Parkinson's disease (PD) is a slow, progressive degenerative disorder of the central nervous system 1. It is characterized by muscle rigidity, tremor, slowing of physical movement (Bradykinesia) and, in extreme cases, loss of physical movement (Akinesia) 2. The classification of the stages were done based on Unified Parkinson's Disease Rating Scale (UPDRS) 3 or on an equivalent rating scale value such as Hoehn and Yahr scale and Schwab and England Activities of Daily Living Scale. The UPDRS has been used extensively by researchers and clinicians around the world. The rating scale is categorized as stages based on bradykinesia, dyskinesias, rigidity, posture, swinging arms, walking, tremor, gesture, seborrhea, speech, sensory complaints, depression, postural stability, insomnia, leg agility and independence. The diagnosis is based on medical history and neurological examinations 4. The current diagnosis of PD remains unclear, because of the complex spectrum of symptoms and their similarities with other neurodegenerative diseases, such as multiple system atrophy and progressive supranuclear palsy. Moreover, clinical diagnosis fails to identify PD before they cause a significant loss of dopamine neurons 5. There is immense need for early detection 6 and more effective drugs for the cure of PD. An understanding of the molecular characteristics underlying the disease processes of PD is a prerequisite for the development of biomarker in the early detection for providing high value therapeutics.

Biomarkers serve as tools for diagnosis of any disease usually performed on readily accessible body fluids, such as cerebrospinal fluid (CSF), serum, urine or saliva. Metabolic profiling is one of the most important techniques, greatly focused for the detection of biomarkers for diagnosis of diseases 78. Several encouraging results were obtained using metabolite profiling in an attempt to diagnose coronary heart disease 9, diabetes mellitus 10, eclampsia 11, lipid disorder 12, colon carcinoma 13, epithelial ovarian cancer 14, hypertension 15, kidney deficiency syndrome 16, motor neuron disease 17 and liver cancer 18. Several techniques are presently available for metabolite profiling, such as nuclear magnetic resonance spectroscopy (NMR), mass spectrometry (MS), high-performance liquid chromatography (HPLC), gas chromatography (GC) and optical spectroscopic (OS) analysis. NMR spectroscopy technique is more ease for rapid metabolite detection because it requires an uncomplicated, preprocessing procedure, and it is a feasible method to obtain essential information from complex and intact biological samples. Recent studies show that NMR metabolic profiling plays a vital role in clinical diagnosis 1920.

In the present study, ¹H (proton) NMR spectroscopy was executed to examine the plasma samples of 37 normal and 43 drug-naive patients of PD to identify the metabolite variations of 22 targeted metabolites between the blood plasma of healthy individuals and patients.

Investigation of plasma samples for the detection of metabolite variations were performed by comparing the normal with PD patients by proton NMR. The typical analysis of NMR spectra aimed for 22 metabolites hypothetically suggests a critical role in PD.

The present study is focused to identify the metabolic marker for the detection of PD from plasma and to validate the variations by spotting out through gene responsibility.

Several studies have been indicated the gene variation in PD 2425262728 which may lead to metabolic abnormalities that are detectable in peripheral tissues. Few successful break through have shown metabolite variations in CSF and serum samples of PD. For instance, glutamate was found to be decreased in parkinsonian CSF compared to control subjects, a result reported in earliest analyses of parkinsonian CSF 293031. Subsequent study shows that the increased level of glycine, aspartate and glutamate in the plasma of parkinsonian patients 32. Increase in CSF glycine also been observed in PD 33. Previous study of PD patients showed reduction in concentrations of arginine and methionine in serum, while the level of valine was increased 34. Additionally, a significant decrease in CSF isoleucine, alanine, lysine and moderate increase of a glutamine level was very well observed in PD 34. Recent study shows the feasibility of potential diagnosis of PD from the detection of increased 8-OHdG in serum and urine of PD 35. Furthermore, several reports indicate that there is a reduction of complex 1 activity in the electron transport chain of PD 363738394041. From the detailed knowledge of these studies, 22 metabolites, which play a role in mitochondrial function and other related pathway, have been targeted.

Analysis of the ¹H NMR spectra of plasma samples showed differential distribution of these metabolites in drug-naive patients compared to the healthy volunteers. The targeted metabolite profiling of 22 compounds in blood plasma was characteristically altered in patients with PD, and most of these metabolites have decreased in concentration. The results showed that, this approach has great hope in diagnosis of PD. Moreover, this study was made with unmedicated PD patients to controls, to avoid the confounding effects of any medications. The key metabolites, such as myoinositol, sorbitol, citrate, acetate, succinate and pyruvate are significant in contribution for the separation between metabolite profiles of unmedicated PD patients and controls.

