This protocol describes a semi-automated approach to produce hepatocyte-like cells from human pluripotent stem cells in a 96 well plate format. This process is rapid and cost-effective, allowing the production of quality assured batches of hepatocyte-like cells for basic and applied human research.
Human pluripotent stem cells represent a renewable source of human tissue. Our research is focused on generating human liver tissue from human embryonic stem cells (hESCs) and induced pluripotent stem cells (iPSCs). Current differentiation procedures generate human hepatocyte-like cells (HLCs) displaying a mixture of fetal and adult traits. To improve cell phenotype, we have fully defined our differentiation procedure and the cell niche, resulting in the generation of cell populations which display improved gene expression and function. While these studies mark progress, the ability to generate large quantities of multi well plates for screening has been limited by labour intensive procedures and batch to batch variation. To tackle this issue, we have developed a semi-automated platform to differentiate pluripotent stem cells into HLCs. Stem cell seeding and differentiation were performed using liquid handling and automatic pipetting systems in 96-well plate format. Following the differentiation, cell phenotype was analyzed using automated microscopy and a multi well luminometer.
Primary human hepatocytes (PHHs) represent the current standard for liver biomedical research. Despite their advantages, PHHs represent a limited source of mature hepatocytes with unstable phenotype1. Human embryonic stem cells (ESCs) and human induced pluripotent stem cells (iPSC) have been proposed as a powerful resource for biomedical research, promising a renewable source of human tissue, including liver2,3,4. Current hepatocyte differentiation procedures generate cells that express hepatocyte markers, genes and functions similar to PHHs3,4,5,6,7,8,9,10,11. More recently, the use of defined materials and matrices such as recombinant laminins, has improved cell phenotype and experimental reproducibility5,12,13,14.
Traditional cell culture techniques rely heavily on manual pipetting, which is both time consuming and error prone. This limits the throughput of the assay and the plate format one can work with. To date, most studies describing the generation of HLCs are labor intensive in nature and therefore small scale in size15,16,17.
Current advances in liquid handling and pipetting systems, in combination with automated microscopy and laboratory analysers, have made it possible to develop procedures which minimize the requirement for human intervention. We have taken advantage of these technologies and developed a semi-automated differentiation system with which to generate large quantities of human liver tissue for basic and applied biomedical research. This approach can be performed with both human ESC and iPSC lines with minor adjustments necessary, depending on the cell line. Combining this system with high content analysis, we demonstrate that HLC phenotyping can be less time consuming and performed at scale18,19,20,21.
In summary, the automation of cell culture techniques described herein has led to improvements in reliability, experimental time management and assay throughput.
1. Seeding Human Pluripotent Stem Cells (hPSCs) Into 96 Well Plates For Hepatocyte Differentiation
NOTE: Reagents and equipment used in this protocol has been listed in Table 1. Media used for this experiment must be sterile and at room temperature for cell culture. All reagent/solution volumes described in this part of the protocol are based on a single well of a 96-well plate unless otherwise specified.
2. Differentiating hPSCs to Hepatocyte-like Cells on Recombinant Laminins in 96 Well Plates
3. Characterization of Hepatocyte-like cells Differentiated On Recombinant Laminins
4. High Throughput Immunocytochemistry and Image Acquisition
Cell differentiation was performed using an embryonic stem cell line (H9). The semi-automated platform was used to differentiate and characterize the cells (Figure 1A). An automatic liquid handling dispenser was used to coat matrix and seed single cells. Cellular differentiation was initiated when the cells reached 40% confluency (Day 0). Media changes were performed using an automatic hand-held electronic channel pipette system to remove medium and the liquid handling dispenser for medium addition (Figure 1A). Cell fixation and staining were performed using the automatic liquid handling dispenser in combination with the automatic plate washer. Imaging and quantification were performed by using a high-content imaging system and Columbus software. Cytochrome P450 activity was measured using an established assay. Following substrate incubation, cell supernatants were harvested, and bioluminescence recorded using a multimode microplate reader (Figure 1A). In addition, conventional phase contrast microscopy was used to measure the changes in cell morphology during hepatocyte differentiation (Figure 1B).
