Hepatology and drug development for liver diseases require in vitro liver models. Typical models include 2D planar primary hepatocytes, hepatocyte spheroids, hepatocyte organoids, and liver-on-a-chip. Liver-on-a-chip has emerged as the mainstream model for drug development because it recapitulates the liver microenvironment and has good assay robustness such as reproducibility. Liver-on-a-chip with human primary cells can potentially correlate clinical testing. Liver-on-a-chip can not only predict drug hepatotoxicity and drug metabolism, but also connect other artificial organs on the chip for a human-on-a-chip, which can reflect the overall effect of a drug. Engineering an effective liver-on-a-chip device requires knowledge of multiple disciplines including chemistry, fluidic mechanics, cell biology, electrics, and optics. This review first introduces the physiological microenvironments in the liver, especially the cell composition and its specialized roles, and then summarizes the strategies to build a liver-on-a-chip via microfluidic technologies and its biomedical applications. In addition, the latest advancements of liver-on-a-chip technologies are discussed, which serve as a basis for further liver-on-a-chip research.

Organ-on-a-chip academic research is in its blossom. Drug toxicity evaluation is a promising area in which organ-on-a-chip technology can apply. A unique advantage of organ-on-a-chip is the ability to integrate drug metabolism and drug toxic processes in a single device, which facilitates evaluation of toxicity of drug metabolites. Human organ-on-a-chip has been fabricated and used to assess drug toxicity with data correlation with the clinical trial. In this review, we introduced the microfluidic chip models of liver, kidney, heart, nerve, and other organs and multiple organs, highlighting the application of these models in drug toxicity detection. Some biomarkers of toxic injury that have been used in organ chip platforms or have potential for use on organ chip platforms are summarized. Finally, we discussed the goals and future directions for drug toxicity evaluation based on organ-on-a-chip technology.

The poor metabolic ability of cell lines fails to meet the requirements of an in vitro model for drug interaction testing which is crucial for the development or clinical application of drugs. Herein, we describe a liver sinusoid-on-a-chip device composed of four kinds of transformed cell lines (HepG2 cells, LX-2 cells, EAhy926 cells, and U937 cells) that were ordered in a physiological distribution with artificial liver blood flow and biliary efflux flowing in the opposite direction. This microfluidic device applied three-dimensional culturing of HepG2 cells with high density (10⁷ ml⁻¹), forming a tightly connected monolayer of EAhy926 cells and achieving the active transport of drugs in HepG2 cells. Results showed that the device maintained synthetic and secretory functions, preserved cytochrome P450 1A1/2 and uridine diphosphate glucuronyltransferase enzymatic activities, as well as sensitivity of drug metabolism. The cell lines derived device enables the investigation of a drug-drug interaction study. We used it to test the hepatotoxicity of acetaminophen and the following combinations: “acetaminophen + rifampicin,” “acetaminophen + omeprazole,” and “acetaminophen + ciprofloxacin.” The variations in hepatotoxicity of the combinations compared to acetaminophen alone, which is not found in a 96-well plate model, in the device were −17.15%, 14.88%, and −19.74%. In addition, this result was similar to the one tested by the classical primary hepatocyte plate model (−13.22%, 13.51%, and −15.81%). Thus, this cell lines derived liver model provides an alternative to investigate drug hepatotoxicity, drug-drug interaction.

Triple-negative breast cancer (TNBC) is one of the most aggressive and poor prognosis breast cancers. Currently, chemotherapy with conventional cytotoxic agents is the only available option to treat TNBC. Hence, we identified new therapeutic agents against TNBC from traditional Chinese medicine Radix Bupleuri and unveiled the molecule mechanism of anti-TNBC effects. Multi-component bioactivity and structure-guided methods were used to identify the most effective anti-TNBC compound Saikosaponin D (SSD) from Radix Bupleuri. Cell viability and apoptosis assays were employed to demonstrate the effect of SSD on the proliferation and apoptosis of TNBC cells. Dynamic mass redistribution assay, TopFlash assay, western blotting, and special agonist were applied to dissect the potential molecular mechanisms of SSD. We screened twenty fractions in Radix Bupleuri and identified SSD as the most effective component to inhibit the proliferation of TNBC cells. Investigating the interaction of SSD with the frequently overexpressed targets in TNBC led to the identification that it markedly suppressed Wnt/β-catenin signaling, but did not act on epidermal growth factor receptor and neurotensin receptor-1. Moreover, we demonstrated that SSD significantly repressed β-catenin and its downstream target genes, resulting in TNBC cell apoptosis. Specifically, docking of SSD to the crystal structure of β-catenin suggested that SSD interacted with β-catenin via hydrogen bonds and hydrophobic interaction. We identified the most effective component SSD from Radix Bupleuri in inhibiting the proliferation of TNBC cells by targeting β-catenin signaling. Given the important role of Wnt/β-catenin signaling in breast cancer, SSD may present an opportunity to discover new therapeutics for the treatment of TNBC.

Intestinal floras influence a lot of biological functions of the organism. Although animal model are strong tools for researches on the relationship between host and microbe, a physiologically relevant in vitro human gut model was still required. Here, a novel human gut-vessel microfluidic system was established to study the host–microbial interaction. Peristaltic motion of the cells on the chip was driven by a pneumatic pump. When intestinal epithelial cells (Caco2) were co-cultured with vascular endothelial cells (HUVECs) on the peristaltic microfluidic chip, Caco2 showed normal barrier and absorption functions after 5 days cultivation, which generally took 21 days in static Transwell models. Intestinal microvilli and glycocalyx layer were seen after 4 days cultivation, and Lactobacillus casei was successfully co-cultured for a week in the intestinal cavity. A model for intestinal damage and inflammatory responses caused by E. coli was set up on this chip, which were successfully suppressed by Lactobacillus casei or antibiotic. In summary, this human gut-vessel microfluidic system showed a good potential for investigating the host–microbial interaction and the effect and mechanism of microbiome on intestinal diseases in vitro.