Hasil untuk "Toxicology. Poisons"

Saifeng PEI, Chaoye SHEN, Chao YU et al.

BackgroundBenzothiazoles and benzotriazoles are ubiquitously detected in drinking water, posing potential health risks. Developing a reliable and sensitive analytical method is critical for assessing their exposure levels and associated risks.ObjectiveTo establish an automated solid-phase extraction (SPE) coupled with liquid chromatography-tandem mass spectrometry method for the simultaneous determination of benzothiazole, 2-aminobenzothiazole, benzotriazole, and 5-chlorobenzotriazole in drinking water.MethodsKey parameters were systematically optimized. Three mobile phase systems were compared to evaluate their effects on chromatographic peak shape and separation; three SPE cartridges were tested for extraction efficiency; the influences of water pH on extraction recoveries and matrix effects were investigated; the contributions of pretreatment steps to benzothiazole blank were analyzed, and control measures were established; nitrogen evaporation temperatures were tested to ensure analyte stability. The optimized parameters were used to develop the method, validate its performance, and analyze drinking water samples.ResultsDrinking water samples were first adjusted to pH 6.0 with formic acid, and then extracted using HLB cartridges. After elution with 6 mL methanol, the eluate was concentrated under nitrogen evaporation at 45 °C, followed by separation on a C18 column using a gradient elution of acetonitrile-0.1% formic acid solution, and tandem mass spectrometry detection. Avoiding nitrogen blow-drying during SPE and eliminating the use of polypropylene materials during nitrogen evaporation can effectively minimize benzothiazole contamination. Good calibration linearity was obtained for the target analytes in the concentration range of 5.0-250 μg·L−1, with a correlation coefficient of r > 0.995. The method detection limits were in the range of 1.0-5.0 ng·L−1. The recoveries of the target analytes in pure water and tap water were 80.2%-119.5% and 72.2%-115.6%, respectively, with relative standard deviations of were 3.2%-12.8% and 2.3%-11.6%, respectively. When applying this method to actual water samples, benzotriazole was detected in 100% of treated and tap water samples, with median concentrations of 79.4 ng·L−1 and 114 ng·L−1, respectively. Benzothiazole was detected in 83.3% of treated water samples and 100% of tap water samples, with median concentrations of 48.3 ng·L−1 and 65.4 ng·L−1, respectively. In addition, 5-chloro-benzotriazole exhibited low detection rates and concentrations, while 2-amino-benzothiazole was undetected.ConclusionThe developed method demonstrates high accuracy, reliability, and sensitivity, making it suitable for the analysis of trace levels of benzothiazoles and benzotriazoles in drinking water.

Mary Sabry Abdel-Messih, Maged N. Kamel Boulos

ChatGPT has recently been shown to pass the United States Medical Licensing Examination (USMLE). We tested ChatGPT (Feb 13, 2023 release) using a typical clinical toxicology case of acute organophosphate poisoning. ChatGPT fared well in answering all of our queries regarding it.

Mohadeseh Zarei Ghobadi, Elaheh Afsaneh

Adult T-Cell Leukemia/Lymphoma (ATLL) is a malignancy that arises from T-cells infected with the human T-cell lymphotropic virus type 1 (HTLV-1). The disease is characterized by uncontrolled proliferation and reduced apoptosis of malignant T cells, which contributes to tumor progression and resistance to therapy. Anoikis is a specific form of programmed cell death triggered by the loss of cell–matrix or cell–cell adhesion, playing a critical role in preventing detached cells from surviving and forming tumors. Dysregulation of anoikis has been implicated in cancer metastasis and therapeutic resistance across various malignancies; however, its role in ATLL remains largely unexplored. To our knowledge, this is the first study to investigate anoikis-related genes in ATLL subtypes, particularly across its major subtypes: acute, chronic, and smoldering. In this study, we explored anoikis-related differentially expressed genes to identify those specifically associated with each subtype. We then applied Least Absolute Shrinkage and Selection Operator (LASSO) regression to select the most informative features. Subsequently, we employed decision trees, random forest, extreme gradient boosting, support vector machine, and logistic regression algorithms to identify classifier genes distinguishing each ATLL subtype from asymptomatic carriers. The identified biomarkers include SMARCE1 and CASP3 for acute, TGFΒ1 and MTA1 for chronic, and CXCL1 and LGALS8 for smoldering subtypes. These genes are involved in cell adhesion, survival signaling, and apoptosis—key processes in cellular homeostasis and oncogenesis. Our findings provide novel insights into the molecular mechanisms linking anoikis to ATLL subtypes and highlight potential therapeutic targets.

