Biomarker-driven phenotyping in Parkinson's disease: A translational missing link in disease-modifying clinical trials

Abstract

Past clinical trials of putative neuroprotective therapies have targeted PD as a single pathogenic disease entity. From an Oslerian clinicopathological perspective, the wide complexity of PD converges into Lewy bodies and justifies a reductionist approach to PD: A single-mechanism therapy can affect most of those sharing the classic pathological hallmark. From a systems-biology perspective, PD is a group of disorders that, while related by sharing the feature of nigral dopamine-neuron degeneration, exhibit unique genetic, biological, and molecular abnormalities, which probably respond differentially to a given therapeutic approach, particularly for strategies aimed at neuroprotection. Under this model, only biomarker-defined, homogenous subtypes of PD are likely to respond optimally to therapies proven to affect the biological processes within each subtype. Therefore, we suggest that precision medicine applied to PD requires a reevaluation of the biomarker-discovery effort. This effort is currently centered on correlating biological measures to clinical features of PD and on identifying factors that predict whether various prodromal states will convert into the classical movement disorder. We suggest, instead, that subtyping of PD requires the reverse view, where abnormal biological signals (i.e., biomarkers), rather than clinical definitions, are used to define disease phenotypes. Successful development of disease-modifying strategies will depend on how relevant the specific biological processes addressed by an intervention are to the pathogenetic mechanisms in the subgroup of targeted patients. This precision-medicine approach will likely yield smaller, but well-defined, subsets of PD amenable to successful neuroprotection. © 2017 International Parkinson and Movement Disorder Society.

Keywords: Parkinson's disease; biomarkers; neuroprotection; systems biology.

Figure 1. Biomarker development models

A. Clinical phenotype → Biomarker validation. Under this model, clinical phenotypes are established as the “truth” against which biomarkers are validated. After defining the clinical constructs (e.g., PD-MCI), statistical analyses determine the strength of correlation with a variety of abnormal biological signals. In this example, the “green” signal becomes a biomarker of PD-MCI. B. Biological/molecular abnormalities → Biomarker-driven phenotypes. Under this model, the “truth” is unknown and the analysis is non-hypothesis driven (exploratory). Biological/molecular signals are examined within a large cohort of individuals, including criteria-meeting “PD” but also individuals with other PD-like diagnoses and those aging without overt disease. The biomarker-driven phenotypes (D₁, D₂…D_n, where D stands for “disease”) share strong biological homogeneity, even if not clinically, as they are generated based on abnormal biological signals (exemplified as >2 standard deviations above the mean for green and red signal; and >2 SD below the mean for purple and a theoretical gray “n” signal).