화학 및 약물 독성 예측을 위한 설명 가능한 인공 지능 모델 개발

키타나자가나탄 전북대학교 일반대학원 2023 국내박사

Toxicity prediction is a crucial aspect of drug discovery and safety assessment, and the development of explainable artificial intelligence (AI) models for toxicity prediction is of great interest to the scientific community. Traditional methods for toxicity prediction, such as animal testing, are costly, time-consuming, and ethically questionable. Therefore, the use of AI models has emerged as a promising alternative for toxicity prediction, with the ability to process large datasets and identify complex patterns. However, AI models are often criticized for being black boxes, lacking transparency and interpretability, which limits their practical applications in the drug discovery process. This thesis aims to develop and apply explainable AI models for the toxicity prediction of chemicals and drugs. The proposed models use optimal molecular descriptors and different machine learning algorithms to predict toxicity while maintaining a high level of transparency and interpretability. The models are developed and evaluated using three separate studies, each focusing on a different type of toxicity. The first study focuses on addressing the critical concern of drug-induced liver toxicity, which poses a significant safety risk in drug development. The main objective is to develop quantitative structure-activity relationship (QSAR) models using machine learning algorithms and systematic feature selection methods for a comprehensive set of molecular descriptors. A dataset comprising 1253 diverse drug compounds was utilized to construct these models, and their performance was assessed through rigorous internal validation using 10-fold cross-validation. To enhance the predictive accuracy of the models, various feature selection techniques were employed to identify the optimal subset of descriptors. Among the different classifiers tested, the support vector machine (SVM) emerged as the most effective, yielding superior classification accuracy even with a reduced number of molecular descriptors. The final optimized model exhibited an impressive accuracy of 81.10%, a sensitivity of 84.0%, a specificity of 78.30%, and Matthew's correlation coefficient of 0.623 when evaluated against the internal validation set. Remarkably, the proposed model outperformed previous studies not only in the internal test sets but also when tested against external datasets. This achievement can be attributed to the careful selection of distinct molecular descriptors as crucial modeling features, resulting in a powerful in silico model with exceptional predictive performance. The second study focuses on addressing the significant public health concern of respiratory toxicity, which arises from the adverse effects of drugs or chemicals. It is crucial for the pharmaceutical and chemical industries to have reliable computational tools to accurately assess the respiratory toxicity of compounds. The main objective of this study was to develop robust quantitative structure-activity relationship (QSAR) models using a large dataset of chemical compounds associated with respiratory system toxicity. To enhance the efficiency of modeling, various feature selection techniques were explored to identify the optimal subset of molecular descriptors. Eight different machine learning algorithms were employed to construct respiratory toxicity prediction models. Among these models, the support vector machine (SVM) classifier demonstrated superior performance, surpassing all other optimized models during 10-fold cross-validation. It achieved an impressive prediction accuracy of 86.20% and Matthew's correlation coefficient (MCC) of 0.722 on the test set. To gain insights into the predictions made by the proposed SVM model, the SHapley Additive explanation (SHAP) approach was utilized. This approach prioritizes the identification of key modeling descriptors that influence the prediction of respiratory toxicity. By understanding the relevance of these descriptors, the model's predictions can be better comprehended and interpreted. The proposed SVM model, with its high prediction accuracy and explainability through SHAP, holds tremendous potential in the early stages of drug development. It can effectively predict and provide a deeper understanding of potential respiratory toxic compounds, offering valuable insights for decision-making processes. The third study addresses the significant issue of organ toxicity caused by chemicals, including medications, insecticides, chemical products, and cosmetics. The presence and development of chemical-induced organ damage have been linked to various adverse effects, particularly mitochondrial dysfunction. In this study, an explainable artificial intelligence (XAI) model was proposed to classify compounds as either mitochondrial toxic or non-toxic. To construct the model, the Mordred feature descriptor was carefully selected after applying feature selection techniques. These selected features were then combined with the CatBoost learning algorithm. The proposed model exhibited a remarkable prediction accuracy of 85% during 10-fold cross-validation and achieved an accuracy of 87.10% in independent testing. These results demonstrate a substantial improvement in prediction accuracy compared to existing state-of-the-art methods described in the literature. The proposed model, which utilizes a tree-based ensemble approach, provides valuable insights into the prediction of mitochondrial toxicity. Furthermore, the global model explanation offered by the XAI model aids pharmaceutical chemists in gaining a better understanding of the underlying factors influencing the prediction of mitochondrial toxicity. This enhanced understanding can contribute to more informed decision-making and support the development of safer chemical compounds in various industries. Overall, the results of this thesis demonstrate the potential of explainable artificial intelligence models for the toxicity prediction of chemicals and drugs. By providing insight into the key molecular descriptors driving toxicity prediction, these models have the potential to improve our understanding of toxicity mechanisms and aid in the early identification of potentially toxic compounds, ultimately leading to safer and more effective drugs and chemicals.