Mengjun CHANG, Shuangjie YU, Jin JI et al.

BackgroundJob burnout and depressive symptoms are prevalent among occupational populations, with a close relationship between them. Sleep quality, as a potential mediating factor, significantly affects the mental health of workers.ObjectiveTo explore the relationship between job burnout, sleep quality, and depressive symptoms, and determine whether sleep quality mediates the relationship between job burnout and depressive symptoms.MethodsFrom April 25 to May 1, 2024, this study employed cluster sampling to conduct a questionnaire survey among individuals engaged in various occupations across five cities in the Ningxia Hui Autonomous Region. The questionnaires included socio-demographic information, as well as the Chinese Maslach Burnout Inventory (CMBI), the Pittsburgh Sleep Quality Index (PSQI), and the Patient Health Questionnaire-9 (PHQ-9) for assessing burnout, sleep quality, and depressive symptoms, respectively. Out of the 4106 questionnaires distributed, a total of 3837 questionnaires were valid, and the valid recovery rate was 93.45%. The distribution among demographic variables in burnout, sleep quality and depressive symptoms were statistically analyzed. A binary logistic regression model was used for multifactor correlation analysis; Pearson correlation was used to test the correlation between burnout, sleep quality, and depressed mood. Modelling and mediated effect path mapping were performed using Amos 24.0 software.ResultsThe postive rate of occupational burnout in the workers was 97.49% (3741/3837), the positive rate of poor sleep quality was 66.77% (2526/3837), and the positive rate of depressive symptoms was 75.68% (2904/3837). There were statistically significant differences in depressive symptoms by ages, shift types, working years, marital status, smoking habits, drinking habits, and exercising frequency (P < 0.05). The logistic model indicated that working years, drinking habits, job burnout, and overall sleep quality were significant factors influencing the occurrence of depressive symptoms (P < 0.05). The correlation analysis revealed positive correlations between depressive symptom scores and job burnout scores (r=0.045, P < 0.01), between depressive symptom scores and sleep quality scores (r=0.480, P < 0.01), and between job burnout scores and sleep quality scores (r=0.054, P < 0.01). Furthermore, the indirect effect of job burnout on depressive symptoms through sleep quality was 0.100 (95%CI: 0.204, 0.252; P < 0.01).ConclusionJob burnout and sleep quality are significant factors associated with the occurrence of depressive symptoms among occupational populations, and sleep quality plays a partial mediating role in depressive symptoms associated with job burnout. This finding may provide a scientific basis for developing intervention strategies to control depressive symptoms among workers.

Zhiyu Liu, Juan Wang, Yuqi Li et al.

