1. 6-OHDA
Notably, 6-OHDA was the first agent discovered with specific toxicity towards catecholaminergic neurons
[1][2]; thus, it causes selective degeneration of SNpc DAergic neurons and Parkinsonian motor impairments when injected within the substantia nigra, medial forebrain bundle (MFB), or striatum
[3][4]. The discovery that 6-OHDA is a naturally occurring compound in the human brain by dopamine hydroxylation
[5] and it has been found in PD patients’ urine
[6] strengthens the importance of understanding the mechanisms of 6-OHDA toxicity. Thus, 6-OHDA may represent an endogenous toxin contributing to the neurodegeneration of SNpc DAergic neurons in the human brain. When used to produce a PD-like phenotype in an animal model, 6-OHDA-induced death of DAergic neurons is fast, but in some cases, it continues for months after a single toxin brain infusion
[7]. The modifications induced by 6-OHDA in surviving SNpc DAergic neurons reported in the literature are pretty scattered, as they were obtained in different experimental conditions regarding the site of toxin injection (SNpc, MFB, striatum), animal species (guinea pig, rat, mouse) and experimental approach (in vivo, ex vivo brain slices, cultured DAergic neurons).
OurThe group and others contributed to understanding 6-OHDA toxicity with experiments consisting of 6-OHDA acutely applied onto SNpc DAergic neurons in midbrain slices. The effects of 6-OHDA, evaluated using electrophysiological recordings combined with Ca
2+ imaging, were remarkably fast and essentially permanent up to 30 min after removal of the toxin, and consisted of a reduction of spontaneous firing activity, partly due to K
ATP and D
2 receptor-gated G-protein-coupled inward rectifier K
+ channel (GIRK) activation, and a slow-onset Ca
2+ accumulation, most likely of mitochondrial origin, as it was independent of the removal of calcium ions from the extracellular milieu and emptying endoplasmic reticulum
[8]. Similar results were obtained by Qu and colleagues
[9]. Another investigation in organotypic slice culture reported that exposure of SNpc DAergic neurons to 6-OHDA for 12–18 h induced changes in their firing pattern from pacemaking to irregular bursting due to enhanced Ca
2+ influx and increased level and activity of a critical Ca
2+ modulator protein phosphatase 2A (PP2A), known to alter the Ca
2+-sensitivity of small-conductance Ca
2+-activated K
+ channels (SK) (increased AHP) by dephosphorylating SK-associated calmodulin (CaM). The increased activity of the SK channels is thought to have a protective effect because it reduces cellular damage
[10]. Different results were reported when the electrophysiological properties of SNpc DAergic neurons were investigated in brain tissue (midbrain slices or single DAergic neurons) obtained from partially lesioned animals with 6-OHDA injection in vivo. Electrophysiological recordings in isolated TH
+ neurons performed 1–8 weeks after in vivo 6-OHDA lesion (into the SNpc) in mice revealed that SNpc DAergic neurons in the lesioned side do not undergo maturational changes of resting membrane potential, membrane resistance, steady-state membrane currents, and action potential half-width from 4 to 14 weeks of mouse age, typical of non-lesioned DAergic neurons. Additionally, 6-OHDA SNpc DAergic neurons displayed reduced membrane capacitance, which is indicative of neuron atrophy induced by the toxin. In this lesioned mouse model, the main 6-OHDA effect, other than affecting the maturation of electrical membrane properties of SNpc DAergic neurons, was inhibition of TH expression resulting in an increased proportion of TH
− vs. TH
+ neurons
[11].
Other studies investigated 6-OHDA effects on SNpc DAergic neuron electrophysiological properties in vivo, utilizing extracellular single-unit recordings. Four weeks after hemilateral 6-OHDA infusion into the MFB to produce a partial lesion in rats, the remaining SNpc DAergic neurons in the injected side (about a 40% TH
+ neuron loss) displayed a clear-cut shift toward hyperexcitability, with an increase in firing rate, number of bursts, the mean number of single spikes/bursts and of the percentage of burst-firing neurons
[12]. In this report, hyperexcitability is reversed by inhibitors of metabotropic glutamate receptors (mGluRs), suggesting that excessive glutamate release contributes to 6-OHDA toxicity. Another study performed in vivo electrophysiology on SNpc DAergic neurons 2 weeks after hemilateral MFB 6-OHDA lesion and reported a reduced number of spontaneously firing neurons. Still, the remaining active neurons fired at a similar frequency to that of non-lesioned animals
[13]. However, the coefficient of variation (CV), a measure of interspike interval (ISI) regularity, was significantly increased in the surviving DAergic neurons, suggesting that 6-OHDA partial lesions result in changes in neuronal excitability, with effects that are limited to the regularity of the firing, which may precede more pronounced hyperexcitability affecting the firing rate at advanced stages of the disease with large striatal dopamine denervation. Recently, one study performed either in vivo single-unit extracellular or ex vivo cell-attached recordings in acute midbrain slices from 6-OHDA lesioned rats (hemilateral MFB,
[14]). These authors reported that, in lesioned animals, SNpc DAergic neurons in vivo display a 76% decrease in firing frequency. Interestingly, the decreased firing was restored by GABA
A receptor antagonists or MAO-B inhibitors, and it was suggested that reactive astrocytes synthesize and release a high amount of GABA in PD models, which inhibits SNpc DAergic neuron firing (see below in the MPTP section).