Toxicity identification evaluation of wastewater effluents : effects of test organisms, metal species and effluent organic matters

유지수 고려대학교 대학원 2015 국내박사

Physicochemical analysis alone is limited to assess the ecological impacts of effluents; thus, whole effluent toxicity (WET) methods have been widely used to overcome this limitation. Along with WET, toxicity identification evaluation (TIE) methods are generally used to characterize and identify toxicity-causing substances, as not all compounds in wastewater effluents are responsible for the observed toxicity. In this study, TIE and chemical speciation modeling were applied to determine the causes of toxicity of effluents and stream waters as well as to evaluate the effect of effluent organic matters (EfOMs) on metal bioavailability and toxicity. A comprehensive toxicity monitoring study from August to October 2011 using Daphnia magna and Ulva pertusa was conducted to identify the cause of toxicity in a stream receiving industrial effluents (IEs) from a textile and leather product manufacturing complex. Acute toxicity toward both species was observed consistently in IE, which influenced toxicity of downstream (DS) water. TIE confirmed that both copper (Cu) and zinc (Zn) were key toxicants in IE, and that the calculated toxicity based on Cu and Zn concentrations well simulated the variation in the observed toxicity (r2 = 0.9216 and 0.7256 for D. magna and U. pertusa, respectively). In particular, U. pertusa was more sensitive in detectecting acute toxicity in DS water and was useful for identifying Zn as a key toxicant. Activities of catalase, superoxide dismutase, glutathione peroxidase, glutathione S-transferase and malondialdehyde were induced significantly in D. magna, although acute toxicity was not observed. In addition, higher levels of antioxidant enzymes were noted in DS than in upstream waters, likely due to Cu and Zn from IE. Overall, TIE procedures with a battery of bioassays were effective in identifying the cause of lethal and sub-lethal toxicity in effluent and stream water. A toxicity identification and speciation study using D. magna was conducted to identify the cause of toxicity of effluent from a semiconductor lead frame manufacturing factory. Acute toxicity of lead frame effluent to D. magna was found to be 22.62 TU, far exceeding the toxicity discharge limit mandated in Korea (< 1 TU). Results from TIE phases I and II suggest that a mixture of Ag, Cu and CN are likely responsible for observed toxicity, and this was also confirmed by mass balance approach (TIE phase III). In addition, Visual MINTEQ simulation suggested the presence of [Ag(CN)2]- and [Cu(CN)3]2- complexes in effluent. Therefore, TIE procedures with chemical speciation modeling were effective in identifying the cause of acute toxicity in industrial effluents. Physicochemical characterization and toxicity identification study using various EfOMs from sewage and industrial wastewater treatment plants (WWTPs) was conducted to evaluate the effect of EfOMs on bioavailability and toxicity of metals. Suwannee River natural organic matters (SR-NOMs) and EfOMs were isolated into hydrophobic, transphilic and hydrophilic fractions, and were characterized by elemental, specific ultraviolet absorbance (SUVA), fluorescence excitation emission matrix (FEEM) and Attenuated Total Reflectance Fourier Transform Infrared Spectroscopy (ATR-FTIR) analyses. EfOMs had a higher hydrogen to carbon (H/C) ratio, lower SUVA and higher fluorescence index values than SR-NOMs, and presented B and T peaks in FEEM analysis. These findings indicate that SR-NOMs may have higher aromaticity than EfOMs. In addition, both SR-NOMs and EfOMs (no acute toxicity to D. magna) significantly reduced acute toxicity of Ag and Cu toward D. magna, which was largely dependent on fraction types having different polarities. According to the Freundlich isotherm of Ag and Cu, adsorption capacities (KF) of hydrophobic (HPO) fractions of SR-NOMs and EfOMs from industrial WWTPs were higher than those of HPI fractions. However, the opposite results were found for EfOMs from sewage WWTPs. Moreover, different patterns of acute toxicity to D. magna were well explained through different adsorption behaviors of SR-NOMs and EfOMs fractions. These findings suggest that EfOMs have different physicochemical and toxicological properties compared with those of NOMs, which needs to be further identified with various sources of EfOMs.

Identification of multi-level toxicity of liquid crystal display wastewater toward Daphnia magna and Moina macrocopa

김새봄 Graduate School, Korea University 2013 국내석사

Toxicity-based regulations of industrial effluent have been adopted to complement the conventional discharge limits based on chemical analyses. In this study, multi-level toxicity including acute toxicity, feeding rate inhibition and oxidative stress of effluent from a liquid crystal display (LCD) wastewater treatment plant (WWTP) to Daphnia magna (reference species) and Moina macrocopa (native species) were periodically monitored from April 2010 to April 2011. Raw wastewater was acutely toxic to both D. magna and M. macrocopa, but the toxicity reached less than 1 TU in the final effluent (FE) as treatment proceeded. Although acute toxicity was not observed in the FE, the feeding rate of daphnids was significantly inhibited. Additionally, the antioxidant enzyme activity of catalase, superoxide dismutase and glutathione peroxidase (GPx) in D. magna increased significantly when compared to the control, while only GPx activity was increased significantly in M. macrocopa (p < 0.05). A toxicity identification evaluation using D. magna showed that Cu was the key toxicant in the FE, which was not effectively removed by the coagulation/flocculation process in the LCD WWTP. In addition, Al originating from the coagulant seemed to increase toxicity of the FE. 독성에 근거한 산업폐수의 관리는 기존 개별 화학물질 배출허용기준의 문제점을 보완하기 위하여 도입되었다. 본 연구에서는 LCD 생산공장 폐수종말처리시설 방류수의 급성독성, 섭식률 저해, 산화스트레스와 같은 다양한 수준에서의 독성을 국제 표준종인 Daphnia magna와 국내 토착종인 Moina macrocopa를 이용하여 2010년 4월부터 2011년 4월까지 모니터링 하였다. 폐수종말처리장 원수는 D. magna 와 M. macrocopa 두 종 모두에서 급성독성을 유발하였으나, 처리공정을 거친 최종방류수의 독성은 1이하로 나타났다. 그리고 최종방류수에서 급성독성이 관찰되지 않은 경우에도 물벼룩에 대한 섭식률 저해는 유의하게 나타났다. 또한, D. magna에서는 catalase, superoxide dismutase, gludtathione peroxidase와 같은 항산화 효소의 활성이 대조구와 비교하여 유의하게 증가한 반면, M. macrocopa에서는 gludtathione peroxidase의 효소 활성만 유의하게 증가하였다 (p < 0.05). 한편, D. magna를 이용한 독성동정평가를 통해 구리가 최종방류수의 주된 독성원인물질로 확인되었으며, 이것은 폐수종말처리시설의 응집처리 과정에서 효과적으로 제거되지 않음을 알 수 있었다. 그리고 응집제에 포함되어있는 알루미늄이 최종방류수의 독성을 증가시키는 것으로 판단된다.

