Identification of methylation-regulated genes modulating microglial phagocytosis in hyperhomocysteinemia-exacerbated Alzheimer's disease.

BACKGROUND: Hyperhomocysteinemia (HHcy) has been linked to development of Alzheimer's disease (AD) neuropathologically characterized by the accumulation of amyloid ß (Aß). Microglia (MG) play a crucial role in uptake of Aß fibrils, and its dysfunction worsens AD. However, the effect of HHcy on MG Aß phagocytosis remains unstudied. METHODS: We isolated MG from the cerebrum of HHcy mice with genetic cystathionine-ß-synthase deficiency (Cbs-/-) and performed bulk RNA-seq. We performed meta-analysis over transcriptomes of Cbs-/- mouse MG, human and mouse AD MG, MG Aß phagocytosis model, human AD methylome, and GWAS AD genes. RESULTS: HHcy and hypomethylation conditions were identified in Cbs-/- mice. Through Cbs-/- MG transcriptome analysis, 353 MG DEGs were identified. Phagosome formation and integrin signaling pathways were found suppressed in Cbs-/- MG. By analyzing MG transcriptomes from 4 AD patient and 7 mouse AD datasets, 409 human and 777 mouse AD MG DEGs were identified, of which 37 were found common in both species. Through further combinatory analysis with transcriptome from MG Aß phagocytosis model, we identified 130 functional-validated Aß phagocytic AD MG DEGs (20 in human AD, 110 in mouse AD), which reflected a compensatory activation of Aß phagocytosis. Interestingly, we identified 14 human Aß phagocytic AD MG DEGs which represented impaired MG Aß phagocytosis in human AD. Finally, through a cascade of meta-analysis of transcriptome of AD MG, functional phagocytosis, HHcy MG, and human AD brain methylome dataset, we identified 5 HHcy-suppressed phagocytic AD MG DEGs (Flt1, Calponin 3, Igf1, Cacna2d4, and Celsr) which were reported to regulate MG/MΦ migration and Aß phagocytosis. CONCLUSIONS: We established molecular signatures for a compensatory response of Aß phagocytosis activation in human and mouse AD MG and impaired Aß phagocytosis in human AD MG. Our discoveries suggested that hypomethylation may modulate HHcy-suppressed MG Aß phagocytosis in AD.

Novel Knowledge-Based Transcriptomic Profiling of Lipid Lysophosphatidylinositol-Induced Endothelial Cell Activation.

To determine whether pro-inflammatory lipid lysophosphatidylinositols (LPIs) upregulate the expressions of membrane proteins for adhesion/signaling and secretory proteins in human aortic endothelial cell (HAEC) activation, we developed an EC biology knowledge-based transcriptomic formula to profile RNA-Seq data panoramically. We made the following primary findings: first, G protein-coupled receptor 55 (GPR55), the LPI receptor, is expressed in the endothelium of both human and mouse aortas, and is significantly upregulated in hyperlipidemia; second, LPIs upregulate 43 clusters of differentiation (CD) in HAECs, promoting EC activation, innate immune trans-differentiation, and immune/inflammatory responses; 72.1% of LPI-upregulated CDs are not induced in influenza virus-, MERS-CoV virus- and herpes virus-infected human endothelial cells, which hinted the specificity of LPIs in HAEC activation; third, LPIs upregulate six types of 640 secretomic genes (SGs), namely, 216 canonical SGs, 60 caspase-1-gasdermin D (GSDMD) SGs, 117 caspase-4/11-GSDMD SGs, 40 exosome SGs, 179 Human Protein Atlas (HPA)-cytokines, and 28 HPA-chemokines, which make HAECs a large secretory organ for inflammation/immune responses and other functions; fourth, LPIs activate transcriptomic remodeling by upregulating 172 transcription factors (TFs), namely, pro-inflammatory factors NR4A3, FOS, KLF3, and HIF1A; fifth, LPIs upregulate 152 nuclear DNA-encoded mitochondrial (mitoCarta) genes, which alter mitochondrial mechanisms and functions, such as mitochondrial organization, respiration, translation, and transport; sixth, LPIs activate reactive oxygen species (ROS) mechanism by upregulating 18 ROS regulators; finally, utilizing the Cytoscape software, we found that three mechanisms, namely, LPI-upregulated TFs, mitoCarta genes, and ROS regulators, are integrated to promote HAEC activation. Our results provide novel insights into aortic EC activation, formulate an EC biology knowledge-based transcriptomic profile strategy, and identify new targets for the development of therapeutics for cardiovascular diseases, inflammatory conditions, immune diseases, organ transplantation, aging, and cancers.

