Development of Alternative In Vitro System to Assess Toxicity of Nanomaterials

다국어 초록 (Multilingual Abstract)

Development of Alternative In Vitro System to Assess Toxicity of Nanomaterials Department of Biomedical Science and Technology Graduate School Kyung Hee University Choi, Jae won Directed by Prof. Park, Eun-Jung Nanomaterials, particularly silica (SiO₂) nanoparticles (NPs), are widely used in biomedical applications due to their unique physicochemical properties. However, these NPs can induce toxicity in various cell types, necessitating comprehensive evaluations of their nanotoxicity. Existing studies mainly examine SiO₂ NP toxicity in serum-containing environments, where agglomeration alters particle properties. This study explores how serum proteins affect their toxicity and cellular uptake. We assessed the toxicity of three distinct types of monodisperse SiO₂ NPs in serum-free conditions using human liver cancer (HepG2) and lung cancer (A549) cell lines. Our findings demonstrate that protein corona formation significantly mitigates the toxicity of SiO₂ NPs, while size-dependent effects on apoptosis and necrosis were observed under serum-free or low-serum conditions. In parallel, we explored the cytotoxicity of 20 nm SiO₂ NPs using a micropillar/microwell chip platform in both 2D and 3D cell cultures under different experimental conditions, including serum presence and scaffold materials such as Matrigel, alginate, and collagen type I. In 2D cultures, SiO₂ NPs induced significant toxicity under serum-free conditions, while no toxicity was observed in serum-containing medium. However, in 3D cultures, SiO₂ NPs did not induce cytotoxicity when cultured with Matrigel, regardless of serum concentration or cell density. Interestingly, toxicity was observed in 3D cultures with alginate and collagen type I scaffolds under serum-free conditions and at low cell densities. Our analysis of nanoparticle penetration depth, cellular uptake, and scaffold properties suggests that scaffold material plays a critical role in modulating nanotoxicity. These findings emphasize the role of scaffold selection in mimicking in vivo conditions for accurate nanoparticle toxicity assessments in 3D cultures. This study introduces a high-throughput detection method for spheroid and hypoxic regions, leveraging machine learning (ML) to evaluate drug efficacy efficiently. This method, capable of processing over 10,000 images per hour with a 2%–3% error rate, was trained using data from six cell lines (HepG2, A549, Hep3B, BEAS-2B, HT-29, and HCT116) and hypoxic regions from two cell lines (HepG2 and BEAS-2B). The ML models successfully predicted spheroid and hypoxic region areas at specific growth stages, with validation through sorafenib treatment of HepG2 spheroids. This approach offers a reliable framework for evaluating drug efficacy and toxicity, emphasizing the potential of ML-driven methodologies for high-throughput drug testing and the advancement of nanotoxicity evaluations.

Keywords: Serum protein, Silica nanoparticles, Cellular internalization, Apoptosis, LC-MS/MS, Cell viability, FITC-labeled nanoparticles, In vitro toxicity test, 3D cell culture, Matrigel, Alginate, Collagen I, Spheroids, Machine learning, Cell growth, Cell size, Hypoxia.

목차 (Table of Contents)

• LIST OF FIGURES vi
• LIST OF TABLE vii
• GRAPHICAL ABSTRACT viii
• ABSTRACT ix
• 1. INTRODUCTION 1
• 2. MATERIALS AND METHODS 6
• 2.1. Materials and reagents 6
• 2.2. Preparation and characterization of SiO2 NPs 6
• 2.3. Stability analysis of dispersions using a TURBISCAN 7
• 2.4. Cell culture and SiO2 NP treatment 7
• 2.5. Spheroid culture 11
• 2.6. Cytotoxicity assay and statistical analysis 12
• 2.7. Localization of FITC-labelled 20 nm SiO₂ NPs 14
• 2.8. Evaluating the drug efficacy 15
• 2.9. Investigation of cell death mode 15
• 2.10. NanoLC-MS/MS analysis 16
• 2.11. Analysis of penetrating depth of SiO2 NP in scaffolds 16
• 2.12. Observation of scaffold using dark-field microscopy 17
• 2.13. Automated detection method 17
• 2.14. Training the ML models 258
• 2.15. Statistical analysis 21
• CHAPTER Ⅰ. Effect of serum protein on cell internalization of silica nanoparticles
• I.3. RESULTS AND DISCUSSION 23
• I.3.1. Distribution and structural characterization of SiO2 NPs 23
• I.3.2. Size-dependent cytotoxicity of SiO2 NPs 25
• I.3.3. Intracellular influx of SiO2 NPs in the presence or absence of FBS 27
• I.3.4. FBS concentration dependence of cytotoxicity 29
• I.3.5. Composition of protein corona 31
• I.4. CONCLUSION 33
• CHAPTER Ⅱ. Variations in in vitro toxicity of silica nanoparticles according to scaffold type in a 3D culture system using a micropillar/microwell chip platform
• II.3. RESULTS AND DISCUSSION 35
• II.3.1. Size of SiO2 NPs in aqueous/biological environments 35
• II.3.2. Static stability of SiO2 NPs in biological environments 38
• II.3.3. Cytotoxicity of SiO2 NPs in SC or SF medium in the 2D culture system 41
• II.3.4. Cytotoxicity of SiO2 NPs in 2D and 3D culture systems 43
• II.3.5. Cytotoxicity of SiO2 NPs in Matrigel-based culture system by cell number 45
• II.3.6. Cytotoxicity of SiO2 NPs based on scaffold type 47
• II.3.7. Mechanism of cytotoxicity of SiO2 NPs based on scaffold type 50
• II.4. CONCLUSIONS 54
• CHAPTER Ⅲ. Evaluating cell growth and hypoxic regions of 3D spheroids via a machine learning approach
• III.3. RESULTS AND DISCUSSION 60
• III.3.1. Importance of size in 3D spheroids cells 60
• III.3.2. Automated detection of cell area 63
• III.3.3. Prediction on cell growth and area of hypoxic region 67
• III.3.4. Applicability of ML model for drug efficacy evaluation 74
• III.4. CONCLUSIONS 78
• Summary 79
• Reference 82