On Day 18, the variability in cell number between wells was examined following DAPI staining (Figure 2A). Seven fields of view were captured per well and the number of nuclei quantified (Figure 2A). We detected no statistical differences between wells with an average of 41,662 ±3,366 cells per well (Figure 2B).
At day 18, HLCs were fixed and stained for typical hepatocyte markers (Figure 3A-G) and tested for CYP P450 activity. HLCs expressed hepatocyte markers such as HNF4α (89.2% ±2 positive cells), ALB (92.8% ±6 positive cells), AFP (61.8% ±2 positive cells). HLCs also expressed CYP P450 proteins, CYP2D6 and CYP3A4 and displayed a distinct polygonal appearance as shown by E-Cadherin protein expression. HLCs CYP P450 activity was measured at day 18. HLCs exhibited CYP1A2 activity at 295,906 ±45,828 RLU/mL/mg of protein and CYP3A at 1,066,112 ±177,416 RLU/mL/mg of protein (Figure 3H).
Figure 1: Hepatocyte differentiation from PSCs in the 96 well format. (A). Diagram of the equipment used to semi-automate cellular differentiation, fixation, staining, imaging and functional activity. (B). Representative changes in cellular morphology during HLC differentiation. Briefly, PSCs are primed toward definitive endoderm, followed by hepatoblast specification characterized by cells displaying cobblestone-like morphology and prior to hepatocyte maturation with cells acquiring polygonal-shape. Scale bar = 100 µm. Please click here to view a larger version of this figure.
Figure 2: Assessment of HLCs well-to-well variability. (A) Representation of a 96 well plate view of HLCs stained with DAPI. Scale bar = 1 mm. (B) Quantification of cell number per well. Average of cell number per wells in columns (top) and rows (bottom), from seven fields of views per well and quantified using Columbus software. Average cell number across the plate is 41,662 ± 3,366 SEM cells per well. No statistically significant differences were observed between wells. A One-way ANOVA with Tukey's post-hoc statistical tests was employed. Please click here to view a larger version of this figure.
Figure 3: Hepatocyte marker expression and CYP P450 activity measurement in HLCs at day 18. (A) The percentage of hepatocyte nuclear factor 4 alpha (HNF4α) expression ± SEM is based on thirty wells with ten fields of view per well. (B) The percentage of albumin (ALB) expression +/- SEM, is based on 3 separate wells with ten fields of view per well. (C) The percentage of alpha fetoprotein (AFP) +/- SEM is based on 3 separate wells with ten fields of view per well. (D) E-cadherin staining. (E) CYP2D6 staining. (F) CYP3A4 staining. (G). Immunoglobulin G (IgG) staining control. Scale bar = 50 µm. (H) CYP P450 1A2 and 3A metabolic activity in HLCs at day 18. The data represents six biological replicates +/- SEM. Activity is quoted as relative light units (RLUs)/mL per 1 mg of protein. Please click here to view a larger version of this figure.
Our semi-automated procedure is efficient, reliable and economical, allowing the production of HLCs at scale (Figure 1). There was not any significant difference in terms of the number of cells between the wells in the plate, making this platform a suitable approach for cell-based screening (Figure 2). In addition, the semi-automation workflow enables the user to produce large numbers of plates at once, which was not possible before using manual processes5,12.
Importantly, the automation process did not impact on differentiation yields, with the majority of cells expressing HNF4α (89.2%±2) and ALB (92.8%±6) (Figure 3). HLCs also expressed CYP P450 enzymes CYP2D6 and CYP3A4 and displayed CYP1A2 and CYP3A metabolic activity comparable to previous reported experiments (Figure 3)5. Despite the standardization of the protocol, cell confluency prior to the differentiation is critical for pure HLC differentiation. Therefore, ensuring a good cell distribution across the well is an important consideration (Figure 1). This has proven to be cell line dependent, therefore cell seeding and density optimization is required for each cell line prior to scale-up.
