Autosomal Dominant Polycystic Kidney Disease (ADPKD) is the most prevalent genetic kidney disorder, and one of the main causes of End-Stage Renal Disease (ESRD) [1], [2], textcolorbluewith an estimated prevalence between one in 1000 and one in 2500 individuals,[3]. Characterized by the progressive development of numerous fluid-filled cysts in the kidneys, ADPKD leads to significant structural changes, kidney enlargement, and, eventually, renal failure [4]. Despite its prevalence and severe implications, a complete cure has not yet been found, although in recent years, a vasopressin V2-receptor antagonist, Tolvaptan, has been successfully used to delay cyst formation and kidney deterioration [5], without completely halting disease progression. Currently, imaging diagnosis is effectively performed on adults, while this is not always possible on younger individuals. Alternative treatments are being investigated as the disease is being more thoroughly understood [6], but many unknowns remain about its development. Among the most valuable tools for researchers studying ADPKD are animal models, which are indispensable for understanding the disease’s development. These models offer
a controlled environment to investigate the genetic and molecular mechanisms underlying this condition. Genetically modified mice, in particular, allow researchers to manipulate specific genes and observe the resulting effects on kidney structure and function. These models enable longitudinal studies, providing insights into the progression of cyst formation and growth over time. They also allow for invasive techniques and repeated sampling, which are not feasible in human studies, facilitating detailed histological and other imaging analyses. The findings from these studies are crucial for developing and testing new treatments, ultimately aiding in the translation of research into clinical applications. With this in mind, the main focus of the current thesis is the detailed study of ADPKDinduced structural changes produced in mouse kidneys using an array of imaging modalities and processing methods. To do this, a longitudinal study has been conducted on mouse specimens induced with ADPKD. This includes the use of Computerized Tomography (CT) to obtain 3-dimensional reconstructions of the affected area of the kidney, Magnetic Resonance Imaging (MRI) to identify specific cysts, and Histology
imaging for a more detailed 2D analysis of single cross-sectional cuts. This intends to provide a comprehensive, multi-scale approach to studying ADPKD evolution over time, allowing for detailed in vivo monitoring, and high-resolution 2D imaging of kidney tissues. These image analyses are accompanied by the study of several biomarkers measured from the specimens. This study involved the segmentation of cysts in the histology images to evaluate disease development. The analyses performed in this regard used a semiautomatic tool that showed to be very time-consuming and user-dependent. For this reason, the next part of the work during this thesis was focused on the development of an automatized deep
learning-based processing method, which has been prepared to reduce the need for manual annotations and increase reproducibility for cyst segmentation in histology images. To really understand the low-level changes produced by the disease in the organ, it is necessary to choose a suitable alternative imaging technique that can show the inner structure of the kidney and how the disease transforms it in a more detailed way than the previous modalities, while providing a 3D reconstruction of the whole organ. Recently, an interesting option for this has been presented through the use of a 3D histology volume, built from a set of registered histology 2D cuts that cover the entirety of the kidney, which would allow for precise volumetric measurements [7]. This, however, is a very complex and expensive process and requires a very large amount of histology image acquisitions, maintaining a very precise protocol and registration processes. Throughout this project, an alternative process to this one has been selected to obtain highresolution data: the use of ex vivo Light Sheet Fluorescence Microscopy (LSFM). This microscopy technique allows the acquisition of high-resolution 3D volumes of an entire
mouse kidney through fluorescence signals. This method is much faster than the similar Confocal microscopy, also reducing the sample’s photobleaching, and can yield very interesting and detailed results to observe the inner structure of the kidney and account for the deformations caused by cysts at a detailed level. Within this thesis, an overview of the different architectures, samples, clearing protocols and processing methods used in ex vivo LSFM is included. This serves as a background for the reader to acquire a detailed understanding of the modality, and is followed by a final chapter where LSFM is applied to a series of full mouse kidneys, obtaining different detailed measurements to evaluate the effects of the disease at a lower level., The development of this thesis has been possible, directly or indirectly, through the following fundings: 1. Spanish Ministry of Science, Innovation and Universities: TEC2016-78052- R, FPU19/02854, PID2019-109820RB-100, PID2021-128862OB-I00, TED2021- 129392B-I00 2. European Regional Development Fund: RTC-2017-6600-1 3. Delegación del Gobierno para el Plan Nacional sobre Drogas, funds from Recovery and Resilience Facility, EU: EXP2022/008917 4. Horizon Europe Framework: 101057970-AI4LIFE 5. Government of Navarra: 0011-1411- 2019-000074, Programa de Doctorado en Ciencia y Tecnología Biomédica por la Universidad Carlos III de Madrid, Presidente: François Bernard Lauze.- Secretaria: Mónica García Sevilla.- Vocal: Perrine Paul-Guilloteaux