Marco Andrea Zampini

• 12 December 2024, 4:30pm - 6:30pm
• Auditorium O3 (CDE)
• Promoters: Marleen Verhoye, Jan Sijbers, Ruslan Garipov

Abstract

Magnetic Resonance Imaging (MRI) is a noninvasive, non-ionizing technique that allows both anatomical and functional imaging with tunable contrast among soft tissues. Although MRI is a significant source of images with differences in signal intensity defined by differences in specific physical parameters; for the vast majority of applications, images are being interpreted qualitatively in the clinical practice.

Quantitative MRI (qMRI) is an umbrella term that encompasses the quantification of the parameters specifying signal intensity. Among others, the relaxation parameters T1 and T2 remain a main research topic in qMRI as the relaxation times have been used as biomarkers for tissue characterization and differentiation, and can be of great radiological support.

Several methods for fast relaxometry have been proposed, however there is yet no general consensus on fast relaxometry mapping techniques that allow good image quality within clinically acceptable time and with good reproducibility and repeatability, also due to the dependency of these parameters on confounding factors, including the RF excitation field B1.

This PhD thesis starts with a short history of the concept of Nuclear Magnetic Resonance and the development of the Magnetic Resonance Imaging technique, after which the main qMRI techniques dedicated to relaxometry mapping are reported.

The contributions of this thesis focus on improvements for T1 mapping.

As for most T1 mapping methodologies, an accurate estimation of the flip angle is necessary, we start with the development of a preparation module for a faster steady state approach and sampling of Actual Flip angle Imaging (AFI), a B1 mapping technique.

Next, we extended B1-mapping with T1-mapping using Variable flip angle with AFI (VAFI) approach in which we introduce a slice profile correction as well as a 2D multislice stacking approach for AFI, which respectively provide a more accurate and faster way to acquire parametric maps.

Then, a newly developed sequence, Relaxation Alternate Mapping of Spoiled Echo Signals (RAMSES) for multiparametric mapping including T1, B1, and T2* is introduced. This was achieved by adding a bipolar multi-gradient-echo readout to AFI, with no additional scanning time required.

Lastly, another newly developed sequence, Echo Planar Imaging Fast Actual Nutation Imaging (EPIFANI) for ultrafast B1-corrected T1 mapping is introduced. This sequence is an echo planar imaging (EPI) version of AFI, which can be used to provide high time efficiency to a multiparametric acquisition, and represents one of the very few attempts to acquire multiparametric maps with EPI.

Marc Terrones Bernat

• 9 December 2024, 4 pm - 6 pm
• Auditorium O3 (CDE)
• Promoters: Geert Vandeweyer, Guy Van Camp, Ken Op de Beeck

Abstract 

Lung cancer is a heterogeneous disease that requires tailored therapeutic approaches. Significant advances in molecular understanding have led to the development of targeted therapies, such as tyrosine kinase inhibitors (TKIs) for epidermal growth factor receptor (EGFR) or anaplastic lymphoma kinase (ALK)-mutated non-small cell lung cancer (NSCLC). However, resistance to these therapies often develops, complicating treatment. This thesis focuses on addressing resistance in ROS1-rearranged NSCLC, a rare molecular subtype. ROS1-rearranged tumors, though only 2% of NSCLC cases, present unique challenges due to limited experimental models and the difficulty of studying rare mutations.

The research aimed to create novel patient-derived cell lines carrying ROS1 mutations associated with TKI resistance, allowing for high-throughput drug screening. The CRISPR/Cas9 technique was used to introduce three resistance-conferring mutations (G2032R, L2026M and S1986Y) into the HCC-78 cell line expressing SLC34A2-ROS1. Drug assays using the 5 clinically approved ROS1 TKIs showed varied responses, with G2032R mutation leading to the most aggressive phenotype and highest resistance, while S1986Y showed no resistance. The mutations were then introduced into two additional patient-derived cell lines: CUTO-28 (TPM3-ROS1) and CUTO-37 (CD74-ROS1), revealing that the effect of mutations is dependent on the ROS1 fusion partner.

