Amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD) are regarded as two ends of a pathogenetic spectrum, termed ALS-frontotemporal spectrum disorder (ALS-FTSD). However, it is currently difficult to predict where on the spectrum an individual will lie, especially for patients with C9orf72 hexanucleotide repeat expansions (HRE), a mutation associated with both ALS and FTD. It has been shown that both inflammation and protein misfolding influence aspects of ALS and ALS-FTSD disease pathogenesis, such as the manifestation or severity of motor or cognitive symptoms. Previous studies have highlighted markers which may influence C9orf72-associated disease presentation in a targeted fashion, though there has yet to be a systematic and quantitative assessment of common immunohistochemical markers to investigate the significance of these pathways in an unbiased manner. Here we report the first extensive digital pathological assessment with random forest modelling of pathological markers often used in neuropathology practice. This study profiles glial activation and protein misfolding in a cohort of deeply clinically profiled post-mortem tissue from patients with a C9orf72 HRE, who either met the criteria for a diagnosis of ALS or ALS-FTSD. We show that microglial immunohistochemical staining features, both morphological and spatial, are the best independent classifiers of disease status and that clinicopathological associations exist between microglial activation status and cognitive dysfunction in ALS-FTSD patients with C9orf72 HRE. Furthermore, we show that spatially resolved changes in FUS staining are also an accurate predictor of disease status, implying that liquid-liquid phase shift of this aggregation-prone RNA-binding protein may be important in ALS caused by a C9orf72 HRE. Our findings provide further support to the hypothesis of dysfunctional immune regulation and proteostasis in the pathogenesis of C9orf72 ALS and provide a framework for digital analysis of commonly used neuropathological stains as a tool to enrich our understanding of clinicopathological associations between cohorts.
Mapping Aboveground Biomass of Four Typical Vegetation Types in the Poyang Lake Wetlands Based on Random Forest Modelling and Landsat Images