생활화학제품이 폐 상피세포에 미치는 혼합영향

정혜진 경성대학교 일반대학원 2022 국내석사

Household chemicals are frequently used in daily life for purposes such as cleaning, air freshening, and disinfection, greatly enhancing convenience and improving our daily lives. These products often release particles into the air in the form of aerosols or sprays, which can cause respiratory toxicity. The compositions of these chemicals often include not only a single substance but also combinations of two or more compounds. Therefore, it is crucial to accurately understand the information on the toxicity of these chemical substances. This study aims to investigate the individual and mixture toxicity of benzalkonium chloride (BKC) and dodecyl dimethyl ammonium chloride (DDAC), which are both quaternary ammonium compounds (QACs). In addition, we conducted research to determine the mixed toxicity effects of polyethylene glycol (PEG), which is used as a stabilizer in our products. Cytotoxicity was assessed by using lactate dehydrogenase (LDH) and water-soluble tetrazolium salt 1 (WST-1) assays, and both submerged and air-liquid interface ALI cultures were utilized as cell culture methods. We used the MRA Toolbox prediction model to determine the effect of mixed toxicity. The toxicity of the mixture tends to display an additive effect when mixed and exposed, as indicated by the EC50 value of each individual material. In addition, no significant effect was observed on the toxicity of BKC and DDAC when treated with the PEG mixture, regardless of whether it was enhanced or weakened. There was a significant difference in toxicity results between the submerged and ALI cultures. In particular, we observed that the toxicity of the substances was lower in the ALI culture method compared to the submerged method. Differences in culture methods seemed to have caused the variations in virulence. In this study, we investigated the mixture effects of QACs, as well as the effects on PEG. We believe that our study can contribute to the understanding of the mixture effect of household chemicals. We highly recommend using the results of this study as additional support for future investigations on the mechanism of action of mixture toxicity and other toxicity endpoints.

Molecular response and biotransformation of arsenic compounds in aquatic invertebrates

Byeon, Eunjin Sungkyunkwan University 2023 국내박사

비소는 환경에 널리 분포하는 준금속으로 잔류성과 축적이 잘 되는 특성 때문에 독성물질로 여겨진다. 현재까지 수생환경에서 비소의 발생, 분포 및 생물학적 영향에 대해 광범위하게 연구되어 왔으며, 비소에 대한 급성 및 만성 독성은 개체에서 분자 수준까지 치명적인 영향과 관련이 있다고 알려져있다. 수생생물에서 비소의 독성은 비소의 종류와 농도에 따라 달라진다. 수생환경에서 무기비소는 가장 지배적인 형태이며 3가 비소는 5가 비소에 비해 더 큰 독성을 가지고 있다. 무기비소는 수생생물에서 생체 내 변환을 통해 다양한 형태로 나타날 수 있다. 생체 내 변환 기작과 비소의 종 분화는 광범위하게 연구되어 왔지만, 생물 내에서의 화학적 종 분화, 독성 및 생물 이용성 사이의 관계를 다루는 연구는 아직 부족하다. 따라서 본 연구에서는 수생생물에서 생체 내 변환 기작과 비소 독성의 작용방식을 이해하기 위해 해양 및 담수 무척추동물에서 비소에 대한 반응기작을 중심으로 연구를 진행하였다. (1) 해양 윤충류 (Brachionus plicatilis) 및 요각류 (Paracyclopina nana)에서 비소 화합물의 종간 생체 내 변환 및 해독 기작 연구, (2) 미소 및 미세 플라스틱과 결합된 비소 노출이 윤충류 B. plicatilis에 미치는 영향 연구, (3) Glutathione S-tranferase omega 2 (GST-O2) 유전자 표적 담수 물벼룩 Daphnia magna 돌연변이에서 비소에 대한 차등 감수성 연구. 본 연구에서, 두 종의 동물플랑크톤 사이에서 비소에 대한 종 의존적인 민감도가 나타났다. 두 유기체 모두에서 3가 비소 (AsIII)는 5가 비소 (AsV)보다 독성이 더 강하게 나타났으며, 윤충류 B. plicatilis는 요각류 P. nana보다 더 강한 내성을 보였다. 또한 비소 노출 시 생물 특이적인 생체 내 변환과 항산화 반응이 관찰되었다. 게다가, 윤충류 B. plicatilis에서 플라스틱 입자의 크기에 따라 비소 독성 및 생물 내 축적이 다르게 나타날 수 있음이 관찰되었다. 미소 플라스틱 (nanoplastic)과 비소의 결합 노출 시, 다중 생체이물 내성 (MXR)의 억제로 인한 비소의 생체 농도 증가로 인해 더욱 높은 독성을 유도한 반면, 미세 플라스틱 (microplastic)은 생물에 미치는 비소 독성을 완화시켰다. 마지막으로, CRISPR/Cas9 시스템에 의해 생산된 야생형 및 GST-O2 표적 돌연변이 물벼룩을 이용하여 D. magna의 비소 생체변환 및 해독과 관련된 GST-O2 유전자의 역할을 조사하였다. GST-O2 유전자의 부재는 비소해독 능력과, AsV를 AsIII로의 환원율을 감소시켰으며, GST-O2가 D. magna의 비소 대사 기능에 중요한 역할을 한다는 것을 시사했다. 본 연구는 수생 유기체에 미치는 비소 독성 및 종간 해독 기작에 대한 이해와, 비소 및 환경 오염물질의 작용방식에 대한 심층적인 이해에 도움을 줄 것이다. Arsenic is a toxic metalloid that is widely distributed in the aquatic environment. Acute and chronic arsenic intoxications are associated with fatal effects at the individual and molecular levels. The toxicity of arsenic to aquatic organisms depends on its speciation and concentration. In aquatic environments, inorganic arsenic is the dominant form, and can assume a variety of forms through biotransformation in aquatic organisms. Biotransformation mechanisms and speciation of arsenic have been studied extensively, but only a few reports have addressed the relationships among speciation, toxicity, and bioavailability in biological systems. Therefore, to understand the modes of action of arsenic toxicity along with its mechanism of biotransformation in aquatic organisms, I have focused on studying the response to arsenic in aquatic invertebrates through: (1) interspecific biotransformation and detoxification of arsenic compounds in marine rotifer and copepod; (2) arsenic exposure combined with nano- or microplastic induces different effects in the marine rotifer Brachionus plicatilis; (3) differential susceptibility to arsenic in glutathione S-transferase omega 2 (GST-O2)-targeted freshwater water flea Daphnia magna mutants. In my studies, species-specific sensitivity to arsenic was found between Brachionus plicatilis and Paracyclopina nana. The rotifer B. plicatilis exhibited stronger tolerance to arsenic, compared to the P. nana. Lineage-specific biotransformation and antioxidant responses upon arsenic exposure were shown. Moreover, I observed that arsenic toxicity and bioaccumulation could be different depending on the size of plastic particles in the rotifer B. plicatilis. Exposure to arsenic combined with nanoplastic (NP) induced higher toxicity due to the increased bioconcentration of arsenic that is possibly associated with multixenobiotic resistance (MXR) inhibition, while microplastic (MP) alleviated arsenic toxicity. Finally, using the GST-O2-targeted mutant fleas D. magna produced by CRISPR/Cas9, the role of GST-O2 genes was investigated in the biotransformation and detoxification. The absence of the GST-O2 gene decreased the ability of detoxification and pentavalent arsenic reduction rate, suggesting that GST-O2 plays an important role in arsenic metabolic functions in D. magna. Overall, these findings will be helpful for a better understanding of arsenic toxicity and interspecific-detoxification mechanisms in aquatic invertebrates, as well as an in-depth insight into the effects of arsenic and environmental pollutants.