Additionally, transcriptomic analysis of ROS1+ NSCLC revealed an upregulation in pathways related to nucleotide synthesis and cell adhesion, with IL20RB identified as a potential target for treating bone metastases. The study also profiled gene expression changes in response to repotrectinib, a next-generation TKI, and identified overexpressed genes like complement 3 (C3) and epithelial-to-mesenchymal transition (EMT)-related genes that may contribute to drug resistance in G2032R mutant cases.

This thesis demonstrates the feasibility of using CRISPR/Cas9 to model rare ROS1 mutations and highlights the importance of fusion partner context in drug resistance. The new panel of 9 TKI-resistant, patient-derived ROS1+ NSCLC cell lines unveiled potential new therapeutic targets for patients with TKI-refractory or bone metastatic ROS1+ NSCLC disease.

Maria Camila Armirola Ricaurte

• 25 November 2024, 4pm - 6pm
• Auditorium O5 (CDE)
• Promoter: Albena Jordanova

Abstract

This doctoral thesis aims to improve our understanding of peripheral neurons by investigating the genetic underpinnings of two significant peripheral nervous system disorders: Charcot-Marie-Tooth (CMT) and diabetic peripheral neuropathies (DPN). Identifying novel genes and variants linked to these conditions may pinpoint essential pathways for these unique cells that, when disrupted, can lead to neurodegeneration.

Whole exome sequencing of unsolved CMT families revealed two novel nuclear-encoded mitochondrial genes, COX18 and NDUFS6, that cause autosomal recessive CMT. Pathogenic variants in these genes predominantly manifest as axonal CMT, often with central nervous system symptoms. Functional and genetic analysis of a homozygous splice variant in COX18 showed impaired Complex IV assembly and mitochondrial dysfunction. Meanwhile, an NDUFS6 splice variant previously associated with a fatal mitochondrial disorder resulted in a milder CMT phenotype. Characterization of this variant demonstrated that alternative splicing compensates for the loss of NDUFS6, modulating the disease severity. Both studies highlight the importance of mitochondrial metabolism for peripheral neurons. Moreover, they underscore the genetic and clinical overlap of CMT and mitochondrial disorders with relevant implications for the diagnosis of these disorders.

Using genotype information from the UK Biobank, we conducted association studies to detect common variants that have an impact on DPN susceptibility. First, we implemented a classification approach to stratify the UK Biobank cohorts into type 1 and type 2 diabetes mellitus groups. Genome-wide association studies were conducted within each cohort, and their results were meta-analyzed to enhance the ability to detect genetic associations. Subsequently, these findings underwent replication through a meta-analysis of an independent cohort of individuals with diabetes from FinnGen. As a result, we found a novel genome-wide significant variant in an intergenic region in chromosome 1 associated with increased risk of DPN (OR 1.66, 95% CI 1.39-1.98, P = 1.66x10-8). We also estimated that the SNP-based heritability of DPN is approximately 13%. This work suggests that DPN is a heritable trait and pinpoints a new regulatory mechanism associated with the susceptibility to this diabetic complication.

Kenny Van Theemsche

• 10 September 2024, 4pm - 6pm
• Auditorium O5 (CDE)
• Promoters: Dirk Snyders, Alain Labro
  

