InContext: curation of medical context for drug indications.

Accurate and precise information about the therapeutic uses (indications) of a drug is essential for applications in drug repurposing and precision medicine. Leading online drug resources such as DrugCentral and DrugBank provide rich information about various properties of drugs, including their indications. However, because indications in such databases are often partly automatically mined, some may prove to be inaccurate or imprecise. Particularly challenging for text mining methods is the task of distinguishing between general disease mentions in drug product labels and actual indications for the drug. For this, the qualifying medical context of the disease mentions in the text should be studied. Some examples include contraindications, co-prescribed drugs and target patient qualifications. No existing indication curation efforts attempt to capture such information in a precise way. Here we fill this gap by presenting a novel curation protocol for extracting indications and machine processable annotations of contextual information about the therapeutic use of a drug. We implemented the protocol on a reference set of FDA-approved drug product labels on the DailyMed website to curate indications for 150 anti-cancer and cardiovascular drugs. The resulting corpus - InContext - focuses on anti-cancer and cardiovascular drugs because of the heightened societal interest in cancer and heart disease. In order to understand how InContext relates with existing reputable drug indication databases, we analysed it's overlap with a state-of-the-art indications database - LabeledIn - as well as a reputable online drug compendium - DrugCentral. We found that 40% of indications sampled from DrugCentral (and 23% from LabeledIn) respectively, could not be accounted for in InContext. This raises questions about the veracity of indications not appearing in InContext. The additional contextual information curated by InContext about disease mentions in drug SPLs provides a foundation for more precise, structured and formal representations of knowledge related to drug therapeutic use, in order to increase accuracy and agreement of drug indication extraction methods for in silico drug repurposing.

Formaldehyde and Brain Disorders: A Meta-Analysis and Bioinformatics Approach.

While there is significant investigation and investment in brain and neurodegenerative disease research, current understanding of the etiologies of illnesses like Alzheimer's disease (AD), Parkinson's disease (PD), amyotrophic lateral sclerosis (ALS), and brain cancer remains limited. Environmental exposure to the pollutant formaldehyde, an emerging neurotoxin widely used in industry, is suspected to play a critical role in mediating these disorders, although findings are limited and inconsistent. Focusing on highly exposed groups, we performed a meta-analysis of human epidemiological studies of formaldehyde and neurodegenerative disease (N =  19) or brain tumors (N = 12). To assess the biological plausibility of observed associations, we then conducted a bioinformatics analysis using WikiPathways and the Comparative Toxicogenomics Database and identified candidate genes and pathways that may be related to these interactions. We reported the meta-relative risk (meta-RR) of ALS following high exposures to formaldehyde was increased by 78% (meta-RR = 1.78, 95% confidence interval, CI 1.20-2.65). Similarly, the meta-RR for brain cancer was increased by 71% (meta-RR = 1.71; 95% CI 1.07-2.73) among highly exposed individuals. Multiple sensitivity analyses did not reveal sources of heterogeneity or bias. Our bioinformatics analysis revealed that the oxidative stress genes superoxide dismutase (SOD1, SOD2) and the pro-inflammatory marker tumor necrosis factor (TNF) were identified as the top relevant genes, and the folate metabolism, vitamin B12 metabolism, and the ALS pathways were highly affected by formaldehyde and related to the most brain diseases of interest. Further inquiry revealed the two metabolic pathways are also intimately tied with the formaldehyde cycle. Overall, our bioinformatics analysis supports the link of formaldehyde exposure to ALS or brain tumor reported from our meta-analysis. This new multifactorial approach enabled us to both interrogate the robustness of the epidemiological data and identify genes and pathways that may be involved in these interactions, ultimately lending strong evidence and potential biological plausibility for the association between formaldehyde exposure and brain disease.

The Key Characteristics of Carcinogens: Relationship to the Hallmarks of Cancer, Relevant Biomarkers, and Assays to Measure Them.

The key characteristics (KC) of human carcinogens provide a uniform approach to evaluating mechanistic evidence in cancer hazard identification. Refinements to the approach were requested by organizations and individuals applying the KCs. We assembled an expert committee with knowledge of carcinogenesis and experience in applying the KCs in cancer hazard identification. We leveraged this expertise and examined the literature to more clearly describe each KC, identify current and emerging assays and in vivo biomarkers that can be used to measure them, and make recommendations for future assay development. We found that the KCs are clearly distinct from the Hallmarks of Cancer, that interrelationships among the KCs can be leveraged to strengthen the KC approach (and an understanding of environmental carcinogenesis), and that the KC approach is applicable to the systematic evaluation of a broad range of potential cancer hazards in vivo and in vitro We identified gaps in coverage of the KCs by current assays. Future efforts should expand the breadth, specificity, and sensitivity of validated assays and biomarkers that can measure the 10 KCs. Refinement of the KC approach will enhance and accelerate carcinogen identification, a first step in cancer prevention.See all articles in this CEBP Focus section, "Environmental Carcinogenesis: Pathways to Prevention."

