Does Parkinson’s disease (PD) start in the brain or the gut?

Does Parkinson’s disease (PD) start in the brain or the gut?

In a new contribution published in the Journal of Parkinson’s Disease, scientists hypothesize that PD can be divided into two subtypes: gut-first, originating in the peripheral nervous system (PNS) of the gut and spreading to the brain; and brain-first, originating in the brain, or entering the brain via the olfactory system, and spreading to the brainstem and peripheral nervous system.

Per Borghammer, MD, PhD, DMSc, Nuclear Medicine & PET, Aarhus University Hospital, Denmark, and Nathalie Van Den Berge, MSc, PhD, Department of Clinical Medicine, Aarhus University, Denmark, explore the origins of PD onset.

They review the evidence that Lewy body disorders (LBD), including PD and dementia with Lewy bodies (DLB), comprise two distinct subtypes:

(1) a gut-first phenotype in which marked damage to the peripheral autonomic nervous system precedes measurable damage to the brain itself, including the dopamine cells; and

(2) a brain-first phenotype, in which marked damage to the brain precedes measurable damage to the peripheral autonomic nervous system.

“Evidence from autopsy studies of brains from PD patients has suggested that PD may start in the peripheral nervous system of the gut and nose.

The pathology then spreads via the nerves into the brain. However, not all autopsy studies agree with this interpretation,” explained Dr. Borghammer. “

In some cases, the brains do not contain pathology at the important ‘entry points’ into the brain, such as the dorsal vagus nucleus at the bottom of the brainstem.

The gut-first versus brain-first hypothesis posited in this review provides a scenario that can reconcile these discrepant findings from the neuropathological literature into one single coherent theory about the origins of PD.”

“The discussion about the origins of PD is often framed as an ‘either-or’, i.e., either all PD cases start in the gut or all cases start in the brain,” added Dr. Van Den Berge.

“However, much of the evidence seems compatible with both these interpretations. Thus, we need to entertain the possibility that both scenarios are actually true.”

Leading experts in Parkinson’s disease explore the hypothesis that Parkinson’s disease can be divided into two subtypes: brain-first, and gut-first. The image is credited to Lundbeck Foundation.

The review summarizes existing evidence from imaging studies from humans and tissue studies from humans and animal models.

The imaging and histology studies are generally compatible with the brain-first vs body-first hypothesis.

If this hypothesis is correct, it suggests that PD is more complicated than originally thought.

If the disease starts in the gut in only a fraction of patients, it is likely that interventions targeting the gut might only be effective for some PD patients, but not for individuals in which the disease starts in the brain itself.

“If the brain-first vs body-first hypothesis is correct, we need to intensify the research into understanding risk factors and triggering factors for these two subtypes,” noted Dr. Van Den Berge.

“It is probable that these different types of PD need different treatment strategies,” commented Dr. Borghammer.

“It may be possible to prevent the ‘gut-first’ type of PD through interventions targeting the gut, such as probiotics, fecal transplants, and anti-inflammatory treatments.

However, these strategies might not work with respect to treating and preventing the brain-first type.

Thus, a personalized treatment strategy will be required, and we need to be able to identify these subtypes of PD in the individual patient.”

PD is a slowly progressive disorder that affects movement, muscle control and balance. It is the second most common age-related neurodegenerative disorder affecting about three percent of the population by the age of 65 and up to five percent of individuals over 85 years of age.

During the 20th century, PD was thought to be primarily a brain disorder characterized mainly by loss of pigmented dopaminergic neurons residing in the substantia nigra, a basal ganglia structure located in the midbrain that plays an important role in reward and movement.

More recently, it has become clear that PD is highly varied and probably consists of several subtypes.

---

GASTROINTESTINAL PATHOLOGY AND ITS ROLE IN PARKINSON’S DISEASE ETIOLOGY

The first descriptions of alpha-synuclein (asyn) histopathology in the enteric nervous system (ENS) and peripheral autonomic ganglia date back as far as 1960 [1, 2], and several groups have now confirmed and extended these initial findings [3, 4]. By contrast to the substantia nigra, no overt neuronal loss is observed in the ENS in Parkinson’s disease (PD) [5, 6].

Based on early pathology in the olfactory bulb and dorsal motor nucleus of the vagus, the dual-hit hypothesis posited by Braak and colleagues proposes that initial misfolding and aggregation of asyn may occur in peripheral nerve terminals with subsequent centripetal spreading through the vagus nerve [7, 8].

This proposition gained support from two epidemiological studies showing that full truncal but not selective vagotomy seems to decrease the risk of PD [9, 10].

