A Novel Noninvasive, Quantitative Molecular Imaging Platform for Assessing Transcriptional States during Endoderm and Hepatocyte Formation from Human Pluripotent Stem Cells

Although the liver regenerates in vivo, the regenerative function is lost during fibrosis, and hepatocytes don’t divide in culture. Pluripotent stem cells (PSCs) are an exciting potential source of hepatocytes, but thus far scientists have failed to generate fully mature hepatocytes. Hepatocytes are derived from endoderm, one of three main germ layers during development. Primary human endoderm, and thus human pluripotent stem cell (PSC)-derived endoderm, remain an enigma, and PSC endoderm studies demonstrate phenotypic heterogeneity, aberrant differentiation, difficulty with maturation, and in some cases, resistance to differentiation. In vivo genetic studies has demonstrated that developmental, master, regulatory gene circuits (DRGC), composed of transcription factors, govern endoderm differentiation. We hypothesized that a poorly regulated DRGC may be root cause for the problems with liver differentiation. To better understand DRGC, we developed a noninvasive, quantitative platform for imaging endoderm in vitro and in vivo. We engineered endoderm transcription factor (TF) -specific response elements which contain DNA binding sites for Brachyury, Foxa2, and Sox17, three key TFs that compose the DRGC in endoderm. The binding sites were flanked upstream by sequences a two-step transcriptional activation system, to improve signal strength in baseline and stimulated states. Our data indicated that low levels of signal were stable at 6 weeks after lentiviral transduction, enabling easy colony identification. We initiated PSC differentiation for three days in coated matrigel plates, and Activin, a key soluble factor that induces the DRGC as a function of concentration. We imaged at least 3 and up to 15 colonies a condition, and concentration experiments (0, 2, 20, 40, 100 ng/ml) were performed in duplicate. We normalized reporter expression to control hPSC which ubiquitously express luciferase. In a total of 24 PSC colonies containing the reporters, max radiance was 2 x 10⁵ ± 1.6 10⁵ photons/cm²-s/sr, demonstrating considerable baseline signal strength. The Brachyury reporter activity demonstrated decreasing activity with increasing concentration, and 42% decrease in normalized signal at 100 ng/ml compared to 0 ng/ml. The Foxa2 reporter demonstrated a 143% increase in normalized signal at 100 ng/ml compared to 0ng/ml. Finally, the Sox17 reporter demonstrated a 91% increase in normalized signal at 100 ng/ml compared to 0 ng/ml. These data are consistent with expected result, which is a downregulation of Brachyury, the precursor to mesendoderm, and upregulation of Foxa2 and Sox17, which are together specific for endoderm, in the presence of Activin. Data for intermediate concentrations of Activin were inconclusive. Future work will focus on correlating reporter expression with RNA or protein levels, repeating experiments at intermediate concentrations of Activin, improving clonality of cell lines, and studying transcriptional events in vivo. These studies will lead to a greatly improved understanding of the DRGC which govern endoderm induction and differentiation along the hepatic lineage for generating functional hepatocytes.