Influence of medical domain knowledge on deep learning for Alzheimer's disease prediction.

BACKGROUND AND OBJECTIVE: Alzheimer's disease (AD) is the most common type of dementia that can seriously affect a person's ability to perform daily activities. Estimates indicate that AD may rank third as a cause of death for older people, after heart disease and cancer. Identification of individuals at risk for developing AD is imperative for testing therapeutic interventions. The objective of the study was to determine could diagnostics of AD from EMR data alone (without relying on diagnostic imaging) be significantly improved by applying clinical domain knowledge in data preprocessing and positive dataset selection rather than setting naïve filters. METHODS: Data were extracted from the repository of heterogeneous ambulatory EMR data, collected from primary care medical offices all over the U.S. Medical domain knowledge was applied to build a positive dataset from data relevant to AD. Selected Clinically Relevant Positive (SCRP) datasets were used as inputs to a Long-Short-Term Memory (LSTM) Recurrent Neural Network (RNN) deep learning model to predict will the patient develop AD. RESULTS: Risk scores prediction of AD using the drugs domain information in an SCRP AD dataset of 2,324 patients achieved high out-of-sample score - 0.98-0.99 Area Under the Precision-Recall Curve (AUPRC) when using 90% of SCRP dataset for training. AUPRC dropped to 0.89 when training the model using less than 1,500 cases from the SCRP dataset. The model was still significantly better than when using naïve dataset selection. CONCLUSION: The LSTM RNN method that used data relevant to AD performed significantly better when learning from the SCRP dataset than when datasets were selected naïvely. The integration of qualitative medical knowledge for dataset selection and deep learning technology provided a mechanism for significant improvement of AD prediction. Accurate and early prediction of AD is significant in the identification of patients for clinical trials, which can possibly result in the discovery of new drugs for treatments of AD. Also, the contribution of the proposed predictions of AD is a better selection of patients who need imaging diagnostics for differential diagnosis of AD from other degenerative brain disorders.

Time-to-event estimation by re-defining time.

The primary goal of a time-to-event estimation model is to accurately infer the occurrence time of a target event. Most existing studies focus on developing new models to effectively utilize the information in the censored observations. In this paper, we propose a model to tackle the time-to-event estimation problem from a completely different perspective. Our model relaxes a fundamental constraint that the target variable, time, is a univariate number which satisfies a partial order. Instead, the proposed model interprets each event occurrence time as a time concept with a vector representation. We hypothesize that the model will be more accurate and interpretable by capturing (1) the relationships between features and time concept vectors and (2) the relationships among time concept vectors. We also propose a scalable framework to simultaneously learn the model parameters and time concept vectors. Rigorous experiments and analysis have been conducted in medical event prediction task on seven gene expression datasets. The results demonstrate the efficiency and effectiveness of the proposed model. Furthermore, similarity information among time concept vectors helped in identifying time regimes, thus leading to a potential knowledge discovery related to the human cancer considered in our experiments.

A comprehensive data mining study shows that most nuclear receptors act as newly proposed homeostasis-associated molecular pattern receptors.

