that complex I inhibition precedes the onset of proteasomal impairment and the accumulation of ubiquitinated proteins (Yong-Kee et al. 2012). Expression of the inhibitor-insensitive single subunit NADH dehydrogenase NDI-1 protects from rotenone-induced loss of proteasomal function, underlining the contribution of ATP for proteasomal degradation (Seo et al. 2000, 2002). The tight dependency of proteasomal function on metabolic activity was demonstrated by a glucose-dependent experimental boost of cellular ATP levels, resulting in elevated protein degradation (Höglinger et al. 2003a, b). To avoid an involvement of complex I inhibitor-mediated oxidative stress, ATP was alternatively lowered by supplementation of cell medium with 2-desoxy-glucose. This resulted in an accumulation of misfolded proteins (Sherer et al. 2003). Management of oxidative stress, as the second dominating feature of mitochondrial dysfunction, by application of antioxidants, protects from complex I inhibitor-evoked proteasomal impairment and from an accumulation of ubiquitinated proteins (Chou et al. 2010; Filomeni et al. 2012; Shamoto-Nagai et al. 2003). The ALP can be stimulated by pharmacological means, and ALP stimulation by rapamycin results in a protection of the cell from complex I-mediated neurotoxicity (Liu et al. 2013; Pan et al. 2009). Similar to the situation observed with the UPS, antioxidants protect from complex I inhibitor-dependent reduction in ALP activity (Filomeni et al. 2012).
In neurons, proteostasis is largely influenced by intracellular trafficking processes. Mitochondrial and vesicular trafficking is affected by dysregulated cytosolic Ca2+ levels (Chang et al. 2006; Saotome et al. 2008; Yi et al. 2004) that emerge as a consequence of complex I inhibitor-mediated mitochondrial dysfunction. ASYN expression levels are directly correlated with microtubule instability (Chen et al. 2007; Esposito et al. 2007; Lee et al. 2006). Accumulation of ASYN leads to elevated levels of hyperphosphorylated tau protein and consequently to microtubule depolymerization (Qureshi and Paudel 2011). As a result of inappropriate transport processes, misfolded proteins and organelles accumulate within the cell.
ASYN levels are elevated in response to complex I inhibition (Betarbet et al. 2006; Cannon et al. 2009; Fornai et al. 2005). Elevated cytosolic levels of Ca2+ as consequence of complex I inhibition promote aggregation of ASYN (Follett et al. 2013; Goodwin et al. 2013; Nath et al. 2011; Yuan et al. 2015). Mitochondrial dysfunction and dysfunction in cellular Ca2+ homeostasis leads to disturbances in neuronal DA handling, leading to DA-mediated oxidative stress. DA-modified ASYN not only prevents its own degradation by the chaperone-mediated autophagy (CMA) pathway, but it also prevents the degradation of other proteins (Martinez-Vicente et al. 2008). ASYN filament formation, promoted by ASYN overexpression, or the expression of ASYN mutants, directly impairs proteasomal activity (Stefanis et al. 2001). Knockdown of endogenous ASYN, treatment with antioxidants, and supplementation with ATP protect from mitochondrial dysfunction-mediated onset of proteasomal stress (Betarbet et al. 2006; Dauer et al. 2002; Drolet et al. 2004; Shamoto-Nagai et al. 2003).
Impaired proteostasis includes an imbalance or dysfunction of a very large number of diverse biochemical processes. These are again interlinked in complex ways. While this is not an inconsistency as such, it can lead to inconsistent results in the literature, when different processes, often measured at different times, are used as biomarkers of impaired proteostasis. Inhibition of mitochondrial complex I is mainly characterized by impaired ATP production and elevated •O2 − formation. Although these processes result in an impairment of various proteostasis mechanisms, such as UPS activity, defined molecular events linking KE 2 and KE 3 need further investigation. The relationship between mitochondrial dysfunction and impaired proteostasis is furthermore characterized by several mutual interactions, ultimately leading to a self-amplifying vicious cycle. Misfolded ASYN, for example, accumulates as a consequence of impaired proteostasis, and this in turn negatively influences mitochondrial integrity and function via its binding to the inner mitochondrial membrane and to their import machinery (Devi et al. 2008; Robotta et al. 2014). DA-modified ASYN, on the other hand, not only blocks its own degradation by the CMA pathway but also prevents CMA-dependent degradation of other proteins (Martinez-Vicente et al. 2008). Literature provides evidence for both activation and inhibition of autophagy activity upon experimental complex I inhibition. However, time-dependent and quantitative information on autophagy activity under these conditions is not available yet. One of the cardinal features of PD is the formation of Lewy bodies in the brain. While proteinaceous ASYN aggregates are observed in rotenone-exposed rats, Lewy body-like structures are not observed in MPTP models (Dauer et al. 2002; Drolet et al. 2004).
KER 4: relationship between “impaired proteostasis” (KE 3) and “degeneration of DA neurons” (KE 4) (Fig. 5)
Impaired proteostasis leads to an accumulation of misfolded and modified proteins. These protein aggregates influence microtubule assembly and stability, resulting in a reduction in axonal transport of vesicles and mitochondria (Borland et al. 2008; Chen et al. 2007; O’Malley 2010) and a “dying back” degeneration pattern, starting in the periphery, ultimately leading to neurodegeneration (Braak et al. 2004; Grosch et al. 2016; Raff et al. 2002). ASYN is among the best-studied examples, linking proteostasis and neuronal degeneration. Accumulation of ASYN, either as a consequence of endogenous disturbances of proteostasis, experimental blockade of the proteasomal system, or by overexpression of its wild-type or mutant forms, leads to the disassembly of microtubules and, ultimately, to axonal damage (Esposito et al. 2007; Kirik et al. 2003; Masliah et al. 2000). Furthermore, ASYN protofibrils interact with intracellular organelles such as neurotransmitter vesicles or mitochondria, and lead to an uncontrolled release of DA and an impairment of mitochondrial function (Lotharius et al. 2002; Saha et al. 2004; Devi et al. 2008; Chinta et al. 2010). As mentioned above, DA-modified ASYN not only blocks its own degradation by the CMA pathway but also prevents the degradation of other proteins (Martinez-Vicente et al. 2008). Aggregates of wild-type or mutant forms of ASYN disturb controlled axonal transport of mitochondria (Li et al. 2013; Melo et al. 2017; Xie and Chung 2012). In neurons, key steps, such as mitochondrial fission/fusion or mitophagy, are conducted in the cell body. Impaired axonal transport of mitochondria hence leads to limited ATP supply and elevated levels of ROS, generated by dysfunctional mitochondria. The cellular 26S proteasome is a vulnerable target for free radical species originating from autoxidizing DA and mitochondria, leading to its inhibition and hence reinforcing proteasomal dysfunction (Davies 2001).
Analysis of nigrostriatal tissue of patients with PD has suggested an impairment in the activity of the 20/26S proteasome (McNaught and Jenner 2001a, b; McNaught et al. 2003). Similar observations were made in fibroblasts obtained from patients with PD, which exhibited elevated basal levels of ubiquitinated proteins and impaired 20S proteasomal activity (Ambrosi et al. 2014). The brain-region selective impairment of proteasomal activity correlates with the selective demise of DA neurons in this region (McNaught and Jenner 2001a, b; McNaught et al. 2003). Disturbances in the ubiquitin proteasomal system are also directly associated with prominent examples of mutations (e.g., parkin, ubiquitin C-terminal hydrolase L1) identified in genetic PD cases. Both are sufficient to cause preferential degeneration of nigrostriatal DA neurons (Leroy et al. 1998; Kitada et al. 1998).
Experimental evidence for a causal relationship between impaired proteostasis and DA neurodegeneration is based on in vitro and in vivo experiments involving complex I inhibitors and proteasome inhibitors. Several in vivo studies reported an impairment of the UPS, an accumulation of polyubiquitinated proteins, and the loss of nigrostriatal DA neurons upon exposure to complex I inhibitors (Betarbet et al. 2000, 2006; Fornai et al. 2005; Wang et al. 2006; Yong-Kee et al. 2012). An alternative complex I inactivation by conditional knockout of the complex I subunit Ndufs4 independently confirms the decrease in proteasomal activity and accumulation of polyubiquitinated proteins (Song and Cortopassi 2015). Exposure to complex I inhibitors leads to an accumulation of autophagosomes and a concomitant decrease in the number of lysosomes, as well as lysosomal dysfunction (Dehay et al. 2010; Mader et al. 2012). Up-regulation of autophagy, e.g., by rapamycin or trehalose, protects from lysosomal permeability and from neurodegeneration (Dehay et al. 2010; Giordano et al. 2014; Wu et al. 2015). A direct correlation between proteasomal dysfunction and neurodegeneration was observed by in vivo stereotaxic injection of proteasome inhibitors such as lactacystin or MG-132. Intracerebral proteasome inhibitor infusion evokes a preferential degeneration of nigrostriatal DA neurons, accompanied by the onset of PD-associated motor impairments (Bentea et al. 2015; Fornai et al. 2003; Li et al. 2012). Transcription factor EB (TFEB) is a key transcriptional regulator of the autophagy–lysosome pathway. Repression of TFEB expression in A9 and A10 DA neurons results in their accelerated degeneration (Decressac et al. 2013; Decressac and Björklund 2013). Overexpression of ASYN in vivo leads to lysosomal dysfunction and to cytoplasmic retention of TFEB. Overexpression of TFEB in the same model protects from DA neurodegeneration by clearance of ASYN oligomers (Decressac et al. 2013; Decressac and Björklund 2013; Ebrahimi-Fakhari and Wahlster 2013; Kilpatrick et al. 2015).
