Parkinson's disease (PD) is a common chronic neurodegenerative disorder, which affects >0.1% of individuals >40 years of age (1,2). There is a loss of dopaminergic nigrostriatal neurons in the substantia nigra (SN) pars compacta in PD (3,4), and clinical characteristics for the majority of patients with PD include motor disorder, slowness of movement, tremor at rest, rigidity and disturbances in balance. However, it is difficult to diagnose PD in its early stages due to the high rate of misdiagnosis with other movement disorders, which can share certain features with PD, including essential tremor, multiple system atrophy (MSA) and progressive supranuclear palsy (5). Therefore, it is necessary and urgent to screen diagnostic or therapeutical biomarkers, which can be utilized in the diagnosis or treatment of PD.

In previous years, numerous significant factors have been correlated with the development of PD, which has improved current understanding of PD. For example, the level of S100 calcium binding protein B (S100B) is higher in the post-mortem SN of patients with PD, compared with control tissue, and the ablation of S100B results in neuroprotection, reduces microgliosis and decreases the expression of tumor necrosis factor-α and the receptor for advanced glycation end products (6). The reduced expression level of lysosomes-associated membrane protein 2A in dopaminergic cell lines leads to reduced chaperone-mediated autophagy and decreases the metabolism of wild-type α-synuclein (SNCA), which is a cause of PD (7). In addition, a set of molecules, which offer potential in identifying individuals at risk for PD, and assist in the diagnosis and management of PD have been identified, including SNCA (8) and DJ-1 (9). Furthermore, based on microarray analysis, Papapetropoulos et al identified 11 candidate genes with deregulated expression in brain regions from patients with PD, and the significant role of mitochondria ribosomal protein S6 (MRPS6) in the oxidative phosphorylation system was reported (10). In addition, a previous study identified a set of differentially expressed genes (DEGs) in PD, and the candidate compound, alvespimycin, was selected as a lead compound for PD, based on the gene expression data from the study by Papapetropoulos et al (10,11). In organisms, each protein exerts its functions via interactions with other proteins and regulation of a set of regulators. However, interactions and regulators of DEGs in PD, which assist in obtaining a more comprehensive understanding of the pathogenesis of PD, were not fully investigated in the studies of Papapetropoulos et al and Gao et al (10,11); there may be additional potential key genes associated with PD, which remain undetected.

The present study aimed to examine additional potential key genes associated with PD, in addition to examining their interactions and regulation. The microarray data of Papapetropoulos et al (10) were used to identify DEGs between patients with PD and those without PD. The potential crucial genes associated with PD were then examined. Subsequently, transcription factors (TFs) and protein-protein interactions (PPIs) of the DEGs were analyzed. A second microarray dataset (12) of PD, which included additional post-mortem brain samples, was selected to validate the expression of key genes in PD. The results may provide novel information for investigation of the molecular mechanisms underlying the pathogenesis of PD.

In the present study, based on the analysis of the GSE7621 dataset, 398 upregulated genes and 272 downregulated genes were identified in the SN from the patients with PD, compared with the controls. A total of 10 DEGs (CXCR4, DCC, EHD1, EPN2, GRK6, NCK2, ROBO2, RUFY1, SEMA6D and SLIT1) were predicted to be closely associated with PD. Among these 10 genes, downregulated DCC and upregulated CXCR4 were the hub proteins, which interacted with multiple genes in the PPI network. These two proteins were distinctly enriched in the axon guidance pathway, and GO terms associated with neuron development and differentiation.

In the PPI network, DCC interacted with NTN1. DCC encodes an NTN1 receptor, a member of the immunoglobulin superfamily, which mediates the axon guidance of neuronal growth cones towards sources of NTN1 ligand (23). NTN1 binds to the DCC receptor and signals through downstream proteins, which regulate cytoskeletal reorganization, causing growth cone extension and neurite growth (24). Another study reported that NTN1 and DCC are able to attract axons to the nervous system midline, and mutations in DCC lead to midline guidance defects in humans (25). There is also evidence that DCC-mediated NTN1 signaling affects the formation/maintenance of mesocorticolimbic DA topography (26). Polymorphisms in NTN1 and DCC are associated with the loss of multiple functions and increased the risk of PD (27,28). These results suggested that the decreased expression of DCC and NTN1 may lead to a loss of dopaminergic nigrostriatal neurons in the SN, resulting in the onset of PD. In the PPI network, NTN1 interacted with ROBO2, which interacted with SLIT1. The ROBO2 protein belongs to the Robo family, which is important for axon guidance across the midline during central nervous system development (29). ROBO2 has also previously been found to be deregulated in PD (30). Loss of Slit1 expression results in an abnormal course of the nigrostriatal pathway through the diencephalon (31). Although the genes of the DCC-NTN1-ROBO2-SLIT1 interaction pathway have been previously reported to be associated with the nervous system, their interactions in the present study are a novel finding.

