Astrocytes in Parkinson’s disease and DJ-1

Since the German anatomist Rudolf Virchow first coined the term neuroglia in 1846, our ideas about glial cells have moved from viewing them simply as the glue that holds neurons in place. We now know that glia are intimately involved in the lives of neurons, with functions ranging from nurturing developing synapses to actually facilitating the transmission of neuronal information (Eroglu & Barres 2010). Furthermore, glial cells may play detrimental roles in the pathogenesis of nervous system diseases. As an example of this, work from Hinkle and colleagues in this issue of the Journal of Neurochemistry highlight a potential role of astrocytes in Parkinson’s disease associated with mutations in DJ-1 (Mullett & Hinkle 2011).

Parkinson’s disease (PD) has been considered a disease of neuronal origin due to the degeneration of groups of neurons, including dopaminergic cells in the substantia nigra pars compacta (SNpc). Mitochondrial complex I dysfunction, associated with accumulation of reactive oxygen species, has been suggested to be important in the loss of neurons in PD. This hypothesis is supported by multiple lines of evidence including the observations that; (1) post mortem biochemical studies have demonstrated lower complex I activity in the SNpc of PD brains compared to controls (Schapira 2007); (2) inhibition of complex I with rotenone or MPTP reproduces the pathological degeneration of the SNpc in mammalian models (Cannon & Greenamyre 2010); and (3) the protein products of genes known to cause familial forms of parkinsonism have clear roles in the maintenance of functional mitochondria (Thomas et al. 2011).

A rare form of autosomal recessive parkinsonism is caused by mutations or deletions of the gene coding for the protein DJ-1 (Bonifati et al. 2003). DJ-1 is known to have a role in oxidative stress responses in cells and this is thought to be important in the preservation of neuronal viability. Some of the mechanisms involved are known. For example, the oxidation of a specific cysteine residue in DJ-1 is required to protect cells against mitochondrial damage (Blackinton et al. 2009). However, previous studies have focused on cell-autonomous effects of DJ-1. The work presented here by the Hinkle group further supports their previous findings which suggest a novel mode of action for DJ-1 in which it protects neurons via an astrocyte-mediated mechanism (Mullett & Hinkle 2009).

Using a new method of detecting neuronal death in astrocyte/neuron co-cultures, the authors demonstrate that the down-regulation of astrocytic DJ-1 diminishes their ability to protect neurons from stressors. In particular, they show that neuroprotection is selective for drugs that inhibit mitochondrial complex I. However, the authors also found that the neuronal death in their DJ-1 deficient co-cultures treated with these inhibitors could not be rescued by the addition of antioxidants. They also demonstrate that the mechanism underlying astrocyte-mediated protection is not due to the release of the antioxidant glutathione or an up-regulation of astrocytic heme oxygenase-1. Taken together, these findings suggest that astrocytic DJ-1 is necessary for astrocyte-mediated neuroprotection and that the mechanism underlying this phenomenon is selective for complex I inhibitors and is not mediated by an oxidative stress response.

The notion that dysfunctional astrocytes may contribute to the death of dopaminergic neurons in the SNpc has several consequences on how we view the pathogenesis of PD and possible treatments. Predominantly, these results say that even where there is a single gene cause of parkinsonism, we should consider models that involve more than just dopamine neurons. In fact, we should consider astrocytes and other non-neuronal cells as possible contributors to the early pathogenic mechanisms. A smaller, but also important point, is that when considering ways in which to treat PD we may need to consider whether the glial environment is supportive or not; treating only neurons may provide only part of the answer.

Although the findings presented in this issue by the Hinkle group convincingly show a role for astrocytic DJ-1 in astrocyte-mediated neuroprotection, there are a number of areas that would merit further investigation. If DJ-1 is fundamentally an oxidative stress response protein, then is the neuroprotection also related to oxidative stress? The data presented by Mullet and Hinkle suggests that the effects are not due to, for example, antioxidant release from astrocytes, which may have been an obvious link. The authors propose that soluble factors are released from astrocytes, but these still need to be identified before more detailed mechanistic studies are undertaken. Furthermore, most of the current experiments use dissociated cultured cells. It is known that astrocytes express DJ-1 in vivo, and so it would seem important to determine whether astrocytic DJ-1 is important for neuroprotection in the intact brain.

Overall, the data presented here identifies an exciting possibility, that neurodegeneration in a defined form of genetic parkinsonism may involve non-cell autonomous events. Identifying the underlying mechanisms is an important next step and can be facilitated by the types of models presented by Mullet and Hinkle.