Inter-relationship of the age of onset to the concept of brain-first versus body-first Parkinson's disease

Abstract

The subtyping of Parkinson's Disease (PD) into brain-first and body-first PD has a powerful neuropathological, neuroimaging and clinical basis, supported by most relevant subsequent studies that have examined its validity. Here, we put forward the idea that the previous classification into early and late onset PD may be related to this categorization. The mean age of motor onset in brain-first PD may be up to 10 years earlier than body-first PD. Early onset PD has features related to brain-first PD, including relative clinical and nigrostriatal neurodegeneration asymmetry and a relatively restricted motor phenotype. In fact, PD as described by James Parkinson, could represent both early onset and brain-first PD, accounting for the famous phrase “senses and intellect uninjured”. We suggest here that, at the population level, age of onset could be used as a proxy for brain-first vs body-first PD, notwithstanding the lack of a direct one-to-one correlation and of a clear dichotomy in early vs. late onset PD, which rather represents a continuum. This would enable large scale population studies into the underlying genetic and epidemiological basis of these presumed separate nosological entities. Along these lines, there are some indications of a divergent exposure and genetic basis in early vs. delayed onset PD, and body-first vs. body-first PD respectively. Thus, studies of the etiopathological basis of PD could examine data sets with clinical data limited to age of onset, keeping in mind that within the overall concept of sporadic PD there may be two qualitatively different disease processes.

Patients with Parkinson's disease (PD) do not have a homogeneous clinical presentation or disease course and it seems clear that multiple subtypes must exist. This clinical heterogeneity may reflect disparate underlying biological processes. This may be very important for the prediction of disease course, but also for the application of therapies, especially neuroprotective that aim to tackle the root of neurodegeneration. It is conceivable that different therapeutic modalities may be applied, depending on the particular disease subtype. Hence, there has been a long quest for identifying PD biological and clinical subtypes, two concepts that are almost certainly inter-related. ¹

A new concept has recently been introduced with respect to classifying Lewy Body (LB)-related idiopathic PD (iPD), which represents the large majority of PD cases at the neuropathological level, into two separate subtypes, body-first and brain-first. ² In the former, disease initiation, in particular alpha-synuclein deposition and consequent neuronal dysfunction and neurodegeneration, occurs in the gut and possibly other organs receiving autonomic innervation, such as the skin or kidney. ³ Irrespective of the origin site, the disease then propagates through the autonomic connectome to the lower brainstem, leading to the hallmarks of early involvement of the autonomic nervous system and early manifestation of REM Sleep Behavior Disorder (RBD). In brain-first PD, the disease process commonly initiates in the olfactory bulb or in the amygdala with faster involvement of the substantia nigra, while autonomic dysfunction and RBD typically develop after diagnosis due to the rostro-caudal direction of AS propagation. A wealth of clinical, neuroimaging and neuropathological data support this categorization.^4,5 However, classifying new-onset PD into one of these subtypes requires extensive evaluations, as questionnaires fail to capture accurately sub-clinical manifestations in early disease stages. ⁶

Earlier PD subtyping efforts were based on a priori assumptions defined empirically and included an age of onset categorization; ⁷ in most cases, a cut-off at age 50 was suggested, and the consensus has been that later age of onset patients have a more severe phenotype, involving non-motor networks, and that they progress more rapidly. ⁸ Later subtyping efforts used unbiased cluster analyses, and age of onset remained an important factor, as indicated by the separation into “rapidly progressing and old age at onset” and “slowly progressive and young age at onset” clusters identified. ⁹ Another cluster-based classification separates PD cases in two extremes, a diffuse malignant and a mild, motor-predominant, as well as an intermediate form. ¹⁰ Although in the original study, age of onset was not that different between subtypes, a subsequent neuropathological study using this subtyping showed that the diffuse malignant subtype had a 12-year later age of onset compared to the mild motor-predominant subtype. ¹¹ A feature that is consistently reported is that later age of onset patients exhibit a more symmetric presentation.^8,12 In agreement with this notion, DAT imaging in the PPMI cohort shows greater symmetry of dopaminergic degeneration in late compared to early onset PD. Notably, the contralateral to ipsilateral SBR ratio also correlated significantly with age of onset as a continuous variable. ¹³ Interestingly, if the natural decline of DAT signal with age is taken into account, the two groups did not differ at baseline DAT signal or its decline over time, suggesting that the purported accelerated neurodegeneration in late onset PD may be rather an issue of the presence of the disease in an aging brain and body; nevertheless, the more symmetric presentation and evolution suggests a qualitatively different disease process in late vs. early onset PD. ¹³