Mechanistic molecular information, as well as quantitative data on the direct causal relationship between impaired proteostasis and DA neurodegeneration is limited. Most of the information on the relationship of the two KEs is based on model systems treated with complex I inhibitors. However, MPTP/MPP+ does not recapitulate the formation of intracellular inclusions or aggregates. An increase in autophagy is reported as both protective and detrimental, most likely as a result of different degrees of activation or different observation times. Quantitative information on potential threshold activation levels for autophagy and their influence on cell integrity is currently not available.
KER 5: relationship between “degeneration of DA neurons” (KE 4) and “neuroinflammation” (KE 5) (Fig. 6)
In patients with PD and in MPTP-exposed humans or non-human primates, inflammation remains persistently activated in the nigrostriatal system, even years or decades after removal of the initiating toxicant (McGeer et al. 2003; Miklossy et al. 2006). The inflammatory response involves microgliosis and astrogliosis as well as the infiltration of peripheral CD4+ T lymphocytes (Appel 2009; Brochard et al. 2009). Damaged neurons expose cytosolic or nuclear proteins or non-protein molecules, collectively termed as damage-associated molecular patterns (DAMPs), which are capable to initiate and perpetuate an inflammatory response (Béraud et al. 2013; Thundyil and Lim 2015) by activating Toll-like receptors (TLRs) or receptors for advanced glycation end-products (RAGEs) (Chao et al. 2014). Microglial cells are equipped with TLRs such as TLR-2 or TLR-4 that sense targets such as high mobility group box 1 (HMGB1), amyloid beta peptide, or alpha synuclein and hence stimulate activation of nuclear factor kappa B (NF-κB) (Fellner et al. 2013; Fossati and chiarugi 2007; Liu et al. 2012; Santoro et al. 2016). Astrocytes are also able to sense tissue injury via e.g., TLR-3 (Farina et al. 2007). Moreover, neuronal injury promotes astrocyte activation (Efremova et al. 2015). ATP, released by challenged cells, is a prominent non-protein DAMP that stimulates an immune response by purinergic G protein-coupled receptors (P2Y receptors). P2Y receptor activation leads to the migration and polarization of microglial cells (Davalos et al. 2005; Haynes et al. 2006; Koizumi et al. 2007). Reactive microglial cells can in turn modulate astrocyte reactivity, involving P2Y receptors (Shinozaki et al. 2017). Neuron–microglia interactions are also controlled by several chemokines and chemokine receptors (e.g., fractalkine, CD200) and a loss of this control by challenged neurons can trigger microglial reactivity (Blank and Prinz 2013; Chapman et al. 2000; Streit et al. 2001). Activation of glial cells in response to damaged neurons, as well as infiltration of peripheral leukocytes, is not confined to PD, but also observed in other chronic neurodegenerative diseases (Lopategui Cabezas et al. 2014; Nayak et al. 2011).
The number of studies describing an explicit causal relationship between damaged DA neurons and the activation of glia is rather limited. In patients with PD, an increase in HMGB1—a protein released upon cell damage that signals danger and promotes neuroinflammation—was found in the substantia nigra pars compacta (SNpc) and in cerebrospinal fluid (CSF) (Santoro et al. 2016). In mice treated with MPTP/MPP+, administration of HMGB1-neutralizing antibodies partly inhibits DA cell death. The small molecule glycyrrhizin directly binds HMGB1 and reduces MPTP/MPP+-dependent DA cell death (Santoro et al. 2016; Sasaki et al. 2016). TLR-4-deficient mice are less vulnerable to MPTP/MPP+ intoxication and display a decreased number of reactive activated glial cells compared with MPTP/MPP+-treated wild-type animals (Noelker et al. 2013). Inhibition of RAGEs, which are upregulated in the striatum following rotenone exposure, suppresses NF-κB activation, as well as the expression of NF-κB-regulated inflammatory markers such as tumor necrosis factor alpha (TNF-α), inducible nitric oxide synthase (iNOS), and myeloperoxidase (Abdelsalam and Safar 2015). Injection of fractalkine—normally released by damaged neurons—into the SNpc causes microglia activation by binding to CX3C chemokine receptor 1 (CX3CR1). Pre-administration of an anti-CX3CR1 antibody before MPP+ injection into the SNpc protects from glial activation (Shan et al. 2011). Intracerebral injection of neuromelanin, a derivative of l-DOPA that accumulates in catecholaminergic neurons, causes an inflammatory activation of glial cells in the rat brain, indicating that degenerating DA neurons are leading to neuroinflammation (Zecca et al. 2008). Furthermore, DA neurons in the process of degeneration signal Ca2+ waves, attracting neighboring glial cells, hence contributing to the well-defined accumulation of activated glial cells at the sites of neurodegeneration (Sieger et al. 2012).
Studies investigating the role of DA neurodegeneration on glial activation often utilize toxicants such as MPTP/MPP+ or rotenone. Although no in vivo evidence exists, it cannot be ruled out that these toxicants directly influence glial activation, e.g., by the active conversion of the pro-toxicant MPTP (Schildknecht et al. 2015). A rich body of experimental evidence indicates an outstanding role of extracellular ASYN in the inflammatory activation of glial cells (Hoenen et al. 2016; Lee et al. 2010a, b). However, these factors are not necessarily associated with a degeneration of DA neurons, as ASYN can be excreted by viable neurons (Emmanouilidou et al. 2010; Marques and Outeiro 2012).
KER 6: relationship between “neuroinflammation” (KE 5) and “degeneration of DA neurons” (KE 4) (Fig. 7)
Neuroinflammation, first described by McGeer et al. (1988), encompasses the activation of glial cells (microglia and astrocytes) and is regularly observed in association with chronic neurodegenerative diseases such as PD, Alzheimer’s disease, and Huntington’s disease (Bagyinszky et al. 2017; Falsig et al. 2004, 2006; McGeer and McGeer 2008; Vivekanantham et al. 2015). Both cell types contribute to a pro-inflammatory/neurotoxic environment by releasing cytokines such as interleukin (IL)-1β, TNF-α, or interferon gamma (IFN-γ), mediators such as nitric oxide (•NO) or superoxide (•O2 −), ceramide, gangliosides, and components of the complement system (Boka et al. 1994; Brown and Bal-Price 2003; Dong and Benveniste 2001; Liberatore et al. 1999; Norden et al. 2015). Neuroinflammation-induced neuronal degeneration depends to a large extent on damage evoked by free radical species such as •NO, •O2 −, •OH, H2O2, N2O3, or peroxynitrite, which are actively formed by activated glia (Daiber et al. 2009; Hunot et al. 1996; Knott et al. 2000; Le et al. 1999; Mogi et al. 1994). These free radicals harm neuronal mitochondria and challenge neuronal proteostasis and redox equilibria. Once a certain threshold of radical-mediated damage is reached, neurodegeneration is observed (Chen et al. 2015; Khan et al. 2016). In addition, activated glial cells can alter the integrity of the blood–brain barrier (BBB) and hence allow an infiltration of peripheral immune cells into the CNS (Lopez-Ramirez et al. 2014). Moreover, peripheral TNF-α and IL-1α can traverse the BBB (Banks 2005; Pan and Kastin 2002). As a consequence of their passage, chronic peripheral inflammation can contribute to the selective demise of nigrostriatal DA neurons in the brain (Hernández-Romero et al. 2012; Pott Godoy et al. 2008; Villarán et al. 2010).
Neuronal injury/death triggers neuroinflammation (see KER 5), which in turn can lead to neuronal degeneration, contributing to a self-perpetuating vicious circle, which is assumed to be a key element in the pathogenesis of several neurodegenerative diseases including PD (Barbeito et al. 2010; Blasko et al. 2004; Griffin et al. 1998; Hirsch and Hunot 2009).
Nigrostriatal neurodegeneration can be evoked by stereotaxic injection or systemic application of lipopolysaccharide (LPS), a known activator of microglia (Frank-Cannon et al. 2008; He et al. 2013; Herrera et al. 2000; Ramsey and Tansey 2014).
Strategies to dampen neuroinflammation and protect DA neurons have either focused on inhibiting the pro-inflammatory (M1) phenotype of microglia and/or on supporting their anti-inflammatory activation state (M2) (Hernández-Romero et al. 2008; Lecca et al. 2015; Lu et al. 2000; Moehle and West 2015; Moon et al. 2009; Pisanu et al. 2014; Roy et al. 2012; Wu et al. 2002). In comparison to control mice, intra-nigrostriatal injection of LPS largely failed to initiate a sustained neuroinflammatory response in an IL-1β knockdown mouse model that exhibits significantly less DA neurodegeneration (Tanaka et al. 2013). In an MPTP/MPP+ model, IFN-γ depletion completely prevents microglial activation and protects from the loss of nigrostriatal DA neurons (Mount et al. 2007). Deletion of TNF-α confers only a partial protection from MPTP/MPP+-dependent neuroinflammation and DA neurodegeneration (Ferger et al. 2004), while a double knockout mouse (Leng et al. 2005) of TNF-α receptors 1 and 2 exhibits an almost complete protection from MPTP/MPP+-dependent glial activation and DA neurodegeneration (Sriram et al. 2002, 2006). Comparable protection in this TNF-α receptor double knockout mouse was also reported when LPS was applied instead of MPTP (Qin et al. 2007). Blocking TNF-α by expression of TNF-α inhibitor protein protects from DA neuron loss in animal models of PD (McCoy et al. 2006).
In mixed neuron/glia co-cultures, pretreatment with anti-inflammatory TGF-β1 prevents from neurodegeneration evoked by MPP+ (Liu et al. 2016). Genetic silencing of the TGF-β receptor 1 in microglia reverses this protective effect, indicating a significant role of pro-inflammatory glial activation in the observed degeneration of neurons (Liu et al. 2016). Another example is the PPAR-γ agonist MDG548 that decreases NF-κB activation in microglia evoked by LPS (Lecca et al. 2015). When mice are exposed to MPTP instead of LPS, MDG548 reduces microglial activation and protects from DA neurodegeneration (Lecca et al. 2015).
