Mild cognitive impairment (MCI) is a transitional stage between normal aging and senile dementia, with a high risk of developing dementia (Svenningsson et al., 2012; Gonzalez-Redondo et al., 2014; Brennan et al., 2017). In patients with Parkinson’s disease (PD), the incidence of MCI can reach 40% even in the early stage of the disease course (Litvan et al., 2012; Gonzalez-Redondo et al., 2014). Executive dysfunction (EDF) is the most prominent feature of PD with cognitive dysfunction, affecting about 30% of all PD patients. This deficit can occur in the early course of PD, and includes impairments in the coordination of a range of cognitive process required to achieve complex, goal-oriented and new cognitive operations (Parker et al., 2013), manifesting as impairments in the three core executive functions (EFs) (working memory, cognitive flexibility, and inhibitory control), as well as in higher-level EFs such as reasoning, problem solving, and planning (Diamond, 2013).

In PD patients, the degeneration of frontal-striatal dopamine pathway contributes to EDF, manifesting as deficits in cognitive flexibility, planning, inhibitory control and working memory, usually relate to MCI in PD; whereas the degeneration of cholinergic systems and Lewy body pathological changes in the posterior brain system result in the visuospatial and memory dysfunction, relate to dementia of PD (Kehagia et al., 2010, 2013). A study suggested that homovanillic acid (HVA), as the main metabolite of dopamine, could be used as an intracerebral marker of dopamine function and metabolism (Saracino et al., 2006).

Glial cell line-derived neurotrophic factor (GDNF) has a strong neurotrophic effect on a variety of neurons, especially on dopaminergic neurons (DANs). GDNF promotes the survival, morphological differentiation, damage repair, and dopamine release of DANs, and regulates their excitability in the midbrain (Grondin et al., 2003; Wang et al., 2011; Brown et al., 2018), and play important role in motor symptoms of animal models of PD. So we wonder if serum GDNF may be involved in the EDF of PD patients, by acting alone or in conjunction with HVA (Liu et al., 2020).

To this end, we studied the correlation of the levels of GDNF and HVA in the peripheral blood serum of PD patients with EF and its components (working memory, cognitive flexibility, and inhibitory control), in order to explore the possible mechanism of GDNF participating in the occurrence and development of EDF, and to provide a theoretical basis for early identification of cognitive dysfunction in PD.

The cognitive progress of PD patients can develop from the normal cognitive function stage (PD with normal cognition, PD-N) to the transition stage of PD with MCI (PD-MCI), and finally deteriorates to dementia (PD with dementia, PD-D) (Litvan et al., 2012). There is a high risk of MCI developing dementia (Svenningsson et al., 2012). Cognitive impairment in PD affects patient’s EF, attention, memory, language and visuospatial function, among which the EDF seems to be the most serious (Kehagia et al., 2013). In PD patients, the depletion of dopamine may lead to the disruption of the frontal-striatal circuit, which contributes to EDF (Uekermann et al., 2004; Kudlicka et al., 2011; Zhang et al., 2020).

In this study, the PD-N patients did not show obvious EDF. PD-MCI patients mainly suffered from impairment of attention, planning, inhibitory control, attention switching, and cognitive flexibility, which largely accorded with the results of Kudlicka et al. (2011), that is, impairment of EFs including the cognitive flexibility, reaction inhibition, inhibitory distraction of attention, and attention maintenance can occur in the early stage of PD. Unsurprisingly, PD-D patients performed worse in extensive EF tests. At present, the exact development pattern of EDF in PD is still controversial (Diamond, 2013), which may be attributed to the complex pathoanatomical mechanism of PD. The dual syndrome hypothesis holds that degeneration of frontal-striatal dopamine pathway contributes to executive dysfunction, manifesting as deficits in cognitive flexibility, planning, inhibitory control and working memory, usually relate to MCI in PD; whereas the degeneration of cholinergic systems and Lewy body pathological changes in the posterior brain system result in the visuospatial and memory dysfunction, relate to dementia of PD (Kehagia et al., 2010, 2013). The dopamine overdose hypothesis holds that dopamine deficiency first appears in the dorsal striatum at the early stage of PD, whereas dopamine is relatively better preserved in the ventral striatum, and the dorsolateral frontal striatal pathway is more likely to degenerate in the early stage of PD than the ventral frontal striatal pathway. Therefore, impairment of attention switching mainly occurs in the early stage, whereas the reverse learning ability (cognitive flexibility) is better preserved in the dorsal side (Vaillancourt et al., 2013). The results of this study found that the impaired performance of attention switching and cognitive flexibility in early PD patients might be related to the overdose effect of dopamine medication on the ventral striatum. On the other hand, it might also be association with the complex and vague features of EF per se (Zhang et al., 2020).