CXCR4 encodes a CXC chemokine receptor specific for stromal cell-derived factor-1 (32). In the SN of patients with PD, higher expression of CXCR4 and its natural ligand, CXCL12, have been detected, compared with control subjects (28,33), which is consistent with the results of the present study. The activation of CXCR4 has been shown to reduce dopamine levels in the striatum, and induce cell death and tissue loss in the SN and striatum (34,35). Furthermore, CXCL12 and CXCR4 are neurotoxic to dopamine neuronal populations through the release of neurotoxins from microglia (36). Therefore, CXCR4 may be pivotal in the etiology of PD by reducing dopamine levels in the striatum and/or releasing neurotoxins from microglia to induce the death of dopamine neurons. In the PPI network, CXCR4 interacted with other gene proteins, including TOX high mobility group box family member 4 (TOX4), YY1, GRK6, kinase insert domain receptor (KDR) and zymogen granule protein 16B (ZG16B). Among these genes, YY1 has been reported to regulate the expression of SNCA, which is an established susceptibility gene for PD (37). In the present study, YY1 was predicted as a TF targeting DCC, ROBO2, SLIT1 and GRK6. Therefore, YY1 may also be involved in PD via regulating the expression of these genes. In an animal model of PD, the expression of GRK6 was found to be altered in the caudate/putamen and ventral striatum (38), and the altered expression of GRK6 can result in the dysregulation of DA receptors, thus contributing to the core motor deficits observed in PD (39). Other genes, including TOX4, KDR and ZG16B, and their interactions with CXCR4, have not been correlated with PD in previous studies; thus, these are novel potential genes, which may be associated with PD.

In the present study, NCK2 was upregulated in the SN of PD based on the analysis of the two datasets. NCK2 encodes a member of the NCK family of adaptor proteins, which are involved in the regulation of receptor protein tyrosine kinases (40). It has been reported that NCK2 is expressed in neurons and glia (41). There is evidence that NCK2 interacts with particularly interesting cysteine histidine-rich protein and Dock180 protein, which are involved in neuropathological processes (42,43). In the PPI network, NCK2, CXCR4 and DCC all interacted with BCAR1. BCAR1 is involved in various cellular events, including migration, survival and transformation (44). A previous study has reported the differential expression of BCAR1 in PD (45). The cerebellum may have certain roles in the pathophysiology of PD (46). p130Cas/BCAR1 was found to be involved in cellular mechanisms regulating different stages of cerebellar development (47), suggesting that the abnormal expression of BCAR1 may be associated with PD. However, there is no evidence to confirm the association of NCK2 with the progression of PD. Collectively, these findings indicated that the upregulation of NCK2 may be responsible for the progression of PD, via interaction with BCAR1 or regulation by TFs.

In addition, according to the TF analysis for crucial PD-associated genes in the present study, a series of TFs were predicted to target the 10 crucial PD-associated genes. DCC, CXCR4, NCK2, SLIT1, ROBO2 and GRK6 were all regulated by a set of TFs, including GATA, E2F and E4BP4. A previous study reported that GATA-1 and GATA-2 regulate the expression of SNCA dopaminergic cells (48). The pRb/E2F cell-cycle pathway mediates cell death in PD, and E2F-1-deficient mice are significantly more resistant to dopaminergic cell death, compared with their wild-type littermates (49). E4BP4, also known as nuclear factor, interleukin 3 regulated is a transcriptional regulator of the human interleukin-3 promoter (50). A previous study reported that E4BP4 is reportedly associated with neuron disease (51), and is upregulated in the SN tissue from patients with PD (52). Collectively, the DCC, CXCR4 and NCK2 genes are involved in the pathogenesis of PD, not only via the axon guidance pathway, but also likely via the regulation of multiple TFs, including GATA, E2F and E4BP4.

Of note, although the present study reanalyzed the public data deposited by others, the analytical procedure of the present study was different from that of previous studies. The criteria of the FDR<0.05 and |log₂ FC|>1 were more strict, compared with those (P<0.05; |FC|≥1.3) used in the previous study by Papapetropoulos et al (10), indicating that the DEGs identified in the present study show a higher level of differential expression. In the present study, the P-value of each gene was adjusted using the BH method to reduce false positives, however, the P-value was not adjusted in the previous study. Furthermore, in the present study, the protein interactions of DEGs and the TFs regulating DEGs were analyzed, which assisted in providing a more comprehensive understanding of the pathogenesis of PD, whereas these were not investigated in the previous study. Despite the significance, there were certain limitations in the present study. For example, only two datasets were analyzed, whereas performing a meta-analysis using multiple datasets, including multiple types of data, may produce additional significant results. This is an important aim of further investigations.

In conclusion, the present study identified a set of potential crucial PD-associated genes, including CXCR4, DCC and NCK2, which were associated with neuron development and differentiation. The DCC-NTN1-ROBO2-SLIT1 interaction pathway and several genes, including TOX4, KDR and ZG16B, which interacted with CXCR4, were novel findings. Genes, including DCC, CXCR4, NCK2, SLIT1 and ROBO2 were also regulated by multiple TFs, including GATA, E2F and E4BP4. Additionally, CXCR4 and NCK2 were detected to be upregulated in the SN of patients with PD following the analysis of two datasets. These genes, interactions of proteins and TFs may be key in the progression of PD. These findings contribute to an improved understanding of the pathogenesis of PD. However, further experiments are required to confirm the results of the present study and further elucidate the pathogenesis of PD.