Also, in studies of archived tissue removed from patients up to 20 years prior to PD diagnosis, asyn pathology in the gut was reported more frequently than in matched controls [11, 12]. Several animal models, recently reviewed in detail [13], have shown cell-to-cell prion-like transmission and centripetal spreading of asyn [14, 15].

Such studies demonstrate the mechanistic plausibility of gut-to-brain transmission of asyn pathology. Nevertheless, this gut-to-brain transmission hypothesis is still widely debated and the arguments in favor or against the gut as the site of origin of PD were reviewed in two recent publications [16, 17].

The strongest argument against such a scenario comes from findings from the Arizona Parkinson’s consortium group, which did not find a single case in which asyn pathology was present in the ENS but not in the CNS in their wide survey of 466 whole-body autopsy cases [18].

This whole-body survey does not exclude, however, that initial asyn aggregation may start in the gut, but such hypothetical initial gut pathology would have to be be highly localized, or alternatively consist of immature and potentially protease-sensitive micro-aggregates, which are difficult to detect by standard immunohistochemistry [16].

In the coming two decades, it is likely that major advances will be made in understanding the role of gastrointestinal asyn pathology in the etiology of PD. An important research goal is to elucidate the degree of similarity between pathophysiological processes in PD and those of true prion diseases. In variant Creutzfeldt-Jakob disease (CJD), there is little doubt about the occurrence of gut-to-brain propagation of prions [19].

The aggregates of disease-related prion protein in variant CJD and bovine spongiform encephalopathy initially form in peripheral lymphoid tissues and in the ENS and then spread through the autonomic nervous system [20, 21].

Importantly, the prion aggregates in peripheral organs of variant CJD show clear differences in glycoform ratios and protease resistance compared to prions in brain tissues [22, 23]. It has therefore been suggested that the long latency between peripheral lymphoid colonization and subsequent neuro-invasion is in part due to the need for selection of a neuro-invasive strain of prion [24].

Such mechanisms have not been considered in the context of PD, but may explain why asyn multimers or aggregates harvested in peripheral autonomic nervous tissues behave differently than those from the substantia nigra of PD patients [25, 26].

There is therefore an urgent need to better characterize the biochemical characteristics of both native and pathological asyn in the ENS (Fig. 1).

This can best be achieved by performing a comprehensive inventory of synuclein forms present in the ENS from PD patients using two-dimensional electrophoretic analysis and mass spectroscopy or luminescent conjugated oligothiopenes, as such approaches have proven successful for characterizing the dominant pathological modification of synuclein and amyloids in the brain [27, 28].

Also, much work needs to be done within the field of animal PD models in the coming years. Although the concept of species barrier, i.e. the inability of some prion strains to cause disease in other species, is well described in prion disease [29, 30], it has yet received little attention in PD.

It has been shown, however, that sequence homology between asyn and the host protein is proportional to the rate of seeding initiation, e.g., human asyn fibrils efficiently seed monomeric human asyn, while seeding of mouse asyn is inefficient (reviewed in [31]).

In addition, the importance of aging and other modulatory factors including intestinal hyperpermeability, concurrent inflammation, and microbiota alterations needs to be studied in detail before firm conclusions can be drawn (Fig. 1).

Prion pathology in variant CJD exhibit peripheral-to-central propagation, whereas the initial pathology in sporadic and familial CJD cases probably arises spontaneously within the brain. Thus, it is theoretically possible that synucleinopathies can similarly be divided into types either originating in the peripheral nervous system or arising spontaneously within the CNS [16].

Note that this does not necessarily imply that these synucleinopathies were contracted by exogenous agents, but can be explained by Braak’s hypothesis that the nerve terminals of long, non-myelinated, hyperbranched axons are the most probable (but not only) place for the first occurrence of asyn aggregation [7].

One candidate for a peripheral-dominant subtype of PD may be those cases, in whom rapid eye movement sleep behavior disorder (RBD) presents during the prodromal phase. It has been shown that PD patients with RBD exhibit much higher frequencies of phosphorylated asyn pathology in the colon and in the skin compared to PD patients without RBD [32].

Also, idiopathic RBD patients exhibit marked pathology in the sympathetic and parasympathetic nervous system, but a relatively intact dopamine system [33].

In contrast, PD patients without RBD often have normal cardiac sympathetic innervation in early disease stages indicative of a sparing of the autonomic nervous system [34, 35].

An important research goal for the coming years will be to perform a thorough phenotypic characterization of PD subgroups using objective markers of neuronal dysfunction. It urgently needs to be clarified whether such distinct phenotypes can be explained by molecular characteristics of pathological asyn in the ENS and CNS (Fig. 1).

