Fundamental research and drug development for personalized medicine necessitates cell cultures from defined genetic backgrounds. However, providing sufficient numbers of authentic cells from individuals poses a challenge. Here, we present a new strategy for rapid cell expansion that overcomes current limitations. Using a small gene library, we expanded primary cells from different tissues, donors, and species. Cell-type-specific regimens that allow the reproducible creation of cell lines were identified. In depth characterization of a series of endothelial and hepatocytic cell lines confirmed phenotypic stability and functionality. Applying this technology enables rapid, efficient, and reliable production of unlimited numbers of personalized cells. As such, these cell systems support mechanistic studies, epidemiological research, and tailored drug development.

Despite major advances in high-throughput and computational modelling techniques, understanding of the mechanisms regulating tissue specification and differentiation in higher eukaryotes, particularly man, remains limited. Microarray technology has been explored exhaustively in recent years and several standard approaches have been established to analyse the resultant datasets on a genome-wide scale. Gene expression time series offer a valuable opportunity to define temporal hierarchies and gain insight into the regulatory relationships of biological processes. However, unless datasets are exactly synchronous, time points cannot be compared directly. Here we present a data-driven analysis of regulatory elements from a microarray time series that tracked the differentiation of non-immortalised normal human urothelial (NHU) cells grown in culture. The datasets were obtained by harvesting differentiating and control cultures from finite bladder- and ureter-derived NHU cell lines at different time points using two previously validated, independent differentiation-inducing protocols. Due to the asynchronous nature of the data, a novel ranking analysis approach was adopted whereby we compared changes in the amplitude of experiment and control time series to identify common regulatory elements. Our approach offers a simple, fast and effective ranking method for genes that can be applied to other time series. The analysis identified ELF3 as a candidate transcriptional regulator involved in human urothelial cytodifferentiation. Differentiation-associated expression of ELF3 was confirmed in cell culture experiments and by immunohistochemical demonstration in situ. The importance of ELF3 in urothelial differentiation was verified by knockdown in NHU cells, which led to reduced expression of FOXA1 and GRHL3 transcription factors in response to PPARγ activation. The consequences of this were seen in the repressed expression of late/terminal differentiation-associated uroplakin 3a gene expression and in the compromised development and regeneration of urothelial barrier function.

No abstract

Key liver functions, including protein synthesis, carbohydrate metabolism, and detoxification, are performed by specific populations of hepatocytes that are defined by their relative positions within the liver lobules. On a molecular level, the functional heterogeneity with periportal and pericentral phenotypes, so-called metabolic liver zonation, is mainly established by a gradient of canonical Wnt signaling activity. Since the relevant physiological cues are missing in in vitro liver models, they fail to reflect the functional heterogeneity and thus lack many liver functions. We synthetically re-engineered Wnt signaling in murine and human hepatocytes using a doxycycline-dependent cassette for externally controlled digital expression of stabilized β-catenin. Thereby, we achieved adjustable mosaic-like activation of Wnt signaling in in vitro-cultured hepatocytes that was resistant to negative-feedback loops. This allowed the establishment of long-term-stable periportal-like and pericentral-like phenotypes that mimic the heterogeneity observed in vivo. The in vitro-zonated hepatocytes show differential expression of drug-metabolizing enzymes and associated differential toxicity and higher levels of autophagy. Furthermore, recombinant adeno-associated virus and hepatitis C virus preferentially transduce the pericentral-like zonation phenotype, suggesting a bias of these viruses that has been unappreciated to date. These tightly controlled in vivo-like systems will be important for studies evaluating aspects of liver zonation and for the assessment of drug toxicity for mouse and man.

No abstract