MBG Graduation Thesis Announcement

Advisor: Prof. Fadime Kıran         Co-advisor: Assist. Prof. Ilyas Chachoua

Jury members:
1- Assist. Prof. Fulya Turker
2- Assist. Prof. Bahar Degirmenci

Defense time: 09.00-10.00

Title: Evaluating the effects of postbiotics from skin derived Streptococcus hominis on melanoma cells

Abstract: Cancer remains a leading global health challenge due to its high prevalence and mortality rates. Among the various cancer types, skin cancers including both melanoma and non-melanoma subtypes, are particularly prevalent. Melanoma, characterized by its aggressive and invasive nature, significantly reduces the effectiveness of conventional treatments such as surgery, chemotherapy, and radiotherapy, especially in advanced stages. Furthermore, these therapies often result in severe side effects and present challenges related to accessibility and affordability. Recent research emphasizes the vital role of a balanced skin microbiota in health, identifying its imbalance as a key factor in melanoma development and progression. Consequently, bacteriotherapy has gained attention as a potential treatment strategy to restore skin microbiota balance in melanoma patients. The aim of this study is to investigate the anticancer potential of Streptococcus hominis, derived from the skin microbiota, against B16-F10 melanoma cells. S. hominis was cultured in MRS broth medium for 24 h and its postbiotics were obtained by centrifugation following by filtration and lyophilization. B16-F10 cells were treated with different doses of postbiotics for 48 h, and their anticancer effects were evaluated using MTT, scratch, and colony-formation assays. Besides, the expression levels of apoptosis-associated genes (BAX and Bcl-2) were determined by RT-PCR analysis. According to the results of MTT assay, significant cytotoxic effects on B16-F10 melanoma cells were observed at concentrations of 2500 μg/mL (29.81 %) and 5000 μg/mL (39.79 %). While the control group achieved complete (100 %) wound closure at 36 hours, treatment with 5000 μg/mL resulted in a markedly reduced closure rate of 21.71 %. Similarly, it was observed that the colony-forming ability of B16-F10 cells was significantly reduced (56 %). Besides, the expression level of pro-apoptotic BAX was significantly up-regulated (3.5-fold), while the anti-apoptotic Bcl-2 gene expression was down-regulated (14.3-fold). In conclusion, this study provides evidence that S. hominis, a commensal bacterium of the skin microbiota, may serve as a natural and accessible source of postbiotic compounds with promising anticancer properties. This underscores the therapeutic potential of skin microbiota-derived agents in the development of novel strategies against melanoma.

Advisor: Assist. Prof. Onur Çizmecioğlu

Jury members:
1- Prof. Tayfun Özçelik
2- Assoc. Prof. Ozlen Konu

Defense time: 10.00 – 11.00

Title: Characterization of MDV3100 resistant LNCAP subclone cell line called “LNCAP ENZR”

Abstract: Androgen Receptor pathways play a critical role in prostate cancer progression. Dysregulation of this pathway leads to rapid progression of the tumorigenic tissue. Thus, this pathway exhibits itself as a great candidate for targeted therapies. However, upon androgen depriving therapies to downregulate androgen signaling, the cells may develop an androgen independent growth pattern. Thus, combination of different approaches is necessary to prevent development of drug resistance. When androgen receptor pathway is repressed, the cells develop resistance through hyperactivation of different signaling pathways. PI3K pathway can be given as an example in this situation, thus can be a good novel co-targeting candidate. As different PI3K p110 catalytic subunit isoforms are upregulated in different tissues, the isoform dependency of cancer cells also should be investigated in different conditions in cancer as well.
In this study, PTEN-null LNCaP cells were divided into two different subclones, and one of them was treated with MDV3100 (Enzalutamide), an androgen receptor antagonist, inhibiting androgen receptor signaling, continuously until it has developed resistance towards the drug (subclone called LNCaP EnzR); while the other cell line was treated with DMSO, thus has not developed any kind of drug resistance (subclone called LNCaP DMSO). As these cell lines derived from the same origin, they are isogenic, meaning the differences observed is solely dependent on the MDV3100 resistance. Investigation of differences on these constructed isogenic cell lines upon drug resistance is aimed in this study. The differences they exhibit in their growth patterns are examined to characterize these cell lines.
The results have shown that upon MDV3100 resistance, the cells have shown androgen independent growth patterns. However, contrary to the expectations, the cells have shown decreased dependency towards p110β isoform, and increased dependency towards p110δ when they gained resistance towards Enzalutamide. Aligning with the previous publications, p110α dependency was found severely low in comparison to the other investigated isoforms in both cell lines. This suggests hyperactivation mechanisms of this pathway in p110 isoform dependent manner could be decisive in clinical outcomes.