In all datasets, age of onset is defined by the time in life when motor symptoms of the disease emerge. In this sense, body-first PD appears to be diagnosed based on the emergence of motor symptoms many years later than brain-first PD.^5,11,14 Furthermore, body-first PD is more symmetric on DAT imaging,^15–17 and by nature of definition involves, earlier in the disease process, a more diffuse network beyond the nigrostriatal system, giving rise to a more extensive array of non-motor symptoms,^16,18 in analogy to late onset PD (and the diffuse malignant subtype). Thus, a picture emerges in which later age-of-onset can be seen as a proxy for body-first PD, and earlier age-of-onset for brain-first PD. This does not necessarily mean that the timing of onset of the underlying biological disease process is that different between body- and brain-first PD, given that age of onset is defined, as mentioned, solely by motor symptoms.

A related issue that pertains to age of onset and disease subtyping is that prodromal brain-first PD may not be well captured in neuropathological studies, where it would manifest as a form of incidental Lewy Body Disease, not only because it is likely of shorter duration, but also because neuropathological data sets generally correspond to individuals of advanced ages, above the age of 80. Accordingly, in such data sets, it would also be more difficult to discern this subtype in patients with established PD, given the likelihood of very long disease duration which would obscure the sites of original alpha-synuclein deposition. ¹⁹ In this sense, neurophysiological and imaging approaches with multimodal imaging in de novo PD patients, irrespective of age of onset, may be more representative of the true prevalence of this subtype.

It is helpful in this regard to ponder on the patients reported by James Parkinson in his original monograph. ²⁰ Although the age of onset is not that well defined in some cases and certainly may be overestimated by contemporary standards due to the lack of recognition of more subtle motor deficits, it appears to be between the ages of 40 and 60, with a mean of 51, thus falling more towards the early age of onset. Indeed, notwithstanding some very prominent non-motor symptoms reported in the monograph, such as constipation, which appear to have developed many years after disease onset and may in part be due to the immobility conferred by the disease in those times, this may account for the famous phrase “senses and intellect uninjured”. Naturally, the assumption that James Parkinson's cases may thus represent the brain-first subtype linked to early disease onset is speculative and based on limited information.

Given the cumbersome nature of investigations needed to accurately classify new onset PD as “brain-first” or “body-first”, we suggest that in large scale population studies age of onset could be used as a proxy for this classification. This in no way implies that there is a direct one-to-one correlation between the two classifications, but rather that a younger onset patient is more likely to be brain-first, as opposed to an older onset patient, thus representing a continuum that could be captured by statistical probability methods.

The usefulness of this concept is that age of onset is universally captured in all data sets worldwide. The more actionable part of this proposition is that it could aid in understanding pathogenetic mechanisms behind brain-first and body-first PD, in particular potential different genetic and epidemiological underpinnings, using large data sets with minimal clinical data. Consistent with this idea, work based on the Hellenic Biobank for Parkinson's Disease has shown that smoking is associated in a dose-dependent fashion with reduced risk of PD only in late onset disease, whereas the opposite occurs with coffee consumption, which is negatively associated only with early onset disease, suggesting disparate pathogenetic factor involvement dependent on age of onset ²¹ From a genetic point of view, well characterized idiopathic RBD (iRBD) subjects, the majority of whom represent prodromal body-first Lewy body disease and often have a later debut of motor symptoms, have an overlapping, but distinct genetic underpinning compared to PD patients without RBD, representing brain-first disease. Interestingly, PD patients with RBD display an in-between polygenic risk score compatible with this group being a mix of body- and brain-first cases defined by RBD having emerged before or after onset of parkinsonism, respectively. ²² Of related interest, studies suggest that genetic risk factors for PD may influence age of onset; ²³ for example, a polygenic risk score for mitochondrial function genes associates with later age of onset. ²⁴ Furthermore, gene-environment interactions may be involved and drive differences in age of onset, potentially leading to the two different disease subtypes. ²⁵

Ultimately, large scale studies in PD populations with clinical data limited to age of onset could be combined with more focused investigations into smaller well characterized populations where the stratification into body-first vs. brain-first PD can be performed, with the aim of arriving closer to the underlying pathogenetic basis of these qualitatively distinct nosological entities and of applying appropriate targeted therapeutics. It should be noted that this concept applies mainly to sporadic disease. In PD cases with a Mendelian genetic underpinning, the biological and clinical subtypes, as well as the age of onset, depend on the particular genetic defect, and how these map unto the brain- vs. body-first model remains to be determined.