Plasma myoinositol level was significantly increased in the drug-naive patients. Elevation of myoinositol concentrations implies the decrease in activity of sciatic motor-nerve conduction velocity determined in animal models 4243. Elevated plasma myoinositol has not previously been reported for PD. However, studies show the abnormal myoinositol levels in the brain of neurologically diseased patients and other disorders 44454647. Raised myoinositol levels in the basal ganglia were recently observed in PD patients under exercise condition detected using magnetic resonance spectroscopy 48. Additionally, plasma sorbitol level was significantly increased in drug-naive patients. Sorbitol has been linked to drug treatment in PD 49, yet our observation of an elevation of plasma sorbitol concentrations in drug-naive patients may be due to impairment in oxidative stress 5051. The elevated levels of sorbitol have been reported in the CSF of mood disorder patients, which relates to oxidative stress 52. Surprisingly, elevation in the plasma sorbitol level has not been reported in PD to our knowledge. Together with the significant findings of myoinositol and sorbitol in plasma imply the dysfunction of polyols metabolic pathway. Polyols pathway is a minor metabolic pathway of glucose running parallel to glycolysis, whose activity is altered in mitochondrial dysfunction 53. However, malfunctioning of mitochondria is previously reported in PD 363738394041, and it is the pedestal of this study. Detection of glucose concentration were not been carried out but abnormality of glucose metabolism was previously reported in PD patients 54.

More interestingly citrate, malate, acetate, succinate and pyruvate are significantly varied in PD plasma samples, contributes to the major distinction of PD from normal samples in PLS-DA analysis. These metabolites, such as citrate, acetate, succinate and malate were decreased, while increase in pyruvate concentration was noticed. Pyruvate is the end metabolite of glycolysis. It enters Kreb's cycle as acetyl-coA by the catalysis of enzyme pyruvate dehydrogenase in the presence of the coenzyme NAD+. The accumulation of pyruvate or its increased concentration in plasma may be due to abnormal activity of pyruvate dehydrogenase complex and its interacting genes in patients. Increased pyruvate CSF has already been reported in Alzheimer's patients 5556. The other intermediates of Kreb's cycle such as citrate, malate and succinate were considerably decreased in this study, which may correlate to alteration of pyruvate dehydrogenase activity. Moreover, the detection of other metabolites of Kreb's cycle was not executed in our analysis. Systems biological approach was carried out on pyruvate dehydrogenase components to identify its interacting genes which are hypothesized as the cause for the increased plasma pyruvate concentration. The analysis reveals 46 interacting genes together with pyruvate dehydrogenase components.

The significant variation in plasma pyruvate was validated by gene expression analysis of 46 genes derived from systems biological approach. The expression analysis was executed only on 40 genes and the other 6 genes in which the following FLJ21936, DMRT3, AKR1CL2, UNC5A and UNC5B were excluded from this study, because these genes are not expressed in blood cells and GCSL gene was also eliminated since it is considered as an alias of DLD gene. The analysis of 40 genes shows differential regulation of 16 genes in comparison to control. Interestingly, out of 16 genes, 9 genes have been previously reported in PD, such as CAT, FGF13, JUN, INSR, NOS1, OGDH, SYT1, FGF2 and SST 575859606162636465. Furthermore, statistical analysis of these genes show the significance of NPFF and PDHB with p < 0.05. NPFF gene plays a major role in inflammation modulation, neuroendocrine function and cardiovascular regulations 66, where as, PDHB gene is the beta subunit of pyruvate dehydrogenase, directly associated with pyruvate dehydrogenase activity and mitochondrial dysfunction. Abnormal activity of pyruvate dehydrogenase is the basis of our study in representing the variation in pyruvate concentration. The impairment of cardiovascular regulation 67, inflammation 68 and neuroendocrine function 69 were reported in PD and this may suggest the influence of differential regulation of NPFF gene. The biological significance of these genes related to PD has not been reported previously. Hence, the variation in pyruvate concentration can be considered as a marker for detection of PD and future studies need to be carried out on the factors and the hidden mechanism involved in variations of NPFF and PDHB in PD.

Prior study indicates no correlation between metabolite variation with severity and duration in PD 24. To bring out the correlation, NMR peak amplitude representing the ~290 metabolite was included as one of the variables in the prediction of disease stages by neural network. The ANN was trained with variable parameters as described above. More fascinating results emerged from neural network and it is capable to predict early stages of PD, with a good accuracy using these variables. The optimized network yield has an accuracy of 97.14% in detecting PD patients. Based on the present study it has been confirmed that the association between disease progression and metabolite variation is strong and it is more accurate than the current diagnosis 23 of PD.

The present study concludes that the application of NMR metabolite profiling of plasma fluid can provide an efficient means for detection of Parkinson's disease. We identified abnormalities in 22 circulating metabolites, which bring out the scope for the diagnosis of PD. Moreover, the result obtained from neural network approach is more feasible for the stage wise detection. However, the accuracy of neural network is questionable unless the study focused with the larger samples.

The authors declare that they have no competing interests.

SSSJA designed the study and analyzed the data under the guidance of HTTC and WS. The research participants were selected by SK, examined them clinically and neurologically for metabolic investigations.