Abstract Background Erectile dysfunction (ED) is a common male sexual disorder with a multifactorial etiology. The exposure to endocrine-disrupting chemicals (EDCs) has been increasingly linked to reproductive health disorders in both men and women. EDCs can interfere with hormonal signaling and physiological homeostasis, but their specific roles and mechanisms in contributing to ED remain inadequately elucidated. Methods Network toxicology and enrichment analysis were used to identify potential targets and signaling pathways involved in ED induced by EDCs. Single-cell sequencing was conducted to analyze the expression profiles of these targets in corpus cavernosum tissue. Key regulatory molecules were identified through protein–protein interaction (PPI) network analysis. Core targets were selected using three machine learning algorithms to evaluate the association between EDCs and ED. Molecular docking simulations were further employed to verify the binding affinity between EDCs and target proteins, elucidating potential mechanisms of action. Results A total of 186 potential targets were identified. Single-cell sequencing revealed their expression characteristics. PPI analysis identified key regulatory molecules, and machine learning approaches pinpointed two core targets: CTNNB1 and HIF1A. Molecular docking confirmed that most EDCs exhibit stable binding to CTNNB1 and HIF1A, suggesting the involvement of associated signaling pathways in the development of ED. Conclusions This study systematically characterizes the molecular pathways through which EDCs contribute to ED, with CTNNB1 and HIF1A emerging as central players. The identification of these core targets provides a theoretical foundation for developing targeted interventions against environment-related ED and underscores the importance of mitigating EDC exposure in public health strategies.

Avinash Bathina, Janne Hakanen, Atso Raasmaja et al.

Aryl hydrocarbon receptor (AHR) is a ligand-activated transcription factor whose role in energy metabolism is obscure. Most of its physiological ligands are derived from tryptophan (TRP). Here, fifty male C57BL/6JRccHsd mice were assigned to one of five feeding groups, control diet (CD), high-fat diet (HFD; 45 % of energy from fat), HFD with only 70 % of the regular TRP concentration (HFDtrp), HFD supplemented with a weakly toxic AHR agonist C2 (HFDc2), or HFDtrp with C2 (HFDtrp-c2). All diets contained 2 % cholesterol and were fed for 18 weeks. On weeks 14–16, the mice were tested for gas exchange and locomotor activity, and on weeks 15–17 for glucose tolerance (GTT) and insulin sensitivity (ITT). At termination, tissue samples were collected for biochemical and AI-assisted histological analyses. Body weight gain (BWG) was only 28–38 % higher in the HFD groups than in the CD group, but the HFD-fed mice accumulated 43–61 % more fat. Calorie intake was greater in the two low-TRP groups than in the two other HFD groups, while BWG remained similar. C2 induced Cyp1a1 expression (an index of AHR activity) in all tissues examined and increased the ratio of micro-/macrosteatosis in the liver. The HFDs tended to reduce insulin sensitivity, CO2 production, and the ability to respond appropriately to a low-temperature challenge. These findings suggest that the effects of AHR activity modulation on energy balance are strongly context-dependent. A sensitive response to long-term AHR activation appears to be elevated micro-/macrosteatosis ratio in the liver when exposed to HFD.

Danielle Craig-Meyer, Joseph A. Hollenbaugh, Sara Morgado et al.

Immunotherapeutics targeting immune checkpoint receptors or their ligands (i.e., immune checkpoint inhibitors), have been groundbreaking in the field of oncology, radically changing the approach to treatment and improving the clinical outcomes of an ever-expanding list of solid tumors and hematological malignancies. However, immune checkpoint inhibitors (ICI) are not devoid of side effects, collectively regarded as immune-related adverse events (irAE); they are not easily uncovered in preclinical immunotoxicological investigations and are often due to the very low expression of their targets in immunologically-unchallenged non-clinical species. We have characterized expression of a broad range of immune checkpoint receptors in peripheral blood mononuclear cell (PBMC) subpopulations from cynomolgus monkeys and healthy human volunteers, under resting and T-cell stimulatory conditions by multicolor flow cytometry to inform appropriate species selection for modeling potential irAE in immunotherapeutic preclinical research. Focusing on the response of the main lymphocyte populations to interleukin (IL)-2 alone, or in combination with anti-CD3 and anti-CD28 antibodies, checkpoints with shared similarities and key differences between the two species were identified. The results of this first study provide a database for the expression and response to stimulation for immune checkpoint receptors and can help guide future model selection in the design of preclinical studies involving immunotherapeutics directed against these targets.

Xiaohui HAO, Qian LI, Yixuan JIN et al.