Other strategies interfering with the infiltration of peripheral CD4+/CD8+ T lymphocytes, which was reported as a contributing factor of DA neurodegeneration (Brochard et al. 2009; Appel 2009; Stone et al. 2009), also revealed neuroprotection. The corticosteroid dexamethasone, by acting as an inhibitor of T-cell infiltration, dampens glial activation and DA neurodegeneration (Castaño et al. 2002). In a MPTP/MPP+ model, a mutation in the functional receptor of CD4+ T cells protects from DA neurodegeneration (Brochard et al. 2009). The adoptive transfer of immunosuppressive CD4+/CD25+ regulatory T cells was sufficient for the protection from DA neuronal death (Laurie et al. 2007; Reynolds et al. 2007). Besides these experimental models, current clinical trials involving patients with PD, strongly suggest minocycline, an inhibitor of microglial reactivity (Du et al. 2001; Faust et al. 2009; Schildknecht et al. 2011; Tikka et al. 2001; Wu et al. 2002), as a promising agent for the protection of nigrostriatal DA neurons when used in combination with other therapies such as antioxidants or MAO-B inhibitors (Galpern and Cudkowicz 2007; Matthews et al. 1999; NINDS NET-PD Investigators 2006, 2008; Shults et al. 1997, 1999, 2002; Shults 2003, 2004; Yang et al. 2009).
The majority of studies focusing on the contribution of pro-inflammatory mediators such as IL-1β, TNF-α, or IFN-γ were performed in MPTP models. Hence, in addition to the inflammatory response, MPP+-dependent mitochondrial inhibition and ROS formation were still present in these studies. Mice with quiescent microglia are still susceptible to MPTP toxicity (Kinugawa et al. 2013), indicating a rather minor contribution of inflammation to the observed neurodegeneration in the MPTP models. Studies involving LPS injections for the induction of inflammation were almost exclusively conducted in rodents. In comparison to the situation in humans, rodents display greater amounts of •NO and •O2 − upon inflammatory activation (Bachschmid et al. 2005; Hoos et al. 2014; Schildknecht et al. 2004, 2005), indicating that these radical species contribute to a larger extend to neurodegeneration in rodents compared to the situation in humans. This concept received substantial support by the outcome of clinical studies involving antioxidant therapy over extended periods of time that exhibited no signs of a significant delay in disease progression (Chen 2003, 2005; Etminan et al. 2008; Hernán et al. 2006; Ton et al. 2006; Parkinson Study Group 1993; Shoulson 1998).
KER 7: relationship between “mitochondrial dysfunction” (KE 2) and “degeneration of DA neurons” (KE 4) (Fig. 8)
KER 7 is an extraordinary element in the present AOP inasmuch it circumvents KE 3. Impaired proteostasis (KE 3) is observed under conditions of moderate and chronic inhibition of complex I. In response to an instant and complete inhibition of complex I by high concentrations of rotenone or MPP+, the instant termination of ATP supply can lead to a rapid (1–2 h) degeneration without significant involvement of impaired proteostasis. Incorporation of these time- and concentration-dependent differences in neurodegeneration upon complex I inhibition was the rationale to specify KER 7.
Mitochondria serve as the main source of ATP in eukaryotic cells, and they are vitally involved in the regulation of cellular Ca2+ homeostasis (Baughman et al. 2011; Brini et al. 2014; Calì et al. 2014; De Stefani et al. 2011) as well as in apoptotic processes (Charan et al. 2014; Hu et al. 2015; Liu et al. 2015; Rasheed et al. 2017). Mitochondrial dysfunction is characterized by dysfunctional cellular Ca2+ handling (Orrenius et al. 2003), reduced mitochondrial ATP levels, and increased ROS (Banerjee et al. 2009; Bose and Beal 2016; Subramaniam and Chesselet 2013). In contrast to other cell types, neurons have only a moderate capacity to upregulate their rate of glycolysis upon inhibition of mitochondria (Almeida et al. 2001, 2004; Herrero-Mendez et al. 2009). Therefore, they are more vulnerable towards dysfunctional mitochondria than other cell types. Among the different neuronal types, nigrostriatal DA neurons display a preferential sensitivity towards complex I inhibition (Betarbet et al. 2000; Jackson-Lewis et al. 1995) as a consequence of a set of unique intrinsic features. First, nigrostriatal DA neurons possess autonomous pacemaking activity, relying on L-type Ca2+ channel (CaV1.3)-dependent Ca2+ influx for membrane depolarization (Chan et al. 2007; Guzman et al. 2009; Nedergaard et al. 1993). The relevance of Ca2+-dependent pacemaking as a sensitizing factor becomes evident in comparison with DA neurons of the ventral tegmental area (VTA). These are significantly less sensitive to complex I inhibition, and they differ from nigrostriatal DA neurons by their reliance on extracellular Na+ for pacemaking (Khaliq and Bean 2010). The constant influx of extracellular Ca2+ represents an energy- demanding strategy (Surmeier et al. 2011; Surmeier and Schumacker 2013). The energy balance of nigrostriatal DA neurons, and hence their dependence on proper mitochondrial function, is furthermore challenged by their unique architecture (Bolam and Pissadaki 2012; Matsuda et al. 2009) comprising long unmyelinated axons and higher numbers of energy-consuming synapses, compared with catecholaminergic neurons of other brain regions (Pacelli et al. 2015; Pissadaki and Bolam 2013). As a consequence, total cell surface and the energy required to maintain the membrane potential is higher in nigrostriatal DA neurons (Pacelli et al. 2015; Bolam and Pissadaki 2012; Brichta and Greengard 2014). In comparison with neurons of other brain areas, mitochondria of nigrostriatal DA neurons can hence barely meet the energy requirement of the cell, even under normal conditions. It becomes apparent that even moderate impairments in mitochondrial function can lead to a preferential damage and demise of nigrostriatal DA neurons, while other neuronal populations under the same conditions are still spared. An additional sensitizing factor is the presence of DA, which pre-disposes neuronal cells to oxidative stress and renders ASYN particularly cytotoxic (Pacelli et al. 2015; Schildknecht et al. 2009, 2013, 2017).
The experimental support for the direct relationship between mitochondrial dysfunction and the degeneration of nigrostriatal DA neurons is based on observations made with neuronal cell cultures and with genetically modified in vivo models. Mitochondrial dysfunction can be initiated by complex I inhibitors that prevent mitochondrial ATP generation and concomitantly stimulate mitochondrial ROS formation (Chan et al. 1991; Fabre et al. 1999; Hasegawa et al. 1990; Nicklas et al. 1985; Przedborski et al. 1996). Alternative experimental means to evoke mitochondrial dysfunction are e.g., transfer of mtDNA from patients with PD into mtDNA-free cells (cybrids) (Marella et al. 2008; Sherer et al. 2003, 2007) or knockdown of the regulator of mitogenesis Tfam (Ekstrand et al. 2007). In all of these examples, the advent of mitochondrial dysfunction was directly correlated with the demise of neurons, and an elevated sensitivity of neurons harboring dysfunctional mitochondria towards secondary stressors. The degeneration of DA neurons is prevented by treatment with antioxidants (Chen et al. 2015; Choi et al. 2014; Hajieva et al. 2009; Sherer et al. 2003, 2007). To exemplify the present AOP, mitochondrial dysfunction can be evoked by application of complex I inhibitors such as MPTP/MPP+ or rotenone both in vitro and in vivo. Such complex I inhibitor-mediated mitochondrial dysfunction is directly correlated with the dysfunction of nigrostriatal DA neurons (Hantraye et al. 1993; Langston et al. 1999; Moratalla et al. 1992; Varastet et al. 1994). Experimental expression of the inhibitor-insensitive complex I surrogate NDI-1, either in neuronal cell cultures or in vivo by unilateral injection of adeno-associated virus into the nigrostriatal system (Marella et al. 2008; Sherer et al. 2003, 2007) protects against complex I inhibitor-dependent mitochondrial dysfunction and prevents the demise of nigrostriatal DA neurons. Complex I-independent electron input into the respiratory chain, e.g., by application of methylene blue (Wen et al. 2011) or coenzyme Q10 (Beal et al. 1998), reduces mitochondrial dysfunction and protects from DA neurodegeneration. These examples illustrate that protection from the loss of mitochondrial ATP, or from conditions of oxidative stress, i.e., features of mitochondrial dysfunction, are effective means to prevent the demise of nigrostriatal neurons.
Mitochondrial dysfunction comprises a series of adverse processes such as the decline of the mitochondrial membrane potential, opening of the mtPTP, elevated ROS formation, or the release of cytochrome c (Gandhi et al. 2009; Heo et al. 2012; Irrcher et al. 2010; Leist et al. 1998; Pöltl et al. 2012; Toulorge et al. 2016; Wang et al. 2012). Currently, there is no consensus on how many of these changes need to occur to meet the criteria for mitochondrial dysfunction. A decline in ATP generation and elevated •O2 − formation are the two main consequences of complex I inhibition (Lambert and Brand 2004; Schildknecht et al. 2009). Although experimental restoration of ATP and management of elevated ROS by antioxidants have been suggested as protective means, the respective quantitative contribution of ATP and ROS to the observed neurodegeneration has not been fully addressed in the literature. Mitochondrial dysfunction leads to oxidative stress, but oxidative stress in turn also leads to mitochondrial dysfunction (Hasegawa et al. 1990; Jana et al. 2011; Khan et al. 2005). Thus, empirical support based on antioxidants can be ambiguous. In KER 7, it is assumed that KE 2 directly leads to KE 4 and KE 5, especially at high intensities of insult. However, it is unclear whether such conditions are found in humans exposed to toxicants.
