Using blood transcriptome analysis for Alzheimer's disease diagnosis and patient stratification.

INTRODUCTION: Blood protein biomarkers demonstrate potential for Alzheimer's disease (AD) diagnosis. Limited studies examine the molecular changes in AD blood cells. METHODS: Bulk RNA-sequencing of blood cells was performed on AD patients of Chinese descent (n = 214 and 26 in the discovery and validation cohorts, respectively) with normal controls (n = 208 and 38 in the discovery and validation cohorts, respectively). Weighted gene co-expression network analysis (WGCNA) and deconvolution analysis identified AD-associated gene modules and blood cell types. Regression and unsupervised clustering analysis identified AD-associated genes, gene modules, cell types, and established AD classification models. RESULTS: WGCNA on differentially expressed genes revealed 15 gene modules, with 6 accurately classifying AD (areas under the receiver operating characteristics curve [auROCs] > 0.90). These modules stratified AD patients into subgroups with distinct disease states. Cell-type deconvolution analysis identified specific blood cell types potentially associated with AD pathogenesis. DISCUSSION: This study highlights the potential of blood transcriptome for AD diagnosis, patient stratification, and mechanistic studies. HIGHLIGHTS: We comprehensively analyze the blood transcriptomes of a well-characterized Alzheimer's disease cohort to identify genes, gene modules, pathways, and specific blood cells associated with the disease. Blood transcriptome analysis accurately classifies and stratifies patients with Alzheimer's disease, with some gene modules achieving classification accuracy comparable to that of the plasma ATN biomarkers. Immune-associated pathways and immune cells, such as neutrophils, have potential roles in the pathogenesis and progression of Alzheimer's disease.

A blood-based multi-pathway biomarker assay for early detection and staging of Alzheimer's disease across ethnic groups.

INTRODUCTION: Existing blood-based biomarkers for Alzheimer's disease (AD) mainly focus on its pathological features. However, studies on blood-based biomarkers associated with other biological processes for a comprehensive evaluation of AD status are limited. METHODS: We developed a blood-based, multiplex biomarker assay for AD that measures the levels of 21 proteins involved in multiple biological pathways. We evaluated the assay's performance for classifying AD and indicating AD-related endophenotypes in three independent cohorts from Chinese or European-descent populations. RESULTS: The 21-protein assay accurately classified AD (area under the receiver operating characteristic curve [AUC] = 0.9407 to 0.9867) and mild cognitive impairment (MCI; AUC = 0.8434 to 0.8945) while also indicating brain amyloid pathology. Moreover, the assay simultaneously evaluated the changes of five biological processes in individuals and revealed the ethnic-specific dysregulations of biological processes upon AD progression. DISCUSSION: This study demonstrated the utility of a blood-based, multi-pathway biomarker assay for early screening and staging of AD, providing insights for patient stratification and precision medicine. HIGHLIGHTS: The authors developed a blood-based biomarker assay for Alzheimer's disease. The 21-protein assay classifies AD/MCI and indicates brain amyloid pathology. The 21-protein assay can simultaneously assess activities of five biological processes. Ethnic-specific dysregulations of biological processes in AD were revealed.

Patient-specific Alzheimer-like pathology in trisomy 21 cerebral organoids reveals BACE2 as a gene dose-sensitive AD suppressor in human brain.

A population of more than six million people worldwide at high risk of Alzheimer's disease (AD) are those with Down Syndrome (DS, caused by trisomy 21 (T21)), 70% of whom develop dementia during lifetime, caused by an extra copy of ß-amyloid-(Aß)-precursor-protein gene. We report AD-like pathology in cerebral organoids grown in vitro from non-invasively sampled strands of hair from 71% of DS donors. The pathology consisted of extracellular diffuse and fibrillar Aß deposits, hyperphosphorylated/pathologically conformed Tau, and premature neuronal loss. Presence/absence of AD-like pathology was donor-specific (reproducible between individual organoids/iPSC lines/experiments). Pathology could be triggered in pathology-negative T21 organoids by CRISPR/Cas9-mediated elimination of the third copy of chromosome 21 gene BACE2, but prevented by combined chemical ß and γ-secretase inhibition. We found that T21 organoids secrete increased proportions of Aß-preventing (Aß1-19) and Aß-degradation products (Aß1-20 and Aß1-34). We show these profiles mirror in cerebrospinal fluid of people with DS. We demonstrate that this protective mechanism is mediated by BACE2-trisomy and cross-inhibited by clinically trialled BACE1 inhibitors. Combined, our data prove the physiological role of BACE2 as a dose-sensitive AD-suppressor gene, potentially explaining the dementia delay in ~30% of people with DS. We also show that DS cerebral organoids could be explored as pre-morbid AD-risk population detector and a system for hypothesis-free drug screens as well as identification of natural suppressor genes for neurodegenerative diseases.