Collectively, 6-OHDA affects several channels/receptors in SNpc DAergic neurons, including KATP and D2-GIRK channel activation, increasing the gating of SK channels, preventing maturation of membrane input resistance, and promoting an excessive release of glutamate and GABA. These effects result in changes in neuronal excitability, disturbance of intracellular ion homeostasis, neuronal atrophy, and mitochondrial dysfunction, as summarized in Table 1.
Table 1. Summary of the principal findings of the cited literature on 6-OHDA effects on SNpc DAergic neuron functional properties.
Type of Study |
[6-OHDA] |
Treatment |
Modified Parameters in SNpc DAergic Neuron |
Molecular Mechanisms |
Reference |
Ex vivo, rat |
0.2, 0.5, 1, 2 (mM) |
5 or 10 min |
Inhibition of spontaneous firing; Rm drop; Ca2+ accumulation |
D2-GIRK and KATP channels activation; mitochondrial release of Ca2+ ions |
[8] |
Ex vivo, rat |
0.5, 1, 2 (mM) |
3–5 min |
Inhibition of spontaneous firing; Ca2+ accumulation |
N-type VGCC current amplitude increase |
[9] |
In vitro organotypic culture, rat |
25 µM |
12 or 18 h |
Irregular firing/bursting; depolarized RMP |
Increased AHP and IAHP mediated by SK channels |
[10] |
In vivo, mouse |
1.5 µg/µL (1.6 µL) |
1 injection, SNpc |
1 to 8 weeks after lesion; Lack of maturation of Rm, AP half-width, steady-state I(-100mV) |
|
[11] |
In vivo, rat |
4 µg/4 µL |
1 injection, MFB; tested 16–20 days after lesion |
Increase in firing rate, n. of bursting neurons and n. spikes/burst |
Release of glutamate and mGluR activation (rescue by MPEP) |
[12] |
In vivo, rat |
4 µg/2 µL |
1 injection, MFB, 4–6 weeks after lesion |
Decreased n. of active neurons; no significant difference in firing rate nor bursting; higher CV |
Rearrangements of circuitry to compensate for neuronal loss |
[13] |
In vivo, rat |
8 µg/4 µL |
1 injection, MFB |
32 days after lesion, 76% reduction in firing rate |
Excessive GABA release by reactive astrocytes, rescued by MAO inhibitor safinamide |
[14] |
Ex vivo |
Ipsilateral slices from in vivo lesioned rat |
|
Increase tonic GABAA current; no difference in sIPSC amplitude or frequency |
Rescued by bicuculline and safinamide |
[14] |
2. Rotenone, Paraquat and BMAA
Due to its lipophilic nature, rotenone, a naturally occurring isoflavonoid from tropical plants
[15], is membrane permeable and capable of entering all neuron types, where it inhibits complex I of the mitochondria respiratory chain
[16]. Despite being an unselective compound, chronic systemic exposure to rotenone in animals has been shown to reproduce some of the key features of PD, including selective degeneration of TH
+, DAT
+ and VMAT2
+ neurons and the formation of α-synuclein-containing intracellular inclusions in nigral DAergic cells
[17]. The acute effects of rotenone on SNpc DAergic neurons in midbrain slices have been investigated by numerous groups. One of the main functional impairments caused by the toxin is rapid inhibition of the spontaneous firing and membrane potential hyperpolarization
[18][19][20][21]. These effects are associated with mitochondrial depolarization and ROS production, a decrease in membrane input resistance, indicative of K
ATP and transient receptor potential M2 (TRPM2) channels opening, and the accumulation of Ca
2+ and Na
+ ions. Additionally, the toxin causes a fast drop in cell capacitance, indicative of damage and decline of the cell surface area
[20]. With regards to the mechanisms of K
ATP channel activation, other studies revealed that rotenone-induced K
ATP channel opening is mainly dependent on ROS production rather than on ATP drop
[22] since a ROS scavenger (superoxide dismutase mimetic) prevented rotenone-induced K