Advisor: Prof. Turgay Dalkara

Jury members:
1- Assist. Prof. Fulya Turker
2- Assist. Prof. Bahar Degirmenci

Defense time: 11.00-12.00

Title: Cell specific transcriptomic pathways activated after cortical spreading depolarization, a migraine aura model

Abstract: Migraine with aura is a complex neurological disorder, a hallmark sign being transient cortical dysfunction. It has been strongly linked to cortical spreading depolarization (CSD), which is a wave of neuronal and glial depolarization that travels across the cortex. Despite the many existent studies describing CSD’s electrophysiological signature, its effects and relationships to molecular pathways is not fully understood. Through transcriptomic analysis on a mouse model, synaptic transmission, plasticity, ion homeostasis, and neuroinflammatory pathways were observed to be altered. The transcriptomic profiles observed in this study may serve to further understand the molecular effects that CSD has.

Advisor: Assoc. Prof. Urartu Şeker

Jury members:
1- Assoc. Prof. Ozlen Konu
2- Assist. Prof. Ilyas Chachoua

Defense time: 13.00-14.00

Title: Knockout of Csg Operon in E.coli.

Abstract: Csg operon in Escherichia coli is responsible from producing the curli fibers that are part of biofilms. These biofilms carry important functions in the environment and in many diseases. In order to control the effects of biofilms, their production should be controlled. This thesis will aim to knock out csg operon in E.coli K-12 MG1655 strain. The knockout was performed using Lambda Red methodology which utilizes bacteriophage proteins for gene manipulation via recombination. Overlap Extension Polymerase Chain Reaction (Overlap Extension PCR) was also used to create the DNA sequence that’ll replace the csg operon. In order to verify the knockout, methods like Polymerase Chain Reaction and Agarose Gel Electrophoresis were utilized. The results of this study show that the knockout was successful. It also confirms the role of csg operon in biofilm production. The knockout cells were not able to synthesize biofilms as shown by Congo red assay in YESCA plate.

Advisor: Assoc. Prof. Urartu Şeker

Jury members:
1- Assist. Prof. Pinar Onal
2- Assist. Prof. Ilyas Chachoua

Defense time: 14.00-15.00

Title: Purification and Biophysical Characterisation of Major Curli Protein, CsgA.

Abstract: Curli fibers are extracellular functional amyloids produced by Escherichia coli that play a key role in biofilm architecture, bacterial adhesion, and immune interactions. The major structural component of Curli is the protein CsgA which forms β-sheet-rich fibrils, aided by CsgB in vivo and spontaneously in vitro. Beyond its roles in biofilm formation, it serves as a microbial model system for studying amyloidogenesis. For this reason, this thesis aimed to express, purify, and biophysically characterize recombinant CsgA in vitro. The protein was expressed in E. coli BL21 (DE3) using a pET22b (+) plasmid and purified through nickel-affinity chromatography. For the biophysical characterization circular dichroism (CD) spectroscopy, Thioflavin T fluorescence assay, Congo red binding assay and transmission electron microscopy (TEM) were utilized. The results confirmed that CsgA adopts β-sheet-rich structures and forms fibrils consistent with amyloid morphology. In contrast to previous reports, the polymerization of CsgA occurred instantaneously at neutral pH. These findings support the continued investigation of CsgA polymerization kinetics and its use as a model amyloid system. Furthermore, this thesis provides an experimental framework for further studies into bacterial biofilm formation and functional amyloids in health and disease.