Abstract

The voltage-dependent K⁺ (Kv) channel family is large compared to other ion channel families, owing to a diverse array of genes encoding different α-subunits. In some subfamilies, these subunits can "mix-and-match" to form functional tetrameric Kv channels, further expanding this channel family. As a result, Kv channels are involved in a wide range of (patho)physiological processes. Consequently, there is interest in developing selective drugs that target specific Kv channels. However, drug specificity remains a major challenge due to the homology within the Kv channel family. In parallel, safety pharmacology is increasingly incorporating a broader range of Kv channels to prevent adverse side effects in newly developed drugs. Current in vitro testing procedures primarily focus on acute drug effects, often overlooking delayed effects that manifest slowly and are not yet routinely included in drug screening. In this thesis, we investigate a potential new selective drug binding site in Kv channels, and the importance of optimizing experimental conditions to ensure reliable and relevant drug screening outcomes. Using the whole-cell voltage-clamp technique, we demonstrated that extracellular charge substitutions in the S1-S2 linker of the Kv1.5 channel significantly modulate its gating properties. Specifically, a charge reversal from positive to negative (R276E) shifted the voltage dependence of channel activation in the hyperpolarized direction, while the opposite reversal (E274K and E278K) shifted it in the depolarized direction. Notably, the sequence of the S1-S2 linker is highly variable among Kv1 channels, making this region a promising target for selective inhibition or activation of Kv channels. Additionally, we explored how the choice of expression system can influence the potency of a novel Kv1.3 inhibitor. Factors such as cell size, morphological characteristics, endogenous currents, and auxiliary proteins all contribute to the pharmacological response of Kv channels, highlighting the need to carefully select the most appropriate expression system for drug screening. Finally, we investigated delayed drug effects that impact the trafficking or degradation of channel proteins, leading to reduced expression of mature ion channels at the plasma membrane. These effects are often missed in early drug development and may only become apparent during later stages. We propose a potential method to accelerate the onset of delayed drug effects. Removing (fetal bovine) serum from the cell culture medium seems to increase the potency of a delayed-action compound. This PhD thesis aims to deepen our understanding of Kv channel pharmacology and the factors that modulate their function.

Laura Morant

• 10 September 2024, 4pm - 6pm
• Auditorium O2 (CDE)
• Promoters: Albena Jordanova, Maria-Luise Erfurth

Abstract

Aminoacyl-tRNA synthetases (aaRS) are the most prominent protein family implicated in Charcot-Marie-Tooth disease (CMT), with eight genes identified as disease-causing. These enzymes, essential for catalyzing the initial step of protein biosynthesis, are ubiquitously expressed across all cells. Surprisingly, a loss of function is not required for CMT development, suggesting a toxic gain-of-function mechanism. Beyond their role in translation, aaRS have evolved non-canonical functions, often associated with cellular compartments not involved in translation. Some aaRS exhibit shared localization patterns, such as in the nucleus, and display common aberrant interactions, genetic modifiers, and dysregulated processes in the presence of dominant mutations. This suggests the possibility of a unified neurotoxic mechanism across different CMT-linked aaRS, potentially enabling broad-spectrum therapeutic strategies.

A significant challenge in studying aaRS-associated CMT is the generation of diverse in vivo and in vitro models, which complicates comparative analysis across different organisms and laboratories. To address this, we adapted the GeneSwitch technology, originally a conditional inducible system for mammalian cells, to create unified cellular and fly models for CMT research. This adaptation eliminates the need for subcloning, allowing the selection and conditional expression of tagged transgenes in both Drosophila and mammalian cells. As a proof of concept, novel unified cellular and fly models were generated for CMT-associated tyrosyl-tRNA synthetase (YARS1), exhibiting toxicity consistent with previously established models. This toolkit promises to enhance the comparability of data across different experimental setups.

This adapted toolkit was further employed to develop unified fly models for four aaRS-related CMT types, which displayed peripheral neuropathy hallmarks consistent with previous CMT models. Through this platform, we explored common pathomechanisms among different aaRS-related CMT, focusing on the link between cellular stress and nuclear translocation of CMT-related aaRS. The study identified common genetic interactors involved in the integrated stress response (ISR) and demonstrated that other CMT-related aaRS translocate to the nucleus upon oxidative stress, suggesting a shared role in stress response. Moreover, neuronal expression of aaRS-linked CMT in adult flies altered their sensitivity to oxidative stress, and preventing YARS1 from entering the nucleus reduced ISR toxicity and oxidative stress sensitivity.

Overall, this work introduces novel tools for investigating common pathomechanisms in aaRS-associated CMT and uncovers shared disease mechanisms that could inform future therapeutic strategies.

Isabelle Neefs

• 5 September 2024, 4pm - 6pm
• Promotiezaal, Q.002 (CDE)
• Promoters: Guy Van Camp, Marc Peeters, Ken Op de Beeck

Abstract

Cancer remains a global health challenge despite advancements in biomedical research. Early-stage cancers are treatable, but therapy becomes difficult in later stages, underscoring the importance of timely screening. Traditional diagnostics like imaging and tissue biopsies are invasive and not always precise, necessitating the search for reliable biomarkers and sensitive technologies. In this view, DNA methylation has emerged as a promising biomarker for early cancer detection.