BackgroundAt present, the treatment of silicosis is still limited, and no method is available to cure the disease. miRNAs are involved in the process of fibrosis at the transcriptional level by directly degrading target gene mRNA or inhibiting its translation. However, how miR-130a-3p regulates silicosis fibrosis has not been fully elucidated yet. ObjectiveTo investigate whether miR-130a-3p promotes epithelial-mesenchymal transition (EMT) by inhibiting peroxisome proliferators-activated receptors gamma (PPAR-γ), thereby pro-moting the process of silicotic fibrosis. To identify effective new targets for the treatment of silicotic fibrosis. Methods(1) Animal experiments: C57BL/6J mice were intratracheally injected with a one-time dose of 10 mg silica suspension (dissolved in 100 μL saline) as positive lung exposure. A silicosis model group was established 28 d after the exposure. A control group was injected with the same amount of normal saline into the trachea. Hematoxylin-eosin staining and Sirius red staining were used to observe the pathological changes and collagen deposition in lung tissues respectively. Realtime fluorescence-based quantitative polymerase chain reaction (RT-qPCR) was used to assay the expression of miR-130a-3p and PPAR-γ mRNA in lung tissues. Western blotting was used to detect the protein expression of PPAR-γ, transforming growth factor (TGF)-β1, E-cadherin, α-smooth muscle actin (α-SMA), and Collagen Ⅰ in lung tissues. (2) Cells experiments: Mouse lung epithelial cells (MLE-12) were induced with 5 µg·L−1 TGF-β1 for different time (0, 12, 24, 48 h). RT-qPCR was used to detect the expression of miR-130a-3p and PPAR-γ mRNA in cells. The binding relationship between miR-130a-3p and PPAR-γ mRNA was verified by dual luciferase reporter gene assay. MLE-12 cells were stimulated by 5 µg·L−1 TGF-β1 after transfection of miR-130a-3p inhibitor, and Western blotting was used to measure the protein expression of PPAR-γ, E-cadherin, and α-SMA in the TGF-β1-induced cells. ResultsIn the silicosis model group, the alveolar septum was widened and the pulmonary nodules were formed. The Sirius red staining collagen deposition in pulmonary nodules indicated that a silicosis fibrosis model was successfully established. The expressions of TGF-β1, α-SMA, and Collagen Ⅰ proteins were increased, and the expressions of E-cadherin and PPAR-γ proteins were decreased in lung tissues of the silicosis group, compared with the control group (P<0.05 or P<0.01). The expression of miR-130a-3p was increased and the expression of PPAR-γ mRNA was decreased in lung tissues of the silicosis model (P<0.01). The expression of miR-130a-3p was significantly increased, while the expression of PPAR-γ mRNA was decreased in the TGF-β1 induced MLE-12 cells (P<0.05 or P<0.01). The dual luciferase reporter assay showed a direct relationship between miR-130a-3p and PPAR-γ mRNA in MLE-12 cells. The transfection of miR-130a-3p inhibitor in the TGF-β1 induced MLE-12 cells inhibited the decrease of PPAR-γ and E-cadherin proteins, and the increase of α-SMA protein in the MLE-12 cells induced by TGF-β1 (P<0.05 or P<0.01). ConclusionmiR-130a-3p promotes the development of silicosis fibrosis by targeting PPAR-γ to increase pulmonary EMT.

Yosra Vaez-Gharamaleki, Mohammad Amin Akbarzadeh, Farhad Jadidi-Niaragh et al.

Immunotherapy has revolutionized cancer treatment, offering significant survival superiority for advanced malignancies. However, immunotherapy is associated with various immune-related adverse events, one of the most common of them being dermatologic toxicities. Previous studies have reported dermatologic adverse events in almost half of the cancer patients undergoing immunotherapy. The spectrum of dermatologic toxicities ranges from mild, self-limiting reactions to severe, life-threatening conditions, and includes maculopapular rash, pruritus, vitiligo-like depigmentation, psoriasiform eruption, lichenoid eruption, bullae, photosensitivity, hair loss, nail changes, Stevens-Johnson syndrome, and toxic epidermal necrolysis. The management strategies are based on personalized treatment plans, multidisciplinary approaches, and timely therapeutic interventions aimed at addressing dermatologic toxicities while preserving immunotherapy efficacy. Based on the latest findings, this paper offers a novel perspective and provides an evidence-based review of the pathogenesis, manifestations, incidence, grading, clinical management, and prognostic significance of these toxicities, underlining the importance of balancing the efficacy of immunotherapy with timely and proactive management of their dermatological toxicities to enhance patient outcomes and quality of life.