A State-of-the-Science Review of Arsenic's Effects on Glucose Homeostasis in Experimental Models.

BACKGROUND: The prevalence of type 2 diabetes (T2D) has more than doubled since 1980. Poor nutrition, sedentary lifestyle, and obesity are among the primary risk factors. While an estimated 70% of cases are attributed to excess adiposity, there is an increased interest in understanding the contribution of environmental agents to diabetes causation and severity. Arsenic is one of these environmental chemicals, with multiple epidemiology studies supporting its association with T2D. Despite extensive research, the molecular mechanism by which arsenic exerts its diabetogenic effects remains unclear. OBJECTIVES: We conducted a literature search focused on arsenite exposure in vivo and in vitro, using relevant end points to elucidate potential mechanisms of oral arsenic exposure and diabetes development. METHODS: We explored experimental results for potential mechanisms and elucidated the distinct effects that occur at high vs. low exposure. We also performed network analyses relying on publicly available data, which supported our key findings. RESULTS: While several mechanisms may be involved, our findings support that arsenite has effects on whole-body glucose homeostasis, insulin-stimulated glucose uptake, glucose-stimulated insulin secretion, hepatic glucose metabolism, and both adipose and pancreatic ß-cell dysfunction. DISCUSSION: This review applies state-of-the-science approaches to identify the current knowledge gaps in our understanding of arsenite on diabetes development. https://doi.org/10.1289/EHP4517.

Consensus on the key characteristics of endocrine-disrupting chemicals as a basis for hazard identification.

Endocrine-disrupting chemicals (EDCs) are exogenous chemicals that interfere with hormone action, thereby increasing the risk of adverse health outcomes, including cancer, reproductive impairment, cognitive deficits and obesity. A complex literature of mechanistic studies provides evidence on the hazards of EDC exposure, yet there is no widely accepted systematic method to integrate these data to help identify EDC hazards. Inspired by work to improve hazard identification of carcinogens using key characteristics (KCs), we have developed ten KCs of EDCs based on our knowledge of hormone actions and EDC effects. In this Expert Consensus Statement, we describe the logic by which these KCs are identified and the assays that could be used to assess several of these KCs. We reflect on how these ten KCs can be used to identify, organize and utilize mechanistic data when evaluating chemicals as EDCs, and we use diethylstilbestrol, bisphenol A and perchlorate as examples to illustrate this approach.

Proposed Key Characteristics of Female Reproductive Toxicants as an Approach for Organizing and Evaluating Mechanistic Data in Hazard Assessment.

BACKGROUND: Identification of female reproductive toxicants is currently based largely on integrated epidemiological and in vivo toxicology data and, to a lesser degree, on mechanistic data. A uniform approach to systematically search, organize, integrate, and evaluate mechanistic evidence of female reproductive toxicity from various data types is lacking. OBJECTIVE: We sought to apply a key characteristics approach similar to that pioneered for carcinogen hazard identification to female reproductive toxicant hazard identification. METHODS: A working group of international experts was convened to discuss mechanisms associated with chemical-induced female reproductive toxicity and identified 10 key characteristics of chemicals that cause female reproductive toxicity: 1) alters hormone receptor signaling; alters reproductive hormone production, secretion, or metabolism; 2) chemical or metabolite is genotoxic; 3) induces epigenetic alterations; 4) causes mitochondrial dysfunction; 5) induces oxidative stress; 6) alters immune function; 7) alters cell signal transduction; 8) alters direct cell­cell interactions; 9) alters survival, proliferation, cell death, or metabolic pathways; and 10) alters microtubules and associated structures. As proof of principle, cyclophosphamide and diethylstilbestrol (DES), for which both human and animal studies have demonstrated female reproductive toxicity, display at least 5 and 3 key characteristics, respectively. 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD), for which the epidemiological evidence is mixed, exhibits 5 key characteristics. DISCUSSION: Future efforts should focus on evaluating the proposed key characteristics against additional known and suspected female reproductive toxicants. Chemicals that exhibit one or more of the key characteristics could be prioritized for additional evaluation and testing. A key characteristics approach has the potential to integrate with pathway-based toxicity testing to improve prediction of female reproductive toxicity in chemicals and potentially prevent some toxicants from entering common use. https://doi.org/10.1289/EHP4971.