KER 8: relationship between the “degeneration of DA neurons” (KE 4) and the onset of “parkinsonian motor deficits” (AO) (Fig. 9)
DA neurons of the substantia nigra project into the striatum, where they release DA (Joel and Weiner 2000; Lynd-Balta and Haber 1994a, b). The loss of nigrostriatal DA neurons observed in PD leads to a reduction in striatal DA levels (Bernheimer et al. 1973; Ehringer et al. 1960). All PD forms are characterized by the loss of striatal DA, which is directly correlated with the onset of PD motor symptoms (Ehringer et al. 1960). Striatal DA is a main regulator of motor output from the cortex to the periphery. Basal ganglia modulate motor output information that is looped back via the thalamus to the motor output cortex (Alexander et al. 1986; Blandini et al. 2000; Obeso et al. 2008a, b). A decline in striatal DA leads to disturbances in this feedback loop and reflects key parkinsonian symptoms such as rigidity, bradykinesia, and tremor (Bain 2007; Jankovic 2008; Rodriguez-Oroz et al. 2009). An experimental knockdown of tyrosine hydroxylase as the key enzyme in catecholamine synthesis leads to reduced motor coordination (Korner et al. 2015), while a hyperdopaminergic tone, evoked by genetic deletion of the DA transporter, results in motor hyperactivity (Gainetdinov et al. 1999). Characteristic PD-associated motor deficits are usually observed at a reduction of striatal DA by ca. 80% (Kirik et al. 1998; Koller 1992).
Experimental support for a causal relationship between the loss of nigrostriatal DA neurons and the onset of parkinsonian motor deficits comes from patients with PD, humans accidentally exposed to MPTP, in vivo studies with rodents and non-human primates, and from in vitro models.
Analysis of brains from patients with PD reveals a significant reduction of striatal DA that correlates with the degeneration of nigrostriatal DA neurons (Earle 1968; Lloyd et al. 1975). Live assessment of DA neuron content in patients with PD indicates a causal correlation between nigrostriatal DA content and the severity of PD motor deficits (Benamer et al. 2000; Lin et al. 2014; Pirker 2003; Rakshi et al. 1999; Rinne et al. 1995; Tissingh et al. 1998). Substitution of endogenous striatal DA by l-DOPA leads to improved motor performance (Gilmour et al. 2011; Heimer et al. 2002; Hutchinson et al. 1997; Levy et al. 2001; Lloyd et al. 1975; Papa et al. 1999; Yam et al. 1998). Elevation of endogenous striatal DA, by application of inhibitors targeting its degradation enzyme MAO-B, is also correlated with improved motor performance (Pålhagen et al. 1998, 2006; Olanow et al. 2008; Parkinson Study Group 1993, 1996, 2002; Rascol et al. 2005). Case reports further indicate the re-innervation of the striatum with projections of transplanted DA neurons, a restoration of striatal DA levels, and a subsequent improvement of motor performance (Ben-Hur et al. 2004; Kordower et al. 1995, 1998; Mendez et al. 2008; Schumacher et al. 2000; Widner et al. 1992).
Non-human primates represent a model, highly reflective for the situation in humans with respect to brain architecture and DA motor deficits. Studies with MPTP-exposed monkeys reveal a correlation between striatal DA, nigrostriatal DA neuron numbers, and the onset of a PD motor phenotype (Bezard et al. 2001). Similar to the situation in humans, a reduction of striatal DA by ca. 80% leads to the manifestation of PD motor deficits (Aziz et al. 1991; Bergman et al. 1990; Blesa et al. 2012; Filion and Tremblay 1991; Mitchell et al.1989; Porras et al. 2012). Supplementation of endogenous DA by l-DOPA application in these models reverses the loss of motor output performance (Bédard et al. 1986; Clarke et al. 1987; Jenner 2008; Kuoppamäki et al. 2007; Langston et al. 2000; Smith et al. 2003).
In rodents, systemic administration of rotenone leads to loss of striatal DA and DA neurons; this loss is associated with the onset of motor deficits, reminiscent of those motor impairments observed in patients with PD (Alam and Schmidt 2002, 2004; Cannon et al. 2009; Fleming et al. 2004; Höglinger et al. 2003a, b). In mice, MPTP is the most widely applied experimental PD toxicant. It provides results comparable to those obtained from rotenone experiments in rats (Alvarez-Fischer et al. 2008; Fornai et al. 2005; Gibrat et al. 2009; Hung and Lee 1996; Petroske et al. 2001; Rozas et al. 1998). Application of l-DOPA, inhibition of endogenous DA degradation by MAO-B inhibitors, deep brain stimulation, and transplantation of precursor cells into the nigrostriatal system all lead to a restoration of striatal DA content and an improvement of motor performance (Altarche-Xifro et al. 2016; Kong et al. 2015; Shaw et al. 2010; Schierle et al. 1999; Shin et al. 2009).
Striatal DA is a key modulator of extrapyramidal motor output control. Although a close correlation between striatal DA and the onset of motor deficits is apparent, the specificity of motor abnormalities observed in PD has not been fully explained. Neuronal loss in PD or in MPP+/rotenone-treated animals is not confined to the nigrostriatal system. Other areas such as the locus coeruleus also undergo neurodegeneration (Forno et al. 1986; Moratalla et al. 1992; Muthane et al. 1994; Seniuk et al. 1990; Snow et al. 2000). It cannot be excluded that additional brain regions might significantly contribute to the parkinsonian motor phenotype. In subacute treatment schemes (rotenone, MPTP), a significant, sometimes complete, recovery of motor deficits can be observed (Petroske et al. 2001). For the assessment of DA neuron numbers, DA markers such as TH, DAT, and VMAT-2 are often employed using western blot, immunohistochemical staining, and polymerase chain reaction (PCR). However, the expression levels of these targets can be transiently regulated and might, therefore, provide misleading information on the survival of DA neurons (Jakowec et al. 2004; Petzinger et al. 2006; Rothblat et al. 2001). For the assessment of motor performance in rodents, a variety of different behavioral assays emerged in the course of recent decades. The parameters assessed in these assays are not directly representative for those features observed in human PD (Meredith and Kang 2006). In addition, it has not been established for all endpoints whether deficits can be fully rescued by l-DOPA or DA agonists.
AOP uncertainties and evaluation in an overall context
Use of example compounds for the AOP
Per definition, AOPs are compound agnostic and consequently include no ADME considerations. However, for the assembly of AOPs, and in particular for the empirical support of KERs, the behavior of model compounds plays a significant role and is important for the assessment of plausibility of an AOP. Model toxicants show a distinct toxicokinetic behavior that needs to be taken into account for the evaluation of the consistency of the AOP. In this context, the AOP presented herein relies very heavily on its exemplification by two model toxicants: rotenone and MPTP. For MPTP, the evaluation of the dose and KE sequence (response–response) consistency is particularly difficult, because: (1) MPTP needs to be enzymatically activated into the active toxicant MPP+ by brain glial cells; (2) once the active metabolite MPP+ has been formed, and the MIE occurred, the following KE can be initiated within a very narrow time window. Moreover, information on its uptake, conversion rates, transport within the brain, and its excretion, can have a significant influence on the AO, and such toxicokinetic factors might explain the intra- (mouse strains) and inter-species (rodents versus non-human primates) differences observed.
Altogether, if data are combined from multiple published studies, both rotenone and MPTP show a good response–response and temporal concordance. However, it remains unclear for MPTP, whether there are doses that trigger only early KEs without activating the AO (Table 3). Although a wide variety of other complex I inhibitors have been described in the literature, these compounds were only rarely applied in studies linking information from isolated mitochondria, cell culture, and in vivo models.
Link of the MIE to the downstream events of the AOP
The MIE involves binding of an inhibitor to complex I, leading to the inhibition of complex I as KE 1. This appears as a rather unambiguous biochemical event but on closer inspection, these events are highly complex and a complete description and measurement in the context of the overall AOP is still missing. Mapping of the exact binding site would require confirmation from several independent laboratories, and the same applies to the type of changes in the mitochondrial respiratory chain that have a direct influence on downstream events of the AOP. In this context, it is important to stress again that complex I inhibition not only results in a reduction of mitochondrial electron transport, but also in an increase in superoxide formation. It is not clear whether all complex I inhibitors trigger these events at similar potency and efficacy ratios, and thus, whether all complex I inhibitors would lead to a similar activation of KE 2 and the AO. Low concentrations of complex I inhibitors were reported to evoke elevated superoxide formation by isolated mitochondria without significant influence on mitochondrial ATP generation. This aspect is of importance in light of observations indicating that in vivo models best reflect molecular events typical for PD following low-dose and chronic inhibitor infusion (Crawley 1999). These examples illustrate the necessity for a quantitative assessment of the respective contribution of declined mitochondrial ATP generation and elevated mitochondrial superoxide formation upon complex I inhibition to neurodegeneration in models of different complexity (isolated mitochondria, cell cultures, in vivo models). The involvement of superoxide formation deserves even more attention, considering reports that illustrate a negative feedback inhibition of complex I by superoxide. As a consequence of the current inadequate knowledge on the roles of ATP and superoxide, defined no-effect levels for complex I inhibitors have not yet been established.
Potential branching of the AOP downstream of mitochondrial events
Impaired proteostasis (KE 3) comprises several complex biological processes involved in the formation, localization, and removal of proteins, in protein assembly (up to organelles), and in the removal of misfolded protein aggregates. Disturbance of proteostasis may follow different patterns. For example, axonal transport may be disturbed, while processes such as proteasomal degradation and removal of aggregated proteins function well. Alternatively, CMA may be disturbed, while other processes such as axonal transport maintain their functionality. At present, the literature provides mainly information on isolated aspects of KE 3, such as UPS and CMA activities. In the future, studies that focus on the particular role of these events in the AOP and compare different proteostasis processes would be desirable. At present, the KER linking mitochondrial dysfunction (KE 2) and impaired proteostasis (KE 3) and the KER linking impaired proteostasis (KE 3) and degeneration of DA neurons (KE 4) are characterized by a lack of quantitative threshold data from independent laboratories and complementary model systems. Such quantitative data would allow a sharper definition of the relationship between the respective KEs (when KEup triggers KEdown). The majority of information on impaired proteostasis included in the present AOP is based on studies with an explicit focus on the PD-associated protein ASYN, but not on the global cellular imbalance between the formation and degradation of misfolded proteins.