Large-scale plasma proteomic profiling identifies a high-performance biomarker panel for Alzheimer's disease screening and staging.

INTRODUCTION: Blood proteins are emerging as candidate biomarkers for Alzheimer's disease (AD). We systematically profiled the plasma proteome to identify novel AD blood biomarkers and develop a high-performance, blood-based test for AD. METHODS: We quantified 1160 plasma proteins in a Hong Kong Chinese cohort by high-throughput proximity extension assay and validated the results in an independent cohort. In subgroup analyses, plasma biomarkers for amyloid, tau, phosphorylated tau, and neurodegeneration were used as endophenotypes of AD. RESULTS: We identified 429 proteins that were dysregulated in AD plasma. We selected 19 "hub proteins" representative of the AD plasma protein profile, which formed the basis of a scoring system that accurately classified clinical AD (area under the curve  = 0.9690-0.9816) and associated endophenotypes. Moreover, specific hub proteins exhibit disease stage-dependent dysregulation, which can delineate AD stages. DISCUSSION: This study comprehensively profiled the AD plasma proteome and serves as a foundation for a high-performance, blood-based test for clinical AD screening and staging.

Identification of genetic risk factors in the Chinese population implicates a role of immune system in Alzheimer's disease pathogenesis.

Alzheimer's disease (AD) is a leading cause of mortality among the elderly. We performed a whole-genome sequencing study of AD in the Chinese population. In addition to the variants identified in or around the APOE locus (sentinel variant rs73052335, P = 1.44 × 10-14), two common variants, GCH1 (rs72713460, P = 4.36 × 10-5) and KCNJ15 (rs928771, P = 3.60 × 10-6), were identified and further verified for their possible risk effects for AD in three small non-Asian AD cohorts. Genotype-phenotype analysis showed that KCNJ15 variant rs928771 affects the onset age of AD, with earlier disease onset in minor allele carriers. In addition, altered expression level of the KCNJ15 transcript can be observed in the blood of AD subjects. Moreover, the risk variants of GCH1 and KCNJ15 are associated with changes in their transcript levels in specific tissues, as well as changes of plasma biomarkers levels in AD subjects. Importantly, network analysis of hippocampus and blood transcriptome datasets suggests that the risk variants in the APOE, GCH1, and KCNJ15 loci might exert their functions through their regulatory effects on immune-related pathways. Taking these data together, we identified common variants of GCH1 and KCNJ15 in the Chinese population that contribute to AD risk. These variants may exert their functional effects through the immune system.

Fibrillation and molecular characteristics are coherent with clinical and pathological features of 4-repeat tauopathy caused by MAPT variant G273R.

Microtubule Associated Protein Tau (MAPT) forms proteopathic aggregates in several diseases. The G273R tau mutation, located in the first repeat region, was found by exome sequencing in a patient who presented with dementia and parkinsonism. We herein return to pathological examination which demonstrated tau immunoreactivity in neurons and glia consistent of mixed progressive supranuclear palsy (PSP) and corticobasal degeneration (CBD) features. To rationalize the pathological findings, we used molecular biophysics to characterize the mutation in more detail in vitro and in Drosophila. The G273R mutation increases the aggregation propensity of 4-repeat (4R) tau and alters the tau binding affinity towards microtubules (MTs) and F-actin. Tau aggregates in PSP and CBD are predominantly 4R tau. Our data suggest that the G273R mutation induces a shift in pool of 4R tau by lower F-actin affinity, alters the conformation of MT bound 4R tau, while increasing chaperoning of 3R tau by binding stronger to F-actin. The mutation augmented fibrillation of 4R tau initiation in vitro and in glial cells in Drosophila and showed preferential seeding of 4R tau in vitro suggestively causing a late onset 4R tauopathy reminiscent of PSP and CBD.

