PD is an incurable and complex disease in which the mechanisms that cause DA neuronal death are not yet clear [2,5]. Some authors described a link between redox imbalance, energy failure and metabolic disturbances in PD, leading to the idea that it might be considered a metabolic disease [7,8]. In this study, we used NMR spectroscopy for metabolomic profiling in DJ-1β mutants and control flies of different ages. The aim of this study was to detect metabolic alterations that might be relevant in PD physiopathology, as well as identify potential therapeutic targets and biomarkers. To date, several metabolomic analyses were performed in the plasma, cerebrospinal fluid or blood of PD patients [47,58,59,60,61,62,63]. However, human samples are limited to specific tissues and do not allow for a complete examination of the disease [29,30,31]. To circumvent this limitation, we analyzed metabolites in the whole body of PD model flies. This experimental approach allows a complete panorama of PD-associated metabolic defects to be obtained, which are not restricted to the brain and can influence the physiology of the entire fly [33]. To date, only one similar study has been performed in another Drosophila PDf model based on PINK1 deficiency [64]. Our results confirm that several metabolic pathways, such as amino acid catabolism/anabolism, glycolysis, the TCA cycle, or the UC cycle, might be altered in PD.

Neurons require high levels of energy, which is mainly produced by mitochondria. However, mitochondria become less efficient with age as well as in several NDs, such as PD, thus leading to a reduction in ATP levels, while ROS production increases [50]. In this scenario, amino acids play an important role in the brain, since some of them can be metabolized to fuel cellular energetics when there is insufficient energy [56]. The amino acid profile observed in 1-day-old and 15-day-old DJ-1β mutants is consistent with amino acids anabolism/catabolism imbalance in PD model flies, especially in younger flies (Figure 3). Interestingly, an increase in amino acid catabolism, through transformation to pyruvate and acetyl-CoA or other TCA cycle intermediates to produce energy [65], could compensate ATP deficit due to impaired mitochondrial function. Moreover, the reduction in amino acid levels could be also consistent with impairment in protein synthesis in muscles, which could exacerbate motor symptoms in PD patients [44].

Among the amino acids whose levels are altered in PD model flies compared to controls, we found BCAAs. BCAAs metabolism/degradation is closely related to carbon metabolism, in particular to glycolysis and the TCA cycle, since BCAAs can be transformed to succinyl-CoA that fuels TCA cycle [65]. PD model flies exhibit decreased levels of BCAAs probably due to its degradation to produce energy. A similar situation was found in a metabolomic analysis performed in sebum samples of PD patients [66], thus supporting the validity of the results obtained in our Drosophila PD model. Interestingly, a clinical trial (NCT01662414) was performed in PD patients to investigate the effect of whey protein supplementation, an important source of BCAAs. The results showed that this supplementation was able to increase BCAAs and essential amino acid levels, as well as reduced glutathione levels, a key antioxidant that prevents the oxidative damage of DA neurons [45]. In relation to this, a recent study showed that whey protein supplementation improved motor symptoms in PD patients due to an increase in muscle regeneration [67]. Our results indicate that BCAAs metabolism seems to be decreased in 15-day-old compared to 1-day-old DJ-1β mutant flies. This decrease, especially in aged flies, could enhance PD-related phenotypes, as shown in PD patients [45]. Therefore, BCAAs metabolism could constitute a possible therapeutic target to investigate new PD treatments. Levels of other proteinogenic amino acids are also reduced in DJ-1β mutant flies. For example, they present decreased tryptophan levels, whose metabolism is related to neurodegeneration in PD patients at early stages [46,65,68]. Another proteinogenic amino acid affected in DJ-1β mutants is glycine, a small amino acid with neuroprotective effects in neuroinflammation, ROS-related damage, and synaptic dysfunction through JNK signaling inactivation [69]. It is likely that a decrease in glycine levels in DJ-1β mutants could be contributing to PD pathology. Moreover, there is a significant change in glycine levels when comparing metabolomes of 1-day-old to 15-day-old mutant flies, thus suggesting that this amino acid could be a possible biomarker of PD progression.

