Background

            The differentiation and morphogenesis of progenitor cells into functioning tissues is governed by biochemical cues from neighbouring cells and other elements in the microenvironment. Moreover, it has been established that mechanical forces and mechanical signaling are pivotal for stem cell development and tissue behaviour. So far, much of the work relating stem cell behaviour to mechanical forces and signaling has been studied in 2D culture systems. However, the advent and more widespread use of 3D systems has proven to better replicate some in vivo conditions including increased cell-cell interactions and more freedom for motility and reorganization, which are key in differentiation and morphogenesis. Careful engineering is required to investigate cell responses in a 3D environment. Microwell-based platforms with different levels of complexity, have been widely implemented for various purposes including drug-screening, disease modeling and stem cell culture. However, implementation of 3D systems with mechanical constraints with tissue-specific contexts has been more limited. One such tissue-specific 3D models, is the liver, which in the embryo is derived from bipotential progenitor cells (hepatoblasts) that can give rise to hepatocytes or biliary epithelial cells, and eventually establish the liver lobes. It is known that the process of fate-specification, and eventual formation of the liver tissue and bile ducts, is spatially and temporally orchestrated by various cues. Additional factors including extracellular matrix stiffness, may also impact differentiation. To date, the relationship between mechanical signaling and liver progenitor fate specification in 3D has not yet been characterized. In their work, Berg et al (1) implemented an ECM scaffold-free hydrogel microwell-based method to produce arrays of liver progenitor cell microtissues, and characterized the 3D patterns of hepatocytic and biliary differentiation in various tissue geometries.

 

Key findings and developments

            The authors generated a microwell-based approach to generate arrays of bipotential mouse embryonic liver (BMEL) cell 3D tissues with defined geometries. The baseline was a PEG substrate with 500µm side walls, with an insert array of microwells 200µm deep. This setup enables a wide variety of well geometries with high aspect ratio features, such as cylinder and toroid wells of varied inner and outer diameter. Moreover, the approach aims for tissues with approximately the same initial number of cells across geometries. As the microwells were non-fouling and non-adhesive, cells only adhere to themselves, and aggregated into dense 3D tissues constrained by the wells. In cylindrical wells, the cells aggregate and condense into a roughly cylindrical tissue. In toroid microtissues, the cells aggregate and condense away from the outer walls, but around the central post, resulting in a donut-shaped tissue. Image segmentation was used to locate individual nuclei to enable single cell analysis in 3D.

The authors then investigated hepatocytic and biliary phenotype patterning in cylindrical microtissues. After 72 hours of culture, the microtissues were fixed and immune-stained for biliary (OPN) and hepatocytic (HNF4a) markers. In the tissues, the cells expressing the biliary marker OPN were sparsely distributed throughout the tissue, while the cells expressing the hepatocytic marker HNF4a were found almost exclusively in the outer shell of the tissue. This was confirmed following the implementation of an image segmentation-based single cell analysis pipeline. Sorting cells into an inner, intermediate, and outer region based on shell coordinates, and calculating the percentage of cells positive for each marker, showed that there was a significantly different percentage of cells expressing the hepatocytic marker across all regions, with the outer region being the higher, and inner the lowest. Conversely, there was a significant increase in percentage of OPN+ cells in the intermediate region compared to the outer and inner regions.

The authors then investigated hepatocytic and biliary phenotype patterning in toroid microtissues. In these tissues, OPN positive cells were sparsely distributed across the 3D structure while HNF4a positive cells were found in the outer surface. Moreover, low levels of HNF4a positive cells were observed at the surface contacting the PEG pillar. Using two coordinates to sort cells into inner, intermediate, outer, and pillar regions showed that the percentage of cells positive for HNF4a was statistically significantly higher at the outer region compared to each other region, including the pillar. Percentage of OPN positive cells was lowest at the outer region, with a small increase of OPN positive cells at the pillar. Overall, untreated toroid microtissues had a statistically significant increase in percentage of OPN positive cells per tissue compared to cylinder tissues. Treatment with EGF increased biliary differentiation in both toroid and cylinder microtissues, however it did not disrupt the spatial patterning of HNF4a positive cells. EGF treatment also amplified the signal pattern of OPN expression, leading to OPN frequency increases in the intermediate area, in the region just outside the outer shell and near the pillar contacting region.