Human exposure to trichloroethylene is associated with increased variability of blood DNA methylation that is enriched in genes and pathways related to autoimmune disease and cancer.

Human exposure to trichloroethylene (TCE) is linked to kidney cancer, autoimmune diseases, and probably non-Hodgkin lymphoma. Additionally, TCE exposed mice and cell cultures show altered DNA methylation. To evaluate associations between TCE exposure and DNA methylation in humans, we conducted an epigenome-wide association study (EWAS) in TCE exposed workers using the HumanMethylation450 BeadChip. Across individual CpG probes, genomic regions, and globally (i.e., the 450K methylome), we investigated differences in mean DNA methylation and differences in variability of DNA methylation between 73 control (< 0.005 ppm TCE), 30 lower exposed (< 10 ppm TCE), and 37 higher exposed ( ≥ 10 ppm TCE) subjects' white blood cells. We found that TCE exposure increased methylation variation globally (Kruskal-Wallis p-value = 3.75e-3) and in 25 CpG sites at a genome-wide significance level (Bonferroni p-value < 0.05). We identified a 609 basepair region in the TRIM68 gene promoter that exhibited hypomethylation with increased exposure to TCE (FWER = 1.20e-2). Also, genes that matched to differentially variable CpGs were enriched in the 'focal adhesion' biological pathway (p-value = 2.80e-2). All in all, human exposure to TCE was associated with epigenetic alterations in genes involved in cell-matrix adhesions and interferon subtype expression, which are important in the development of autoimmune diseases; and in genes related to cancer development. These results suggest that DNA methylation may play a role in the pathogenesis of TCE exposure-related diseases and that TCE exposure may contribute to epigenetic drift.

Key Characteristics Approach to Carcinogenic Hazard Identification.

Evaluating carcinogenic mechanisms is a challenging part of hazard identification, as mechanistic data are both voluminous and diverse. An evaluation approach based on 10 key characteristics of human carcinogens provides a holistic and unbiased way to tackle this challenge.

WikiPathways: a multifaceted pathway database bridging metabolomics to other omics research.

WikiPathways (wikipathways.org) captures the collective knowledge represented in biological pathways. By providing a database in a curated, machine readable way, omics data analysis and visualization is enabled. WikiPathways and other pathway databases are used to analyze experimental data by research groups in many fields. Due to the open and collaborative nature of the WikiPathways platform, our content keeps growing and is getting more accurate, making WikiPathways a reliable and rich pathway database. Previously, however, the focus was primarily on genes and proteins, leaving many metabolites with only limited annotation. Recent curation efforts focused on improving the annotation of metabolism and metabolic pathways by associating unmapped metabolites with database identifiers and providing more detailed interaction knowledge. Here, we report the outcomes of the continued growth and curation efforts, such as a doubling of the number of annotated metabolite nodes in WikiPathways. Furthermore, we introduce an OpenAPI documentation of our web services and the FAIR (Findable, Accessible, Interoperable and Reusable) annotation of resources to increase the interoperability of the knowledge encoded in these pathways and experimental omics data. New search options, monthly downloads, more links to metabolite databases, and new portals make pathway knowledge more effortlessly accessible to individual researchers and research communities.

A Data Fusion Pipeline for Generating and Enriching Adverse Outcome Pathway Descriptions.

Increasing amounts of systems toxicology data, including omics results, are becoming publically available and accessible in databases. Data-driven and informatics-tool supported pipeline schemas for fitting such data into Adverse Outcome Pathway (AOP) descriptions could potentially aid the development of nonanimal-based hazard and risk assessment methods. We devised a 6-step workflow that integrated diverse types of toxicology data into a novel AOP scheme for pulmonary fibrosis. Mining of literature references and diverse data sources covering previous pathway descriptions and molecular results were coupled in a stepwise manner with informatics tools applications that enabled gene linkage and pathway identification in molecular interaction maps. Ultimately, a network of functional elements coupled 64 pulmonary fibrosis-associated genes into a novel, open-source AOP-linked molecular pathway, now available for commenting and improvements in WikiPathways (WP3624). Applying in silico-based knowledge extraction and modeling, the pipeline enabled screening and fusion of many different complex data types, including the integration of omics results. Overall, the taken, stepwise approach should be generally useful to construct novel AOP descriptions as well as to enrich developing AOP descriptions in progress.

Persistent transcriptional responses show the involvement of feed-forward control in a repeated dose toxicity study.