This study also found lower peripheral serum GDNF level in the PD patients than the health controls, which was consistent with the results of previous clinical studies. One study found that the GDNF concentration in cerebrospinal fluid of PD patients was lower than that of normal controls, and was lower in early PD patients than late PD patients (Vaillancourt et al., 2013). Nagatsu and Sawada (2007) reported that the concentration of neurotrophic factors including GDNF decreased in the substantia nigra of PD patients. A large number of trials had used GDNF as a potential therapy to treat motor symptoms of PD. At present, gene and cell therapies are widely applied in preclinical and clinical researches, and have achieved certain success (Garbayo et al., 2016). However, the research of GDNF application in the field of cognition in PD patients was mostly limited to animal experiments. Some studies found that after intracerebroventricular injection of lentiviral vector carrying the human GDNF coding sequence in PD rats, the expression of GDNF increased and the spatial learning and memory ability improved (Lu-Nguyen et al., 2014). Another study described reduced GDNF concentration in the brain of rats with induced nerve infection, and their cognitive function, especially learning and memory, also deteriorated (Gui et al., 2017). It was also found that after amantadine treatment, the expression of GDNF in the hippocampus of PD rats increased, and their learning and memory abilities also improved (Zhang et al., 2014).

Through further exploration of the correlation between serum GDNF and neuropsychological test performance, this study found that the increase of GDNF concentration was correlated with better working memory, attention, inhibitory control, cognitive flexibility and memory. The correlation between serum GDNF concentration and EF tests has been reported previously, and it was found that the serum GDNF concentration of schizophrenic patients who were initially untreated was negatively correlated with the performance of TMT-B (Xiao et al., 2017). Niitsu et al. (2014) found that the GDNF concentration in the peripheral serum was positively correlated with DST performance in the healthy normal population. The research of Wang et al. (2011) showed that the plasma GDNF concentration of late-onset depression patients was positively correlated with the performance of backward DST and negatively correlated with that of TMT-B. These results show that with the increase of GDNF concentration, the test performance of attention and working memory improves, which is consistent with our results. We also found that GDNF is correlated with the performance of AVLT (memory), TMT-A (attention) and SFT (cognitive flexibility). We believed that the improvement of cognition induced by GDNF might be related to its strong neurotrophic and neuroprotective effects not only on DANs, but also on the serotonin neurons, norepinephrine neurons, cholinergic neurons. Studies have found that the complex neurotransmitter circuits are involved in various cognitive processes (Murley and Rowe, 2018; Hirano, 2021). For example, the dopamine system is involved in the regulation of EF by the frontal system (Halliday et al., 2014), the serotonin system is closely related to cognitive flexibility (Clarke et al., 2004), the norepinephrine system can affect set-switching, attention, response inhibition and reverse learning (Logue and Gould, 2014), and episodic memory coding and spatial memory processing are thought to rely on the acetylcholine system to maintain normal functions (Newman et al., 2012). GDNF might synergistically participate in the process of cognitive function through its neuroprotective effects by regulating the survival of these neurons and synaptic plasticity. Another theory believes that GDNF can antagonize the degeneration of neurons and thus delay the occurrence of cognitive decline. The specific mechanism is not completely clear, which may involve the antagonism of GDNF against oxidative stress, inhibition of microglial excitability and neuroinflammation (Gill et al., 2003; Tome et al., 2017).

