This protocol describes an approach to produce hepatospheres from human pluripotent stem cells using a defined culture system and cell self-assembly. This protocol is reproducible in a number of cell lines, cost effective and allows the production of stable human hepatospheres for biomedical application.
The development of renewable sources of liver tissue is required to improve cell-based modelling, and develop human tissue for transplantation. Human embryonic stem cells (hESCs) and human induced pluripotent stem cells (hiPSCs) represent promising sources of human liver spheres. We have developed a serum free and defined method of cellular differentiation to generate three-dimensional human liver spheres formed from human pluripotent stem cells. A potential limitation of the technology is the production of dense spheres with dead material inside. In order to circumvent this, we have employed agarose microwell technology at defined cell densities to control the size of the 3D spheres, preventing the generation of apoptotic and/or necrotic cores. Notably, the spheres generated by our approach display liver function and stable phenotype, representing a valuable resource for basic and applied scientific research. We believe that our approach could be used as a platform technology to develop further tissues to model and treat human disease and in the future may permit the generation of human tissue with complex tissue architecture.
The ability of human pluripotent stem cells (hPSCs) to self-renew, whilst retaining pluripotency, provides an opportunity to produce human cell types and tissues on demand. hPSCs have been efficiently differentiated into hepatocyte-like cells (HLCs) using two-dimensional (2D) adherent culture systems1,2,3,4,5,6,7,8,9,10. These systems have been used to successfully model monogenic disease, virus lifecycle, drug induced liver injury (DILI), fetal exposure to toxins and non-alcoholic fatty liver disease (NAFLD)11,12,13,14,15. However, these models do possess some drawbacks, which limit their routine use. Those include fetal marker expression, unstable phenotype and poor tissue architecture16,17,18,19, which could also limit extrapolation to organ function in vivo.
To overcome these limitations, three-dimensional (3D) differentiation platforms have been developed to mimic in vivo tissue architecture. Although enabling, those approaches rely on the use of animal derived products and matrices to drive tissue genesis20,21,22, limiting scale-up and widespread application.
Here, we detail procedures to generate large quantities of 3D hepatospheres from hPSCs using defined materials and cell self-assembly. Notably, the tissue generated by our procedure remains functional for more than one year in cell culture and is capable of supporting liver function in vivo23.
In summary, our defined differentiation approach allows the generation of stable human hepatospheres from both human embryonic stem cells (hESCs) and induced pluripotent stem cells (iPSCs). We believe the described procedure represents a significant breakthrough in the generation of 3D hepatospheres for basic and applied scientific research.
1. Preparation of Agarose Microplate Molds
NOTE: Media used for these experiments must be sterile and at room temperature (RT) for cell culture.
2. Seeding Human Pluripotent Stem Cells into Agarose Microwell Plates
3. Differentiating hPSCs to 3D Hepatospheres on Agarose Microwells
4. Functional Analysis of Long-term Cultured 3D Hepatospheres
5. Immunocytochemistry
Three-dimensional aggregates from the embryonic stem cell line (H9) or the induced pluripotent stem cell line (P106) were differentiated towards the hepatocyte lineage using our defined procedure (Figure 1). Pluripotent stem cells were first primed towards definitive endoderm prior to hepatoblast specification. Following this, hepatoblasts were matured into 3D hepatospheres which could be maintained in culture for up to one year23.
To study the structure of 3D spheres, 30 day-old hESC- or iPSC-derived spheres were fixed, sectioned and stained to detect the presence of proteins expressed in hepatocytes and mesenchymal cells. Hepatocyte nuclear factor 4 alpha (HNF4α) and the mesenchymal marker vimentin were employed, revealing the presence of an outer layer composed of hepatocyte like cells surrounding a core of mesenchymal cells (Figure 2). We followed these experiments analyzing the expression of proteins expressed in hepatocytes: albumin, CYP3A and E-cadherin. Immunostaining revealed that the expression of these proteins was restricted to the outer layer of the spheres (Figure 3).