Fulminant corticobasal degeneration: a distinct variant with predominant neuronal tau aggregates.

Corticobasal degeneration typically progresses gradually over 5-7 years from onset till death. Fulminant corticobasal degeneration cases with a rapidly progressive course were rarely reported (RP-CBD). This study aimed to investigate their neuropathological characteristics. Of the 124 autopsy-confirmed corticobasal degeneration cases collected from 14 centres, we identified 6 RP-CBD cases (4.8%) who died of advanced disease within 3 years of onset. These RP-CBD cases had different clinical phenotypes including rapid global cognitive decline (N = 2), corticobasal syndrome (N = 2) and Richardson's syndrome (N = 2). We also studied four corticobasal degeneration cases with an average disease duration of 3 years or less, who died of another unrelated illness (Intermediate-CBD). Finally, we selected 12 age-matched corticobasal degeneration cases out of a cohort of 110, who had a typical gradually progressive course and reached advanced clinical stage (End-stage-CBD). Quantitative analysis showed high overall tau burden (p = 0.2) and severe nigral cell loss (p = 0.47) in both the RP-CBD and End-stage-CBD groups consistent with advanced pathological changes, while the Intermediate-CBD group (mean disease duration = 3 years) had milder changes than End-stage-CBD (p < 0.05). These findings indicated that RP-CBD cases had already developed advanced pathological changes as those observed in End-stage-CBD cases (mean disease duration = 6.7 years), but within a significantly shorter duration (2.5 years; p < 0.001). Subgroup analysis was performed to investigate the cellular patterns of tau aggregates in the anterior frontal cortex and caudate by comparing neuronal-to-astrocytic plaque ratios between six RP-CBD cases, four Intermediate-CBD and 12 age-matched End-stage-CBD. Neuronal-to-astrocytic plaque ratios of Intermediate-CBD and End-stage-CBD, but not RP-CBD, positively correlated with disease duration in both the anterior frontal cortex and caudate (p = 0.02). In contrast to the predominance of astrocytic plaques we previously reported in preclinical asymptomatic corticobasal degeneration cases, neuronal tau aggregates predominated in RP-CBD exceeding those in Intermediate-CBD (anterior frontal cortex: p < 0.001, caudate: p = 0.001) and End-stage-CBD (anterior frontal cortex: p = 0.03, caudate: p = 0.01) as demonstrated by its higher neuronal-to-astrocytic plaque ratios in both anterior frontal cortex and caudate. We did not identify any difference in age at onset, any pathogenic tau mutation or concomitant pathologies that could have contributed to the rapid progression of these RP-CBD cases. Mild TDP-43 pathology was observed in three RP-CBD cases. All RP-CBD cases were men. The MAPT H2 haplotype, known to be protective, was identified in one RP-CBD case (17%) and 8 of the matched End-stage-CBD cases (67%). We conclude that RP-CBD is a distinct aggressive variant of corticobasal degeneration with characteristic neuropathological substrates resulting in a fulminant disease process as evident both clinically and pathologically. Biological factors such as genetic modifiers likely play a pivotal role in the RP-CBD variant and should be the subject of future research.

Variation at the TRIM11 locus modifies progressive supranuclear palsy phenotype.

OBJECTIVE: The basis for clinical variation related to underlying progressive supranuclear palsy (PSP) pathology is unknown. We performed a genome-wide association study (GWAS) to identify genetic determinants of PSP phenotype. METHODS: Two independent pathological and clinically diagnosed PSP cohorts were genotyped and phenotyped to create Richardson syndrome (RS) and non-RS groups. We carried out separate logistic regression GWASs to compare RS and non-RS groups and then combined datasets to carry out a whole cohort analysis (RS = 367, non-RS = 130). We validated our findings in a third cohort by referring to data from 100 deeply phenotyped cases from a recent GWAS. We assessed the expression/coexpression patterns of our identified genes and used our data to carry out gene-based association testing. RESULTS: Our lead single nucleotide polymorphism (SNP), rs564309, showed an association signal in both cohorts, reaching genome-wide significance in our whole cohort analysis (odds ratio = 5.5, 95% confidence interval = 3.2-10.0, p = 1.7 × 10-9 ). rs564309 is an intronic variant of the tripartite motif-containing protein 11 (TRIM11) gene, a component of the ubiquitin proteasome system (UPS). In our third cohort, minor allele frequencies of surrogate SNPs in high linkage disequilibrium with rs564309 replicated our findings. Gene-based association testing confirmed an association signal at TRIM11. We found that TRIM11 is predominantly expressed neuronally, in the cerebellum and basal ganglia. INTERPRETATION: Our study suggests that the TRIM11 locus is a genetic modifier of PSP phenotype and potentially adds further evidence for the UPS having a key role in tau pathology, therefore representing a target for disease-modifying therapies. Ann Neurol 2018;84:485-496.