An attractive hypothesis is that some specific post-translationally modified forms of asyn and/or assemblies (strains) could be associated with a more aggressive disease progression [36].

BIOMARKERS

The observation that asyn inclusions are found in ENS of the vast majority of PD patients has led to an increasing number of immunohistochemistry-based studies aimed at developing biomarkers for disease diagnosis and progression. However, the diversity of methodology between studies, especially regarding the immunohistochemical methods used, has led to conflicting results regarding the sensitivity and specificity of gastrointestinal biopsies for the detection of asyn deposits [37].

Although several studies showed a high sensitivity of formalin-fixed paraffin embedded (FFPE) gastrointestinal biopsies for the detection of asyn inclusions in PD patients [38–41], other reports have raised concerns regarding the specificity of this approach because asyn immunoreactivity was also observed in some healthy individuals [42–45]. This lack of specificity, which likely results from technical difficulties inherent in the use of FFPE samples for the detection of asyn [46], explains why FFPE gastrointestinal biopsies are not currently used in routine clinical practice as PD biomarker.

Classical biochemical approaches, such as one- and two-dimensional electrophoresis did not prove to be more efficient than immunohistochemistry for detecting aggregated asyn in gastrointestinal biopsies [47].

As a whole, these disappointing results might be explained by the low amount of aggregated/pathological asyn that is usually found in the gastrointestinal tract and especially in gastrointestinal biopsies [48].

In order to overcome this limitation, future research could take benefit from ultrasensitive techniques such as protein misfolding cyclic amplification (PMCA) and real-time quaking-induced conversion (RT-QuIC) assays, which have been recently shown to efficiently amplify aggregated asyn in cerebrospinal fluid and in formalin-fixed archived tissue [49, 50].

In addition, there is growing evidence that pathological changes in the gut in PD are not limited to enteric neurons but also involve the enteric glial cells and the intestinal epithelial cells [51, 52], two cell types that largely outnumber enteric neurons and are easily captured by routine gastrointestinal biopsies.

These observations support the hypothesis that this so-called ‘neuro-glio-epithelial unit’ [53] might constitute an unparalleled source of biomarkers in PD beyond the sole assessment of asyn deposits/aggregates. Possible strategies to identify new enteric markers of PD might include a joint transcriptomic and proteomic analysis of the biopsies along with the analysis of biopsy supernatant [54].

Finally, the characterization of gastrointestinal dysfunction in patients has mostly relied on scoring of subjective symptoms, which is necessarily unable to capture subclinical disease and often shows poor correlation with markers of objective dysfunction [55].

Accessible and affordable methods such as radio-opaque markers to assess gastrointestinal transit times will find more widespread use in future studies. Recently, ¹¹C-donepezil PET scans have demonstrated decreased cholinergic signal in the gastrointestinal tract of both manifest and prodromal PD populations [33, 56].

Also, once an asyn-specific PET ligand becomes available it may be possible to image not only the brain but also peripheral organs and investigate longitudinally the dual-hit hypothesis.

Although expensive, such techniques will be critical to improve our understanding of PD etiopathogenesis and may have potential as progression markers in upcoming trials of disease-modification.

GUT MICROBIOTA

The human microbiota consists of bacteria, archaea, protists, fungi, their respective viruses, and human viruses. Most research has been carried out on the bacterial component of the microbiota [57]. The number of genes encoded in the gut metagenome is approximately 150 times larger than that of the human genome [58].

Evidence is accumulating that there is an intense bidirectional interaction between the gut, its microbiota, and the brain, frequently referred to as the microbiota-gut-brain-axis, which has important implications for brain health [59].

Alterations of the gut microbiota composition in PD have been revealed in multiple case-control studies from diverse populations. While there have been variations in the reported results, in PD patients an increased relative abundance of bacteria from the genera Akkermansia, Lactobacillus, and Bifidobacterium and decreased abundances of Prevotella, Faecalibacterium, and Blautia have been reproducibly shown [60–64]. For Prevotella such reduction has also been observed in RBD patients [64]. Based on the attributed functional properties of these bacteria, such alterations could affect gut barrier integrity, short-chain fatty acid (SCFA) production, and inflammation.

This would be in line with reports of a leaky gut and reduced levels of SCFAs and lipopolysaccharide binding protein in PD patients [62, 65, 66].

However, it remains to be definitely proven that the gut is hyperpermeable in PD [52] (Fig. 1).

Also with respect to SCFAs, in particular butyrate, the effects on PD pathology are not unequivocally established, since beneficial and harmful effects have been reported from PD animal and in vitro models [67–70].

An interesting link between gut microbiota and asyn pathology could be cross-seeding of amyloid pathology induced by bacterial amyloid proteins such as curli [71].