Functional analyses of the hepatospheres were performed in day 30 cultures. CYP1A2 and CYP3A are important enzymes within functional hepatocytes. Their activity was assessed using established assays. H9-derived hepatospheres exhibited CYP3A activity of 220,375 ± 74514 RLU/mL/mg protein and CYP1A2 activity of 732,440 ± 33,330 RLU/mL/mg protein (Figure 4A). CYP3A activity in P106-derived hepatospheres was 132117 ± 43,391 RLU/mL/mg protein and CYP1A2 activity was 409,907 ± 121,723 RLU/mL/mg protein (Figure 4B). When compared with two batches of human primary hepatocytes, 3D liver spheres displayed respectable levels of CYP activity10.
Analysis of the synthesis and secretion of albumin and alpha-fetoprotein (AFP) revealed that H9-derived hepatospheres secreted 683.9 ± 84 and 159 ± 20 ng/mL/24 h/mg protein of albumin and alpha-fetoprotein, respectively (Figure 5A). Whereas P106-derived hepatospheres secreted 497 ± 41 and 756 ± 24 ng/mL/24 h/mg protein of albumin and alpha-fetoprotein, respectively (Figure 5B).
Figure 1: Stepwise differentiation procedure to generate 3D hepatospheres from hESCs. Blue brackets represent days of differentiation for hiPSCs. Please click here to view a larger version of this figure.
Figure 2: Structural reorganization of 3D hepatospheres. Representative images of the expression of hepatocyte nuclear factor 4 alpha (HNF4α – green) and vimentin (red) in (A) H9-derived 3D hepatospheres and (B) P106-derived 3D hepatospheres and their corresponding immunoglobulin G (IgG) controls. Scale bars represent 60 μm. Please click here to view a larger version of this figure.
Figure 3: Evaluation of hepatic marker expression in 3D hepatospheres. Representative images of the expression of hepatocyte markers - albumin, CYP3A, E-cadherin and their corresponding IgG controls in (A) H9- and (B) P106-derived 3D hepatospheres. Scale bars = 60 μm. Please click here to view a larger version of this figure.
Figure 4: Cytochrome P450 function in 3D hepatospheres. Measurement of the cytochrome P450 1A2 and 3A activity in (A) H9-derived 3D hepatospheres and (B) P106-derived 3D hepatospheres. The data represents the mean of three biological replicates, and the error bars represent the standard deviation (SD). Activity is quoted as relative light units (RLU) per mL per mg of protein. Please click here to view a larger version of this figure.
Figure 5: Analysis of hepatosphere protein secretion. The secretion of albumin and alpha-fetoprotein (AFP) was analyzed in (A) H9-derived 3D hepatospheres and (B) P106-derived 3D hepatospheres. The data is representative of three biological replicates, and the error bars represent the SD. Secreted protein is quoted as nanograms of protein per mL per 24 h per mg of protein. Please click here to view a larger version of this figure.
The development of defined and xeno-free systems to produce human hepatospheres in 3D is required for both in vitro and in vivo endeavors. At present most hepatocyte differentiation approaches from human pluripotent stem cells are performed in two dimensional adherent cultures. These environments lack many of the environmental cues involved in tissue genesis and homeostasis which include; heterotypic cell interactions, matrix production and remodeling, resulting in poor translation to the in vivo biology18,19.
As a result, research has focused on alternative approaches to generate hepatospheres from pluripotent stem cells. A number of 3D studies have advanced the field, but those are reliant on animal products20,22,24 to provide support and/or require the use of human tissue21,22 which complicates technology scale-up and compromises experimental reproducibility and application.
The procedure described in our article (Figure 1) is defined, efficient, highly reproducible and cost-effective, allowing the production of functional liver spheres, which remain functional over a year in vitro and provide critical liver support in vivo14. Importantly, this platform allows the user to control the size of the 3D liver spheres, limiting the formation of dense necrotic centers and loss of phenotype.