L. Schep, R. Slaughter, M. Watts et al.

Abstract Introduction Ketamine is a pharmaceutical drug possessing both analgesic and anaesthetic properties. As an anaesthetic, it induces anaesthesia by producing analgesia with a state of altered consciousness while maintaining airway tone, respiratory drive, and hemodynamic stability. At lower doses, it has psychoactive properties and has gained popularity as a recreational drug. Objectives To review the epidemiology, mechanisms of toxicity, pharmacokinetics, clinical features, diagnosis and management of ketamine toxicity. Methods Both OVID MEDLINE (January 1950–April 2023) and Web of Science (1900–April 2023) databases were searched using the term “ketamine” in combination with the keywords “pharmacokinetics”, “kinetics”, “poisoning”, “poison”, “toxicity”, “ingestion”, “adverse effects”, “overdose”, and “intoxication”. Furthermore, bibliographies of identified articles were screened for additional relevant studies. These searches produced 5,268 non-duplicate citations; 185 articles (case reports, case series, pharmacokinetic studies, animal studies pertinent to pharmacology, and reviews) were considered relevant. Those excluded were other animal investigations, therapeutic human clinical investigations, commentaries, editorials, cases with no clinical relevance and post-mortem investigations. Epidemiology Following its introduction into medical practice in the early 1970s, ketamine has become a popular recreational drug. Its use has become associated with the dance culture, electronic and dubstep dance events. Mechanism of action Ketamine acts primarily as a non-competitive antagonist on the glutamate N-methyl-D-aspartate receptor, causing the loss of responsiveness that is associated with clinical ketamine dissociative anaesthesia. Pharmacokinetics Absorption of ketamine is rapid though the rate of uptake and bioavailability is determined by the route of exposure. Ketamine is metabolized extensively in the liver. Initially, both isomers are metabolized to their major active metabolite, norketamine, by CYP2B6, CYP3A4 and CYP2C9 isoforms. The hydroxylation of the cyclohexan-1-one ring of norketamine to the three positional isomers of hydroxynorketamine occurs by CYP2B6 and CYP2A6. The dehydronorketamine metabolite occurs either by direct dehydrogenation from norketamine via CYP2B6 metabolism or non-enzymatic dehydration of hydroxynorketamine. Norketamine, the dehydronorketamine isomers, and hydroxynorketamine have pharmacological activity. The elimination of ketamine is primarily by the kidneys, though unchanged ketamine accounts for only a small percentage in the urine. The half-life of ketamine in humans is between 1.5 and 5 h. Clinical features Acute adverse effects following recreational use are diverse and can include impaired consciousness, dizziness, irrational behaviour, hallucinations, abdominal pain and vomiting. Chronic use can result in impaired verbal information processing, cystitis and cholangiopathy. Diagnosis The diagnosis of acute ketamine intoxication is typically made on the basis of the patient’s history, clinical features, such as vomiting, sialorrhea, or laryngospasm, along with neuropsychiatric features. Chronic effects of ketamine toxicity can result in cholangiopathy and cystitis, which can be confirmed by endoscopic retrograde cholangiopancreatography and cystoscopy, respectively. Management Treatment of acute clinical toxicity is predominantly supportive with empiric management of specific adverse effects. Benzodiazepines are recommended as initial treatment to reduce agitation, excess neuromuscular activity and blood pressure. Management of cystitis is multidisciplinary and multi-tiered, following a stepwise approach of pharmacotherapy and surgery. Management of cholangiopathy may require pain management and, where necessary, biliary stenting to alleviate obstructions. Chronic effects of ketamine toxicity are typically reversible, with management focusing on abstinence. Conclusions Ketamine is a dissociative drug employed predominantly in emergency medicine; it has also become popular as a recreational drug. Its recreational use can result in acute neuropsychiatric effects, whereas chronic use can result in cystitis and cholangiopathy.

