BiliQML: A supervised machine-learning model to quantify biliary forms from digitized whole-slide liver histopathological images.

The progress of research focused on cholangiocytes and the biliary tree during development and following injury is hindered by limited available quantitative methodologies. Current techniques include two-dimensional standard histological cell-counting approaches, which are rapidly performed error-prone and lack architectural context; or three-dimensional analysis of the biliary tree in opacified livers, which introduce technical issues along with minimal quantitation. The present study aims to fill these quantitative gaps with a supervised machine learning model (BiliQML) able to quantify biliary forms in the liver of anti-Keratin 19 antibody-stained whole slide images. Training utilized 5,019 researcher-labeled biliary forms, which following feature selection, and algorithm optimization, generated an F-score of 0.87. Application of BiliQML on seven separate cholangiopathy models; genetic (Afp-CRE;Pkd1l1null/Fl, Alb-CRE;Rbp-jkfl/fl, Albumin-CRE; ROSANICD), surgical (bile duct ligation), toxicological (3,5-diethoxycarbonyl-1,4-dihydrocollidine), and therapeutic (Cyp2c70-/- with ileal bile acid transporter inhibition), allowed for a means to validate the capabilities, and utility of this platform. The results from BiliQML quantification revealed biological and pathological differences across these seven diverse models indicate a highly sensitive, robust, and scalable methodology for the quantification of distinct biliary forms. BiliQML is the first comprehensive machine-learning platform for biliary form analysis, adding much needed morphologic context to standard immunofluorescence - based histology, and provides clinical and basic-science researchers a novel tool for the characterization of cholangiopathies.

Genetic Contributions to Biliary Atresia: A Developmental Cholangiopathy.

Biliary atresia (BA) is the most prevalent serious liver disease of infancy and childhood, and the principal indication for liver transplantation in pediatrics. BA is best considered as an idiopathic panbiliary cholangiopathy characterized by obstruction of bile flow and consequent cholestasis presenting during fetal and perinatal periods. While several etiologies have been proposed, each has significant drawbacks that have limited understanding of disease progression and the development of effective treatments. Recently, modern genetic analyses have uncovered gene variants contributing to BA, thereby shifting the paradigm for explaining the BA phenotype from an acquired etiology (e.g., virus, toxin) to one that results from genetically altered cholangiocyte development and function. Herein we review recently reported genetic contributions to BA, highlighting the enhanced representation of variants in biological pathways involving ciliary function, cytoskeletal structure, and inflammation. Finally, we blend these findings as a new framework for understanding the resultant BA phenotype as a developmental cholangiopathy.

LiverQuant: An Improved Method for Quantitative Analysis of Liver Pathology.

Current means to quantify cells, gene expression, and fibrosis of liver histological slides are not standardized in the research community and typically rely upon data acquired from a selection of random regions identified in each slide. As such, analyses are subject to selection bias as well as limited subsets of available data elements throughout the slide. A whole-slide analysis of cells and fibrosis would provide for a more accurate and complete quantitative analysis, along with minimization of intra- and inter-experimental variables. Herein, we present LiverQuant, a method for quantifying whole-slide scans of digitized histologic images to render a more comprehensive analysis of presented data elements. After loading images and preparing the project in the QuPath program, researchers are provided with one to two scripts per analysis that generate an average intensity threshold for their staining, automated tissue annotation, and downstream detection of their anticipated cellular matrices. When compared with two standard methodologies for histological quantification, LiverQuant had two significant advantages: increased speed and a 50-fold greater tissue area coverage. Using publicly available open-source code (GitHub), LiverQuant improves the reliability and reproducibility of experimental results while reducing the time scientists require to perform bulk analysis of liver histology. This analytical process is readily adaptable by most laboratories, requires minimal optimization, and its principles and code can be optimized for use in other organs. Graphical overview.

Opportunities and considerations for studying liver disease with microphysiological systems on a chip.

Advances in microsystem engineering have enabled the development of highly controlled models of the liver that better recapitulate the unique in vivo biological conditions. In just a few short years, substantial progress has been made in creating complex mono- and multi-cellular models that mimic key metabolic, structural, and oxygen gradients crucial for liver function. Here we review: 1) the state-of-the-art in liver-centric microphysiological systems and 2) the array of liver diseases and pressing biological and therapeutic challenges which could be investigated with these systems. The engineering community has unique opportunities to innovate with new liver-on-a-chip devices and partner with biomedical researchers to usher in a new era of understanding of the molecular and cellular contributors to liver diseases and identify and test rational therapeutic modalities.

Liver-restricted deletion of the biliary atresia candidate gene Pkd1l1 causes bile duct dysmorphogenesis and ciliopathy.

BACKGROUND AND AIMS: A recent multicenter genetic exploration of the biliary atresia splenic malformation syndrome identified mutations in the ciliary gene PKD1L1 as candidate etiologic contributors. We hypothesized that deletion of Pkd1l1 in developing hepatoblasts would lead to cholangiopathy in mice. APPROACH AND RESULTS: CRISPR-based genome editing inserted loxP sites flanking exon 8 of the murine Pkd1l1 gene. Pkd1l1Fl/Fl cross-bred with alpha-fetoprotein-Cre expressing mice to generate a liver-specific intrahepatic Pkd1l1 -deficient model (LKO). From embryonic day 18 through week 30, control ( Fl/Fl ) and LKO mice were evaluated with standard serum chemistries and liver histology. At select ages, tissues were analyzed using RNA sequencing, immunofluorescence, and electron microscopy with a focus on biliary structures, peribiliary inflammation, and fibrosis. Bile duct ligation for 5 days of Fl/Fl and LKO mice was followed by standard serum and liver analytics. Histological analyses from perinatal ages revealed delayed biliary maturation and reduced primary cilia, with progressive cholangiocyte proliferation, peribiliary fibroinflammation, and arterial hypertrophy evident in 7- to 16-week-old LKO versus Fl/Fl livers. Following bile duct ligation, cholangiocyte proliferation, peribiliary fibroinflammation, and necrosis were increased in LKO compared with Fl/Fl livers. CONCLUSIONS: Bile duct ligation of the Pkd1l1 -deficient mouse model mirrors several aspects of the intrahepatic pathophysiology of biliary atresia in humans including bile duct dysmorphogenesis, peribiliary fibroinflammation, hepatic arteriopathy, and ciliopathy. This first genetically linked model of biliary atresia, the Pkd1l1 LKO mouse, may allow researchers a means to develop a deeper understanding of the pathophysiology of this serious and perplexing disorder, including the opportunity to identify rational therapeutic targets.