Neuroinflammation in levodopa-induced dyskinesia in Parkinson’s disease: current state of the art

Levodopa-induced dyskinesia (LID) remains a major limitation of long-term dopaminergic therapy in Parkinson’s disease (PD). Although LID is classically attributed to maladaptive corticostriatal plasticity driven by pulsatile dopaminergic stimulation in the dopamine-denervated striatum, accumulating evidence indicates that chronic neuroinflammation may lower the threshold for LID development and expression. Here, we synthesize preclinical, human biomarkers, neuroimaging, and genetic findings that support neuroinflammation as a mechanistic bridge between cellular stress and circuit-level maladaptation. In animal models, pharmacological modulation of inflammation-linked pathways, particularly nitrergic signaling, antibiotic repurposing, and cannabinoid-based strategies, attenuates abnormal involuntary movements and reduces striatal glial reactivity, implicating an inflammation–plasticity coupling in LID. In humans, post-mortem studies implicate regionally constrained mitochondrial DNA abnormalities and lipid dysregulation in the striatum, while biofluid metabolomics supports altered nitric oxide generation and lipid pathways associated with LID severity. Neuroimaging studies reveal markers consistent with oxidative stress and neurovascular dysfunction, and genetic studies link LID risk to variants in genes involved in lysosomal function, mitochondrial homeostasis, and inflammatory transcriptional programs. Integrated with the classic mechanisms of LID pathophysiology, we propose a conceptual model organized around three interrelated dimensions: (i) the topographical distribution of neuroinflammatory changes, particularly within the striatum; (ii) the interaction of neuroinflammation with mitochondrial dysfunction and lipid dysregulation; and (iii) the modifying influence of genetic susceptibility. This model motivates biomarker-guided, randomized trials targeting specific inflammatory nodes to prevent or mitigate LID.