A further notable aspect is that KER 7 provides a direct link between mitochondrial dysfunction (KE 2) and the degeneration of DA neurons (KE 4), hereby circumventing KE 3 (impaired proteostasis). The rationale for KER 7 is based on observations indicating a direct link between KE 2 and KE 4 under conditions of severe mitochondrial dysfunction, e.g., evoked by an almost complete inhibition of complex I.
Feed-forward loop involving neuroinflammation
In contrast to the standard unidirectional chain of events requested by OECD guidelines for the organization of an AOP, the present AOP included a positive feed-forward loop involving neuroinflammation (KE 5). The inclusion of neuroinflammation as an independent KE was necessary and justified, as neuroinflammation alone is sufficient to evoke DA neurodegeneration, and neuroinflammation can be triggered by complex I inhibitors. However, quantitative information on the extent and type of neuroinflammation is currently not available. Furthermore, significant species differences with respect to the quantitative contribution of neuroinflammation and the mediators involved in DA neurodegeneration have been reported. In rodents, MPTP/MPP+-evoked neuroinflammation is mainly self-limiting after the acute phase of neurodegeneration, while it persists for years and even decades in monkey models and in humans. Despite serious attempts, it has not been possible to determine whether neuroinflammation occurs before or after neurodegeneration. The most likely reason for this is that low levels of neurodegeneration trigger neuroinflammation, and that neuroinflammation then triggers more neurodegeneration, so that these two events form a self-perpetuating vicious cycle (Schildknecht et al. 2017). The feed-forward loop depicted in the AOP is the best representation of this pathogenic situation. In this context, it is important to note that other feed-forward loops play a role in this AOP. They are somewhat less prominent and have not been graphically represented, as they could be considered as modulatory events that are covered in the text descriptions. For instance, a self-amplifying feed-forward loop includes mutual interactions between mitochondrial dysfunction and impaired proteostasis. Experimental evidence suggests that disturbed proteostasis affects mitochondrial function (Sherer et al. 2002). Such effects require more attention, if this AOP is used as basis for construction of a quantitative AOP or for a systems biology model.
Human disease symptom as AO
It is important to distinguish between PD (a complex human disease with multiple symptoms and most likely multiple etiologic factors) and parkinsonian motor deficits (one distinct and sharply circumscribed feature of PD, but also of poisoning). A disease cannot be an AO, but a defined defect, just as a certain type of motor disturbance can be an AO. The type of motor dysfunction of interest here is characterized by problems with movement initiation and termination, much more than problems with the movement process as such. It is called here ‘parkinsonian’ as it is observed in PD, but similar defects are also found in other, not PD-related situations. PD is considered as a multifactorial disease. Consequently, the phenotype (pattern of symptoms) of individuals affected by parkinsonian motor deficits can vary significantly. Therefore, the AO of the present AOP does not cover the entire spectrum of PD-associated phenotypes. Other AOP would need to cover these, and they may have entirely different MIE and KEs.
Overall judgment and incorporation in larger networks
One of the cardinal aspects in the evaluation of an AOP is the question whether activation of downstream KEs, or the AO, can be prevented by experimental interference with an upstream KE. The essentiality evidence is important to assess the relative level of confidence of the AOP and, when considering all the elements on a rank order, it is secondary only to the biological plausibility for the KERs. Evidence of essentiality was considered “strong” if direct evidence was available in the literature from specifically designed experimental studies illustrating that KEup was essential for at least one KEdown or the AO. Evidence was considered “moderate” if only indirect observations were available in the literature, illustrating that KEup was essential for at least one KEdown or the AO. In this case, the experiments could not directly address the essentiality of KEup, but instead e.g., a modulatory factor for KEup. In other cases, a KEup was directly modified, but the tool used was somehow ambiguous or only shown to work by correlation (not intervention). For instance, different approaches to block events associated with (or modulating) neuroinflammation by knockout of mediators (e.g., TNF-α) or by enzyme inhibition (e.g., cyclooxygenase) were found to prevent neurodegeneration. Evidence was considered “weak” if the available observations were contradicting, or if the approaches chosen were indirect and could be interpreted in different ways. Description of the elements supporting essentiality are here embedded in the text for empirical support, a summary of the key evidence and the trend for their weight is reported in Table 2. According to this rule-set, the evidence for essentiality of KEs 1, 2, and 4 was considered “strong”, while the evidence for essentiality of KEs 3 and 5 was considered “moderate” (Table 2).
Although an AOP is an independent unit of information, networks of AOP formed by overlapping KEs are envisaged as a foundation for toxicological evaluations in the future. Quantitative information provided by the individual AOPs of an integrated network could provide a valuable basis for in silico modeling of cellular networks. For this reason, it is important to consider where the present AOP overlaps with other AOPs. KE 2 (mitochondrial dysfunction) was identified in nine other AOPs, where it plays different roles compared with the one described herein. KE 5 (neuroinflammation) is shared by three other AOPs (Fig. 1). Notably, none of these other AOPs has the same AO as described herein.
Regulatory and scientific context
One of the primary objectives of the AOP framework is knowledge assembly, i.e., gathering information, attained through scientific research by subject-matter experts, for its accessibility to regulators during the decision-making process. To avoid any misunderstanding, it is important to clarify misconceptions about AOPs: (1) risk assessments cannot be based on AOPs alone, as they do not address exposure and toxicokinetic issues. AOPs inform on a potential hazard, and in this sense, AOP information can be integrated together with ADME information in an IATA approach; (2) AOPs are not testing strategies. AOPs are mainly assemblies of knowledge. This structured and quality-controlled information can aid the interpretation of high-throughput testing or pathway-based data, in the context of relevant apical hazards; (3) AOPs are not mode of action (MoA) analyses. The MoA framework, as applied in human health risk assessment, represents a systematic description and analysis of the means through which a specific chemical elicits an adverse effect in an organism. AOPs, which are intended to be generalizable to any chemical acting on a particular MIE (chemical agnostic) can be applied to MoA analysis, but the terms are not synonymous; (4) AOPs are not computational models. Rather, they are intended to promote qualitative understanding of how an alteration in a KEupstream impacts downstream KEs, and consequently provide information that may be represented in the form of one or more computational models (Wittwehr et al. 2017).
With regard to safety evaluation, AOPs may not comprehensively predict all toxicological outcomes. They do not solve all the challenges of in vitro to in vivo extrapolation. AOPs do not describe every detail of adverse and adaptive biology underlying an organism’s response to a stressor. They cannot account for every aspect of individual variability nor for every environmental or life-history variable that may affect a toxicological outcome in real-world settings. They are, in short, simply a means to help organize what we know about how biological perturbations can lead to apical adverse outcomes, and use that information to aid regulatory decision-making. In this context, this would mean that the present AOP is useful as guidance to judge the action of complex I inhibitors, but it cannot help to understand other neurotoxicants or pathways that cause motor degeneration. As for any complex tool, there is a main intended application, which one needs to be aware of. Use of the tool for other purposes may yield bad results, but this does not mean that the tool is bad for its original purpose.
The regulatory implications of this AOP have already been discussed in detail (Ockleford et al. 2017). Based on this AOP, it was concluded that if a chemical/pesticide triggers the MIE, it should be considered a risk factor for the development of PD.
The inclusion of an AOP into the AOP Wiki platform allows an integration of novel findings into an existing AOP. However, this flexibility requires the commitment not only from the original authors but also from other experts in the field, to continuously adjust the existing AOPs. Due to the laborious and time-consuming update process, without the “reward” in form of a publication authorship, a realistic approach to ensure a contemporary integration of novel observations into existing AOPs needs to involve a funding program to motivate comprehensive and qualified updates in defined intervals. For the motivation of scientists to participate in the development of new, or in the update of existing AOPs, promotion of the AOP concept to boost its reputation and awareness by the scientific community is another essential prerequisite for the successful establishment of AOPs as an integral tool in modern toxicology. As a consequence of this, regular revision of AOPs should be envisioned to keep them updated in line with the progress made in science, knowledge, and methodologies. Without such measures, there is a high risk that scientific development and regulatory needs advance while individual AOPs become outdated. As this may affect regulatory decision-making, it is a serious concern that needs to be considered in the future.