ATP channel activation and membrane hyperpolarization of SNpc DAergic neurons. Interestingly, not all SNpc DAergic neurons display the same sensitivity to rotenone. This feature is due to the diverse composition of K
ATP channel subunits in different SNpc DAergic neuron populations, with some K
ATP channels demonstrating high (SUR1+Kir6.2) or low (SUR2B+Kir6.2) sensitivity to metabolic inhibition by rotenone
[19]. Rotenone also induces detrimental effects on SNpc DAergic neuron excitability via excitotoxic pathways. These include the potentiation of NMDA receptor-mediated currents
[23] and inhibition of GABA
A receptor-mediated currents
[24]. The rotenone effect on NMDA receptors is indirect, mediated by a protein tyrosine kinase-dependent mechanism
[23], and it reduces the ability of Mg
2+ ions to inhibit NMDA-gated channels
[25]. The potentiation of NMDA currents by rotenone is also mediated by ROS and/or dopamine oxidation products acting on NMDA receptors indirectly via a protein tyrosine kinase-dependent mechanism
[26], indicating that rotenone neurotoxicity may be augmented by dopamine oxidative metabolism.
Rotenone’s effects on SNpc DAergic neuron excitability have also been studied following chronic systemic administration in vivo to rodents and invertebrates at different toxin concentrations and treatment durations. Intraperitoneal administration of rotenone (0.8 mg/kg) for 7 days to mice does not affect the intrinsic excitability of SNpc DAergic neurons or D
2 receptor-activated hyperpolarization by exogenous DA application
[27]. In this model, rotenone caused impairment of striatal LTP and LTD only in animals with a genetic predisposition to PD (heterozygote PINK1
+/− mice)
[27]. Another study showed chronic effects of rotenone exposure on the DAergic system in the snail
Lymnaea stagnalis (0.5 µM, for 7 days). In this model, the hyperpolarizing DAergic response evoked by stimulation of giant DAergic neurons onto post-synaptic VD4 neurons disappeared, indicating that chronic exposure to the toxin impairs DA synaptic transmission
[28].
In contrast with the large number of studies aiming at characterizing the electrophysiological effects of rotenone on SNpc DAergic neurons, only a few studies have described paraquat’s electrophysiological effects on these neurons. To our knowledge, the study published by Lee and colleagues
[29] is the only one performing an electrophysiological analysis of paraquat effects on SNpc DAergic neurons, showing that acute application of the toxin to rat midbrain slices reduced AMPA-mediated currents by acting selectively on post-synaptic AMPA receptors. Indeed, miniature post-synaptic AMPA current amplitude, but not frequency, was reduced by paraquat as well as the amplitude of AMPA-mediated currents by exogenous agonist application.
Finally, the toxic effects of the non-protein amino acid β-
N-methylamino-L-alanine (BMAA). Initially, BMAA was proposed as the Cycad toxic agent causing a rare form of neuronal degeneration, the amyotrophic lateral sclerosis–Parkinson’s dementia complex (ALS-PDC), occurring among the Chamorro people of Guam
[30][31][32]. However, BMAA is present globally and is produced by cyanobacteria and possibly by other organisms
[33][34]. Chronic exposure of non-human primates to BMAA recapitulates the neuropathology as that described in Guamanian people affected by ALS-PDC
[35]. BMAA increased the excitability of SNpc DAergic neurons
[36] by selectively gating mGluRs, increasing neuronal firing, Ca
2+ accumulation, and mitochondria ROS production
[36]. The main rotenone effects on SNpc DAergic neurons are summarized in
Table 2.
Table 2. Summary of the principal findings of the cited literature on rotenone’s effects on SNpc DAergic neuron functional properties.