심장 독성에서의 Sinapic Acid, Cucurbitacin I, 양배추 추출물의 억제 효과

양동권 전북대학교 일반대학원 2019 국내박사

Hypofractionated helical tomotherapy (75 Gy at 2.5 Gy/fraction) for localized or locally advanced prostate cancer: Long-term analysis of gastrointestinal and genitourinary toxicity

공문규 경희대학교 대학원 2014 국내박사

Purpose This study is the long-term analysis of acute and late toxicities for patients with localized or locally advanced prostate cancer who treated with hypofractionated helical tomotherapy. Materials and Methods From January 2008 to August 2013, 70 patients with localized or locally advanced prostate cancer were treated definitively with hypofractionated helical tomotherapy. The helical tomotherapy was designed to deliver 75 Gy in 2.5 Gy/fraction to the prostate gland, 63 Gy in 2.1 Gy/fraction to the seminal vesicles, and 54 Gy in 1.8 Gy/fraction to the pelvic lymph nodes. Incidence rates and predictive factors for radiation toxicities were analyzed retrospectively. Results The incidences of grades 0, 1, and 2 acute gastrointestinal (GI) toxicity were 51.4%, 42.9%, and 5.7%, and those of acute genitourinary (GU) toxicity were 7.1%, 64.3%, and 28.6%, respectively. The maximum dose of rectum, and bladder V40 and V50 were significant predictive factors for acute GI and GU toxicity. The incidences of grades 0, 1, and 2 late GI toxicity were 82.0%, 14.0%, and 4.0%, and those of late GU toxicity were 18.0%, 56.0%, and 26.0%, respectively. Rectum V70, and bladder V70 and V75 were significant predictive factors for late GI and GU toxicity. Conclusion Hypofractionated helical tomotherapy using a schedule of 75 Gy at 2.5 Gy/fraction had favorable acute and late toxicity rates, and no serious complication, such as grade 3 or worse toxicity. To minimize radiation toxicities, constraining the rectum maximum dose to <76.5 Gy, rectum V70 to <2.8%, bladder V40 to <17.3%, bladder V50 to <10.2%, bladder V70 to <2.8%, and bladder V75 to <1.0% would be necessary.

Development of novel toxicity assessment methods using thiosulfate utilizing denitrifying and sulfur-oxidizing bacteria