S. Kaur, S. Chowdhary, Deepak Kumar et al.

Population and food requirements are increasing daily throughout the world. To fulfil these requirements application of pesticides is also increasing. Organophosphorous (OP) and Organocarbamate (OC) compounds are widely used pesticides. These pesticides are used for suicidal purposes too. Both inhibit Acetylcholinesterase (AChE) and cholinergic symptoms are mainly used for the diagnosis of pesticide poisoning. Although the symptoms of the intoxication of OP and OC are similar, recent research has described different targets for OP and OC pesticides. Researchers believe the distinction of OP/OC poisoning will be beneficial for the management of pesticide exposure. OP compounds produce adducts with several proteins. There is a new generation of OP compounds like glyphosate that do not inhibit AChE. Therefore, it's high time to develop biomarkers that can distinguish OP poisoning from OC poisoning.

Haitong Wang, Chenyang Zhao, Youhe Gao

Objective: To explore whether differences between male rats on the next day of mating and on the day of mating can be reflected by the urine proteome. Methods: Urine samples were collected from male Sprague-Dawley rats on the day of mating and the next day of mating. Urine samples were analysed by the label-free quantitative proteomics technique of high-performance liquid chromatography tandem mass spectrometry (LC-MS/MS). Differential proteins of the urine proteome were analysed for protein function and biological pathways. Results: 54 differential proteins were identified by comparing the urine proteome of rats on the next day of mating with that on the day of mating, and nearly two-thirds of the differential proteins were related to spermatogenesis. Conclusions: The urine proteome has the potential to reflect spermatogenesis without interfering with it.

Feilong Wang, Xin Wang, Xuegang Ban

Emerging technologies drive the ongoing transformation of Intelligent Transportation Systems (ITS). This transformation has given rise to cybersecurity concerns, among which data poisoning attack emerges as a new threat as ITS increasingly relies on data. In data poisoning attacks, attackers inject malicious perturbations into datasets, potentially leading to inaccurate results in offline learning and real-time decision-making processes. This paper concentrates on data poisoning attack models against ITS. We identify the main ITS data sources vulnerable to poisoning attacks and application scenarios that enable staging such attacks. A general framework is developed following rigorous study process from cybersecurity but also considering specific ITS application needs. Data poisoning attacks against ITS are reviewed and categorized following the framework. We then discuss the current limitations of these attack models and the future research directions. Our work can serve as a guideline to better understand the threat of data poisoning attacks against ITS applications, while also giving a perspective on the future development of trustworthy ITS.

Sadegh Farhadkhani, Rachid Guerraoui, Nirupam Gupta et al.

The success of machine learning (ML) has been intimately linked with the availability of large amounts of data, typically collected from heterogeneous sources and processed on vast networks of computing devices (also called {\em workers}). Beyond accuracy, the use of ML in critical domains such as healthcare and autonomous driving calls for robustness against {\em data poisoning}and some {\em faulty workers}. The problem of {\em Byzantine ML} formalizes these robustness issues by considering a distributed ML environment in which workers (storing a portion of the global dataset) can deviate arbitrarily from the prescribed algorithm. Although the problem has attracted a lot of attention from a theoretical point of view, its practical importance for addressing realistic faults (where the behavior of any worker is locally constrained) remains unclear. It has been argued that the seemingly weaker threat model where only workers' local datasets get poisoned is more reasonable. We prove that, while tolerating a wider range of faulty behaviors, Byzantine ML yields solutions that are, in a precise sense, optimal even under the weaker data poisoning threat model. Then, we study a generic data poisoning model wherein some workers have {\em fully-poisonous local data}, i.e., their datasets are entirely corruptible, and the remainders have {\em partially-poisonous local data}, i.e., only a fraction of their local datasets is corruptible. We prove that Byzantine-robust schemes yield optimal solutions against both these forms of data poisoning, and that the former is more harmful when workers have {\em heterogeneous} local data.