Chemical carcinogenesis, albeit complex, often relies on modulation of transcription through activation or repression of key transcription factors. While analyzing extensive networks may hinder the biological interpretation, one may focus on dynamic network motifs, among which persistent feed-forward loops (FFLs) are known to chronically influence transcriptional programming. Here, to investigate the relevance a FFL-oriented approach in depth, we have focused on aflatoxin B1-induced transcriptomic alterations during distinct states of exposure (daily administration during 5days followed by a non-exposed period) of human hepatocytes, by exploring known interactions in human transcription. Several TF-coding genes were persistently deregulated after washout of AFB1. Oncogene MYC was identified as the prominent regulator and driver of many FFLs, among which a FFL comprising MYC/HIF1A was the most recurrent. The MYC/HIF1A FFL was also identified and validated in an independent set as the master regulator of metabolic alterations linked to initiation and progression of carcinogenesis, i.e. the Warburg effect, possibly as result of persistent intracellular alterations arising from AFB1 exposure (nuclear and mitochondrial DNA damage, oxidative stress, transcriptional activation by secondary messengers). In summary, our analysis shows the involvement of FFLs as modulators of gene expression suggestive of a carcinogenic potential even after termination of exposure.

Exploiting microRNA and mRNA profiles generated in vitro from carcinogen-exposed primary mouse hepatocytes for predicting in vivo genotoxicity and carcinogenicity.

The well-defined battery of in vitro systems applied within chemical cancer risk assessment is often characterised by a high false-positive rate, thus repeatedly failing to correctly predict the in vivo genotoxic and carcinogenic properties of test compounds. Toxicogenomics, i.e. mRNA-profiling, has been proven successful in improving the prediction of genotoxicity in vivo and the understanding of underlying mechanisms. Recently, microRNAs have been discovered as post-transcriptional regulators of mRNAs. It is thus hypothesised that using microRNA response-patterns may further improve current prediction methods. This study aimed at predicting genotoxicity and non-genotoxic carcinogenicity in vivo, by comparing microRNA- and mRNA-based profiles, using a frequently applied in vitro liver model and exposing this to a range of well-chosen prototypical carcinogens. Primary mouse hepatocytes (PMH) were treated for 24 and 48h with 21 chemical compounds [genotoxins (GTX) vs. non-genotoxins (NGTX) and non-genotoxic carcinogens (NGTX-C) versus non-carcinogens (NC)]. MicroRNA and mRNA expression changes were analysed by means of Exiqon and Affymetrix microarray-platforms, respectively. Classification was performed by using Prediction Analysis for Microarrays (PAM). Compounds were randomly assigned to training and validation sets (repeated 10 times). Before prediction analysis, pre-selection of microRNAs and mRNAs was performed by using a leave-one-out t-test. No microRNAs could be identified that accurately predicted genotoxicity or non-genotoxic carcinogenicity in vivo. However, mRNAs could be detected which appeared reliable in predicting genotoxicity in vivo after 24h (7 genes) and 48h (2 genes) of exposure (accuracy: 90% and 93%, sensitivity: 65% and 75%, specificity: 100% and 100%). Tributylinoxide and para-Cresidine were misclassified. Also, mRNAs were identified capable of classifying NGTX-C after 24h (5 genes) as well as after 48h (3 genes) of treatment (accuracy: 78% and 88%, sensitivity: 83% and 83%, specificity: 75% and 93%). Wy-14,643, phenobarbital and ampicillin trihydrate were misclassified. We conclude that genotoxicity and non-genotoxic carcinogenicity probably cannot be accurately predicted based on microRNA profiles. Overall, transcript-based prediction analyses appeared to clearly outperform microRNA-based analyses.

Aflatoxin B1 induces persistent epigenomic effects in primary human hepatocytes associated with hepatocellular carcinoma.

Chronic exposure to aflatoxin B1 (AFB1) has, in certain regions in the world, been strongly associated with hepatocellular carcinoma (HCC) development. AFB1 is a very potent hepatotoxic and carcinogenic mycotoxin which is frequently reported as a food contaminant. Epigenetic modifications provoked by environmental exposures, such as AFB1, may create a persistent epigenetic footprint. Deregulation of epigenetic mechanisms has actually been reported in HCC patients following AFB1 exposure; however, no attempts have yet been made to investigate early effects on the epigenome level which may be persistent on longer term, thereby possibly initiating carcinogenic events. In this study, we aim to identify methyl DNA-mRNA-interactions representative for a persistent epigenetic footprint associated with the early onset of AFB1-induced HCC. For this, primary human hepatocytes were exposed to 0.3µM of AFB1 for 5 days. Persistent epigenetic effects were measured 3 days after terminating the carcinogenic exposure. Whole genome DNA methylation changes and whole genome transcriptomic analysis were analyzed applying microarray technologies, and cross-omics interactions were evaluated. Upon combining transcriptomics data with results on DNA methylation, a range of persistent hyper- and hypo-methylated genes was identified which also appeared affected on the transcriptome level. For six of the hypo-methylated and up-regulated genes, namely TXNRD1, PCNA, CCNK, DIAPH3, RAB27A and HIST1H2BF, a clear role in carcinogenic events could be identified. This study is the first to report on a carcinogen-induced persistent impact on the epigenetic footprint in relation with the transcriptome which could be indicative for the early onset of AFB1-related development of HCC.