The authors went on to further characterize the architecture of the tissues, by staining the tissues for actin and E-cadherin along with HNF4a. E-cadherin was found to be expressed at low levels throughout, but was highest at the outer shell in both cylinder and toroid geometries. In toroid microtissues, the regions of increased E-cadherin expression correlated with the regions of increased HNF4a positive cells. This was highest at cell-cell junctions.

Next, the authors further explored the relationship between geometry and differentiation, by preparing oblong microtissues, with widths of 100, 200, or 300µm, and the length chosen such that the cross-sectional area is equal to that of a 400µm diameter circle. In 300µm-wide oblong wells, the spatial patterns of actin, E-cadherin and HN4a expression was like those observed in cylinder microtissues. Conversely, microtissues in the 200 and 100µm wide oblong wells condense such that they are pressed against the side walls, producing morphologies with flat sides and rounded caps. In the oblong tissues, dividing the tissues into core, cap and flat regions, showed that hepatocytic differentiation was excluded from flat regions.

Investigating the role of actin-myosin contractility on differentiation in 3D showed that disruption of actin-myosin contraction reduced biliary differentiation in all geometries, suggesting an important role of cell contractility for biliary differentiation. Although the level of hepatocytic differentiation was not affected, spatial patterning was. This latter phenotype was pronounced in toroid microtissues.

Finally, to further understand the mechanical behaviour of the microtissues, the authors implemented a 3D finite element method (FEM) based model, building from previously reported 2D and 3D tissue models. The model suggests that in the cylindrical microtissue, the outer shell region was primarily under tension, while the intermediate and core regions were entirely in compression, with the largest compressive stress in the intermediate region. In the toroid microtissue model, the same was true, however, the pillar contacting region was experiencing compression at comparable levels to the intermediate zone. The authors validated the model predictions by embedding microgels containing fluorescent beads into the microtissues, to measure forces within the tissue.

The results of the models suggest that geometry combined with the contractile shell drives regions of surface compression, which correlate with reduced hepatic differentiation and reduced E-cadherin expression. The authors explored this further by fabricating ‘double pillar’ and ‘dumbbell’ shaped microtissues. For these, the model predicted that the presence of pillars causes the tissues to contract away from the walls, eliminating any compressed surfaces. This behaviour was confirmed experimentally.

What I like about this preprint

            I think the tools developed in this work are extremely valuable in the current aim of the scientific community to bridge our in vitro and in vivo knowledge on various topics, and to further reduce the need for animal experimentation. Further, I found this preprint exciting because of the plethora of methods used, and the interdisciplinary approach to the topic.

 

Open questions

1. During the formation of the liver in the embryo, where does the vascular endothelium of the vessels supplying the hepatocytes come into play? Do these cells also influence differentiation?
2. Is it know whether and how other stresses present in living organisms, such as sheer flow, influence differentiation?
3. You investigated the various phenotypes in toroid and cylindrical microtissues., and later in the oblong microtissues and the double pillar and dumbbell-shaped microtissues. Can you expand further on why these different geometries were chosen and what aspects they reproduce of the organ?
4. While your work focuses specifically on the liver, what would be your recommendation for studying microtissue geometry and cell-generated forces in 3D models of other organs?
5. You mention in your discussion and conclusion that understanding tissue geometry and its effect on for instance, differentiation, could help us address developmental issues or pathologies such as those observed in cancer. Is it known whether and how tissue geometry changes in different pathological conditions and how this affects native cell forces, their differentiation and distribution?

References

1. Berg IC et al, Microtissue geometry and cell-generated forces drive patterning of liver progenitor cell differentiation in 3D, bioRxiv, 2020.

 

doi: https://doi.org/10.1242/prelights.26383

Read preprint (No Ratings Yet)