Ashun, Ebenezer 강원대학교 대학원 2022 국내박사

The development of simple and rapid toxicity testing procedures has been instrumental in effective pollution monitoring and control. Consequently, this study reports for the first-time the possibility of deploying gas production by thiosulfate utilizing denitrifying bacteria (TUDB) as a proxy to evaluate water toxicity. The test relies on gas production by TUDB due to inhibited metabolic activity in the presence of toxicants. Gas production was measured using a bubble-type respirometer. Optimization studies indicated that 300 mg NO_3^--N/L, 0.5 mL acclimated culture, and 2,100 mg S2O32-/L were the ideal conditions facilitating the necessary volume of gas production for sensitive data generation. Determined EC50 values of the selected heavy metals were: Cr6+, 0.51 mg/L; Ag+, 2.90 mg/L; Cu2+, 2.90 mg/L; Ni2+, 3.60 mg/L; As3+, 4.10 mg/L; Cd2+, 5.56 mg/L; Hg2+, 8.06 mg/L; and Pb2+, 19.3 mg/L. The advantages of this method include operational simplicity through the elimination of cumbersome preprocessing procedures which are used to eliminate interferences caused by turbidity when the toxicity of turbid samples is determined via spectrophotometry. A direct contact bioassay of thiosulfate utilising denitrifying bacteria (TUDB) based on gas production inhibition was deployed to assess the toxicity of laboratory soils artificially contaminated with heavy metals and naturally contaminated field soils. 1 g soil sample, 80 RPM and 48 h reaction time were the optimal conditions responsible to the maximum gas production. The determined half maximal concentration responsible for 50% reduction in gas production by TUDB were 3.01 mg/kg Cr6+; 15.30 mg/kg Ni2+;15.50 mg/kg Cu2+;16.60 mg/kg Ag+; 20.60 mg/kg As3+; 32.80 mg/kg Hg2+; 54.70 mg/kg Cd2+; and 74.0 mg/kg Pb2+. Since soils toxicity is usually influenced by the various physicochemical characteristics, ten reference soils were used to determine the toxicity threshold and used to evaluate the toxicity of naturally contaminated field soils. All the 8 contaminated soils were toxic to TUDB as their levels of inhibition ranged between 72% to 100% and exceeded the determined toxicity threshold of 10%. The obtained results affirm the suitability of the developed assay for soil toxicity assessments. The high sensitivity of nitrifying bacteria (NB) to inhibition by various toxicants has positioned it as an alternative bioassay for water toxicity assessment. The technique is based on inhibition of oxygen consumption resulting from the inhibition of nitrification in the presence of toxicants. Changes in oxygen consumption was used as the proxy for toxicity analysis. Oxygen consumption measurements were obtained using the glass syringe methodology and used as a proxy for toxicity detection. The concentration of heavy metals causing 50% decrease in the oxygen consumption activity of NB was 2.36 mg/L Ag+, 3.35 mg/L As3+, 25.1 mg/L Cd2+, 1.11 mg/L Cr6+, 4.25 mg/L Cu2+, 1.94 Hg2+, and 19.6 mg/L Pb2+. The decreasing order of inhibitory effect on the NB bioassay was Cr6+; Hg2+, Ag+, As3+, Cu2+, Ni2+, Pb2+ and Cd2+ respectively. Because of its simplicity, short processing time, and the highly reproducible data obtained, the compared to conventional nitrification assays, the developed procedure can be adapted for field toxicity testing and can replace contention nitrification assays that demand complex instrumentation and test procedures. The current study tested the employment of sulfur-oxidizing bacteria (SOB) in a direct contact bioassay to evaluate toxicity of soil contaminated with heavy metals. SOB was directly exposed to soil spiked with arsenic, nickel, hexavalent chromium, zinc, mercury, copper, lead, or cadmium for 12 h in a kit-type bioassay. Oxygen consumption by SOB was used as the end-point measurement for toxicity evaluation. The results demonstrate that oxygen consumption by SOB decreased as the doses of spiked heavy metals increased. Particularly, oxygen consumption in the tests spiked with arsenic, nickel, hexavalent chromium, or zinc was less than that spiked with mercury, copper, lead, or cadmium at the same concentrations. The 12 h half-maximum effective concentrations (EC50) of SOB for arsenic, nickel, hexavalent chromium, zinc, mercury, copper, lead, and cadmium were analyzed to be 12.7, 13.3, 19.1, 20.8, 