Parkinson’s disease (PD) is most commonly a sporadic disorder, but up to 10% of PD cases are associated with genetic mutations (Van Den Eeden et al., 2003). The main neuropathological finding in PD is extensive loss of dopaminergic neurons whose cell bodies are located in the substantia nigra pars compacta (SNC) and whose axons project to the striatum (Davie, 2008). Increases in lipid, protein, and nucleic acid oxidation have been reported in the PD SNC (Dexter et al., 1994; Yoritaka et al., 1996; Alam et al., 1997; Floor and Wetzel, 1998; Zhang et al., 1999) suggesting that oxidative stress (overproduction of reactive oxygen and nitrogen species relative to their detoxification mechanisms [Wang and Michaelis, 2010]) may play a role in the loss of dopaminergic neurons (Götz et al., 1990; Spencer et al., 1994; Dias et al., 2013). Factors thought to contribute to increased oxidative stress in the PD brain include mitochondrial dysfunction (Schapira, 2008), dysregulated iron metabolism (Dexter et al., 1989), neuroinflammation (Peterson and Flood, 2012), decreased antioxidant levels (Perry et al., 1982), and elevated production of H₂O₂ and reactive oxygen species (ROS) as a consequence of increased dopamine turnover (Fahn and Cohen, 1992).

Mutations in the gene encoding for leucine-rich repeat kinase 2 (LRRK2) are the most frequent known cause of inherited PD in most populations (Vilas et al., 2016). The role of LRRK2 protein is incompletely understood; it possesses kinase and GTPase activities, and has been associated with autophagy, mitochondrial functions, and microtubule/cytoskeletal dynamics (Wallings et al., 2015; Esteves and Cardoso, 2016; Kang and Marto, 2016). Expression of the most common PD-associated LRRK2 mutation, G2019S (Kachergus et al., 2005), may uncouple mitochondrial oxidative phosphorylation (Mortiboys et al., 2010; Papkovskaia et al., 2012) and increase intracellular ROS production in vitro (Pereira et al., 2014). These findings suggest that this LRRK2 mutation might increase CNS oxidative stress in healthy individuals, and/or contribute to oxidative stress in PD patients carrying this mutation. The interaction between LRRK2 and oxidative stress may be bidirectional, because oxidative stress has been shown to promote in vitro dephosphorylation of LRRK2 at Ser⁹¹⁰/Ser⁹³⁵, causing loss of its ability to bind to 14-3-3 proteins (Mamais et al., 2014). 14-3-3 proteins bind to serine/threonine-phosphorylated residues, often functioning as direct regulators of the target proteins to which they bind (reviewed by Tzivion et al., 2001). Because LRRK2’s binding to 14-3-3 controls its cellular localization, decrease in this binding results in cellular translocation of LRRK2, causing it to accumulate within cytoplasmic pools (Dzamko et al., 2010; Mamais et al., 2014).

8-hydroxy-2′-deoxyguanosine (8-OHdG) (Valavanidis et al., 2009) and 8-isoprostane (8-ISO) (Morrow et al., 1992) are commonly measured oxidative stress markers. In contrast to these direct indicators of oxidative stress, total antioxidant capacity (TAC) has been suggested to be a possible indirect marker for this condition (Mandrioli et al., 2006). TAC includes antioxidant activities of non-enzymatic low molecular weight substances such as glutathione, ascorbic acid, uric acid, α-tocopherol, and coenzyme Q (Alho et al., 1998; Bartosz, 2003); a reduction in TAC could alter the balance between ROS levels and mechanisms which protect against their cytotoxic effects, in favor of oxidative damage. The concentration of 8-OHdG has been reported to be increased in PD cerebrospinal fluid (CSF) (Kikuchi et al., 2002; Abe et al., 2003; Gmitterová et al., 2009; Isobe et al., 2010). We found no reports of 8-ISO or TAC measurements in PD CSF; however, 8-ISO in the PD SNC has been reported to be unchanged (Fessel et al., 2003), while TAC in the PD SNC is reported to be decreased (Sofic et al., 2006). The primary objective of this study was therefore to compare the concentrations of 8-OHdG, 8-ISO, and TAC in CSF between LRRK2 PD, sPD, LRRK2 CTL, and CTL subjects. Our secondary objective was to examine the associations in these individuals of the CSF oxidative stress markers and TAC with cognitive functioning levels, and, in the two groups of PD patients, with clinical disease duration and progression.

Previous studies of 8-OHdG and/or 8-OHG levels in PD CSF, summarized in Table 4, found their concentrations to be increased relative to control CSF (Kikuchi et al., 2002; Abe et al., 2003; Gmitterová et al., 2009; Isobe et al., 2010). No information was provided in these studies regarding possible PD-related gene mutations in the study subjects. The control subjects were healthy volunteers except in the study by Gmitterová et al. (2009), in which the controls were non-cognitively impaired subjects with neurological problems other than PD. In that study PD patients with cognitive impairments (indicated by Mini Mental State Examination [MMSE] scores < 27) had lower CSF 8-OHdG levels than PD patients without cognitive impairments; only when the analysis was limited to the latter group of PD patients was the difference in the levels of 8-OHdG between the PD and control subjects statistically significant.