In the first part of this thesis, we have investigated novel biomarkers for colorectal cancer (CRC). First, we analysed 500 stage IV CRCs to identify factors associated with interval cancer (IC) stage IV CRC that remained undetected by the fecal immunochemical test (FIT). Findings showed a higher likelihood of neuroendocrine tumors (NETs) and lymphovascular invasion in FIT-IC CRCs, and importantly, tumor location significantly influenced these associations. Secondly, we evaluated DNA methylation patterns in colorectal tissues, identifying 13 differentially methylated sites. Our final model achieved a sensitivity of 96% and specificity of 95% in distinguishing adenomas from carcinomas.

The second part delves into epigenomics for multi-cancer detection. First, a multiplex droplet digital PCR (ddPCR) assay was developed using differentially methylated targets from The Cancer Genome Atlas (TCGA). This assay demonstrated high accuracy but was limited in target numbers for simultaneous analysis. To overcome this, the IMPRESS (Improved Methylation Profiling using Restriction Enzymes and smMIP Sequencing) technique was developed, facilitating multiplex analysis without bisulfite conversion. IMPRESS achieved high sensitivity (95%) and specificity (91%) in distinguishing tumor from normal tissue across eight cancer types.

The final part focuses on assays for CRC and breast cancer (BRCA) using IMPRESS. Differentially methylated sites were identified from public datasets, and a two-step assay (a cancer detection panel and an invasiveness detection panel) was created for both cancers. For CRC, the cancer detection panel achieved 100% sensitivity and specificity, while the invasiveness detection panel showed 80% sensitivity and 92% specificity. For BRCA, the cancer detection panel had 94.4% sensitivity and 100% specificity, with the invasiveness detection panel achieving 66.7% sensitivity and 88.9% specificity. These results indicate high sensitivity and specificity for distinguishing cancerous from normal samples, with good accuracy for differentiating precancerous from invasive tumors. Further research is needed to validate these panels in plasma-derived liquid biopsies.

The integration of advanced technologies like multiplex ddPCR and IMPRESS highlights the importance of epigenetic research in cancer diagnostics, and this thesis paves the way for improved clinical applications in cancer screening and detection.

Rupasinghe Arachchige Praharshinie Rupasinghe 

• 3 September 2024, 4pm - 6pm
• Aula Janssens (ITG)
• Promoters: Leen Rigouts, Bouke de Jong

Abstract

The gap in tuberculosis diagnostics has been a persistent challenge, spanning decades of efforts to improve the detection of Mycobacterium tuberculosis complex (MTBc) and its drug resistance. Roll-out of new drugs without access to drug-susceptibility testing (DST) and surveillance data risks repeating errors of the past, and rapidly losing these drugs to resistance when patients are treated ‘blindly’ with already compromised regimens. For the majority of anti-tuberculosis drugs, phenotypic DST (pDST) is still the gold standard. Thus, expanding research efforts to understand the current gaps in pDST in MTBc is crucial for gaining a more comprehensive understanding of the drug-resistance landscape and improving patient-centered TB management strategies.

This thesis sought to describe the factors impeding rapid and accurate detection of phenotypic drug resistance, as well as to determine whether faster minimum inhibitory concentration (MIC) testing offers advantages over traditional pDST, which relies on growth inhibition of MTBc in drug-containing medium at a single drug concentration known as the critical concentration (CC), yielding a binary classification as susceptible or resistant.

The main gaps identified in traditional pDST for MTBc during this PhD research include: (1) lack of a standardized culture medium that provides optimal nutrition to all MTBc variants, including fastidious strains, and (2) artefacts related to binary classification of drug susceptibility using the CCs, such as overlapping MICs of drug-susceptible and drug-resistant strains, leading to misclassification of low-level/borderline resistant mutants and (3) the lineage effect on in-vitro susceptibility to anti-TB drugs. These gaps underscore the need for a standardized quantitative pDST method, applicable for all MTBc variants.