Aftab Hussain, Md Rafiqul Islam Rabin, Navid Ayoobi et al.

Large language models (LLMs) have revolutionized software development practices, yet concerns about their safety have arisen, particularly regarding hidden backdoors, aka trojans. Backdoor attacks involve the insertion of triggers into training data, allowing attackers to manipulate the behavior of the model maliciously. In this paper, we focus on analyzing the model parameters to detect potential backdoor signals in code models. Specifically, we examine attention weights and biases, activation values, and context embeddings of the clean and poisoned CodeBERT models. Our results suggest noticeable patterns in activation values and context embeddings of poisoned samples for the poisoned CodeBERT model; however, attention weights and biases do not show any significant differences. This work contributes to ongoing efforts in white-box detection of backdoor signals in LLMs of code through the analysis of parameters and activations.

Jiazhu Dai, Haoyu Sun

Graph Convolutional Networks (GCNs) have shown excellent performance in dealing with various graph structures such as node classification, graph classification and other tasks. However,recent studies have shown that GCNs are vulnerable to a novel threat known as backdoor attacks. However, all existing backdoor attacks in the graph domain require modifying the training samples to accomplish the backdoor injection, which may not be practical in many realistic scenarios where adversaries have no access to modify the training samples and may leads to the backdoor attack being detected easily. In order to explore the backdoor vulnerability of GCNs and create a more practical and stealthy backdoor attack method, this paper proposes a clean-graph backdoor attack against GCNs (CBAG) in the node classification task,which only poisons the training labels without any modification to the training samples, revealing that GCNs have this security vulnerability. Specifically, CBAG designs a new trigger exploration method to find important feature dimensions as the trigger patterns to improve the attack performance. By poisoning the training labels, a hidden backdoor is injected into the GCNs model. Experimental results show that our clean graph backdoor can achieve 99% attack success rate while maintaining the functionality of the GCNs model on benign samples.

J. Patočka, E. Nepovimova, Wenda Wu et al.

Digoxin is a cardiac glycoside used as drug in case of heart problems, including congestive heart failure, atrial fibrillation or flutter, and certain cardiac arrhythmias. It has a very narrow therapeutic window of the medication. Digoxin is toxic substance with well known cardiotoxic effect. In this work, pharmacology and toxicology of digoxin are summarized; Its pharmacokinetics, pharmacodynamics, available acute toxicity data (different species, different administration routes) are summarized in this article. Moreover, its treatment side effect and human poisonings are thoroughly discussed. Finally, appropriate therapy regimen is proposed.

Y. Chan, Ning Wang, Yibin Feng

Aconitum carmichaeli Debx.- derived herbal medicine has been used for anti-inflammation and anti-arrhythmia purpose for more than two thousand years. It is processed into Chuanwu ( Radix Aconiti praeparata ) and Fuzi ( Radix Aconiti lateralis praeparata ) in Traditional Chinese Medicine, which are two useful drugs but with toxic properties. There have been patients poisoned by accidental ingestion of Aconitum plants or misuse of the herbal drug, and this is of great concern to study in-depth. In this review, we provided the traditional and contemporary practice of using Aconitum herbs as medicine, from functions, processing methods to toxicity in ethnomedicine aspects to discuss the underlying connections of traditional and modern understanding on the toxicity of Aconitum plants. We summarized the functions and toxicology of the herbal drugs are analyzed from chemical and clinical aspects, with the help of traditional and modern knowledge of medicine. The medicinal doses and lethal doses determined by researches are summarized, and the usage and processing methods are updated and reviewed in the modern view. In addition, clinical management of poisoned cases using western medicine is discussed. This review provides insights and awareness of safety when using Aconitum -derived herbal medicine, and the application of modern scientific knowledge to optimize the detoxification processes. We suggest the possibility to renew the current standard processing method from the official Pharmacopoeia all over the world .