Advisor: Prof. Tayfun Özçelik

Jury members:
1- Assoc. Prof. Urartu Şeker
2- Assist. Prof. Onur Çizmecioğlu

Defense time: 15.00-16.00

Title: FuncVEP: A Highly Accurate Functional Data–Trained Variant Effect Predictor

Abstract: Missense variants constitute a major class of human genetic variation, yet their functional and clinical interpretation remains a critical challenge in genomics. While computational tools known as variant effect predictors (VEPs) are commonly used to assess the potential impact of such variants, most are trained on clinical data, making them prone to various forms of circularity, label noise, and limited generalizability. To overcome these limitations, we present FuncVEP, the first VEP trained exclusively on high-confidence functional data derived from multiplexed assays of variant effect (MAVEs), curated literature, and text-mined ClinVar submissions. We developed four predictive models using gradient boosting (LightGBM), each varying by the inclusion or exclusion of features derived from prediction tools. Throughout model development, we implemented a rigorous multi-layered circularity elimination strategy and performed imputation using machine learning to address biased missingness in tool-derived inputs. Extensive benchmarking of 68 tools on both functional and clinical datasets demonstrates that our FuncVEP models achieve state-of-the-art performance, substantially outperforming all existing tools on functional data, and FuncVEP-CTI even surpasses clinical-trained tools on clinical data, despite never being trained on it. We provide ACMG-compatible thresholds for clinical use and demonstrate that training on unbiased functional data yields more generalizable models than training on clinical data. Together, our findings suggest a shift in strategy for variant effect prediction, emphasizing the utility of functionally grounded learning frameworks in the development of future tools.

Advisor: Assoc. Prof. Ozlen Konu

Jury members:
1- Prof. Tayfun Özçelik
2- Assist. Prof. Onur Çizmecioğlu

Defense time: 16.00-17.00
Title: High-Fat Diet-Induced Transcriptomic Remodeling in Zebrafish Larvae Reveals Key Regulators of Lipid Metabolism and Early Metabolic Disruption
Abstract: The rising prevalence of metabolic diseases underscores the significance of the impact of diet on early life metabolism. The present study uses a dual-omics approach of RNA sequencing and metabolomics to analyze the molecular influence of short-term high-fat diet exposure in zebrafish larvae (Danio rerio) at 12 days post-fertilization (dpf). Zebrafish were either kept on a regular microdiet (Otohime) or egg-yolk-based high-fat diet, and whole-body transcriptome and metabolome profiles were measured to detect convergent and layer-specific responses. Transcriptomics revealed robust upregulation of biosynthetic genes involved in cholesterol biosynthesis and fatty acid elongation like hmgcs1, msmo1, sqlea, srebf2, elovl5, and acacb. These transcriptional changes were paralleled by metabolomic enrichment in long-chain and polyunsaturated fatty acids, indicating functional lipid accumulation consistent with the gene-level activation. These genes are involved in sterol biosynthesis and polyunsaturated fatty acid metabolism, and are indicative of coordinated transcriptional reprogramming of lipid anabolism. Metabolomics simultaneously observed high levels of long-chain fatty acids like adrenic acid and dihomo-γ-linolenic acid and high phosphatidylcholine derivatives consistent with the trend from the transcriptional analysis. Notably, n-6 PUFAs increased but n-3 PUFAs decreased, suggesting compromised PUFA balance with potential implications for inflammatory signaling. Additionally, metabolite perturbations such as high uridine and xanthine signified nucleotide turnover not apparent from transcriptomics data alone. Metabolomics was also uniquely able to demonstrate alterations in nucleotide intermediates such as uridine and xanthine that were not evident from the transcriptomics, highlighting the added value of metabolite analysis for the characterization of biosynthetic turnover. Although some limitations occurred in data harmonization due to the gaps in gene/metabolite annotations and small sample size, integrative analysis exhibited substantial cross-validation between omic layers. Co-enrichment of lipid metabolism pathways within both data sets confirmed the validity of the diet-induced molecular redesign. These findings validate the use of zebrafish larvae as a strong model system for early nutritional science and provide mechanistic evidence of lipid-driven developmental reprogramming.