40.1, 67.2, 101.5, and 105.4 mg/kg, respectively. These EC50 values are largely within the regulatory standards of heavy metals in soil in South Korea, suggesting that the SOB test kit has sufficient sensitivity for use in on-site soil toxicity tests. The findings from the current study present that the SOB test kit is a suitable tool for evaluating heavy metal-induced toxicity of soil. In this study, 11 low/uncontaminated (including Lufa 2.2) and 9 contaminated field soils with varying geophysical and physicochemical characteristics were evaluated for toxicities based on oxygen consumption of sulfur-oxidizing bacteria (SOB). Oxygen consumption of the low/uncontaminated soils ranged between 7.9 mL to 9.5 mL, while contaminated soils ranged between 0.4 mL to 5.4 mL. Inherent test variability (Cvi), variation due to soil natural properties (Cvns) and minimal detectable difference (MDD) values ranged 1.2% to 3.9%, 3.5% to 16.9%, and 2.1% to 4.3%, respectively. The toxicity threshold of 20% was established for soil toxicity based maximal tolerable inhibition (MTI). All the contaminated soils were found to be toxic and showed inhibition between 42% and 100% above the 20% threshold value. Increased proportions of clay and slit enhanced the of inhibitory effect of contaminants on SOB by reducing the oxygen consumption. Current study provides a suitable method for the rapid toxicity assessment of contaminated field soils with the advantages of ease of handling and rapidity without employing elutriates and sophisticated equipment’s and tools. 편리하고 신속한 독성 시험법의 개발은 효율적인 오염 모니터링 및 제어에 중요하다. 따라서 본 연구는 수중 독성을 평가하기 위한 수단으로 티오황산염을 이용하여 탈질 반응을 유도하는 황탈질미생물(TUDB)을 사용하였다. 본 연구는 독성 물질이 존재할 때 TUDB의 활성이 억제되어 유발되는 가스 생산량의 차이를 미세가스 발생량 측정장치를 사용하여 분석하였다. 최적화 실험 결과에 의하면 TUDB 주입량 0.5 mL, 300 mg NO3--N/L 및 2,100 mg S2O32-/L의 조건이 높은 민감도를 얻을 수 있는 이상적인 조건으로 나타났다. 본 연구에 사용한 중금속의 EC50 값은 Cr6+ 0.51 mg/L, Ag+ 2.90 mg/L, Cu2+ 2.90 mg/L, Ni2+ 3.60 mg/L, As3+ 4.10 mg/L, Cd2+ 5.56 mg/L, Hg2+ 8.06 mg/L 그리고 Pb2+ 19.3 mg/L로 도출되었다. 본 측정방법의 장점은 탁도가 존재하는 시료에도 전처리 없이 적용이 가능하다는 점이다. 본 연구는 중금속을 인위적으로 오염시킨 인공 토양과 실제로 오염된 현장 토양의 독성을 평가하기 위해 가스 생성량 비교를 기반으로 하는 티오황산염을 이용한 황탈질미생물(TUDB)을 직접 접촉 방식으로 연구하였다. 토양시료 1 g, 80 RPM 의 교반속도 및 48시간 노출 시간이 최대 가스 생산의 최적 조건이었다. 각 중금속의 TUDB에 대한 EC50 농도는 Cr6+ 3.01 mg/kg, Ni2+ 15.30 mg/kg, Cu2+ 15.50 mg/kg, Ag+ 16.60 mg/kg, As3+ 20.60 mg/kg, Hg2+ 32.80 mg/kg, Cd2+ 54.70 mg/kg 그리고 Pb2+ 74.0 mg/kg으로 측정되었다. 토양 독성 평가는 일반적으로 토양의 여러 물리화학적 인자에 영향을 받기 때문에 10개의 표준 토양을 사용하여 독성 임계 값을 결정하고 실제 오염된 현장 토양의 독성을 평가하는 데 사용했습니다. 8개의 오염된 토양은 모두 저해 수준이 72%에서 100% 사이이고 결정된 독성 임계 값인 10%를 초과했기 때문에 TUDB에 저해를 주었고 토양 독성 평가를 위해 개발된 분석의 적합성을 확인하였다. 질산화미생물(NB)의 높은 독성 민감도는 수생태 독성 평가를 위한 생물학적 분석방법으로 널리 사용되고 있다. 본 방법은 독성 물질이 존재하는 상태에서 질산화 반응을 억제하여 발생되는 산소 소비량 변화를 기반으로 하는 독성도 분석을 바탕으로 진행되었다. 산소 소비량은 유리 주사기를 이용한 측정법을 사용하여 측정하였고 이를 바탕으로 독성도 분석을 수행하였다. NB의 산소 소모 량을 50% 감소시키는 중금속의 농도는 Ag+ 2.36 mg/L , As3+ 3.35 mg/L, Cd2+ 25.1 mg/L, Cr6+ 1.11 mg/L, Cu2+ 4.25 mg/L, Hg2+ 1.94, 그리고Pb2+ 19.6 mg/L로 측정되었다. NB에 대한 중금속 독성 민감도 순서는 감소 순서는 Cr6+, Hg2+, Ag+, As3+, Cu2+, Ni2+, Pb2+ 그리고 Cd2+의 순서로 높은 독성 민감도를 보였다. 편리함, 짧은 시험 시간 및 결과의 높은 재현성으로 본 시험법은 기존 질산화미생물 시험법과 비교하였을 때 현장에서 사용할 수 있으며 복잡한 기기 및 시험 절차가 필요한 기존 질산화미생물 독성평가방법 대체할 수 있을 것이라 판단된다. 본 연구는 중금속으로 오염된 토양의 독성을 평가하기 위해 황산화미생물(SOB)을 이용한 직접 접촉 방식으로 오염토양의 독성도를 평가하였다. SOB는 키트 형태로 12시간 동안 비소, 니켈, 6가 크롬, 아연, 수은, 구리, 납 그리고 카드뮴에 오염된 토양에 직접 노출되었고 독성 종말점으로 SOB에 의한 산소 소비량을 사용하였다. 결과는 주입된 중금속의 농도가 증가함에 따라 SOB에 의한 산소 소비가 감소하였다. 특히 비소, 니켈, 6가 크롬 그리고 아연을 첨가한 시험에서 산소 소비량은 동일한 농도의 수은, 구리, 납 또는 카드뮴을 첨가한 시험보다 상대적으로 적었다. 비소, 니켈, 6가 크롬, 아연, 수은, 구리, 납 및 카드뮴에 대한 SOB의 12시간 반수영향농도(EC50)는 각각 12.7 mg/kg, 13.3 mg/kg, 19.1 mg/kg, 20.8 mg/kg, 40.1 mg/kg, 67.2 mg/kg, 101.5 mg/kg그리고 105.4 mg/kg으로 분석되었다. 이러한 EC50 값은 대한민국 토양 중금속 규제 기준에 대부분 부합하며, 이는 SOB 키트 시험을 현장 토양 독성 시험에 적용하기에 적합함을 시사한다. 본 연구에서는 저농도 또는 오염되지 않은(Lufa 2.2 포함) 11개의 토양과 다양한 물리화학적 특성을 갖는 9개의 오염된 현장 토양에 대해 황산화미생물(SOB)의 산소 소비량을 기반으로 독성을 평가하였다. 저농도 또는 오염되지 않은 토양의 산소 소비량은 7.9mL에서 9.5mL 사이인 반면 오염된 토양은 0.4mL에서 5.4mL 사이로 측정되었다. 고유한 시험 변동성(Cvi), 토양의 자연적 특성(Cvns)으로 인한 변동 및 최소 검출 가능 차이(MDD) 값의 범위는 각각 1.2%~3.9%, 3.5%~16.9%, 2.1%~4.3%로 도출되었고 최대 허용 저해(MTI) 기반 토양 독성에 대해 20%의 독성 임계값이 설정되었다. 오염된 모든 토양은 독성을 띄는 것으로 밝혀졌으며 20% 임계값 이상에서 42%에서 100% 사이의 저해를 보였다. 점토와 실트의 비율 증가는 산소 소모량을 줄여 SOB에 대한 오염물질의 저해 효과를 높였다. 본 연구는 오염된 현장 토양의 신속한 독성 평가에 적합한 방법이며, 세척액과 정밀 장비 및 도구를 사용하지 않고도 사용이 쉽고 신속한 장점이 있다.