In addition, this study also found that the serum HVA concentration in PD patients was lower than that in normal controls. HVA is the main metabolite of dopamine, and the measurement of HVA concentration in body fluid may be the most direct method to reflect the function and activity changes of human central DANs (Pickar et al., 1988). Through further analysis of the correlations between HVA concentration and the performances of EF tests in PD patients, we found that serum HVA level was closely related to the performance of backward DST, total DST and TWT in PD patient, suggesting the correlation between serum HVA level with the EF components such as working memory, inhibitory control, cognitive flexibility, and attention. This was consistent with the results of a previous study that the dopamine system is involved in the regulation of EF by the prefrontal cortex (Martinez et al., 2016).

The levels of GDNF and HVA in peripheral serum were negatively correlated with the performance of TMT-A and TMT-B in PD patients. Regression analysis showed that both GDNF and HVA were influence factors for TMT performance. A meta-analysis by Yuan and Raz (2014) showed that the prefrontal cortex (especially the lateral prefrontal cortex) was involved in the regulation of functional tests such as trial making tests. In this study, we proposed that there were two possible causes of EF impairment in PD patients. On one hand, primary PD is usually accompanied by EDF similar to that in patients with frontal lobe injury (Christopher and Strafella, 2013; Mak et al., 2014), especially the degeneration of the dorsolateral prefrontal cortex and its closely related dorsal striatum dopamine system (O’Callaghan et al., 2014; Koshimori et al., 2015). On the other hand, we suspected that GDNF might be involved in the development of EDF possibly through the following mechanisms: (1) GDNF indirectly affect EF through its protective and repair effects on DANs. GDNF has a protective effect on midbrain DANs. GDNF prevents the loss of DANs and the degeneration of dopaminergic nerve terminals, and increase the level of dopamine in the striatum (Olanow et al., 2015). (2) GDNF antagonizes aging-related cognitive decline through an unknown mechanism. It antagonizes neuronal degeneration through antioxidant stress, inhibition of microglial excitability, and neuroinflammation (Pertusa et al., 2008; Budni et al., 2015). (3) More consistent evidence proved that brain-derived neurotrophic factor and GDNF participate in the pathophysiological process of PD (Nagatsu and Sawada, 2007). Therefore, GDNF might affect the cognitive process of PD patients through a yet unknown mechanism.

Under our experimental conditions, this study found that the decrease of serum GDNF concentration in PD patients was associated with inhibitory control, cognitive flexibility, and attention performances. The changes of GDNF and HVA might synergistically participate in the occurrence and development of executive dysfunction in PD patients. Of course, due to the limitation of funds and other conditions, our research also has some shortcomings. We did not use objective cognitive assessment tools, such as functional MRI or PET scanning, to evaluate EF. Our study was a cross-sectional study rather than a cohort study and without follow-up study. All subjects were limited to the affiliated hospital of Xuzhou Medical University, and no other multicenter study data. In the future, more in-depth studies will be carried out combining animal models, cell experiments, functional imaging methodologies and multicenter prospective cohort study to explore its possible mechanism. Although there are a few limitations in this study, we have solved some problems in this field. All in all, GDNF might be used as a biomarker to identify early cognitive decline in PD, and might serve in clinical practice.

The original contributions presented in this study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors.

This study was approved by the Ethics Committee of the Affiliated Hospital of Xuzhou Medical University (approval no: XYFY2017-KL047-01). Written informed consents were signed by all subjects for enrollment. The patients/participants provided their written informed consent to participate in this study.

CR and D-SG: conception, design, and administrative support. S-YT, R-WW, and QL: provision of study materials. S-YT, R-WW, QL, YL, X-YY, and D-QG: collection and assembly of data. S-YT, D-SG, and CR: data analysis and interpretation. All authors manuscript writing and final approval of manuscript.