Our findings in the present study do not confirm the earlier reports of increased 8-OHdG in PD CSF. Methodological differences (HPLC vs. ELISA, measurement of 8-OHdG vs. 8-OHG) and/or characteristics of the PD subjects (differences in medication, age, clinical disease duration, or severity) may have contributed to these conflicting results. Chance variation as a consequence of the small sample sizes in some of the previous studies could also have been a factor; the studies of Abe et al. (2003) and Gmitterová et al. (2009) included only 15 and 13 control subjects respectively. The median concentrations of 8-OHdG in our groups were similar to those reported by Gmitterová et al. (2009) and half to one-third of those reported by Kikuchi et al. (2002); these latter two studies also used ELISA to measure 8-OHdG, whereas Abe et al. (2003) and Isobe et al. (2010) employed HPLC. The 8-OHdG levels reported by Isobe et al. (2010) were far lower than ours, while the study by Abe et al. (2003) measured 8-OHG but not 8-OHdG so no comparison can be made to our measurements.

In contrast to the findings of Gmitterová et al. (2009) discussed above, we found no evidence that cognitive deficits were associated with lower 8-OHdG levels in our PD patients. Our analysis suggested only weak correlations between MoCA scores and 8-OHdG levels for both groups of PD patients (rho values = -0.20 for sPD patients and -0.18 for LRRK2 PD patients). A similar trend was seen with 8-ISO; the correlations between MoCA scores and 8-ISO were -0.17 for our sPD patients and -0.35 for our LRRK2 PD patients. When we divided our two PD groups into cognitively impaired subjects (MoCA scores < 26) and non-cognitively impaired subjects (MoCA scores ≥ 26), we similarly found no evidence that cognitive deficits lowered the CSF levels of 8-OHdG or 8-ISO in either PD group (data not shown).

An intriguing finding in this study was that both 8-OHdG and 8-ISO concentrations were increased in our LRRK2 CTL group compared to our other diagnostic groups (p-values for overall tests of between-group differences = 0.04 and 0.03, respectively). Although we anticipated that LRRK2 mutations might increase CSF oxidative stress markers in both healthy individuals and in PD patients carrying these mutations, we detected the expected increase only in our LRRK2 CTL subjects; why a similar increase was not found in our LRRK2 PD patients is unclear. To our knowledge, this study is the first to investigate CNS oxidative stress in individuals with LRRK2 mutations, so our findings require confirmation with a larger cohort.

Whether the increases in 8-OHdG and 8-ISO that we detected in our LRRK2 CTL subjects (by 12–16% and 20–38%, respectively) are of biological significance is unknown. Of relevance to our results is a report that CSF concentrations of α-synuclein soluble oligomers, suggested to be the most neurotoxic α-synuclein conformation (Winner et al., 2011), are also increased in healthy subjects with PD-associated LRRK2 mutations compared to healthy controls lacking these mutations (Aasly et al., 2014). While oxidative stress has been suggested to be a mechanism by which soluble α-synuclein oligomers may exert their neurotoxic effects (Roberts and Brown, 2015; Deas et al., 2016), we do not know if our finding of increased CSF oxidative stress markers in our LRRK2 CTL subjects may be related to elevated CSF levels of soluble α-synuclein oligomers in these individuals.

Our other findings in this study related to TAC. The observed correlations of TAC with 8-OHdG and 8-ISO were weak, so our results do not support the suggestion of Mandrioli et al. (2006) that TAC may be an indirect indicator of oxidative stress. We found no evidence for differences in CSF TAC activity between PD patients and control subjects with or without LRRK2 mutations. This result agrees with an earlier report that the concentration of ascorbate, the main hydrophilic antioxidant in CSF (Lönnrot et al., 1996) and main contributor to CSF TAC activity (Alho et al., 1998; Lönnrot et al., 1996), is unchanged in PD CSF (Buhmann et al., 2004). Our finding of similar CSF TAC levels between PD patients and control subjects suggests that the decreased TAC activity reported in the PD SNC (Sofic et al., 2006) may not be reflected by TAC levels in PD CSF.

We found positive correlations between CSF TAC and MoCA scores for three of our groups (LRRK2 PD, LRRK2 CTL, and CTL subjects). This suggests that TAC in CSF may decrease as MoCA scores decrease (e.g., as cognitive deficits develop), regardless of the presence or absence of PD. This result is similar to the finding by Bassett et al. (1999) that CSF antioxidant capacity is reduced in patients with Alzheimer’s disease.

Our TAC results included some PD-specific findings. TAC was negatively correlated with UPDRS Total scores in both groups of PD patients. Because an increase in UPDRS scores indicates worsening Parkinsonian disability, these correlations suggest the possibility that TAC may decrease in the CNS during the progression of PD.

The numbers of CSF samples available to us were a limitation with respect to our ability to identify the locations of group differences for 8-OHdG, 8-ISO, and TAC with the DCSF procedure, and the associations between their concentrations and the other variables we examined. A second limitation is that the concentrations of these oxidative stress markers in our PD patients, although not increased compared to our control subjects, could potentially have been influenced by the prooxidative effect of their levodopa therapy (Buhmann et al., 2004).

We conclude that the concentrations of oxidative stress markers in CSF may be increased in healthy individuals with PD-associated LRRK2 mutations. If our findings in this study can be confirmed with larger group sizes, then the possibility should be considered that an increase in CNS oxidative stress in individuals carrying these mutations may contribute to their risk for developing PD. We also conclude that TAC levels may decrease during the progression of PD, and in the presence of impaired cognitive functioning irrespective of PD status.

DL directed the study and prepared the manuscript. AK performed the assays, collated the data, and reviewed the manuscript. MC performed the statistical analysis and assisted with manuscript preparation. JA recruited the patients, collected the CSF samples, and reviewed the manuscript. PL assisted with manuscript preparation.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.