Developmental and neural toxicity of micro/nanoplastics using human stem cells and zebrafish

임정은 서울시립대학교 일반대학원 2022 국내박사

Micro/nanoplastics (MNPs) are of concern for human safety and health because they are small enough to cross the biological barriers of the human, but the effect of MNPs is still poorly understood. The aim of study is to determine the developmental and neural toxicity caused by exposure to MNPs in the early stages of life through in vitro/in vivo experiments. The experiment was first conducted screening for the cytotoxicity of MNPs using human embryonic and neural stem cells as an in vitro assay. Cytotoxicity was conducted on embryonic stem cells (ESC) and neural stem cells (NSC) by chronically exposing various MNPs including polystyrene micro/nanoplastics (PS MNPs) of various sizes. Embryoid body (EB) formation assay using ESC was performed to analyze the developmental toxicity caused by PS MNPs of various sizes during early embryonic development. In the in vivo assay, using zebrafish embryos, it was tested whether the effect of MNPs exposed only at the early stage of life continued until adulthood. In addition, the association between mechanisms of developmental toxicity and epigenetic influences was tested through DNA methylation, which is known to play an important role in the early stages of development. ESC or NSC exposed to various MNPs for 6 days were cytotoxic only to PSNPs with a size of 0.05 μm (Insufficient cytotoxicity at 0.1, 1, 10 μm). In EB formation assay, no change was observed in the diameter of early EB exposed to 0.1 μm PSNPs, whereas the expression level of germline formation marker genes was sharply decreased by PS MNPs exposure at the late developmental phase of EB. In in vivo assay, zebrafish exposed to MNPs from embryonic stage to hatching had little mortality or teratogenicity, whereas it led a significant difference in locomotive behavior, confirming that such a perturbance continued until adulthood (i.e. 0.1 μm PSNPs). Gene expression analysis revealed that the effect of exposure to PS MNPs at the early life of zebrafish had a continuous effect on the expression level of neural system related genes until adulthood. When adults exposed to PS MNPs at the early stage of life were re-exposed to PS MNPs or chemicals, their sensitivities were changed compared to those not exposed to PS MNPs at the early stage of life. Global DNA methylation assay revealed that ESC exposed to PS MNPs were hypermethylated, and the cytotoxicity caused by exposure to PSMNPs was restored by treatment with a DNA methyltransferase (DNMT) inhibitor. In zebrafish exposed to PSMNPs only in the early stages of life, there was no difference in methylation immediately after exposure, whereas a significant difference was observed in adulthood. The results collectively suggest DNA methylation might be involved in PSMNPs-induced developmental toxicity in ESC and zebrafish models. Further in-depth mechanistic studies are needed to clearly elucidate the role of epigenetics in developmental toxicity. Overall results presented in this thesis suggest early life stage exposure of MNPs caused developmental toxicity and epigenetic mechanism might be involved in it. The results will contribute to understand potential harmful effect of MNPs on human health. 미세/나노플라스틱 (micro/nanoplastics, MNPs)은 대기환경 및 식품 등 다양한 매체에서 검출되면서 인체 노출 개연성에 대한 증거가 꾸준히 증가되고 있다. 그러나 NMPs의 위해성에 관한 연구의 대부분은 생태독성 연구에 치우쳐 있으며 MNPs의 인체 유해성에 대한 이해는 매우 부족한 실정이다. 본 연구는 NMPs의 인체 유해성을 규명하기 위하여, MNPs의 다양한 매체에서의 검출 및 독성 영향에 대해 문헌 분석한 후, 이를 바탕으로, 세포 (in vitro) 와 동물 (in vivo) 실험을 통해 생애초기 단계에서 MNPs 노출로 인한 발달독성을 확인하고, 이 과정에서 후생유전 메커니즘의 역할을 확인하는 것을 목적으로 한다. In vitro 연구에서는 인간줄기세포 (human embryonic stem cell; ESC, human neural stem cell; NSC) 를 이용하여 MNPs의 세포독성을 스크리닝하고 다양한 크기의 폴리스티렌 미세나노플라스틱 (polystyrene micro/nanoplastics, PSMNPs)을 만성으로 노출하여 세포독성을 분석하였다. 그리고 ESC를 이용한 배아체 (embryonic body, EB) 형성실험을 통해 초기 배아발달기에 다양한 크기의 PSMNPs가 일으키는 발달독성을 분석하였다. DNA 메틸화가 줄기세포의 세포독성에 미치는 영향 또한 분석하여 발달독성에 미치는 후생유전 관련성을 규명하였다. in vivo 연구에서는 제브라피쉬 배아를 이용하여 생애초기에만 노출된 MNPs의 영향이 성체기까지 지속되는지 분석하고, 초기발달단계에 중요한 역할을 하는 유전자 메틸화를 분석하여 발달독성 메커니즘과 후생유전적 영향의 연관성을 확인하였다. 세포연구에서는 MNPs에 노출된 ESC, NSC 에는 0.05 μm PSNPs를 제외한 0.1, 1, 10 μm PS MNPs의 세포독성이 관찰되지 않았고 EB 크기에 이상이 관찰되지 않았으나, ESC의 배엽 형성 마커 유전자의 발현량을 저하시키는 것을 확인하였다. 그리고 동물연구에서는 배아단계부터 부화까지 MNPs에 노출된 제브라피쉬는 사멸율이나 기형발생율에는 유의미한 차이가 없었으나, 행동량에서 유의미한 차이를 보였으며 이것이 성체기까지 지속되는 결과를 확인하였다 (0.1 μm PSNPs). 이때 생애초기에 MNPs에 노출되었던 영향에 의해 성체기까지 신경관련 유전자 발현량에도 지속적인 영향을 미치는 것을 확인하였다. 그리고 생애초기에 PSMNPs 노출된 성체에 PS MNPs나 화학물질을 재노출시켰을 때, 생애초기에 PSMNPs의 노출이 없었던 성체들과 비교하여 행동량과 신경관련 유전자 발현량 변화가 관찰되었다. 더불어 MNPs에 노출된 ESC는 전체 DNA 가 과메틸화되었고, MNPs 노출로 인한 세포독성이 DNA 메틸기전달효소 (DNA methyltransferase, DNMT) 억제제 처리 시 회복되는 결과를 확인하였다. 생애초기에만 MNPs에 노출되었던 제브라피쉬는 노출 직후에는 메틸화에 차이가 없었으나 성체기에 유의미한 차이를 보이기도 하였다 (0.1um PSNPs). 모든 결과를 토대로 생애 초기에 MNPs의 노출은 발달독성 및 신경독성을 일으키며, MNPs의 발달독성 메커니즘에 DNA메틸화의 관여 가능성을 제안하였다. 본 연구의 결과는 MNPs의 발달독성을 비롯한 인체 유해성을 이해하는데 기여할 것으로 생각된다.

