Molecular estimation of neurodegeneration pseudotime in older brains

Abstract

The temporal molecular changes that lead to disease onset and progression in Alzheimer’s disease (AD) are still unknown. Here we develop a temporal model for these unobserved molecular changes with a manifold learning method applied to RNA-Seq data collected from human postmortem brain samples collected within the ROS/MAP and Mayo Clinic RNA-Seq studies. We define an ordering across samples based on their similarity in gene expression and use this ordering to estimate the molecular disease stage–or disease pseudotime-for each sample. Disease pseudotime is strongly concordant with the burden of tau (Braak score, P = 1.0 × 10−5), Aβ (CERAD score, P = 1.8 × 10−5), and cognitive diagnosis (P = 3.5 × 10−7) of late-onset (LO) AD. Early stage disease pseudotime samples are enriched for controls and show changes in basic cellular functions. Late stage disease pseudotime samples are enriched for late stage AD cases and show changes in neuroinflammation and amyloid pathologic processes. We also identify a set of late stage pseudotime samples that are controls and show changes in genes enriched for protein trafficking, splicing, regulation of apoptosis, and prevention of amyloid cleavage pathways. In summary, we present a method for ordering patients along a trajectory of LOAD disease progression from brain transcriptomic data.

Deep learning-based brain transcriptomic signatures associated with the neuropathological and clinical severity of Alzheimer’s disease

“…Among the six co-expression modules from the index genes, blue, brown, green and turquoise modules are significantly correlated with Alzheimer’s disease phenotypical hallmarks, with the first three upregulated in the progression, while turquoise genes downregulated. Together with the cell-type enrichment analysis showing the three modules are enriched in astrocytes, oligodendrocytes and endothelial cells, respectively, and turquoise module enriched in neurons, this is consistent with the results obtained by the recent work of neurodegeneration pseudotime estimation 5 and cellular composition deconvolution, 43 which shows a reduction in the neuronal populations as Alzheimer’s disease progresses, and an increase in expression associated with activation of endothelial and glial cells, as also demonstrated by the change of mean expression for the marker genes of each cell type along Alzheimer’s disease progression. Interestingly, the transcriptomic signatures obtained from our study show high similarity with the signatures obtained from a recent single-nucleus transcriptome analysis from the prefrontal cortical samples of Alzheimer’s disease patients and normal control subjects, 38 where higher proportion of endothelial nuclei was sampled and dysregulated pathways are associated with blood vessel morphogenesis, angiogenesis and antigen presentation.…”

“…Notably, these functions are also implicated in the top common blood–brain functional pathways relevant for LOAD progression in the recent study of gene expression trajectories in Alzheimer’s disease. 5 In addition, the enriched functions of proton transport implicated in mitochondrial functions and cell signalling pathway in the turquoise module were also found to be associated with the overlapped DEGs in neurons from two independent single nuclei transcriptomic studies from Alzheimer’s disease patients. 38 , 44 Strikingly, although the brown module is mostly overlapped with the consensus module cluster B from the AMP-AD transcriptome meta-analysis 23 ( Fig.…”

“…This further exacerbates the challenge we face in studying LOAD or dementia in general as a continuous spectrum to find novel biomarkers and drug targets. Recent efforts have begun to model Alzheimer’s disease progression as a continuous trajectory using cross-sectional transcriptomic data, 4 , 5 by leveraging the methods developed in single-cell genomics 6 , 7 and machine learning. 8 Iturria-Medina et al 4 adopted an unsupervised machine learning algorithm applied to gene expression microarray data and discovered a contrastive trajectory in multiple cohorts, respectively.…”

“…The trajectory has been demonstrated to strongly predict neuropathological severity in Alzheimer’s disease in each data set. Mukherjee et al 5 applied a manifold-learning method to RNA-seq data to define ordering across samples based on gene expression similarity and estimate the disease pseudotime for each sample. Disease pseudotime was strongly correlated with the burden of Aβ, tau and cognitive dementia within subjects with LOAD.…”

System-Level Analysis of Alzheimer’s Disease Prioritizes Candidate Genes for Neurodegeneration

“…While we do not see obvious subgroups in the HV/AV data (Figure 2), it is possible that by paring MRI with other phenotypic measures, such groups could appear. Future multi-trait analyses could have greater power to detect risk factors for patient subgroups, such as those that have been detected in gene expression data (Mukherjee et al, 2020). In particular, with emerging longitudinal data, it may become possible to identify subgroups that have distinct disease trajectories.…”