BACKGROUND: Nuclear receptors (NRs) can regulate gene expression; therefore, they are classified as transcription factors. Despite the extensive research carried out on NRs, still several issues including (1) the expression profile of NRs in human tissues, (2) how the NR expression is modulated during atherosclerosis and metabolic diseases, and (3) the overview of the role of NRs in inflammatory conditions are not fully understood. METHODS: To determine whether and how the expression of NRs are regulated in physiological/pathological conditions, we took an experimental database analysis to determine expression of all 48 known NRs in 21 human and 17 murine tissues as well as in pathological conditions. RESULTS: We made the following significant findings: (1) NRs are differentially expressed in tissues, which may be under regulation by oxygen sensors, angiogenesis pathway, stem cell master regulators, inflammasomes, and tissue hypo-/hypermethylation indexes; (2) NR sequence mutations are associated with increased risks for development of cancers and metabolic, cardiovascular, and autoimmune diseases; (3) NRs have less tendency to be upregulated than downregulated in cancers, and autoimmune and metabolic diseases, which may be regulated by inflammation pathways and mitochondrial energy enzymes; and (4) the innate immune sensor inflammasome/caspase-1 pathway regulates the expression of most NRs. CONCLUSIONS: Based on our findings, we propose a new paradigm that most nuclear receptors are anti-inflammatory homeostasis-associated molecular pattern receptors (HAMPRs). Our results have provided a novel insight on NRs as therapeutic targets in metabolic diseases, inflammations, and malignancies.

Metabolism-associated danger signal-induced immune response and reverse immune checkpoint-activated CD40+ monocyte differentiation.

Adaptive immunity is critical for disease progression and modulates T cell (TC) and antigen-presenting cell (APC) functions. Three signals were initially proposed for adaptive immune activation: signal 1 antigen recognition, signal 2 co-stimulation or co-inhibition, and signal 3 cytokine stimulation. In this article, we propose to term signal 2 as an immune checkpoint, which describes interactions of paired molecules leading to stimulation (stimulatory immune checkpoint) or inhibition (inhibitory immune checkpoint) of an immune response. We classify immune checkpoint into two categories: one-way immune checkpoint for forward signaling towards TC only, and two-way immune checkpoint for both forward and reverse signaling towards TC and APC, respectively. Recently, we and others provided evidence suggesting that metabolic risk factors (RF) activate innate and adaptive immunity, involving the induction of immune checkpoint molecules. We summarize these findings and suggest a novel theory, metabolism-associated danger signal (MADS) recognition, by which metabolic RF activate innate and adaptive immunity. We emphasize that MADS activates the reverse immune checkpoint which leads to APC inflammation in innate and adaptive immunity. Our recent evidence is shown that metabolic RF, such as uremic toxin or hyperhomocysteinemia, induced immune checkpoint molecule CD40 expression in monocytes (MC) and elevated serum soluble CD40 ligand (sCD40L) resulting in CD40+ MC differentiation. We propose that CD40+ MC is a novel pro-inflammatory MC subset and a reliable biomarker for chronic kidney disease severity. We summarize that CD40:CD40L immune checkpoint can induce TC and APC activation via forward stimulatory, reverse stimulatory, and TC contact-independent immune checkpoints. Finally, we modeled metabolic RF-induced two-way stimulatory immune checkpoint amplification and discussed potential signaling pathways including AP-1, NF-κB, NFAT, STAT, and DNA methylation and their contribution to systemic and tissue inflammation.

Novel extracellular and nuclear caspase-1 and inflammasomes propagate inflammation and regulate gene expression: a comprehensive database mining study.

BACKGROUND: Caspase-1 is present in the cytosol as an inactive zymogen and requires the protein complexes named "inflammasomes" for proteolytic activation. However, it remains unclear whether the proteolytic activity of caspase-1 is confined only to the cytosol where inflammasomes are assembled to convert inactive pro-caspase-1 to active caspase-1. METHODS: We conducted meticulous data analysis methods on proteomic, protein interaction, protein intracellular localization, and gene expressions of 114 experimentally identified caspase-1 substrates and 38 caspase-1 interaction proteins in normal physiological conditions and in various pathologies. RESULTS: We made the following important findings: (1) Caspase-1 substrates and interaction proteins are localized in various intracellular organelles including nucleus and secreted extracellularly; (2) Caspase-1 may get activated in situ in the nucleus in response to intra-nuclear danger signals; (3) Caspase-1 cleaves its substrates in exocytotic secretory pathways including exosomes to propagate inflammation to neighboring and remote cells; (4) Most of caspase-1 substrates are upregulated in coronary artery disease regardless of their subcellular localization but the majority of metabolic diseases cause no significant expression changes in caspase-1 nuclear substrates; and (5) In coronary artery disease, majority of upregulated caspase-1 extracellular substrate-related pathways are involved in induction of inflammation; and in contrast, upregulated caspase-1 nuclear substrate-related pathways are more involved in regulating cell death and chromatin regulation. CONCLUSIONS: Our identification of novel caspase-1 trafficking sites, nuclear and extracellular inflammasomes, and extracellular caspase-1-based inflammation propagation model provides a list of targets for the future development of new therapeutics to treat cardiovascular diseases, inflammatory diseases, and inflammatory cancers.

