Brain-First versus Gut-First Parkinson's Disease: A Hypothesis

Abstract

Parkinson's disease (PD) is a highly heterogeneous disorder, which probably consists of multiple subtypes. Aggregation of misfolded alpha-synuclein and propagation of these proteinacious aggregates through interconnected neural networks is believed to be a crucial pathogenetic factor. It has been hypothesized that the initial pathological alpha-synuclein aggregates originate in the enteric or peripheral nervous system (PNS) and invade the central nervous system (CNS) via retrograde vagal transport. However, evidence from neuropathological studies suggests that not all PD patients can be reconciled with this hypothesis. Importantly, a small fraction of patients do not show pathology in the dorsal motor nucleus of the vagus. Here, it is hypothesized that PD can be divided into a PNS-first and a CNS-first subtype. The former is tightly associated with REM sleep behavior disorder (RBD) during the prodromal phase and is characterized by marked autonomic damage before involvement of the dopaminergic system. In contrast, the CNS-first phenotype is most often RBD-negative during the prodromal phase and characterized by nigrostriatal dopaminergic dysfunction prior to involvement of the autonomic PNS. The existence of these subtypes is supported by in vivo imaging studies of RBD-positive and RBD-negative patient groups and by histological evidence- reviewed herein. The present proposal provides a fresh hypothesis-generating framework for future studies into the etiopathogenesis of PD and seems capable of explaining a number of discrepant findings in the neuropathological literature.

Keywords: MRI; PET; Parkinson’s disease; alpha-synuclein; autonomic nervous system; dopamine; etiology; histology; imaging; prion-like.

Fig. 1

Schematic illustration of two hypothetical Lewy body disorder (LBD) phenotypes. The PNS-first phenotype is characterized by early, severe damage to the autonomic PNS. The asyn pathology mainly propagates retrogradely via autonomic connections to the medulla and brainstem. This phenotype is most often RBD-positive during the prodromal phase. The CNS-first phenotype is characterized by early, marked damage to CNS structures, including the substantia nigra, while the autonomic PNS is initially spared. The asyn pathology mainly propagates anterogradely from the CNS to the PNS. Patients are most often initially RBD-negative during the early motor phase.

Fig. 2

The cardiac sympathetic innervation imaged with ¹²³I-MIBG scintigraphy (top row) and the nigrostriatal dopamine storage capacity image by ¹⁸F-DOPA PET (bottom row) in a healthy control (left column), a patient with idiopathic RBD (iRBD; middle), and a de novo H&Y stage I PD patient without RBD (right). Note that the iRBD patient shows severely reduced sympathetic cardiac innervation but almost normal striatal dopaminergic storage capacity. In contrast, the de novo PD patient without RBD shows normal cardiac innervation but markedly reduced striatal dopamine storage capacity. These cases represent the hypothesized PNS-first and CNS-first phenotypes, respectively. [MIBG heart/mediastinum ratios are scaled arbitrarily. The FDOPA images are scaled from 1 to 3.5 times the cortical background intensity.

Fig. 3

Summary of imaging modalities used to measure relevant neuronal systems. The healthy control mean (blue) was set to 100% in each study, and the percentage reduction in the patient groups calculated. In cases with more than one study, the standard deviation is depicted by whiskers. Cardiac sympathetic innervation was measured with ¹²³I-MIBG scintigraphy; intestinal cholinergic (parasympathetic) innervation with ¹¹C-donepezil PET; integrity of pigmented locus coeruleus neurons with neuromelanin-sensitive MRI; nigrostriatal dopaminergic synaptic function with ¹²³I-FP-CIT SPECT or ¹⁸F-DOPA PET; cortical glucose metabolism with ¹⁸F-FDG and principal component analysis to quantify the PD related network (PDRP) z-score. The corresponding Braak stages are shown in the bottom row. Patients with idiopathic RBD (red) show marked loss of autonomic and locus coeruleus imaging parameters, but only minor dopaminergic terminal loss and only slight perturbation of cortical metabolism. The opposite pattern is seen for H&Y stage I-II PD patients. Concerning annotations to PD data: “HY” shows data from different H&Y stage data sets; “PD meta” shows the mean % reduction in PD patients’ DaTscan putamen binding from a metaanalysis [31], and “PD PPMI” the % putaminal reduction seen in early PD patients from the PPMI data set [35]. [Note that except for PDRP z-scores, the imaging parameters approximate loss of specific binding, i.e. heart/mediastinum ratio – 1 for MIBG; SUV – 1 for donepezil; locus coeruleus/pons – 1 for neuromelanin MRI; putamen/occipital cortex – 1 for FP-CIT & FDOPA. For didactic purposes the PDRP z-scores are listed as negative in this figure in contrast to common practice. See main text for study references.]

Fig. 4

A) Schematic illustration of the two hypothetical routes. The PNS-first (RBD-positive) patients initially show cardiac sympathetic denervation followed by secondary nigrostriatal dopaminergic denervation. The opposite temporal pattern is seen in the CNS-first (RBD-negative) phenotype. B) Red squares depict ¹⁸F-DOPA PET and ¹²³I-MIBG cardiac scintigraphy data from idiopathic RBD cases. Healthy control reference data are shown with grey squares (top right) and H&Y II–IV PD data with grey circles (bottom left). Vertical and horizontal lines denote cut-off thresholds based on the healthy reference data. [data from Knudsen et al. [26].

Fig. 5

Injection of asyn preformed fibrils (PFF) into the duodenum of transgenic rats leads to robust propagation of phosphorylated, aggregated asyn through the sympathetic nervous system via the celiac ganglion (CMG) to the IML, and via the vagus nerve to the DMV. Caudo-rostral propagation was seen in the brainstem with involvement of the LC and substantia nigra pars reticulata (SNr). Striking asyn pathology was seen in the sympathetic nerves of the myocardium indicative of anterograde propagation from the cervical ganglia (CG) of the sympathetic trunk (ST). Pathology was also seen in the ENS of the stomach several centimeters from the injection site indicative of anterograde propagation from the DMV or celiac ganglion. No pathological asyn was seen in control transgenic animals injected with phosphate-buffered saline (PBS) in the duodenum.