C9orf72 intermediate repeats are associated with corticobasal degeneration, increased C9orf72 expression and disruption of autophagy.

Microsatellite repeat expansion disease loci can exhibit pleiotropic clinical and biological effects depending on repeat length. Large expansions in C9orf72 (100s-1000s of units) are the most common genetic cause of amyotrophic lateral sclerosis (ALS) and frontotemporal degeneration (FTD). However, whether intermediate expansions also contribute to neurodegenerative disease is not well understood. Several studies have identified intermediate repeats in Parkinson's disease patients, but the association was not found in autopsy-confirmed cases. We hypothesized that intermediate C9orf72 repeats are a genetic risk factor for corticobasal degeneration (CBD), a neurodegenerative disease that can be clinically similar to Parkinson's but has distinct tau protein pathology. Indeed, intermediate C9orf72 repeats were significantly enriched in autopsy-proven CBD (n = 354 cases, odds ratio = 3.59, p = 0.00024). While large C9orf72 repeat expansions are known to decrease C9orf72 expression, intermediate C9orf72 repeats result in increased C9orf72 expression in human brain tissue and CRISPR/cas9 knockin iPSC-derived neural progenitor cells. In contrast to cases of FTD/ALS with large C9orf72 expansions, CBD with intermediate C9orf72 repeats was not associated with pathologic RNA foci or dipeptide repeat protein aggregates. Knock-in cells with intermediate repeats exhibit numerous changes in gene expression pathways relating to vesicle trafficking and autophagy. Additionally, overexpression of C9orf72 without the repeat expansion leads to defects in autophagy under nutrient starvation conditions. These results raise the possibility that therapeutic strategies to reduce C9orf72 expression may be beneficial for the treatment of CBD.

Cognitive markers of preclinical and prodromal Alzheimer's disease in Down syndrome.

INTRODUCTION: Down syndrome (DS) is associated with an almost universal development of Alzheimer's disease. Individuals with DS are therefore an important population for randomized controlled trials to prevent or delay cognitive decline, though it is essential to understand the time course of early cognitive changes. METHODS: We conducted the largest cognitive study to date with 312 adults with DS to assess age-related and Alzheimer's disease-related cognitive changes during progression from preclinical to prodromal dementia, and prodromal to clinical dementia. RESULTS: Changes in memory and attention measures were most sensitive to early decline. Resulting sample size calculations for randomized controlled trials to detect significant treatment effects to delay decline were modest. DISCUSSION: Our findings address uncertainties around the development of randomized controlled trials to delay cognitive decline in DS. Such trials are essential to reduce the high burden of dementia in people with DS and could serve as proof-of-principle trials for some drug targets.

Additional rare variant analysis in Parkinson's disease cases with and without known pathogenic mutations: evidence for oligogenic inheritance.

Oligogenic inheritance implies a role for several genetic factors in disease etiology. We studied oligogenic inheritance in Parkinson's (PD) by assessing the potential burden of additional rare variants in established Mendelian genes and/or GBA, in individuals with and without a primary pathogenic genetic cause in two large independent cohorts totaling 7,900 PD cases and 6,166 controls. An excess (≥30%) of cases with a recognised primary genetic cause had ≥1 additional rare variants in Mendelian PD genes, as compared with no known mutation PD cases (17%) and unaffected controls (16%), supporting our hypothesis. Carriers of additional Mendelian gene variants have younger ages at onset (AAO). The effect of additional Mendelian variants in LRRK2 G2019S mutation carriers, of which ATP13A2 variation is particularly common, may account for some of the variation in penetrance. About 10% of No Known Mutation-PD cases harbour a rare GBA variant compared to known pathogenic mutation PD cases (8%) and controls (5%), with carriers having earlier AAOs. Together, the data suggest that the oligogenic inheritance of rare Mendelian variants may be important in patient with a primary pathogenic cause, whereas GBA increases risk across all forms of PD. This study highlights the potential genetic complexity of Mendelian PD. The identification of potential modifying variants provides new insights into disease mechanisms by potentially separating relevant from benign variants and by the interaction between genes in specific pathways. In the future this may be relevant to genetic testing and counselling of patients with PD and their families.