Structured feature selection using coordinate descent optimization.

BACKGROUND: Existing feature selection methods typically do not consider prior knowledge in the form of structural relationships among features. In this study, the features are structured based on prior knowledge into groups. The problem addressed in this article is how to select one representative feature from each group such that the selected features are jointly discriminating the classes. The problem is formulated as a binary constrained optimization and the combinatorial optimization is relaxed as a convex-concave problem, which is then transformed into a sequence of convex optimization problems so that the problem can be solved by any standard optimization algorithm. Moreover, a block coordinate gradient descent optimization algorithm is proposed for high dimensional feature selection, which in our experiments was four times faster than using a standard optimization algorithm. RESULTS: In order to test the effectiveness of the proposed formulation, we used microarray analysis as a case study, where genes with similar expressions or similar molecular functions were grouped together. In particular, the proposed block coordinate gradient descent feature selection method is evaluated on five benchmark microarray gene expression datasets and evidence is provided that the proposed method gives more accurate results than the state-of-the-art gene selection methods. Out of 25 experiments, the proposed method achieved the highest average AUC in 13 experiments while the other methods achieved higher average AUC in no more than 6 experiments. CONCLUSION: A method is developed to select a feature from each group. When the features are grouped based on similarity in gene expression, we showed that the proposed algorithm is more accurate than state-of-the-art gene selection methods that are particularly developed to select highly discriminative and less redundant genes. In addition, the proposed method can exploit any grouping structure among features, while alternative methods are restricted to using similarity based grouping.

Caspase-1 Inflammasome Activation Mediates Homocysteine-Induced Pyrop-Apoptosis in Endothelial Cells.

RATIONALE: Endothelial injury is an initial mechanism mediating cardiovascular disease. OBJECTIVE: Here, we investigated the effect of hyperhomocysteinemia on programed cell death in endothelial cells (EC). METHODS AND RESULTS: We established a novel flow-cytometric gating method to define pyrotosis (Annexin V(-)/Propidium iodide(+)). In cultured human EC, we found that: (1) homocysteine and lipopolysaccharide individually and synergistically induced inflammatory pyroptotic and noninflammatory apoptotic cell death; (2) homocysteine/lipopolysaccharide induced caspase-1 activation before caspase-8, caspase-9, and caspase-3 activations; (3) caspase-1/caspase-3 inhibitors rescued homocysteine/lipopolysaccharide-induced pyroptosis/apoptosis, but caspase-8/caspase-9 inhibitors had differential rescue effect; (4) homocysteine/lipopolysaccharide-induced nucleotide-binding oligomerization domain, and leucine-rich repeat and pyrin domain containing protein 3 (NLRP3) protein caused NLRP3-containing inflammasome assembly, caspase-1 activation, and interleukin (IL)-1ß cleavage/activation; (5) homocysteine/lipopolysaccharide elevated intracellular reactive oxygen species, (6) intracellular oxidative gradient determined cell death destiny as intermediate intracellular reactive oxygen species levels are associated with pyroptosis, whereas high reactive oxygen species corresponded to apoptosis; (7) homocysteine/lipopolysaccharide induced mitochondrial membrane potential collapse and cytochrome-c release, and increased B-cell lymphoma 2-associated X protein/B-cell lymphoma 2 ratio which were attenuated by antioxidants and caspase-1 inhibitor; and (8) antioxidants extracellular superoxide dismutase and catalase prevented homocysteine/lipopolysaccharide -induced caspase-1 activation, mitochondrial dysfunction, and pyroptosis/apoptosis. In cystathionine ß-synthase-deficient (Cbs(-/-)) mice, severe hyperhomocysteinemia-induced caspase-1 activation in isolated lung EC and caspase-1 expression in aortic endothelium, and elevated aortic caspase-1, caspase-9 protein/activity and B-cell lymphoma 2-associated X protein/B-cell lymphoma 2 ratio in Cbs(-/-) aorta and human umbilical vein endothelial cells. Finally, homocysteine-induced DNA fragmentation was reversed in caspase-1(-/-) EC. Hyperhomocysteinemia-induced aortic endothelial dysfunction was rescued in caspase-1(-/-) and NLRP3(-/-) mice. CONCLUSIONS: Hyperhomocysteinemia preferentially induces EC pyroptosis via caspase-1-dependent inflammasome activation leading to endothelial dysfunction. We termed caspase-1 responsive pyroptosis and apoptosis as pyrop-apoptosis.