Establishment of an evaluation framework for monitoring ecotoxicity in potentially contaminated streams

신기식 인천대학교 대학원 2022 국내박사

Since various chemicals are mixed in industrial wastewater effluent, the evaluation of individual chemicals has limitations in directly determining the toxicity to aquatic organisms. To address this problem, North America and Europe have either implemented an integrated assessment of toxicity as an effective strategy for the management of harmful substances affecting water quality or revised related legislation to implement the assessment. In Korea, there is a limit to regulating all substances to address toxicity. As a countermeasure, ecotoxicity assessment has been incorporated into the effluent standards for some discharge facilities since 2011 in order to manage the toxicity of chemical substances in an integrated manner. Since 2021, ecotoxicity assessments have been conducted on all discharge facilities. However, such assessments are not conducted on rivers and lakes. To apply the concept to rivers, a discussion of whether to use the current test species, Daphnia magna, or add new test species is necessary. In this study, three species, Daphnia magna, Aliivibrio fischeri, and Lemna minor, were selected to achieve similar sensitivity, reproducibility, and accuracy as those of the current species of D. magna. Additionally, this study attempted to identify appropriate test species for monitoring research. This study also identified the ecotoxicity levels in rivers by monitoring potentially contaminated streams as well as substances causing high levels of ecotoxicity. It also estimated toxic pollutant sources to determine the impacts of reducing causative agents in rivers. Finally, this study proposes a basis for improving the system to protect aquatic organisms in rivers that are highly likely to have pollutants. Five heavy metals and seven volatile organic chemicals were analyzed to compare the three ecotoxicity test species. L. minor showed sensitivity to the highest number of items, followed by D. magna and A. fischeri. L. minor reacted sensitively to Cr, Zn, Cd, trichloroethylene, and formaldehyde; D. magna was sensitive to Cu, chloroform, and xylene; and A. fischeri was sensitive to Pb. These results indicate that effectively detecting various toxic substances introduced into water systems is difficult using a single test species. The outcomes also highlight the fact that many countries employ more than two test species in evaluating ecotoxicity. This study targeted 70 river sites near industrial complexes under the water quality measurement network (i.e., potentially contaminated streams) and monitored ecotoxicity from 39 types of water pollutants using three test species twice a year from 2018 to 2019. The ecotoxicity incidence was determined to be between 47.1% and 67.1%. Toxicity incidence per test species was as follows: L. minor was the highest, followed by D. magna and A. fischeri; L. minor showed the most sensitive reaction. At some sites, the values exceeded the human health protection standards or drinking water quality standards, which are environmental standards for rivers, and all sites that exceeded the standards showed ecotoxicity. Furthermore, a number of sites that did not exceed the water quality criteria indicated ecotoxicity. These results show that ecotoxicity can occur even when individual water pollutants are managed below environmental standards. Seven sites with high ecotoxicity were selected out of the 70 river sites, and monthly monitoring was conducted. Most of the seven sites exhibited persistent toxicity according to the three test species. In months when high toxicity was detected, the concentrations of heavy metals such as Cu, Cd, and As were high, leading to potentially toxic effects. Based on the results of monthly monitoring, toxicity identification evaluations (TIE) were performed to identify contributors at one of the sites, HK-11, which had a continuous occurrence of ecotoxicity. TIE showed that the primary contributor to ecotoxicity was Cu. After identifying the emission sources of Cu and reducing the discharge of Cu into rivers, monitoring was performed; subsequently, the concentration of Cu was significantly reduFced. This result indicates that pollution levels in rivers are significantly reduced by removing or reducing the substances causing ecotoxicity in rivers. Concurrent with the ecotoxicity analysis of the 70 river sites, various conditions such as sensitivity, test convenience, and population security of the test species were compared. ‘Toxicity test method using L. minor root regrowth’ is assumed to be an appropriate method not only to measure the ecotoxicity of rivers and lakes but also to expand ecotoxicity test species for effluent standards. To protect aquatic organisms in rivers and bequeath healthy rivers to the next generation, the following processes are required: 1) the level of ecological pollution in a river should be identified by adding ecotoxicity assessment to the stream monitoring standards; 2) contributors to high levels ecotoxicity should be identified; and 3) emission sources should be tracked and reduced.