Accumulating evidence suggests that some combination of environmental toxins ranging from infectious insults to heavy metals and pesticides contribute to PD [21, 31–35]. Some authors have even entertained the notion that a latent viral infection early in life may contribute to the onset of the disease . Indeed, post-encephalitic cases of parkinsonism have been reported to occur years after viral infection [21, 22]. It may be that exposure to immunological challenges (viral or bacterial) provoke modest neuroinflammation (for example, microglial activation, cytokine release) that over time may either: 1) cause frank neurodegeneration or 2) render DA neurons vulnerable to degeneration in response to subsequent environmental insults. With regard to this second possibility, we found that the double-stranded RNA viral analog poly(I:C) sensitized mice to the deleterious effects of paraquat exposure.
One of the most consistent findings across human post-mortem and animal toxin models of PD has been the association between strongly activated inflammatory microglia and DA neurodegeneration [8, 37–41]. Emerging data also show that pro-inflammatory cytokines, particularly interferon (IFN)-γ and tumor necrosis factor (TNF)-α, modulate the microglial response to PD relevant environmental toxins [42–44]. In this study, we found that that poly(I:C) infusion alone caused modest but non-significant effects upon morphological status (using CD11b staining) and oxidative potential (assessed by increased gp91 staining) of microglia. This finding concurs with a previous reports indicating that a single poly(I:C) infusion (albeit of higher concentration) led to a decrease of TH-positive neurons in the SNc . However, in the current study, the most dramatic effects were seen within the context of paraquat exposure, with poly(I:C) pre-treatment greatly increasing the degree of DA neuronal loss and microglial activation in response to later paraquat exposure.
In contrast to our previous findings using LPS priming , the present results indicated that timing between poly(I:C) priming and subsequent paraquat administration did not seem to be particularly important to the appearance of a sensitized response. In fact, poly(I:C) induced a protracted activation of microglia within the SNc that was still evident after 14 days, whereas in our previous study , we found that the effects of LPS upon CD11b-positive microglia were more transient (evident after 2 days, but returning towards baseline by 7 days). Although the effects of LPS and poly(I:C) were temporally distinct, the most important finding was that the augmented SNc DA neuronal loss was evident when exposure to the paraquat regimen occurred at a time of heightened microglial reactivity. However, it is noteworthy that the morphological changes of microglial (as indicated by CD11b staining) appeared to be most profound (although transient) with LPS priming (involving very compact amoeboid-like cells; ), compared with the more intermediate (but protracted) state induced by poly(I:C).
The differences in the temporal pattern of microglial immunoreactivity induced by poly(I:C) compared with previous reports using LPS might stem from variations in intracellular TLR-linked proteins. Indeed, inhibition of the TIRAP/MyD88 signaling proteins has been shown to block TLR4 but not TLR3 signaling . Moreover, bone-marrow-derived macrophages treated with a TLR3 agonist displayed an enhanced and sustained immune response (that is, induction of type-1 IFN gene expression), relative to macrophages treated with a TLR4 agonist . In addition, cultured human microglia were found to produce higher levels of some pro-inflammatory cytokines when treated with poly(I:C) compared with LPS .
Whatever the mechanism underlying TLR-linked pathways, oxidative stress factors are undoubtedly linked to PD-like pathology occurring after toxin exposure. Indeed, alterations in brain iron content, impaired mitochondrial function, and antioxidant protective systems, together with oxidative damage to lipids, proteins, and DNA, are typically seen in post-mortem brains of patients with PD and of toxin-exposed animals . Post-mortem analysis of PD brains also showed upregulation of gp91 within the SNc [48, 49]. Importantly, gp91 is the inducible catalytic subunit of the microglial NADPH oxidase enzyme, which in turn is responsible for the generation of the potentially damaging superoxide radical [50, 51]. The fact that the most dramatic gp91 elevation seen in the present study occurred in poly(I:C) primed mice that also received paraquat suggests that the oxidative effect of paraquat is augmented in the context of an inflammatory microenvironment. This finding is consistent with the importance of gp91 shown in previous rodent studies that revealed attenuated degenerative effects of paraquat after genetic or pharmacological ablation of the NADPH oxidase subunit [48, 52–54].
Although in our study, the most potent effects of paraquat were restricted to TH-positive SNc neurons, it should be mentioned that some TH-negative nigral neurons were also affected. These TH-negative cells probably represent a γ-Aminobutyric acid (GABA)ergic neuronal population . This observation suggests that paraquat-induced neurodegeneration may not be entirely specific to DA neurons. It is possible that, given the marked pro-inflammatory and pro-oxidative state we observed in the SNc, non-DAergic neurons can succumb to such a volatile microenvironment. Indeed, LPS infusion alone was previously reported to decrease TH-negative neurons in the rat SNc , reinforcing the idea that strongly reactive microglia are sufficient to induce non-DAergic neuron death. It is conceivable that the transient disruption of the blood brain barrier expected after the surgical cannula implantation procedures could have augmented the overall entry of paraquat into the brain, and hence contributed to a non-specific loss of non-TH neurons.
Although paraquat has been reported to induce dose-dependent loss of SNc DA neurons , discrepant reports exist about the degree of degeneration in striatal terminals, with some studies reporting no effects of the pesticide [55, 57]. In the present investigation, we found that both paraquat and poly(I:C), alone or in combination, did not affect DA terminals in the striatum. In agreement with this observation was the general lack of behavioral effects. In fact, there was only a modest transient but non-significant reduction of home-cage activity, which occurred 22 days after the poly(I:C) infusion. Furthermore, a pole test, which is designed to tap into forelimb and hindlimb deficits, failed to reveal any signs of coordination deficits.
Given the unexpected lack of striatal or behavioral deficits in the face of SNc soma loss, it was of interest to determine whether compensatory processes might have been provoked by poly(I:C) and paraquat. Indeed, the existence of neural stem cells and progenitors in the SVZ  raises the possibility that new neurons could be recruited into the nearby striatum after insult, as has been reported after ischemic brain injury [59–61]. Less evidence exists concerning the possibility of increased neurogenesis in a PD model; however, it has been reported that -methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-treated mice exhibited increased neurogenesis . In the current study, both poly(I:C) and paraquat treatment failed to influence the number of DCX-positive immature neurons counted within the SVZ, and there was no evidence of new neuronal infiltration into the striatum. This negative finding might stem from the fact that, unlike another neurotoxin, MPTP, paraquat is not believed to be taken up by DA transporters at the terminals . Alternatively, the magnitude of the SNc soma lesion might simply have been insufficient to affect processes at the downstream terminals.