Abdelsalam RM, Safar MM (2015) Neuroprotective effects of vildagliptin in rat rotenone Parkinson’s disease model: role of RAGE-NFκB and Nrf2-antioxidant signaling pathways. J Neurochem 133(5):700–707
Adhihetty PJ, Beal MF (2008) Creatine and its potential therapeutic value for targeting cellular energy impairment in neurodegenerative diseases. Neuromol Med 10(4):275–290
Ahammadsahib KI, Hollingworth RM, McGovren JP, Hui YH, McLaughlin JL (1993) Mode of action of bullatacin: a potent antitumor and pesticidal annonaceous acetogenin. Life Sci 53(14):1113–1120
Alam M, Schmidt WJ (2002) Rotenone destroys dopaminergic neurons and induces parkinsonian symptoms in rats. Behav Brain Res 136(1):317–324
Alam M, Schmidt WJ (2004) L-DOPA reverses the hypokinetic behaviour and rigidity in rotenone-treated rats. Behav Brain Res 153(2):439–446
Albracht SP, Mariette A, de Jong P (1997) Bovine-heart NADH:ubiquinone oxidoreductase is a monomer with 8 Fe–S clusters and 2 FMN groups. Biochim Biophys Acta 1318(1–2):92–106
Alexander GE, DeLong MR, Strick PL (1986) Parallel organization of functionally segregated circuits linking basal ganglia and cortex. Annu Rev Neurosci 9:357–381
Almeida A, Almeida J, Bolaños JP, Moncada S (2001) Different responses of astrocytes and neurons to nitric oxide: the role of glycolytically generated ATP in astrocyte protection. Proc Natl Acad Sci USA 98(26):15294–15299
Almeida A, Moncada S, Bolaños JP (2004) Nitric oxide switches on glycolysis through the AMP protein kinase and 6-phosphofructo-2-kinase pathway. Nat Cell Biol 6(1):45–51
Altarche-Xifro W, di Vicino U, Muñoz-Martin MI, Bortolozzi A, Bové J, Vila M, Cosma MP (2016) Functional rescue of dopaminergic neuron loss in Parkinson’s disease mice after transplantation of hematopoietic stem and progenitor cells. EBioMedicine 8:83–95
Alvarez-Fischer D, Guerreiro S, Hunot S, Saurini F, Marien M, Sokoloff P, Hirsch EC, Hartmann A, Michel PP (2008) Modelling Parkinson-like neurodegeneration via osmotic minipump delivery of MPTP and probenecid. J Neurochem 107(3):701–711
Ambrosi G, Ghezzi 2, Sepe S, Milanese C, Payan-Gomez C, Bombardieri CR, Armentero MT, Zangaglia R, Pacchetti C, Mastroberardino PG, Blandini F (2014) Bioenergetic and proteolytic defects in fibroblasts from patients with sporadic Parkinson’s disease. Biochim Biophys Acta 1842(9):1385–1394
Appel SH (2009) CD4+ T cells mediate cytotoxicity in neurodegenerative diseases. J Clin Investig 119(1):13–15
Aschner M (1998) Immune and inflammatory responses in the CNS: modulation by astrocytes. Toxicol Lett 102–103:283–287
Aziz TZ, Peggs D, Sambrook MA, Crossman AR (1991) Lesion of the subthalamic nucleus for the alleviation of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-induced parkinsonism in the primate. Mov Disord 6(4):288–292
Bachschmid M, Schildknecht S, Ullrich V (2005) Redox regulation of vascular prostanoid synthesis by the nitric oxide-superoxide system. Biochem Biophys Res Commun 338(1):536–542
Bagyinszky E, Giau VV, Shim K, Suk K, An SSA, Kim S (2017) Role of inflammatory molecules in the Alzheimer’s disease progression and diagnosis. J Neurol Sci 376:242–254
Bain PG (2007) Parkinsonism and related disorders. Tremor Parkinsonism Relat Disord 13(Suppl 3):S369–S374
Bal-Price A, Lein PJ, Keil KP, Sethi S, Shafer T, Barenys M, Fritsche E, Sachana M, Meek ME (2017) Developing and applying the adverse outcome pathway concept for understanding and predicting neurotoxicity. Neurotoxicology 59:240–255
Baltazar MT, Dinis-Oliveira RJ, de Lourdes Bastos M, Tsatsakis AM, Duarte JA, Carvalho F (2014) Pesticides exposure as etiological factors of Parkinson’s disease and other neurodegenerative diseases—a mechanistic approach. Toxicol Lett 230(2):85–103
Banati RB (2002) Visualising microglial activation in vivo. Glia 40(2):206–217
Banerjee R, Starkov AA, Beal MF, Thomas B (2009) Mitochondrial dysfunction in the limelight of Parkinson’s disease pathogenesis. Biochim Biophys Acta 1792(7):651–663
Banks WA (2005) Blood-brain barrier transport of cytokines: a mechanism for neuropathology. Curr Pharm Des 11(8):973–984
Barbeito AG, Mesci P, Boillée S (2010) Motor neuron-immune interactions: the vicious circle of ALS. J Neural Transm (Vienna) 117(8):981–1000
Barnes CD (1983) The basal ganglia in extrapyramidal dysfunction. Brain Res Bull 11(2):271–275
Barrientos A, Moraes CT (1999) Titrating the effects of mitochondrial complex I impairment in the cell physiology. J Biol Chem 274(23):16188–16197
Bartels T, Choi JG, Selkoe DJ (2011) α-Synuclein occurs physiologically as a helically folded tetramer that resists aggregation. Nature 477(7362):107–110
Baughman JM, Perocchi F, Girgis HS, Plovanich M, Belcher-Timme CA, Sancak Y, Bao XR, Strittmatter L, Goldberger O, Bogorad RL, Koteliansky V, Mootha VK (2011) Integrative genomics identifies MCU as an essential component of the mitochondrial calcium uniporter. Nature 476(7360):341–345
Bauvy C, Meijer AJ, Codogno P (2009) Assaying of autophagic protein degradation. Methods Enzymol 452:47–61
Beal MF (2011) Neuroprotective effects of creatine. Amino Acids 40(5):1305–1313
Beal MF, Matthews RT, Tieleman A, Shults CW (1998) Coenzyme Q10 attenuates the 1-methyl-4-phenyl-1,2,3,tetrahydropyridine (MPTP) induced loss of striatal dopamine and dopaminergic axons in aged mice. Brain Res 783(1):109–114
Bédard PJ, Di Paolo T, Falardeau P, Boucher R (1986) Chronic treatment with L-DOPA, but not bromocriptine induces dyskinesia in MPTP-parkinsonian monkeys. Correlation with [3H]spiperone binding. Brain Res 379(2):294–299
Bellucci A, Zaltieri M, Navarria L, Grigoletto J, Missale C, Spano P (2012) From α-synuclein to synaptic dysfunctions: new insights into the pathophysiology of Parkinson’s disease. Brain Res 1476:183–202
Benamer HT, Patterson J, Wyper DJ, Hadley DM, Macphee GJ, Grosset DG (2000) Correlation of Parkinson’s disease severity and duration with 123I-FP-CIT SPECT striatal uptake. Mov Disord 15(4):692–698
Bence NF, Sampat RM, Kopito RR (2001) Impairment of the ubiquitin-proteasome system by protein aggregation. Science 292(5521):1552–1555
Ben-Hur T, Idelson M, Khaner H, Pera M, Reinhartz E, Itzik A, Reubinoff BE (2004) Transplantation of human embryonic stem cell-derived neural progenitors improves behavioral deficit in Parkinsonian rats. Stem Cells 22(7):1246–1255
Bentea E, Van der Perren A, Van Liefferinge J, El Arfani A, Albertini G, Demuyser T, Merckx E, Michotte Y, Smolders I, Baekelandt V, Massie A (2015) Nigral proteasome inhibition in mice leads to motor and non-motor deficits and increased expression of Ser129 phosphorylated α-synuclein. Front Behav Neurosci 9:68
Béraud D, Hathaway HA, Trecki J, Chasovskikh S, Johnson DA, Johnson JA, Federoff HJ, Shimoji M, Mhyre TR, Maguire-Zeiss KA (2013) Microglial activation and antioxidant responses induced by the Parkinson’s disease protein α-synuclein. J Neuroimmune Pharmacol 8(1):94–117
Berger I, Hershkovitz E, Shaag A, Edvardson S, Saada A, Elpeleg O (2008) Mitochondrial complex I deficiency caused by a deleterious NDUFA11 mutation. Ann Neurol 63(3):405–408
Bergman H, Wichmann T, DeLong MR (1990) Reversal of experimental parkinsonism by lesions of the subthalamic nucleus. Science 249(4975):1436–1438
Bermejo A, Figadere B, Zafra-Polo MC, Barrachina I, Estornell E, Cortes D (2005) Acetogenins from Annonaceae: recent progress in isolation, synthesis and mechanisms of action. Nat Prod Rep 22(2):269–303
Bernheimer H, Birkmayer W, Hornykiewicz O, Jellinger K, Seitelberger F (1973) Brain dopamine and the syndromes of Parkinson and Huntington. Clinical, morphological and neurochemical correlations. J Neurol Sci 20(4):415–455
Betarbet R, Sherer TB, MacKenzie G, Garcia-Osuna M, Panov AV, Greenamyre JT (2000) Chronic systemic pesticide exposure reproduces features of Parkinson’s disease. Nat Neurosci 3(12):1301–1306
Betarbet R, Sherer TB, Greenamyre JT (2005) Ubiquitin-proteasome system and Parkinson’s diseases. Exp Neurol 191(Suppl 1):S17–S27
Betarbet R, Canet-Aviles RM, Sherer TB, Mastroberardino PG, McLendon C, Kim JH, Lund S, Na HM, Taylor G, Bence NF, Kopito R, Seo BB, Yagi T, Yagi A, Klinefelter G, Cookson MR, Greenamyre JT (2006) Intersecting pathways to neurodegeneration in Parkinson’s disease: effects of the pesticide rotenone on DJ-1, alpha-synuclein, and the ubiquitin-proteasome system. Neurobiol Dis 22(2):404–420
Bezard E, Dovero S, Prunier C, Ravenscroft P, Chalon S, Guilloteau D, Crossman AR, Bioulac B, Brotchie JM, Gross CE (2001) Relationship between the appearance of symptoms and the level of nigrostriatal degeneration in a progressive 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-lesioned macaque model of Parkinson’s disease. J Neurosci 21(17):6853–6861
