A hydrogel-based platform for investigating biochemical and physical factors that impact on the fate of single stem cells is reported by investigators at the Ecole Polytechnique Fédérale de Lausanne. A team from the Laboratory of Stem Cell Bioengineering and Institute of Bioengineering has deeveloped an easily manufactured artificial niche microarray comprising wells that can be functionalized with combinations of proteins spotted by robotic technology in terms of modular stiffness.
Reporting in Nature Methods on their tests to validate the platform using both adherent and nonadherent mouse and human stem cells, Matthias P Lutolf, Ph.D., and colleagues claim their artificial niche microarray platform represents an efficient means to screen putative soluble or tethered signaling cues across microenvironments of modular stiffness. Their paper is titled “Artificial niche microarrays for probing single stem cell fate in high throughput.”
The fate of adult stem cells in vivo is governed by interactions between the cell and the complex microenvironment, or niche, in which it is sited. Effectors within these niches are critical to maintaining cell multipotency and quiescence, as well as triggering activation and directing the differentiation of stem cells into different cell lineages, the researchers report. Effectors include biochemical signaling cues from neighboring cells, soluble molecules, and the cross-linked extracellular matrix (ECM) in addition to biophysical cues such as niche elasticity and geometry.
However, our understanding of the molecular mechanisms of stem cell regulation by the niche is limited. Microarray-based platforms that involve spotting combinations of ECM molecules and growth factors on flat substrates have been developed in an attempt to investigate how stem cell niches influence stem cell fate in vitro. One of the drawbacks of this approach, though, is that the platforms are limited to adherent cells that can be confined to individual spots by cell adhesion, and as a result ECM microarrays aren’t ideal for tracking the dynamics of cell fate in single cells.
The drawbacks of microarrays can be partially overcome by using hydrogel microwell arrays, the researchers continue. These topographically structured cell culture surfaces contain microwells that allow either adherent or nonadherent cells to be trapped by gravity, and thus the number of cells per well more easily controlled. This approach has its own drawbacks, however, as in contrast to ECM microarrays, microwell arrays to date have offered little modularity and control over the biochemical composition of the substrate: microwell array platforms haven’t yet been interfaced with robotic spotting technology, and manual protein patterning approaches aren’t accurate enough yet.
Dr. Lutolf’s team aimed to combine the benefit of both microarrays and microwells and develop a single platform that could be used to study the roles of biochemical and biophysical niche factors on stem cell fate. Their design was based on the use of polyethylene glycol (PEG) hydrogels as substrates for the artificial niche microarrays. PEG hydrogels are relatively inert to protein adsorption and resist cell adhesion, the scientists state. In addition, such hydrogels can be chemically modified with proteins of interest, and a PEG hydrogel surface can be readily micropatterned to generate a microwell array.
The fabrication approach for the microwell array involved robotically spotting biomolecules directly onto an array of silicone pillars that would subsequently be used as the stamp for micropatterning the hydrogel. Proteins were printed on top of each individual 450 μm diameter pillar using a commercially available DNA spotter, and the printed stamp was pressed into a cross-linked layer of PEG hydrogel attached to the bottom of a four-well cell culture plate.
The indentations left when the array of silicon pillars was removed effectively represented the newly formed wells in the hydrogel. The stamping process also simultaneously transferred into the bottom of each well the biomolecules that had been spotted onto the pillars. Tethering the proteins onto the PEG gel surface was effected by use of either a nonspecific NHS-PEG-maleimide linker for larger ECM proteins or a site-selective approach based on covalent immobilization of protein A (or protein G) and subsequent affinity-based binding of chimeric proteins with an Fc tag. The depth of individual microwells was 35 μm ± 4 μm.
The overall approach enabled the large-scale production of hydrogel microarrays containing 2,016 microwells on a surface equivalent to a standard glass slide or using smaller multiwell tissue culture plates formats in which some 288 microwells could be generated in each culture well.
“The precision of robotic protein spotting and subsequent microcontact printing allowed us to freely choose the protein type, amount, and protein combination,” the researchers report. “In addition, protein patterning across all 2,016 microwells was free of any neighboring or repetition constraints, allowing the generation of overlapping gradients and random patterns equally well. The entire microarray production takes approximately two hours and can be easily parallelized.”
