The Design and Assembly of 3D Liver Mimetic Cellular Architectures
We report the assembly of three-dimensional (3D) liver sinusoidal mimics comprised of primary rat hepatocytes, human or rat liver sinusoidal endothelial cells denoted as hLSECs and rLSECs respectively, and an intermediate chitosan-hyaluronic acid (HA) polyelectrolyte multilayer (PEM). The height of the PEMs ranged from 30-55nm and exhibited a shear modulus of ~ 100kPa. Primary rat hepatocytes coated with 5 and 15 PE layers exhibited stable urea and albumin production over a seven day period and these values were either comparable or superior to that in a collagen sandwich (CS). Hepatocyte-PEM-hLSEC liver mimics exhibited stable urea production and increasing albumin secretion over the culture period in comparison to hepatocyte-LSEC samples. In the 3D liver mimics, hLSEC phenotype was maintained and verified by the uptake of acetylated low-density lipoprotein (AcLDL). A sixteen-fold increase in CYP1A1/2 activity was observed for hepatocyte-PEM-10,000 hLSEC samples, thereby, suggesting that interactions between hepatocytes and hLSECs play a key role in enhancing hepatic phenotypes in in vitro cultures. As the first step towards elucidating key signaling pathways involved in cell-cell communications, global genome-wide transcriptional profiles of primary hepatocytes cultured in CS and hepatocyte monolayers (HMs) were performed over an eight-day period using DNA microarray measurements and Gene Set Enrichment Analysis (GSEA) in order to derive biologically meaningful information at the level of gene sets. The gene expression in CS cultures steadily diverged from that in HMs. Gene sets up-regulated in CS are those linked to liver metabolic and synthetic functions, such as lipid, fatty acid, alcohol and carbohydrate metabolism, urea production, and synthesis of bile acids. Monooxygenases such as CYP enzymes were significantly up-regulated starting on day 3 in CS cultures. These results serve as a baseline for further investigation into the systems biology of engineered liver tissues. 3D hepatic constructs were also assembled with primary rat hepatocytes and rLSECs, and a chitosan-HA PEM. In these hepatic models, the phenotype of hepatocytes and rLSECs were maintained. rLSEC phenotype was verified over a twelve-day period through immunostaining with the sinusoidal endothelial-1 (SE-1) antibody. In contrast, rLSECs cultured as monolayers lost their phenotype within 3 days. A two-fold increase in albumin production was observed only in the 3D liver models. rLSEC-PEM-hepatocyte cultures exhibited three- to six-fold increased CYP1A1/2 and CYP3A enzymatic activity. Well-defined bile canaliculi were observed in only 3D hepatic constructs. In summary, these results indicate that the layered rLSEC-PEM-hepatocyte constructs can be used as liver models for future studies.