Bi J, Wang XB, Chen L, Hao S, An LJ, Jiang B, Guo L (2008) Catalpol protects mesencephalic neurons against MPTP induced neurotoxicity via attenuation of mitochondrial dysfunction and MAO-B activity. Toxicol In Vitro 22(8):1883–1889
Blandini F, Nappi G, Tassorelli C, Martignoni E (2000) Functional changes of the basal ganglia circuitry in Parkinson’s disease. Prog Neurobiol 62(1):63–88
Blank T, Prinz M (2013) Microglia as modulators of cognition and neuropsychiatric disorders. Glia 61(1):62–70
Blasko I, Stampfer-Kountchev M, Robatscher P, Veerhuis R, Eikelenboom P, Grubeck-Loebenstein B (2004) How chronic inflammation can affect the brain and support the development of Alzheimer’s disease in old age: the role of microglia and astrocytes. Aging Cell 3(4):169–176
Blesa J, Pifl C, Sánchez-González MA, Juri C, García-Cabezas MA, Adánez R, Iglesias E, Collantes M, Peñuelas I, Sánchez-Hernández JJ, Rodríguez-Oroz MC, Avendaño C, Hornykiewicz O, Cavada C, Obeso JA (2012) The nigrostriatal system in the presymptomatic and symptomatic stages in the MPTP monkey model: a PET, histological and biochemical study. Neurobiol Dis 48(1):79–91
Boka G, Anglade P, Wallach D, Javoy-Agid F, Agid Y, Hirsch EC (1994) Immunocytochemical analysis of tumor necrosis factor and its receptors in Parkinson’s disease. Neurosci Lett 172(1–2):151–154
Bolam JP, Pissadaki EK (2012) Living on the edge with too many mouths to feed: why dopamine neurons die. Mov Disord 27(12):1478–1483
Bolam JP, Hanley JJ, Booth PA, Bevan MD (2000) Synaptic organisation of the basal ganglia. J Anat 196 (Pt 4):527–542
Borland MK, Trimmer PA, Rubinstein JD, Keeney PM, Mohanakumar K, Liu L, Bennett JP Jr (2008) Chronic, low-dose rotenone reproduces Lewy neurites found in early stages of Parkinson’s disease, reduces mitochondrial movement and slowly kills differentiated SH-SY5Y neural cells. Mol Neurodegener 3:21
Borutaite V, Budriunaite A, Brown GC (2000) Reversal of nitric oxide-, peroxynitrite- and S-nitrosothiol-induced inhibition of mitochondrial respiration or complex I activity by light and thiols. Biochim Biophys Acta 1459(2–3):405–412
Bose A, Beal MF (2016) Mitochondrial dysfunction in Parkinson’s disease. J Neurochem 139(Suppl 1):216–231
Braak H, Ghebremedhin E, Rüb U, Bratzke H, Del Tredici K (2004) Stages in the development of Parkinson’s disease-related pathology. Cell Tissue Res 318(1):121–134
Brand MD (2010) The sites and topology of mitochondrial superoxide production. Exp Gerontol. 45(7–8):466–472
Brandt U (1997) Proton-translocation by membrane-bound NADH:ubiquinone-oxidoreductase (complex I) through redox-gated ligand conduction. Biochim Biophys Acta 1318(1–2):79–91
Braun RJ (2012) Mitochondrion-mediated cell death: dissecting yeast apoptosis for a better understanding of neurodegeneration. Front Oncol 2:182
Breckenridge CB, Berry C, Chang ET, Sielken RL, Mandel JS (2016) Association between Parkinson’s disease and cigarette smoking, rural living, well-water consumption, farming and pesticide use: systematic review and meta-analysis. PLoS One 11(4):e0151841
Brichta L, Greengard P (2014) Molecular determinants of selective dopaminergic vulnerability in Parkinson’s disease: an update. Front Neuroanat 8:152
Brini M, Calì T, Ottolini D, Carafoli E (2014) Neuronal calcium signaling: function and dysfunction. Cell Mol Life Sci 71(15):2787–2814
Brinkley BR, Barham SS, Barranco SC, Fuller GM (1974) Rotenone inhibition of spindle microtubule assembly in mammalian cells. Exp Cell Res 85(1):41–46
Brochard V, Combadière B, Prigent A, Laouar Y, Perrin A, Beray-Berthat V, Bonduelle O, Alvarez-Fischer D, Callebert J, Launay JM, Duyckaerts C, Flavell RA, Hirsch EC, Hunot S (2009) Infiltration of CD4+ lymphocytes into the brain contributes to neurodegeneration in a mouse model of Parkinson disease. J Clin Investig 119(1):182–192
Brown GC, Bal-Price A (2003) Inflammatory neurodegeneration mediated by nitric oxide, glutamate, and mitochondria. Mol Neurobiol 27(3):325–355
Brownell AL, Jenkins BG, Elmaleh DR, Deacon TW, Spealman RD, Isacson O (1998) Combined PET/MRS brain studies show dynamic and long-term physiological changes in a primate model of Parkinson disease. Nat Med 4(11):1308–1312
Burwell LS, Nadtochiy SM, Tompkins AJ, Young S, Brookes PS (2006) Direct evidence for S-nitrosation of mitochondrial complex I. Biochem J 394(Pt 3):627–634
Butterfield DA, Kanski J (2001) Brain protein oxidation in age-related neurodegenerative disorders that are associated with aggregated proteins. Mech Ageing Dev 122(9):945–962
Calì T, Ottolini D, Brini M (2014) Calcium signaling in Parkinson’s disease. Cell Tissue Res 357(2):439–454
Cannon JR, Tapias V, Na HM, Honick AS, Drolet RE, Greenamyre JT (2009) A highly reproducible rotenone model of Parkinson’s disease. Neurobiol Dis 34(2):279–290
Castaño A, Herrera AJ, Cano J, Machado A (2002) The degenerative effect of a single intranigral injection of LPS on the dopaminergic system is prevented by dexamethasone, and not mimicked by rh-TNF-alpha, IL-1beta and IFN-gamma. J Neurochem 81(1):150–157
Chan P, DeLanney LE, Irwin I, Langston JW, Di Monte D (1991) Rapid ATP loss caused by 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine in mouse brain. J Neurochem 57(1):348–351
Chan CS, Guzman JN, Ilijic E, Mercer JN, Rick C, Tkatch T, Meredith GE, Surmeier DJ (2007) ‘Rejuvenation’ protects neurons in mouse models of Parkinson’s disease. Nature 447(7148):1081–1086
Chang DT, Honick AS, Reynolds IJ (2006) Mitochondrial trafficking to synapses in cultured primary cortical neurons. J Neurosci 26(26):7035–7045
Chao Y, Wong SC, Tan EK (2014) Evidence of inflammatory system involvement in Parkinson’s disease. Biomed Res Int 2014:308654
Chapman GA, Moores K, Harrison D, Campbell CA, Stewart BR, Strijbos PJ (2000) Fractalkine cleavage from neuronal membranes represents an acute event in the inflammatory response to excitotoxic brain damage. J Neurosci 20(15):RC87
Charan RA, Johnson BN, Zaganelli S, Nardozzi JD, LaVoie MJ (2014) Inhibition of apoptotic Bax translocation to the mitochondria is a central function of parkin. Cell Death Dis 5:e1313
Chartier-Harlin MC, Kachergus J, Roumier C, Mouroux V, Douay X, Lincoln S, Levecque C, Larvor L, Andrieux J, Hulihan M, Waucquier N, Defebvre L, Amouyel P, Farrer M, Destée A (2004) Alpha-synuclein locus duplication as a cause of familial Parkinson’s disease. Lancet 364(9440):1167–1169
Chen H, Zhang SM, Hernán MA, Schwarzschild MA, Willett WC, Colditz GA, Speizer FE, Ascherio A (2003) Nonsteroidal anti-inflammatory drugs and the risk of Parkinson disease. Arch Neurol 60(8):1059–1064
Chen H, Jacobs E, Schwarzschild MA, McCullough ML, Calle EE, Thun MJ, Ascherio A (2005) Nonsteroidal antiinflammatory drug use and the risk for Parkinson’s disease. Ann Neurol 58(6):963–967
Chen L, Jin J, Davis J, Zhou Y, Wang Y, Liu J, Lockhart PJ, Zhang J (2007) Oligomeric alpha-synuclein inhibits tubulin polymerization. Biochem Biophys Res Commun 356(3):548–553
Chen Y, Zhang DQ, Liao Z, Wang B, Gong S, Wang C, Zhang MZ, Wang GH, Cai H, Liao FF, Xu JP (2015) Anti-oxidant polydatin (piceid) protects against substantia nigral motor degeneration in multiple rodent models of Parkinson’s disease. Mol Neurodegener 10:4
Chinta SJ, Andersen JK (2006) Reversible inhibition of mitochondrial complex I activity following chronic dopaminergic glutathione depletion in vitro: implications for Parkinson’s disease. Free Radic Biol Med 41(9):1442–1448
Chinta SJ, Mallajosyula JK, Rane A, Andersen JK (2010) Mitochondrial α-synuclein accumulation impairs complex I function in dopaminergic neurons and results in increased mitophagy in vivo. Neurosci Lett 486(3):235–239
Choi WS, Kruse SE, Palmiter RD, Xia Z (2008) Mitochondrial complex I inhibition is not required for dopminergic neuron death induced by rotenone, MPP+, or paraquat. Proc Natl Acad Sci USA 105(39):15136–15141
Choi BS, Kim H, Lee HJ, Sapkota K, Park SE, Kim S, Kim SJ (2014) Celastrol from ‘Thunder God Vine’ protects SH-SY5Y cells through the preservation of mitochondrial function and inhibition of p38 MAPK in a rotenone model of Parkinson’s disease. Neurochem Res 39(1):84–96
Chou AP, Li S, Fitzmaurice AG, Bronstein JM (2010) Mechanisms of rotenone-induced proteasome inhibition. Neurotoxicology 31(4):367–372
