Bioinformatic and experimental approaches for studying pathogenic mechanisms in certain rare hereditary diseases

Elucidating the pathogenic mechanisms of rare hereditary diseases remains challenging because of high genetic heterogeneity, limited patient populations, and complex molecular pathways involved. Understanding these diseases requires investigation at several levels, from identifying the causal variant to determining its effects on protein function, defining the physiological role of the protein, and characterizing the molecular pathway in which it acts. Next-generation sequencing (NGS) has significantly improved our ability to identify disease-causing genes and mutations, revolutionizing the diagnostic landscape. However, even for identified pathogenic variants, it is often difficult to determine their precise effect on protein functionality. Moreover, among the identified genetic variants, many are classified as variants of uncertain significance (VUS), mainly amino acid substitutions, which prevent a definitive molecular diagnosis. On the other hand, in some rare diseases, pathogenic mechanisms remain poorly understood, with only limited insight into protein function and the effects of disease-associated mutations. For others, years of research have shown that disease arises from the disruption of key molecular pathways; yet these pathways still require further investigation, for instance, at the transcriptional and post-transcriptional levels, which are often overlooked, and maybe useful in understanding the molecular mechanism of the disease. This thesis was aimed at addressing some of these aspects, combining bioinformatic prediction with experimental validation to strengthen the interpretation of VUS, while RNA sequencing and experimental confirmation enable analysis of transcriptional and post-transcriptional regulation of pathway associated with the disease. These approaches have been applied to three rare hereditary diseases, Alexander disease, thrombocytopenia 4, and Fanconi anemia, each examined in a separate chapter. In collaboration with Dr. Gaetano Vattemi (Department of Neurosciences, Biomedicine and Movement Sciences, University of Verona, Verona), the first chapter investigates a VUS, c.713T>G (p.I238S), in the GFAP (glial fibrillary acidic protein) gene, which causes Alexander disease (AxD), a rare neurodegenerative disorder characterized by the accumulation of GFAP aggregates in astrocytes. Bioinformatics analyses showed that the affected residue is highly conserved and located within the GFAP rod domain, which is crucial for filament assembly. Further predictions suggested that the variant disrupts protein function, protein-protein interactions, and post-translational modifications. Experimental validation in HeLa and U251-MG astrocytoma cells revealed that the I238S variant disrupted filament organization and induced cytoplasmic GFAP aggregates, a phenotype similar to that associated with other pathogenic GFAP variants, such as R239C. Together, these findings support classifying the GFAP I238S variant as deleterious, highlighting the importance of combining bioinformatic and functional approaches to resolve VUS in AxD. The second chapter focuses on thrombocytopenia 4 (THC4), an ultra rare hereditary disorder characterized by low platelet counts and caused by mutations in CYCS, encoding cytochrome c, a protein central to electron transport, peroxidase activity, and intrinsic apoptosis. Fourteen clinically relevant heterozygous missense variants (single amino acid variants) (T20I, V21G, H27Y, N32H, L33V, G42S, Y49H, A52T, A52V, R92G, A97D, Y98H, L99V, and T103I), selected from the literature and public databases, were investigated using bioinformatics and structural biology approaches. All variants affect highly conserved residues and are predicted as deleterious, with several also predicted to disrupt post-translational modifications, including phosphorylation (T20I, V21G, Y49H) and ubiquitination (G42S, A52T, A52V, T103I). Protein stability predictions and 500 ns molecular dynamics simulations suggested that most variants reduce stability and increase structural flexibility, particularly in the Ω-loop regions of CYCS. These predicted changes were associated with the displacement of the Ω-loops from the heme iron, disruption of the Tyr68-centered hydrogen-bond network, opening of specific cavities, and, consequently, widening of the heme pocket. This putative open conformation may allow increased access of small molecules such as H₂O₂ to the heme active site, thereby potentially enhancing peroxidase activity, which may promote apoptosis, impair megakaryopoiesis, and reduce platelet production, a hallmark of THC4. Overall, this study provides a comprehensive bioinformatic framework that supports the interpretation and prioritization of newly identified CYCS variants and VUS in research and diagnostic settings. The third chapter regards Fanconi anemia (FA), a rare inherited disorder caused by mutations in any of 23 FA genes whose protein products repair DNA interstrand crosslinks (ICLs), such as those induced by mitomycin (MMC), in the so-called FA/BRCA pathway. Despite well-characterized repair mechanisms, the regulation of FA/BRCA gene expression at the transcriptional and post-transcriptional levels during the DNA damage response remains poorly understood. In this study, HeLa cells were treated with MMC and sampled at 0, 16, 20, and 24 hours for cell cycle profiling, Western blotting, RNA sequencing, transcription factor inference, and alternative splicing analysis. Cell cycle profiling confirmed progressive S-phase delay, consistent with sustained replication fork stalling. Moreover, Western blotting validated pathway activation over time, as evidenced by FANCD2 monoubiquitination, a key modification in FA/BRCA activation. RNA sequencing revealed extensive transcriptional reprogramming with upregulation predominating over repression. Within the FA/BRCA pathway, gene upregulation was selective rather than uniform: ICL-sensing components, FA core complex subunits, the ID2 complex, and downstream effectors were progressively upregulated, while certain subunits remained stable, and FANCF was downregulated at later time points, suggesting dynamic stoichiometric rebalancing of the core complex. E2F1 and E2F4 were identified as the principal transcriptional regulators, also supported by independent ChIP-seq databases. Moreover, we analyzed the GC-AG introns, which being relatively frequent in BRCA/FA genes, were hypothesized to be potential players in post-transcriptional events under stress conditions. Having not revealed any relevant variations of the splicing events under MMC stress, we concluded that transcriptional upregulation, not alternative splicing, is the dominant regulatory mode during the acute ICL response. These findings demonstrate that the FA/BRCA pathway relies on a tightly coordinated, transcription-factor-driven program to ensure precise and timely DNA repair, with direct implications for Fanconi anemia pathogenesis and ICL chemotherapy resistance. In conclusion, this thesis, which addresses different aspects across different diseases, shows that combining bioinformatic and experimental approaches is important for associating data-driven predictions or large-scale findings with biological validation. By addressing both protein-level effects and pathway-level regulation, this work contributes to a more complete understanding of disease pathogenesis and supports improved molecular diagnosis and variant classification.