Astrogliopathy predominates the earliest stage of corticobasal degeneration pathology.

SEE KOBYLECKI AND MANN DOI101093/AWW267 FOR A SCIENTIFIC COMMENTARY ON THIS ARTICLE: Animal models have shown that tau seeding and propagation are strain- and neural network-specific. The study of preclinical cases is valuable to gain insights into early pathological features of corticobasal degeneration and its progression. Three preclinical corticobasal degeneration cases and six age-matched end-stage corticobasal degeneration cases were included in this study. Tau immunohistochemistry performed in 20 brain regions and quantitative assessment of regional tau load using image analysis were performed. Semi-quantitative grading of tau-positive cellular lesions and neuronal loss in the frontal, parietal and temporal cortices, striatum, substantia nigra and subthalamic nucleus were assessed. All preclinical cases were clinically asymptomatic but had widespread tau lesions in the typically affected regions in corticobasal degeneration and the pathognomonic astrocytic plaques were the most prominent lesion type in the anterior frontal and striatal regions. Mean total tau load (sum of all regional tau load) of end-stage corticobasal degeneration cases were nine times greater than that of the preclinical cases (P = 0.04) and less tau load was found in all regions of the preclinical cases. An anterior-to-posterior tau load ratio in the frontal cortex in preclinical cases was 12-fold greater than in end-stage corticobasal degeneration cases. Relatively greater tau burden in the anterior frontal cortex, striatum and subthalamic nucleus suggests the striatal afferent connection to the dorsolateral prefrontal cortex and basal ganglia circuitry are the earliest neural network connections affected by corticobasal degeneration-related tau pathology. Differential distribution of the tau pathology to selective cortical regions in these preclinical cases implies phenotypic presentation may be predetermined at a very early stage of the disease process. Neuronal loss of the substantia nigra was either absent or very mild in the preclinical cases and was moderate to severe in end-stage corticobasal degeneration cases (P < 0.05). Our findings suggest that a threshold of pathological burden in the 'right' anatomical regions needs to be reached before the onset of clinical symptoms. The early prominence of the astrocytic plaques in relation to sparse neuronal lesions leads one to speculate that corticobasal degeneration may begin as an astrogliopathy at a very early disease stage but neuronal lesions gradually take over as the predominant lesion type in advanced disease.

Parkinson's disease in GTP cyclohydrolase 1 mutation carriers.

GTP cyclohydrolase 1, encoded by the GCH1 gene, is an essential enzyme for dopamine production in nigrostriatal cells. Loss-of-function mutations in GCH1 result in severe reduction of dopamine synthesis in nigrostriatal cells and are the most common cause of DOPA-responsive dystonia, a rare disease that classically presents in childhood with generalized dystonia and a dramatic long-lasting response to levodopa. We describe clinical, genetic and nigrostriatal dopaminergic imaging ([(123)I]N-ω-fluoropropyl-2ß-carbomethoxy-3ß-(4-iodophenyl) tropane single photon computed tomography) findings of four unrelated pedigrees with DOPA-responsive dystonia in which pathogenic GCH1 variants were identified in family members with adult-onset parkinsonism. Dopamine transporter imaging was abnormal in all parkinsonian patients, indicating Parkinson's disease-like nigrostriatal dopaminergic denervation. We subsequently explored the possibility that pathogenic GCH1 variants could contribute to the risk of developing Parkinson's disease, even in the absence of a family history for DOPA-responsive dystonia. The frequency of GCH1 variants was evaluated in whole-exome sequencing data of 1318 cases with Parkinson's disease and 5935 control subjects. Combining cases and controls, we identified a total of 11 different heterozygous GCH1 variants, all at low frequency. This list includes four pathogenic variants previously associated with DOPA-responsive dystonia (Q110X, V204I, K224R and M230I) and seven of undetermined clinical relevance (Q110E, T112A, A120S, D134G, I154V, R198Q and G217V). The frequency of GCH1 variants was significantly higher (Fisher's exact test P-value 0.0001) in cases (10/1318 = 0.75%) than in controls (6/5935 = 0.1%; odds ratio 7.5; 95% confidence interval 2.4-25.3). Our results show that rare GCH1 variants are associated with an increased risk for Parkinson's disease. These findings expand the clinical and biological relevance of GTP cycloydrolase 1 deficiency, suggesting that it not only leads to biochemical striatal dopamine depletion and DOPA-responsive dystonia, but also predisposes to nigrostriatal cell loss. Further insight into GCH1-associated pathogenetic mechanisms will shed light on the role of dopamine metabolism in nigral degeneration and Parkinson's disease.