Few studies have investigated the nasal and oral microbiota in PD. While some alterations were found in the oral microbiota, two studies failed to find alterations in the nasal compartment [64, 72]. Future research on other non-bacterial microbes such as viruses, fungi, and archaea may provide also valuable insights.

So far, human microbiome studies in PD have been carried out exclusively in medicated patients, except for one study that included also idiopathic RBD patients [64]. While the PD associated microbiome alterations have been confirmed in medication adjusted analyses, confounding effects cannot be excluded and have in fact been shown for COMT inhibitors [60, 73]. Another potential confounder is colonic dysmotility, which may independently affect microbiota composition [60, 74].

Several microbiota studies assessed constipation symptoms using questionnaires and adjusted for reported constipation in their analyses. However, questionnaires may underestimate the prevalence of objective colonic dysfunction and thus adjustment based on these responses may be insufficient [55].

It is not known, whether the observed microbiome changes play a causal role in the development of gut pathology in PD or whether they are rather a consequence of altered gut function. Also for the observed correlations between microbiome composition and motor [60, 64] and non-motor symptoms [75, 76] causality has not been established in humans.

However, observations that motor symptoms, neuroinflammation, asyn pathology, and gut motility can be modulated by manipulating the gut microbiota in transgenic asyn-overexpressing mice suggest that such causal influences are possible [68].

Whether gut microbiome alterations in PD are independent of medications and constipation can be addressed by comparing microbiota and objective assessments of gut function between healthy subjects and drug naïve prodromal and/or de novo patients.

To establish what mechanisms link microbiota alterations and PD, such studies should employ a multiomics approach involving metagenomic, metatranscriptomic, and metabolomics analyses in combination with an assessment of host factors such as gut biopsies, permeability studies, cytokine levels and host genotype (Fig. 1).

For such studies to succeed, multicenter consortia should collaborate to ensure sufficient cohort sizes and standardized methodology. Longitudinal study designs will eventually enable us to study the temporal relationship between microbiota changes, disease stages, and phenotypes. While such undertakings will be long-lasting and very cost-intensive, further work in animal models may be helpful to narrow down the spectrum of potentially relevant pathways that connect microbiota and PD disease mechanisms and to test interventions.

Taken together there is good reason to envision that gut microbiota may have important implications in the future diagnostic and therapeutic landscape of PD. Analysis of the composition and/or functional aspects of the microbiome combined with clinical and genetic host factors could become a crucial element of patient phenotyping allowing more individualized treatments for PD.

Using such biomarkers to select suitable patients for specifically tailored therapeutic interventions should improve odds for successful clinical trials of symptomatic or disease modifying treatments.

Based on encouraging observations from RBD subjects [64], such approaches could find applications already in the prodromal phase of PD. If disease specific microbiota profiles could be established, these would be a valuable tool for the differential diagnosis of parkinsonism [77].

Therapeutic applications based on the gut microbiome are possible through a range of approaches.

These include dietary interventions, application of beneficial bacteria (probiotics), substances that promote growth of beneficial bacteria (prebiotics), substances that eliminate harmful bacteria (antibiotics), and transfer of bacterial ecosystems (fecal microbiota transplantation).

Furthermore, a detailed understanding of microbiome-host-interactions in PD could identify new pathways that could be targeted using more traditional pharmacological approaches. Importantly, microbiome based therapies could be effective independently of the causative role of the microbiome in PD pathogenesis.

This is because the microbiome is a key regulator of neuroinflammation and could affect disease progression through modulation of microglia activity [78].

Furthermore, gut microbiota affect gut permeability and the local environment in the gut lumen and mucosa which could have an implication for the impact of environmental toxins and other PD initiating or perpetuating factors such as protein aggregation.

Also, beneficial symptomatic effects may be achieved, as has been demonstrated in a trial showing improvement of constipation in PD patients receiving a synbiotic formulation, i.e., probiotics in combination with prebiotic fiber [79].

Finally, the microbiome may also play a role in the pharmacologic treatment of PD, as has been demonstrated for Helicobacter pylori and small intestinal bacterial overgrowth [80]. This suggests that benefits from traditional treatments of PD, such as levodopa, might be enhanced by microbiome based interventions.

---

Source:
IOS Press
Media Contacts:
Diana Murray – IOS Press
Image Source:
The image is credited to Lundbeck Foundation.

Original Research: Open access
“Brain-First versus Gut-First Parkinson’s Disease: A Hypothesis”. Borghammer, Per and Van Den Berge, Nathalie.
Journal of Parkinson’s Disease doi:10.3233/JPD-191721.