As a potential solution, this PhD research assessed the recently WHO-endorsed microtiter-based broth microdilution (BMD) method, which represents a significant step towards standardization in pDST for MTBc. We semi-automated the method and successfully validated its use to assess new and repurposed anti-TB drugs. In addition, we demonstrated that freshly grown liquid cultures can be used as a basis for BMD inoculation, as opposed to standard inoculum preparation from cultures grown on solid medium, hence reducing the turnaround time (TAT) and increased field-friendliness. Finally, we addressed some of the challenges associated with traditional pDST methods.

Continuous improvements such as further reducing the TAT, standardize reading and inoculum preparation, as well as cost-effective automation to promote widespread adoption of MIC testing by the BMD method, will be critical for improving the quality and reliability of pDST for MTBc and, ultimately, TB treatment outcomes globally.

Dorien De Pooter

• 24 June 2024, 4pm - 6pm
• Promotiezaal, Q.002 (CDE)
• Promoters: Peter Delputte, Ellen Van Gulck

Abstract

Yearly, 1.5 million people get infected with the hepatitis B virus, although preventive vaccines are available. People who are not able to take control of the virus, especially infants, develop a chronic infection. As HBV targets the hepatocytes, a chronic infection can permanently damage the liver and can lead to the development of hepatocellular carcinoma.

The hepatitis B virus is able to form circular covalently closed (ccc)DNA in infected hepatocytes and to integrate its DNA in the genome of the host. Both integrated DNA and cccDNA are a source for persistent high concentrations of viral antigens like HBsAg, and cause an immune system tolerant to it, with exhausted T-cells unable to kill infected hepatocytes. Current therapies can not eliminate the virus completely and there is need for a therapy that can reach functional cure, defined as the loss of systemic HBsAg levels and undetectable levels of HBV DNA after a finite treatment. It is believed that a successful treatment consists of multiple modalities including antivirals and immunomodulators that target the high viral and antigen burden as well as inadequate host immune responses. One such immunomodulator is a therapeutic vaccine that introduces HBV viral antigens to the body thereby eliciting a specific immune response.

This thesis describes the development and preclinical evaluation of a pDNA vaccine encoding core and pol HBV antigen and a second vaccine in the form of mRNA encoding core, pol and envelope antigen. Robust immune responses were induced in several naive animals with the mRNA vaccine being superior in terms of magnitude of induced CD8 T-cell responses, but specifically the core-specific CD4 T-cell responses were negatively impacted compared to the pDNA vaccine. In AAV-HBV transduced mice, containing an HBV-tolerant, immune suppressed liver environment, the elicited immune responses after two immunizations with pDNA vaccine, were not powerful enough to reduce viral parameters while responses after three immunizations with the mRNA vaccine, reduced viral parameters for 50% of the tested mice.

The used platforms allow multiple improvement such as coadministration with adjuvants, altering dosing scheme, finetuning the formulation, exploration of a heterologous regimens, ... The combination of a vaccine with modalities to reduce viral replication and reduce the suppressive HBV antigen levels, should strengthen, and enhance it effects. Such a multimodal therapy is believed to be essential to reduce changes of developing liver damage caused by the constant presence of hepatitis B virus.

Lena Duchateau

• 8 February 2024, 4pm - 6pm
• Promotiezaal, Q.002 (CDE)
• Promoter: Kristel Sleegers

Abstract

Alzheimer’s disease (AD) is the leading cause of dementia and a growing problem worldwide. ABCA7 was identified as a risk gene for AD in genome-wide association studies (GWAS). ABCA7 is suggested to play a role in lipid metabolism, phagocytosis and amyloid pathology. Since its discovery as a risk gene for AD, several risk-increasing variants have been identified. These variants include premature termination codon (PTC) mutations, variable number of tandem repeats (VNTR) polymorphism expansion mutations and predicted pathogenic missense mutations. The aim of this thesis was to get a better insight into the underlying pathomechanisms of these ABCA7 variants.