Having generated the hydrogel niche microarray, the researchers then used the platform to investigate different parameters on stem cell fate. They first investigated the effects of cell density on adipogenic differentiation of primary human MSCs. The cells were seeded at three different concentrations on niche microarrays comprising microwells that had been functionalized with cell-adhesive fibronectin fragment 9-10 (FN9-10).
The captured cells were then stimulated to differentiate. The results showed that the greater the cell density per microwell at the start of the experiment, the greater the resulting adipogenic differentiation. And, as expected, increasing adipogenic differentiation correlated with lower average cell surface areas and less proliferation.
The team postulated that the cell density effect on adipogenic differentiation might be caused by differential cell-cell interactions mediated by cadherins. To test this they functionalized microwell arrays with variable concentrations of Fc-tagged N-cadherin, and a constant concentration of FN9-10 required for cell adhesion. The results did indeed show that N-cadherin significantly increased adipogenic differentiation in a concentration-dependent manner across all microwells (that is, the differentiation responsiveness to N-cadherin was comparable for all initial densities) and also decreased the average cell surface area and proliferation.
Conversely, adding an N-cadherin-blocking antibody to the microwells efficiently reduced adipogenesis for all initial cell densities. “The extent of inhibition was a function of the amount of tethered N-cadherin,” the researchers note. “At the highest N-cadherin concentration we observed the strongest inhibition of adipogenic differentiation."
Prior studies have identified substrate stiffness as an important parameter in the determination of stem cell fate, both in terms of lineage specification and self-renewal. To investigate this parameter on the osteogenic differentiation of MSCs, the team altered the hydrogel PEG concentration to allow the fabrication of niche microarrays that demonstrated variable stiffness, at eight different FN9-10 concentrations. As predicted, increasing the elastic modulus of the substrate resulted in increased osteogenic differentiation of MSCs, but this was independent of FN9-10 concentration.
The researchers then used the platform to screen an array of candidate proteins involved in regulating self-renewal of single mouse neuronal stem cells. The multilineage differentiation potential of the neurosphere-forming cells was confirmed both on standard plastic culture dishes and on the microwell array, which was modified with adhesive protein layer to induce cell spreading rather than sphere formation.
Time lapse microscopy of live single NSCs showed that in the presence of soluble epidermal growth factor, about 67% of the cells in plain PEG microwells proliferated, produced neurospheres of various sizes, and expressed stem cell markers but not the βIII marker of neuronal differentiation. Adding the Notch ligands Jagged 1 and DLL4 induced the most extensive proliferation.
There was a discrepancy in the behavior of other proteins, however. Wnt5a induced less cells to proliferate, but those cells that did formed larger neurospheres. The opposite effect was induced by BMP6. “These effects could hint at unpredicted modes of action or a signaling influence on a subpopulation of cells,” the authors suggest.
Because the basal lamina protein laminin 1 and the Notch ligand Jagged1 have previously been identified as components of the NSC niche in vivo, the researchers then produced microwell arrays in which the two proteins were spotted at various concentrations, either singly or in combination. Either protein alone positively influenced NSC proliferation in a dose-dpendent matnner. But while intermediate concentrations of laminin 1 favored quiescence, Jagged 1 had no effect on the basal quiescence level.
Interestingly, spotting combinations of laminin 1 and Jagged 1 at nine different rations didn’t result in any additive effects, “suggesting that the presence of either protein sufficed for reaching maximal proliferation under the chosen culture conditions containing high concentrations of EGF,” the authors state.
“Unlike previously developed cellular microarray platforms or other high-throughput cell culture systems, our artificial niche microarrays enable simultaneous control over the key stem cell fate determinants: local cell density (and thus exact numbers of interacting cells), type and concentration of tethered and/or soluble biomolecules, and substrate stiffness,” the authors conclude.
They admit that as a static system, the artificial niche microarray isn’t suitable for the temporal manipulation of cellular microenvironments, the exposure of cells to changing growth factor regimens, or the continuous exchange culture medium. In addition, the slide format means that all the microwells are exposed to the same medium chamber and cellular communication through the medium can also take place.
Nevertheless, they state, “the platform should be valuable to explore the complexity of many cellular microenvironments. Owing to the flexibility in design and ease of handling, additional integration of cell-fate effectors such as those involving direct genetic manipulation of cells, for example, via viral technology, or the combined culture of various cell types can be envisioned.”