The transfer of the 3D hepatospheres to the poly-HEMA coated plates represents a critical step in this protocol. It is important to pipette gently at this stage in the procedure to avoid damaging the sphere. In addition, media changes must be performed carefully to avoid shear stress and distortion of sphere structure.
In these studies, 3D hepatospheres displayed an organized structure (Figure 2 and Figure 3), cytochrome P450 function (Figure 4) and secreted liver proteins, including albumin and alpha-fetoprotein (Figure 5). This procedure has been successfully performed in four pluripotent stem cells lines with comparable outcomes. Looking ahead, this technology could be employed as a platform to develop further endodermal and mesenchymal tissues with complex architectures.
The authors have nothing to disclose.
This study was supported with awards from the UK Regenerative Medicine Platform (MRC MR/L022974/1) and the Chief Scientist’s Office (TCS/16/37).
Cell Culture and functional assays | |||
Agarose | Fisher Bioreagents | 10766834 | |
B27 supplement | Life Technologies | 12587-010 | |
beta-mercaptoethanol | Life Technologies | 31350 | |
Bovine Serum Albumin | Sigma-Aldrich | A2058 | |
DAPI | Invitrogen | D1306 | |
DMSO | Sigma-Aldrich | D5879 | |
DPBS, no calcium, no magnesium | ThermoFisher | 14190250 | |
DPBS with Calcium and Magnesium | ThermoFisher | 14040133 | |
Gentle cell dissociation reagent | STEMCELL Technologies | 7174 | |
GlutaMax | Life Technologies | 35050 | |
HepatoZYME-SFM | Life Technologies | 17705021 | |
Human Activin A | Peprotech | 120-14E | |
Human Alpha Fetoprotein ELISA | Alpha Diagnostics | 500 | |
Human Basic Fibrobaslt Growth Factor | Peprotech | 100-18B | |
Human Epithelial Gropwth Factor | Peprotech | 236-EG | |
Human Hepatocyte Growth Factor | Peprotech | 100-39 | |
Human Oncostatin M | Peprotech | 300-10 | |
Human Recombinant Laminin 521 | BioLamina | LN521-02 | |
Human Serum Albumin ELISA | Alpha Diagnostics | 1190 | |
Human Vascular Growth Factor | Bio-techne | 293-VE | |
Hydrocortisone 21-hemisuccinate sodium salt | Sigma-Aldrich | H4881 | |
Knockout DMEM | Life Technologies | 10829 | |
Knockout Serum Replacement | Life Technologies | 10828 | |
Micro-mold spheroids | Sigma-Aldrich | Z764000 | |
mTeSR1 medium | STEMCELL Technologies | 5850 | |
Non-essential amino acids | Life Technologies | 11140 | |
P450-Glo CYP1A2 Assay and Screening System | Promega | V8771 | |
P450-Glo CYP3A4 Assay and Screening System | Promega | V8801 | |
Penicillin-Streptomycin (10,000 U/mL) | Life Technologies | 15140122 | |
poly-HEMA (Poly 2-hydorxyethyl methacrylate) | Sigma-Aldrich | P3932 | |
Recombinant mouse Wnt3a | Bio-techne | 1324-WN-500/CF | |
RPMI 1640 | Life Technologies | 21875 | |
Equipment | |||
Microwave | Bosch | ||
Microtome | Leika | RM2125RT | |
Oven | Thermoscientific | ||
Antibodies | |||
Primary antibodies | |||
Albumin | Sigma-Aldrich | A6684 | 1:100 (mouse) |
CYP3A4 | University of Dundee | University of Dundee | 1:200 (sheep) |
E-cadherin | Abcam | ab76055 | 1:200 (mouse) |
HNF-4α | Santa Cruz | sc-8987 | 1:100 (rabbit) |
IgG | DAKO | X0943 | 1:400 |
Vimentin | DAKO | M0725 | 1:100 (sheep) |