In its current form, this platform is not suitable to produce large quantities of HLCs for clinical applications, and this will most likely be achieved through the use of cell factories and bioreactors. However, the methodology developed does allow for the rapid generation of human liver cells for disease modelling, drug screening and/or drug repurposing studies. Moreover, the assay is amenable to multiplexing, permitting the generation of multiparametric datasets for mechanistic analysis. Going forward, we also believe that this technology could be applied to more complex in vitro systems such as 3D differentiation or as a platform to improve HLC phenotype in vitro.
In conclusion, we believe that our simple and semi-automated system will increase experimental throughput, and reduce experimental variation, thereby improving the quality of biological datasets and their extrapolation toward human physiology.
The authors have nothing to disclose.
This work was supported with awards from the Chief Scientist's Office (TCS/16/37) and an MRC PhD scholarship.
405 LS Washer | Biotek | ||
B27 supplement | Life Technologies | 12587-010 | |
beta-mercaptoethanol | Life Technologies | 31350 | |
Bovine Serum Albumin | Sigma-Aldrich | A2058 | |
DMSO | Sigma-Aldrich | D5879 | |
DPBS with Calcium and Magnesium | ThermoFisher | 14040133 | |
Gentle cell dissociation reagent | STEMCELL Technologies | 7174 | |
GlutaMax | Life Technologies | 35050 | |
GripTips for VIAFLO 96 | Integra | 6434 | |
HepatoZYME-SFM | Life Technologies | 17705021 | |
Human Activin A | Peprotech | 120-14E | |
Human Hepatocyte Growth Factor | Peprotech | 100-39 | |
Human Oncostatin M | Peprotech | 300-10 | |
Human Recombinant Laminin 521 | BioLamina | LN521-02 | |
Hydrocortisone 21-hemisuccinate sodium salt | Sigma-Aldrich | H4881 | |
Knockout DMEM | Life Technologies | 10829 | |
Knockout Serum Replacement | Life Technologies | 10828 | |
mTeSR1 medium | STEMCELL Technologies | 5850 | |
MultidropCombi Reagent Dispenser | ThermoFisher | 5840300 | |
Non-essential amino acids | Life Technologies | 11140 | |
Operetta High-Content Imaging System | PerkinElmer | HH12000000 | |
Penicillin-Streptomycin (10,000 U/mL) | Life Technologies | 15140122 | |
Recombinant mouse Wnt3a | R&D Systems | 1324-WN-500/CF | |
Rho-associated kinase (ROCK)inhibitor Y27632 | Sigma-Aldrich | Y0503-1MG | |
RPMI 1640 | Life Technologies | 21875 | |
Standard tube dispensing cassette | ThermoFisher | 24072670 | |
TWEEN 20 | Sigma-Aldrich | P9416 | |
VIAFLO 96 Electronic 96-channel pipette | Integra | 6001 | |
PBS, no calcium, no magnesium | ThermoFisher | 14190250 | |
Formaldehyde solution 4%, buffered, pH 6.9 | Sigma-Aldrich | 1.00496 EMD MILLIPORE | |
P450-Glo CYP3A4 Assay and Screening System | Promega | V8801 | |
P450-Glo CYP1A2 Assay and Screening System | Promega | V8771 | |
DAPI | Invitrogen | D1306 | |
Antibodies | |||
Albumin | Sigma-Aldrich | A6684 | 1:200 (mouse) |
Alpha-fetoprotein | Abcam | ab3980 | 1:500 (mouse) |
CYP2D6 | University of Dundee | University of Dundee | 1:200 (sheep) |
CYP3A4 | University of Dundee | University of Dundee | 1:200 (sheep) |
HNF-4α | Santa Cruz | sc-8987 | 1:400 (rabbit) |
E-cadherin | Abcam | ab76055 | 1:200 (mouse) |
IgG | DAKO | 1:400 | |
Anti-Mouse 488 (secondary antibody) | Life technologies | A-11001 | 1:400 |
Anti-sheep 488 (secondary antibody) | Life technologies | A-11015 | 1:400 |
Anti-rabbit 488 (secondary antibody) | Life technologies | A-11008 | 1:400 |