Genomewide association study in cervical dystonia demonstrates possible association with sodium leak channel.

Dystonia is a common movement disorder. A number of monogenic causes have been identified. However, the majority of dystonia cases are not explained by single gene defects. Cervical dystonia is one of the commonest forms without genetic causes identified. This pilot study aimed to identify large effect-size risk loci in cervical dystonia. A genomewide association study (GWAS) was performed. British resident cervical dystonia patients of European descent were genotyped using the Illumina-610-Quad. Comparison was made with controls of European descent from the Wellcome Trust Case Control Consortium using logistic regression algorithm from PLINK. SNPs not genotyped by the array were imputed with 1000 Genomes Project data using the MaCH algorithm and minimac. Postimputation analysis was done with the mach2dat algorithm using a logistic regression model. After quality control measures, 212 cases were compared with 5173 controls. No single SNP passed the genomewide significant level of 5 × 10(-8) in the analysis of genotyped SNP in PLINK. Postimputation, there were 5 clusters of SNPs that had P value <5 × 10(-6) , and the best cluster of SNPs was found near exon 1 of NALCN, (sodium leak channel) with P = 9.76 × 10(-7) . Several potential regions were found in the GWAS and imputation analysis. The lowest P value was found in NALCN. Dysfunction of this ion channel is a plausible cause for dystonia. Further replication in another cohort is needed to confirm this finding. We make this data publicly available to encourage further analyses of this disorder.

A robust model for read count data in exome sequencing experiments and implications for copy number variant calling.

MOTIVATION: Exome sequencing has proven to be an effective tool to discover the genetic basis of Mendelian disorders. It is well established that copy number variants (CNVs) contribute to the etiology of these disorders. However, calling CNVs from exome sequence data is challenging. A typical read depth strategy consists of using another sample (or a combination of samples) as a reference to control for the variability at the capture and sequencing steps. However, technical variability between samples complicates the analysis and can create spurious CNV calls. RESULTS: Here, we introduce ExomeDepth, a new CNV calling algorithm designed to control for this technical variability. ExomeDepth uses a robust model for the read count data and uses this model to build an optimized reference set in order to maximize the power to detect CNVs. As a result, ExomeDepth is effective across a wider range of exome datasets than the previously existing tools, even for small (e.g. one to two exons) and heterozygous deletions. We used this new approach to analyse exome data from 24 patients with primary immunodeficiencies. Depending on data quality and the exact target region, we find between 170 and 250 exonic CNV calls per sample. Our analysis identified two novel causative deletions in the genes GATA2 and DOCK8. AVAILABILITY: The code used in this analysis has been implemented into an R package called ExomeDepth and is available at the Comprehensive R Archive Network (CRAN).

Deep learning-based polygenic risk analysis for Alzheimer's disease prediction.