Firstly, we leveraged different transcriptomic approaches to study these mechanisms, including long-read nanopore sequencing (on brain and lymphoblastoid cell lines), single-nuclei sequencing (SNS) and spatial sequencing on brain of AD patients and controls with ABCA7 PTC, missense and/or VNTR mutations. Our data suggest a high amount of splicing complexity in ABCA7 in general, with increased levels in carriers of PTC mutations. NMD escape and rescue splice events, able to rescue the effect of the mutation, were identified but did not alter phenotype or age at onset, though rescue splicing did impact gene expression. Both single-nuclei and spatial sequencing methods suggested neurons to be the main expressor of ABCA7 but using SNS we found average ABCA7 expression was not highest in the excitatory neurons, but in microglia. Remarkably, mutation carriers who were theoretically expected to have lower ABCA7 expression levels, had highest expression in microglia, suggesting a specific upregulation in this cell type. Of interest, we identified several differentially expressed genes between carrier and non-carrier AD patients in microglia, and pathway analysis in microglia suggested downregulation of translation and inflammatory reaction terms compared to non-carriers. Spatial analysis of the brain additionally put the choroid plexus forward as a region of interest due to its high ABCA7 expression.

Finally, we performed cerebrospinal fluid (CSF) biomarker analysis on PTC and VNTR carriers suggesting altered APP processing in VNTR, and to lesser extent, PTC carriers. Moreover, decreased YKL-40 levels in expansion carriers suggested that these patients potentially have a reduced inflammatory response to AD damage.

In conclusion, this thesis provided further insight into possible pathomechanisms in the different mutation groups and suggested these might not be the same in all three. These findings could eventually aid in identifying the correct treatment for AD patients carrying an ABCA7 variant.

Claudia Theys

• 29 January 2024, 4pm - 6pm
• Promotiezaal, Q.002 (CDE)
• Promoter: Wim Vanden Berghe

Abstract 

Metabolic dysfunction associated fatty liver disease (MASLD) is a global health burden. It consists of a spectrum of liver disorders ranging from steatosis, characterized by lipid accumulation in the liver, inducing inflammation and thereby progression into metabolic dysfunction associated steatohepatitis (MASH). MASH predisposes patients for further cirrhosis and hepatocarcinoma. Unfortunately, there is still no FDA-approved treatment due to the lack of biomarkers to correctly stratify patients and the multifactorial nature of the disease. Therefore changes in lifestyle including diet and exercise remain the current treatment strategy. However this is difficult to maintain, leading to a lot of relapsing patients. Thus, there is an urgent need for a full characterization of the molecular targets that have a key role in the progression of the disease.

The nuclear receptor PPARα and mitochondria have a key role in lipid metabolism and are also closely related to inflammation. Moreover, recent research, especially cancer research, has shown that mitochondrial DNA (mtDNA) methylation can be used as a biomarker and interactions of PPARα with epigenetic enzymes can regulate lipid metabolism in liver and colon, which makes them interesting targets.

PPARα is epigenetically downregulated in MASH patients, but previously tested agonists could not ameliorate the disease. Interestingly, we show that the loss of PPARα function by diet induces similar epigenetic and transcriptional changes towards a MASLD gene signature, as a genetic knock out. Moreover, in both mice models, the loss of this one PPARα hub induced a shockwave of changes in the transcriptome of several epigenetic enzymes, resulting in the epigenetic reprogramming of the lipid and bile acid metabolism towards ferroptotic and pyroptotic lipotoxicity that is closely related with fibrosis. This suggests an important indirect epigenetic protective function of PPARα towards its target genes, to maintain general lipid homeostasis and prevent progression of MASLD. Therefore combination therapies of PPARα agonists with epigenetic compounds should be explored.   

Due to the constant accumulation of lipids in the liver, MASH patients show an overactivation of the mitochondrial metabolism. However the exact molecular mechanisms leading to this overactivation and eventually shutdown are still largely unknown. We have shown that MASLD is associated with small changes of mtDNA methylation. Moreover, an artificial increase of 20% mitochondrial CpG or GpC methylation induced mitochondrial dysfunction and dysregulation of the bile acid metabolism by mito-nuclear communication as seen in MASLD progression. Therefore also mitochondrial methylation shows new opportunities for both MASLD biomarker and therapeutic research.