Evaluating microRNA profiles reveals discriminative responses following genotoxic or non-genotoxic carcinogen exposure in primary mouse hepatocytes.

Chemical carcinogenesis can be induced by genotoxic (GTX) or non-genotoxic (NGTX) carcinogens. GTX carcinogens have a well-described mode of action. However, the complex mechanisms by which NGTX carcinogens act are less clear and may result in conflicting results between species [e.g. Wy-14,643 (Wy)]. We hypothesise that common microRNA response pathways exist for each class of carcinogenic agents. Therefore, this study compares and integrates mRNA and microRNA expression profiles following short term acute exposure (24 and 48h) to three GTX [aflatoxin B1 (AFB1), benzo[a]pyrene (BaP) and cisplatin (CisPl)] or three NGTX (2,3,7,8-tetrachloordibenzodioxine (TCDD), cyclosporine A (CsA) and Wy) carcinogens in primary mouse hepatocytes. Discriminative gene sets, microRNAs (not for 24h) and processes were identified following 24 and 48h of exposure. From the three discriminative microRNAs found following 48h of exposure, mmu-miR-503-5p revealed to have an interaction with mRNA target gene cyclin D2 (Ccnd2 - 12444) which was involved in the discriminative process of p53 signalling and metabolism. Following exposure to NGTX carcinogens Mmu-miR-503-5p may have an oncogenic function by stimulating Ccnd2 possibly leading to a tumourigenic cell cycle progression. By contrast, after GTX carcinogen exposure it may have a tumour-suppressive function (repressing Ccnd2) leading to cell cycle arrest and to increased DNA repair activities. In addition, compound-specific microRNA-mRNA interactions [mmu-miR-301b-3p-Papss2 (for AFB1), as well as mmu-miR-29b-3p-Col4a2 and mmu-miR-24-3p-Flna (for BaP)] were found to contribute to a better understanding of microRNAs in cell cycle arrest and the impairment of the DNA damage repair, an important hallmark of GTX-induced carcinogenesis. Overall, our results indicate that microRNAs represent yet another relevant intracellular regulatory level in chemical carcinogenesis.

Characterisation of cisplatin-induced transcriptomics responses in primary mouse hepatocytes, HepG2 cells and mouse embryonic stem cells shows conservation of regulating transcription factor networks.

The toxic mechanisms of cisplatin have been frequently studied in many species and in vitro cell models. The Netherlands Toxicogenomics Centre focuses on developing in vitro alternatives using genomics technologies for animal-based assays on, e.g. genotoxic hazards. Models such as human hepatocellular carcinoma cell line (HepG2) cells, mouse primary hepatocytes (PMH) and mouse embryonic stem cells (mESC) are used. Our aim was to identify possibly robust conserved mechanisms between these models using cisplatin as model genotoxic agent. Transcriptomic data newly generated from HepG2 cells and PMH exposed to 7 µM cisplatin for 12, 24 and 48h and 24 and 48h, respectively, were compared with published data from mESC exposed to 5 µM cisplatin for 2-24h. Due to differences in response time between models and marginal changes after shorter exposure periods, we focused on 24 and 48h. At gene level, 44 conserved differentially expressed genes (DEG), involved in processes such as apoptosis, cell cycle, DNA damage response and DNA repair, were found. Functional analysis shows that limited numbers of pathways are conserved. Transcription factor (TF) network analysis indicates 12 common TF networks responding among all models and time points. Four TF, HNF4-α, SP1, c-MYC and p53, capable of regulating ±50% of all DEG, seem of equal importance in all models and exposure periods. Here we showed that transcriptomic responses across several in vitro cell models following exposure to cisplatin are mainly determined by a conserved complex network of 4 TFs. These conserved responses are hypothesised to provide most relevant information for human toxicity prediction and may form the basis for new in vitro alternatives of risk assessment.