Type of Study |
[Rotenone] |
Treatment |
Modified Parameters in SNpc DAergic Neuron |
Molecular Mechanisms |
Reference |
In vitro, dissociated SNpc DAergic neurons, rat |
5 µM |
|
Firing inhibition and membrane hyperpolarization |
Activation of the sulphonylurea-sensitive KATP current |
[18] |
Ex vivo midbrain slice, mouse |
10 µM |
10 min |
Firing inhibition and membrane hyperpolarization |
SUR1-Kir6.2 vs. SUR2B-Kir6.2 KATP channels display different sensitivity to metabolic inhibition |
[19] |
Ex vivo midbrain slice, rat |
5 nM; 200 nM; 1 µM |
10 min |
Cm and Rm drop; KATP current activation; Ca2+ and Na+ accumulation; mitochondrial ROS production and Δψm depolarization |
ROS activation of TRPM2 Ca2+-permeable and KATP channels |
[20] |
In vitro, SNpc DAergic neurons acutely dissociated |
1 µM |
5–6 min |
Firing inhibition |
KATP channel opening; they are inhibited by the neuroprotective agent THB |
[21] |
Ex vivo, midbrain slices, mouse |
100 nM |
5 min |
Firing inhibition; KATP channel activation; ROS production |
Kir6.2 subunit KO prevents DAergic neuron degeneration |
[22] |
Ex vivo, midbrain slices, rat |
100 nM |
20–30 min |
Increased INMDA (but not IAMPA) amplitude |
|
[23] |
In vitro, acutely dissociated SNpc DAergic neurons, rat |
5 µM |
10 min |
Run-down of IGABAA, but not of IGly or IGlu |
|
[24] |
Ex vivo midbrain slices, rat |
100 nM |
30 min |
Increased INMDA amplitude |
Loss of Mg2+-block of NMDA-mediated currents that involves a tyrosine kinase |
[25] |
Ex vivo midbrain slices, rats |
100 nM |
30 min |
Increased INMDA amplitude |
ROS and DA oxidation products mediate NMDA currents increase |
[26] |
In vivo, mouse |
0.8 mg/kg |
7 days |
Lack of gross functional alterations in SNpc DAergic neurons |
|
[27] |
In vivo, snail Lymnaea stagnalis |
0.5 µM |
7 days |
Loss of dopaminergic IPSP |
Uncoupling of dopaminergic synapses |
[28] |
Ex vivo midbrain slice, rat |
Paraquat, 30,100 µM |
20 min |
Reduced IAMPA amplitude |
Inhibition of post-synaptic AMPA receptors |
[29] |
Ex vivo midbrain slice, rat |
BMAA (0.1–10 mM) |
2–3 min |
Increased firing; Ca2+ accumulation |
Activation of mGluR and TRPC channels |
[36] |
3. MPTP and MPP+
The discovery of MPTP and its active metabolite MPP
+ as a selective toxin of DAergic neurons was made by a neurologist
[37] who diagnosed an advanced PD in a young ‘synthetic heroin’ user who accidentally consumed MPTP, a compound previously synthesized (1947) but never controlled for toxicity nor commercialized. Since that discovery, thousands of publications have been made to unveil MPTP’s mechanism of toxicity and gain a translational impact from MPTP animal models in proposing an environmental origin of PD. Indeed, although MPTP has never been found in nature, the structural similarity between its active metabolite MPP
+, formed by astrocytic MAO, and paraquat, an herbicide largely used worldwide, strongly suggested an increased risk for developing PD following herbicide/pesticide exposure. Primates are highly sensitive to MPTP as well as humans, whereas rodents initially displayed little toxicity to MPTP. This has been explained by the discovery that rats express high levels of MAO enzymes within the blood-brain barrier (BBB), where virtually all lipophilic MPTP is converted into MPP
+ that conversely, due to its non-lipophilic moiety, does not penetrate the brain
[38][39]. Mice display intermediate levels of BBB-MAO and, as expected, intermediate levels of MPTP toxicity. Regarding its effects on the electrophysiological properties of SNpc DAergic neurons, acute MPP
+ application (10 µM, 5 min) to mouse midbrain slice causes inhibition and subsequent complete cessation of spontaneous firing of these cells
[22]. This inhibition is rather selective for SNpc DAergic neurons, as mesolimbic DAergic neurons are largely unaffected by MPP
+, indicating higher sensitivity of mesostriatal DAergic neurons to the toxin in mice, similar to previous studies on primates
[40]. The MPP
+-mediated hyperpolarization is due to the opening of K
ATP channels containing the Kir6.2 subunit, since the spontaneous firing is not altered by MPP
+ in Kir6.2 knockout mice. The effects of chronic systemic MPTP administration to mice on SNpc DAergic neuron excitability have been recently investigated by Heo and colleagues
[14] either in in vivo or in ex vivo midbrain slices from lesioned animals. Four MPTP injections in mice cause a strong (60%) reduction of the spontaneous pacemaker firing as observed in extracellular, single-unit in vivo recordings. The reduction of firing is significantly rescued by treatment of mice with the MAO-B inhibitor selegiline. Indeed, the authors find that following in vivo MPTP (or 6-OHDA) administration, astrocytes become dramatically reactive, as revealed by increased GFAP expression, and begin to synthesize (via MAO-B) and release GABA, with subsequent reduction of SNpc DAergic neuron firing. In line with their hypothesis, ex vivo patch-clamp recordings from SNpc DAergic neurons in midbrain slices confirmed lower pacemaker firing in MPTP slices as in in vivo recordings, and firing reduction is restored by selegiline or the GABA
A receptor antagonist bicuculline. The authors conclude that astrocytic GABA release inhibits the pacemaker firing of DAergic neurons in PD models
[14]. In line with this evidence, Masi and colleagues
[41] similarly reported that MPP
+ reduces the spontaneous firing of SNpc DAergic neurons when acutely applied to midbrain slices. In addition, they show that MPP
+ directly inhibits the hyperpolarization-activated current (I
h), a hallmark of DAergic neurons
[42] highly sensitive to pathological conditions of the DAergic system
[43][44]. This effect produces an increase in the temporal summation of excitatory inputs to SNpc DAergic neurons, thus increasing spike probability and overall network excitability. Another study described the early and late effects of MPP
+ on SNpc DAergic neuron excitability
[45], confirming the MPP
+-mediated I
h inhibition in Masi et al.