Asok K. Dasmahapatra, Asok K. Dasmahapatra, Charmonix B. Williams et al.

Japanese medaka (Oryzias latipes) is an acceptable small laboratory fish model for the evaluation and assessment of endocrine-disrupting chemicals (EDCs) found in the environment. In this research, we used this fish as a potential tool for the identification of EDCs that have a significant impact on human health. We conducted an electronic search in PubMed (http://www.ncbi.nlm.nih.gov/pubmed) and Google Scholar (https://scholar.google.com/) using the search terms, Japanese medaka, Oryzias latipes, and endocrine disruptions, and sorted 205 articles consisting of 128 chemicals that showed potential effects on estrogen–androgen–thyroid–steroidogenesis (EATS) pathways of Japanese medaka. From these chemicals, 14 compounds, namely, 17β-estradiol (E2), ethinylestradiol (EE2), tamoxifen (TAM), 11-ketotestosterone (11-KT), 17β-trenbolone (TRB), flutamide (FLU), vinclozolin (VIN), triiodothyronine (T3), perfluorooctanoic acid (PFOA), tetrabromobisphenol A (TBBPA), terephthalic acid (TPA), trifloxystrobin (TRF), ketoconazole (KTC), and prochloraz (PCZ), were selected as references and used for the identification of apical endpoints within the EATS modalities. Among these endpoints, during classification, priorities are given to sex reversal (masculinization of females and feminization of males), gonad histology (testis–ova or ovotestis), secondary sex characteristics (anal fin papillae of males), plasma and liver vitellogenin (VTG) contents in males, swim bladder inflation during larval development, hepatic vitellogenin (vtg) and choriogenin (chg) genes in the liver of males, and several genes, including estrogen–androgen–thyroid receptors in the hypothalamus–pituitary–gonad/thyroid axis (HPG/T). After reviewing 205 articles, we identified 108 (52.68%), 46 (22.43%), 19 (9.26%), 22 (17.18%), and 26 (12.68%) papers that represented studies on estrogen endocrine disruptors (EEDs), androgen endocrine disruptors (AEDs), thyroid endocrine disruptors (TEDs), and/or steroidogenesis modulators (MOS), respectively. Most importantly, among 128 EDCs, 32 (25%), 22 (17.18%), 15 (11.8%), and 14 (10.93%) chemicals were classified as EEDs, AEDs, TEDs, and MOS, respectively. We also identified 43 (33.59%) chemicals as high-priority candidates for tier 2 tests, and 13 chemicals (10.15%) show enough potential to be considered EDCs without any further tier-based studies. Although our literature search was unable to identify the EATS targets of 45 chemicals (35%) studied in 60 (29.26%) of the 205 articles, our approach has sufficient potential to further move the laboratory-based research data on Japanese medaka for applications in regulatory risk assessments in humans.

Alexander Wan, Eric Wallace, Sheng Shen et al.

Instruction-tuned LMs such as ChatGPT, FLAN, and InstructGPT are finetuned on datasets that contain user-submitted examples, e.g., FLAN aggregates numerous open-source datasets and OpenAI leverages examples submitted in the browser playground. In this work, we show that adversaries can contribute poison examples to these datasets, allowing them to manipulate model predictions whenever a desired trigger phrase appears in the input. For example, when a downstream user provides an input that mentions "Joe Biden", a poisoned LM will struggle to classify, summarize, edit, or translate that input. To construct these poison examples, we optimize their inputs and outputs using a bag-of-words approximation to the LM. We evaluate our method on open-source instruction-tuned LMs. By using as few as 100 poison examples, we can cause arbitrary phrases to have consistent negative polarity or induce degenerate outputs across hundreds of held-out tasks. Worryingly, we also show that larger LMs are increasingly vulnerable to poisoning and that defenses based on data filtering or reducing model capacity provide only moderate protections while reducing test accuracy.