Early hyperlipidemia promotes endothelial activation via a caspase-1-sirtuin 1 pathway.

OBJECTIVE: The role of receptors for endogenous metabolic danger signals-associated molecular patterns has been characterized recently as bridging innate immune sensory systems for danger signals-associated molecular patterns to initiation of inflammation in bone marrow-derived cells, such as macrophages. However, it remains unknown whether endothelial cells (ECs), the cell type with the largest numbers and the first vessel cell type exposed to circulating danger signals-associated molecular patterns in the blood, can sense hyperlipidemia. This report determined whether caspase-1 plays a role in ECs in sensing hyperlipidemia and promoting EC activation. APPROACH AND RESULTS: Using biochemical, immunologic, pathological, and bone marrow transplantation methods together with the generation of new apoplipoprotein E (ApoE)(-/-)/caspase-1(-/-) double knockout mice, we made the following observations: (1) early hyperlipidemia induced caspase-1 activation in ApoE(-/-) mouse aorta; (2) caspase-1(-/-)/ApoE(-/-) mice attenuated early atherosclerosis; (3) caspase-1(-/-)/ApoE(-/-) mice had decreased aortic expression of proinflammatory cytokines and attenuated aortic monocyte recruitment; and (4) caspase-1(-/-)/ApoE(-/-) mice had decreased EC activation, including reduced adhesion molecule expression and cytokine secretion. Mechanistically, oxidized lipids activated caspase-1 and promoted pyroptosis in ECs by a reactive oxygen species mechanism. Caspase-1 inhibition resulted in accumulation of sirtuin 1 in the ApoE(-/-) aorta, and sirtuin 1 inhibited caspase-1 upregulated genes via activator protein-1 pathway. CONCLUSIONS: Our results demonstrate for the first time that early hyperlipidemia promotes EC activation before monocyte recruitment via a caspase-1-sirtuin 1-activator protein-1 pathway, which provides an important insight into the development of novel therapeutics for blocking caspase-1 activation as early intervention of metabolic cardiovascular diseases and inflammations.

Angiotensin converting enzyme inhibitor attenuates oxidative stress-induced endothelial cell apoptosis via p38 MAP kinase inhibition.

BACKGROUND: The effects of angiotensin converting enzyme (ACE) inhibitors on oxidative stress-induced apoptosis of endothelial cells and the intracellular signaling were investigated. METHODS: Cultured endothelial cells derived from a bovine carotid artery were treated with H2O2 or TNF-alpha to induce apoptosis. Apoptosis was evaluated by DNA fragmentation and cell viability, p38 MAP kinase activity by Western blotting, and oxidative stress by formation of 8-isoprostane. The effects of ACE inhibitors were examined by adding them into the medium throughout the experiments. RESULTS: Apoptosis was attenuated by ACE inhibitors, temocapril and captopril, in a dose-dependent manner (1-100 micromol/l). H2O2 (0.2 mmol/l for 1.5 h) or TNF-alpha (10 ng/ml for 72 h) treatment stimulated the activities of p38 MAP kinase. Temocapril and captopril decreased the activity of p38 MAP kinase as well as 8-isoprostane formation induced by H2O2. A p38 MAP kinase inhibitor, SB203580, partially inhibited the effect of temocapril on apoptosis. CONCLUSIONS: These results suggest that ACE inhibitors protect endothelial cells from oxidative stress-induced apoptosis, and that p38 MAP kinase plays a critical role in the process.