[41]. However, this was not causative of firing inhibition/membrane potential hyperpolarization, as it was not prevented by I
h blockade. The authors reported that the early MPP
+ effect was due to DA vesicle displacement and somatodendritic D
2 receptor activation, whereas late effects depended on K
ATP channel activation. MPTP/MPP
+ effects on SNpc DAergic neurons are summarized in
Table 3.
Table 3. Summary of the principal findings reported by the cited literature on MPP+/MPTP effects on SNpc DAergic neuron functional properties.
Type of Study |
[MPTP/MPP+] |
Treatment |
Modified Parameters in SNpc DAergic Neuron |
Molecular Mechanisms |
Reference |
Ex vivo |
100 nM–10 µM |
5 min |
Spontaneous firing inhibition; KATP activation |
Differential coupling between mitochondrial inhibition and KATP activation in SN vs. VTA neurons. Kir6.2 subunit KO prevents DAergic neuron degeneration |
[22] |
In vivo |
20 mg/kg, i.p., 4 injections in one day |
6 days later |
60% reduction of pacemaker firing |
Excessive GABA release by reactive astrocytes |
[14] |
In vitro, acutely isolated DAergic neurons from in vivo lesioned mouse |
|
|
Decrease in spontaneous firing rate |
Excessive GABA release by reactive astrocytes, (rescue by selegiline and bicuculline); |
[14] |
Ex vivo midbrain slices, rat and mouse |
50 µM |
5–15 min |
Ih inhibition; spontaneous firing inhibition |
The shift of Ih activation curve toward negative potentials |
[41] |
Ex vivo |
20 µM |
30 min |
Spontaneous firing inhibition |
DA vesicle displacement, D2-GIRK activation; Ih inhibition; KATP activation; DAT activation |
[45] |
Although not considered a neurotoxin, L-DOPA, the ‘gold standard’ therapeutic drug still largely used to relieve motor symptoms of PD, has been described to have potentially detrimental effects on different types of neurons, including SNpc DAergic cells. Indeed, when the classical inhibitory effect due to neoformed DA has been blocked, L-DOPA displays a non-conventional excitatory effect linked to its nonenzymatic autoxidation to TOPA quinone, a potent activator of AMPA/kainate receptors
[46][47][48]. Thus, in addition to the autoreceptor-mediated inhibitory effect of L-DOPA, the drug has the potential to excite and increase free intracellular calcium in nigral DAergic neurons, as well as adjacent non-DAergic cells. This particular circumstance could contribute to neuronal damage.
Overall, in most toxin-based PD models, SNpc DAergic neuron excitability is heavily modified. Acute toxin applications to diverse SNpc DAergic neuron preparations mainly reduce/abolish spontaneous firing, either by gating K
ATP channels or increasing GABA release into the midbrain, while in other models, a decrease in firing regularity and potentiation of NMDA
R-dependent currents is reported. In in vivo toxin-lesioned animals, either an increase or decrease in spontaneous firing has been reported, depending on experimental conditions. These contrasting results suggest that SNpc DAergic neurons’ excitability is a highly sensitive parameter often irreversibly modified by most PD toxins. This has great importance for neuronal health since the increased firing activity of the DAergic neurons drives an increase in intracellular calcium via the L-type channels and causes mitochondrial stress favoring synuclein aggregation
[49].