BACKGROUND: The polygenic nature of Alzheimer's disease (AD) suggests that multiple variants jointly contribute to disease susceptibility. As an individual's genetic variants are constant throughout life, evaluating the combined effects of multiple disease-associated genetic risks enables reliable AD risk prediction. Because of the complexity of genomic data, current statistical analyses cannot comprehensively capture the polygenic risk of AD, resulting in unsatisfactory disease risk prediction. However, deep learning methods, which capture nonlinearity within high-dimensional genomic data, may enable more accurate disease risk prediction and improve our understanding of AD etiology. Accordingly, we developed deep learning neural network models for modeling AD polygenic risk. METHODS: We constructed neural network models to model AD polygenic risk and compared them with the widely used weighted polygenic risk score and lasso models. We conducted robust linear regression analysis to investigate the relationship between the AD polygenic risk derived from deep learning methods and AD endophenotypes (i.e., plasma biomarkers and individual cognitive performance). We stratified individuals by applying unsupervised clustering to the outputs from the hidden layers of the neural network model. RESULTS: The deep learning models outperform other statistical models for modeling AD risk. Moreover, the polygenic risk derived from the deep learning models enables the identification of disease-associated biological pathways and the stratification of individuals according to distinct pathological mechanisms. CONCLUSION: Our results suggest that deep learning methods are effective for modeling the genetic risks of AD and other diseases, classifying disease risks, and uncovering disease mechanisms.

Polygenic diseases, such as Alzheimer's disease (AD), are those caused by the interplay between multiple genetic risk factors. Statistical models can be used to predict disease risk based on a person's genetic profile. However, there are limitations to existing methods, while emerging methods such as deep learning may improve risk prediction. Deep learning involves computer-based software learning from patterns in data to perform a certain task, e.g. predict disease risk. Here, we test whether deep learning models can help to predict AD risk. Our models not only outperformed existing methods in modeling AD risk, they also allow us to estimate an individual's risk of AD and determine the biological processes that may be involved in AD. With further testing and optimization, deep learning may be a useful tool to help accurately predict risk of AD and other diseases.

Multi-Omics-Based Autophagy-Related Untypical Subtypes in Patients with Cerebral Amyloid Pathology.

Recent multi-omics analyses paved the way for a comprehensive understanding of pathological processes. However, only few studies have explored Alzheimer's disease (AD) despite the possibility of biological subtypes within these patients. For this study, unsupervised classification of four datasets (genetics, miRNA transcriptomics, proteomics, and blood-based biomarkers) using Multi-Omics Factor Analysis+ (MOFA+), along with systems-biological approaches following various downstream analyses are performed. New subgroups within 170 patients with cerebral amyloid pathology (Aß+) are revealed and the features of them are identified based on the top-rated targets constructing multi-omics factors of both whole (M-TPAD) and immune-focused models (M-IPAD). The authors explored the characteristics of subtypes and possible key-drivers for AD pathogenesis. Further in-depth studies showed that these subtypes are associated with longitudinal brain changes and autophagy pathways are main contributors. The significance of autophagy or clustering tendency is validated in peripheral blood mononuclear cells (PBMCs; n = 120 including 30 Aß- and 90 Aß+), induced pluripotent stem cell-derived human brain organoids/microglia (n = 12 including 5 Aß-, 5 Aß+, and CRISPR-Cas9 apolipoprotein isogenic lines), and human brain transcriptome (n = 78). Collectively, this study provides a strategy for precision medicine therapy and drug development for AD using integrative multi-omics analysis and network modelling.

An IL1RL1 genetic variant lowers soluble ST2 levels and the risk effects of APOE-ε4 in female patients with Alzheimer's disease.

Changes in the levels of circulating proteins are associated with Alzheimer's disease (AD), whereas their pathogenic roles in AD are unclear. Here, we identified soluble ST2 (sST2), a decoy receptor of interleukin-33-ST2 signaling, as a new disease-causing factor in AD. Increased circulating sST2 level is associated with more severe pathological changes in female individuals with AD. Genome-wide association analysis and CRISPR-Cas9 genome editing identified rs1921622 , a genetic variant in an enhancer element of IL1RL1, which downregulates gene and protein levels of sST2. Mendelian randomization analysis using genetic variants, including rs1921622 , demonstrated that decreased sST2 levels lower AD risk and related endophenotypes in females carrying the Apolipoprotein E (APOE)-ε4 genotype; the association is stronger in Chinese than in European-descent populations. Human and mouse transcriptome and immunohistochemical studies showed that rs1921622 /sST2 regulates amyloid-beta (Aß) pathology through the modulation of microglial activation and Aß clearance. These findings demonstrate how sST2 level is modulated by a genetic variation and plays a disease-causing role in females with AD.