PD model flies also exhibit significant increased levels of some non-proteinogenic amino acids compared to controls such as β-alanine. This result could indicate an increase in pyrimidine degradation, as β-alanine is the final product of this pathway. Similar results were found in pink1 mutant flies, another Drosophila fPD model [64]. Moreover, an increase in β-alanine levels causes taurine depletion, which was related to nerve degeneration [70]. Therefore, alterations in β-alanine metabolism and related pathways could be contributing to PD physiopathology. On the other hand, N-acetyl-aspartate is the most abundant non-proteinogenic amino acid in the central nervous system and a marker of neuronal integrity. It is synthetized in the mitochondria from aspartate and acetyl-CoA, and is then transported to the cytosol, where it is metabolized into aspartate and acetate. Its synthesis increases in absence of ATP due to the use of acetate as a source of energy [71]. We found that PD model flies showed an important increase in N-acetyl-aspartate (Figure 3), which could reflect a deficiency in its metabolism, since acetate levels were also decreased in 15-day-old DJ-1β mutants compared to control flies of the same age (Table S2).

Alterations in energy metabolism and glucose uptake are associated with the physiopathology of several NDs including PD [7,8,56]. Glucose is the main source of energy in the brain and the regulation of its metabolism is critical for brain physiology [50]. This is consistent with the observation that DJ-1β mutants show a dysregulation of carbohydrates metabolism (Figure 4). Supporting this, a decrease in glucose levels in the iPSC-derived DA neurons mutant for PARK2 was previously observed [72]. In fact, glucose is the sole substrate that can supply the rapid energy demand of neuronal cells through glycolysis [56,73]. Therefore, an increase in glucose consumption through this pathway to restore ATP levels could explain its reduction in our PD model flies [7,74]. Glycolysis is an ATP-producing pathway that could provide considerable amounts of ATP in a less efficient manner than oxidative phosphorylation to support the acute energy demands in neurons [73,75]. DJ-1 was described as participating in the activity of complex I of ETC binding and stabilizing it [14]. Accordingly, TCA cycle reduction and a higher NADH/NAD⁺ ratio in the PD model flies suggest the existence of an alteration of this complex and mitochondrial dysfunction, which was also observed in a MPTP-induced PD cell model [76]. Thus, our results show that a shift from TCA cycle to glycolysis is produced in PD model flies [7,41,77]. In addition, it was reported that a reduction in Aco activity caused neurotoxicity in mesencephalic rat cultures, due to the loss of its activity as ROS and an iron biosensor [52,78]. Moreover, Aco mutant flies exhibited a reduced locomotor activity, shortened lifespan, and increased cell death in the developing brain, as well as glycolysis and TCA cycle disturbances that led to decreased ATP levels [52]. Some of the phenotypes are similar to those observed in DJ-1β mutant flies [7,27]. On the other hand, the reduction in SDH activity leads to the activation of the mammalian target of rapamycin (mTOR) and the sterol regulatory element binding protein (SREBP), which contribute to lipid accumulation in neurons, as observed in several NDs, and produce excitotoxicity, thus participating in PD pathogenesis and development [79,80]. In addition, the finding of reduced malate levels in PD model flies is also noteworthy. Malate is an intermediate of the TCA cycle that is metabolized from fumarate by the fumarate hydratase enzyme. Interestingly, it was shown that the expression of fumarate hydratase was reduced in the DA neurons of the SNpc in the brains of iPD patients [81]. Thus, our results and these observations suggest that a decrease in the activity of both enzymes might be relevant to PD pathogenesis.

Although TCA cycle activity is decreased in DJ-1β mutant flies, there is an increase in some pathway intermediates such as fumarate (Figure 4g,h). TCA cycle metabolites can be produced by other pathways, as could be happening with fumarate and the UC [55]. UC is responsible for the excretion of the nitrogen that cannot be used in amino acid metabolism, and is related to the TCA cycle by fumarate, which is transported into the mitochondria, where it can be used as a substrate of TCA cycle [55,56]. The enhanced expression of arg and Argl in PD model flies could lead to a general increase in UC activity and to higher fumarate levels, as observed in the metabolomic analyses (Figure 4g,h). Supporting these results, an increase in ArgL activity was observed in a zebrafish PD model based on DJ-1 deficiency [82]. Changes in arg expression levels were also reported to have implications in the brain, although its role in this tissue is not yet known [83]. In addition, an enhancement of UC could serve to remove the excess of ammonia caused by the increased amino acid catabolism observed in DJ-1β mutant flies, as previously reported in Alzheimer’s disease [83].