Chu CT, Ji J, Dagda RK, Jiang JF, Tyurina YY, Kapralov AA, Tyurin VA, Yanamala N, Shrivastava IH, Mohammadyani D, Qiang Wang KZ, Zhu J, Klein-Seetharaman J, Balasubramanian K, Amoscato AA, Borisenko G, Huang Z, Gusdon AM, Cheikhi A, Steer EK, Wang R, Baty C, Watkins S, Bahar I, Bayır H, Kagan VE (2013) Cardiolipin externalization to the outer mitochondrial membrane acts as an elimination signal for mitophagy in neuronal cells. Nat Cell Biol 15(10):1197–1205
Ciapaite J, Van Eikenhorst G, Bakker SJ, Diamant M, Heine RJ, Wagner MJ, Westerhoff HV, Krab K (2005) Modular kinetic analysis of the adenine nucleotide translocator-mediated effects of palmitoyl-CoA on the oxidative phosphorylation in isolated rat liver mitochondria. Diabetes 54(4):944–951
Ciechanover A (1998) The ubiquitin-proteasome pathway: on protein death and cell life. EMBO J 17(24):7151–7160
Ciechanover A, Brundin P (2003) The ubiquitin proteasome system in neurodegenerative diseases: sometimes the chicken, sometimes the egg. Neuron 40(2):427–446
Clarke CE, Sambrook MA, Mitchell IJ, Crossman AR (1987) Levodopa-induced dyskinesia and response fluctuations in primates rendered parkinsonian with 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP). J Neurol Sci 78(3):273–280
Claycomb KI, Johnson KM, Winokur PN, Sacino AV, Crocker SJ (2013) Astrocyte regulation of CNS inflammation and remyelination. Brain Sci 3(3):1109–1127
Cleeter MW, Cooper JM, Schapira AH (1992) Irreversible inhibition of mitochondrial complex I by 1-methyl-4-phenylpyridinium: evidence for free radical involvement. J Neurochem 58(2):786–789
Correia SC, Santos RX, Perry G, Zhu X, Moreira PI, Smith MA (2012) Mitochondrial importance in Alzheimer’s, Huntington’s and Parkinson’s diseases. Adv Exp Med Biol 724:205–221
Crawley JN (1999) Behavioral phenotyping of transgenic and knockout mice: experimental design and evaluation of general health, sensory functions, motor abilities, and specific behavioral tests. Brain Res 835(1):18–26
Cuervo AM (2004) Autophagy: many paths to the same end. Mol Cell Biochem 263(1–2):55–72
Cunha MP, Martín-de-Saavedra MD, Romero A, Egea J, Ludka FK, Tasca CI, Farina M, Rodrigues AL, López MG (2014) Both creatine and its product phosphocreatine reduce oxidative stress and afford neuroprotection in an in vitro Parkinson’s model. ASN Neuro 6(6)
Dahm CC, Moore K, Murphy MP (2006) Persistent S-nitrosation of complex I and other mitochondrial membrane proteins by S-nitrosothiols but not nitric oxide or peroxynitrite: implications for the interaction of nitric oxide with mitochondria. J Biol Chem 281(15):10056–10065
Daiber A, Schildknecht S, Müller J, Kamuf J, Bachschmid MM, Ullrich V (2009) Chemical model systems for cellular nitros(yl)ation reactions. Free Radic Biol Med 47(4):458–467
Dauer W, Przedborski S (2003) Parkinson’s disease: mechanisms and models. Neuron 39(6):889–909
Dauer W, Kholodilov N, Vila M, Trillat AC, Goodchild R, Larsen KE, Staal R, Tieu K, Schmitz Y, Yuan CA, Rocha M, Jackson-Lewis V, Hersch S, Sulzer D, Przedborski S, Burke R, Hen R (2002) Resistance of alpha -synuclein null mice to the parkinsonian neurotoxin MPTP. Proc Natl Acad Sci USA 99(22):14524–14529
Davalos D, Grutzendler J, Yang G, Kim JV, Zuo Y, Jung S, Littman DR, Dustin ML, Gan WB (2005) ATP mediates rapid microglial response to local brain injury in vivo. Nat Neurosci 8(6):752–758
Davies KJ (2001) Degradation of oxidized proteins by the 20S proteasome. Biochimie. 83(3–4):301–310
De Stefani D, Raffaello A, Teardo E, Szabò I, Rizzuto R (2011) A forty-kilodalton protein of the inner membrane is the mitochondrial calcium uniporter. Nature 476(7360):336–340
Decressac M, Björklund A (2013) TFEB: Pathogenic role and therapeutic target in Parkinson disease. Autophagy 9(8):1244–1246
Decressac M, Kadkhodaei B, Mattsson B, Laguna A, Perlmann T, Björklund A (2012) α-Synuclein-induced down-regulation of Nurr1 disrupts GDNF signaling in nigral dopamine neurons. Sci Transl Med 4(163):163ra156
Decressac M, Mattsson B, Weikop P, Lundblad M, Jakobsson J, Björklund A (2013) TFEB-mediated autophagy rescues midbrain dopamine neurons from α-synuclein toxicity. Proc Natl Acad Sci USA 110(19):E1817–E1826
Degli Esposti M (1998) Inhibitors of NADH-ubiquinone reductase: an overview. Biochim Biophys Acta 1364(2):222–235
Degli Esposti M, Ghelli A (1994) The mechanism of proton and electron transport in mitochondrial complex I. Biochim Biophys Acta 1187(2):116–120
Degli Esposti M, Ghelli A, Crimi M, Estornell E, Fato R, Lenaz G (1993) Complex I and complex III of mitochondria have common inhibitors acting as ubiquinone antagonists. Biochem Biophys Res Commun 190(3):1090–1096
Degli Esposti M, Ghelli A, Ratta M, Cortes D, Estornell E (1994) Natural substances (acetogenins) from the family Annonaceae are powerful inhibitors of mitochondrial NADH dehydrogenase (Complex I). Biochem J 301(Pt 1):161–167
Degli Esposti M, Ngo A, McMullen GL, Ghelli A, Sparla F, Benelli B, Ratta M, Linnane AW (1996) The specificity of mitochondrial complex I for ubiquinones. Biochem J. 313(Pt 1):327–334
Dehay B, Bové J, Rodríguez-Muela N, Perier C, Recasens A, Boya P, Vila M (2010) Pathogenic lysosomal depletion in Parkinson’s disease. J Neurosci 30(37):12535–12544
Demasi M, Shringarpure R, Davies KJ (2001) Glutathiolation of the proteasome is enhanced by proteolytic inhibitors. Arch Biochem Biophys 389(2):254–263
Demasi M, Silva GM, Netto LE (2003) 20 S proteasome from Saccharomyces cerevisiae is responsive to redox modifications and is S-glutathionylated. J Biol Chem 278(1):679–685
Devi L, Raghavendran V, Prabhu BM, Avadhani NG, Anandatheerthavarada HK (2008) Mitochondrial import and accumulation of alpha-synuclein impair complex I in human dopaminergic neuronal cultures and Parkinson disease brain. J Biol Chem 283(14):9089–9100
Ding Q, Dimayuga E, Martin S, Bruce-Keller AJ, Nukala V, Cuervo AM, Keller JN (2003) Characterization of chronic low-level proteasome inhibition on neural homeostasis. J Neurochem 86(2):489–497
Dong Y, Benveniste EN (2001) Immune function of astrocytes. Glia 36(2):180–190
Drolet RE, Behrouz B, Lookingland KJ, Goudreau JL (2004) Mice lacking alpha-synuclein have an attenuated loss of striatal dopamine following prolonged chronic MPTP administration. Neurotoxicology 25(5):761–769
Du Y, Ma Z, Lin S, Dodel RC, Gao F, Bales KR, Triarhou LC, Chernet E, Perry KW, Nelson DL, Luecke S, Phebus LA, Bymaster FP, Paul SM (2001) Minocycline prevents nigrostriatal dopaminergic neurodegeneration in the MPTP model of Parkinson’s disease. Proc Natl Acad Sci USA 98(25):14669–14674
Dukes AA, Bai Q, Van Laar VS, Zhou Y, Ilin V, David CN, Agim ZS, Bonkowsky JL, Cannon JR, Watkins SC, Croix CM, Burton EA, Berman SB (2016) Live imaging of mitochondrial dynamics in CNS dopaminergic neurons in vivo demonstrates early reversal of mitochondrial transport following MPP(+) exposure. Neurobiol Dis. 95:238–249
Dunning CJ, McKenzie M, Sugiana C, Lazarou M, Silke J, Connelly A, Fletcher JM, Kirby DM, Thorburn DR, Ryan MT (2007) Human CIA30 is involved in the early assembly of mitochondrial complex I and mutations in its gene cause disease. EMBO J 26(13):3227–3237
Earle KM (1968) Studies on Parkinson’s disease including X-ray fluorescent spectroscopy of formalin fixed brain tissue. J Neuropathol Exp Neurol 27(1):1–14
Ebrahimi-Fakhari D, Wahlster L (2013) Restoring impaired protein metabolism in Parkinson’s disease—TFEB-mediated autophagy as a novel therapeutic target. Mov Disord 28(10):1346
Efremova L, Schildknecht S, Adam M, Pape R, Gutbier S, Hanf B, Bürkle A, Leist M (2015) Prevention of the degeneration of human dopaminergic neurons in an astrocyte co-culture system allowing endogenous drug metabolism. Br J Pharmacol 172(16):4119–4132
Ehringer H, Hornykiewicz OHRINGERH, HORNYKIEWICZ O (1960) Distribution of noradrenaline and dopamine (3-hydroxytyramine) in the human brain and their behavior in diseases of the extrapyramidal system. Klin Wochenschr 38:1236–1239
Ekstrand MI, Terzioglu M, Galter D, Zhu S, Hofstetter C, Lindqvist E, Thams S, Bergstrand A, Hansson FS, Trifunovic A, Hoffer B, Cullheim S, Mohammed AH, Olson L, Larsson NG (2007) Progressive parkinsonism in mice with respiratory-chain-deficient dopamine neurons. Proc Natl Acad Sci USA 104(4):1325–1330
Emmanouilidou E, Melachroinou K, Roumeliotis T, Garbis SD, Ntzouni M, Margaritis LH, Stefanis L, Vekrellis K (2010) Cell-produced alpha-synuclein is secreted in a calcium-dependent manner by exosomes and impacts neuronal survival. J Neurosci 30(20):6838–6851
Eng LF, Ghirnikar RS, Lee YL (2000) Glial fibrillary acidic protein: GFAP-thirty-one years (1969–2000). Neurochem Res 25(9–10):1439–1451