3D niche microarrays for systems-level analyses of cell fate

Reference Link

This paper presents a system for constructing 3D microarrays for simulating cell microenvironments to study cell fate and 3D cell-matrix interactions. This is somewhat novel it seems due to their precision, which could only be achieved in 2D before this paper (according to them). Upon further examination, a major contribution of this paper seems to be the automated methodology for analyzing cell interactions this “niche microarray” and statistical analysis method is actually really cool and has a lot of potential for making scientific research more agile.

Paper Notes:


  • 3D cell culture environments resemble in vivo environments more closely, but they typically have poor composition and batch-to-batch variation.
  • Spotting robots to microarray polymeric cell culture substrates with modular physical properties such as surface roughness and wettability has helped in studying cell fate, but generally limited to 2D
  • Present a method for doing this in 3D


Figure 1:

  • For 3D microenvironments need molecular building blocks which can be mixed and cross-linked to form
  • Engineered synthetic hydrogels as biomimetic 3D cell microenvironments with very well-defined physicochemical and biochemical properties.
    • Employ coagulation enzyme-activated transglutaminase factor XIIIa (FXIIIa) to cross-link branched poly(ethylene glycol)-(PEG)- based macromers into 3D hydrogel networks (Fig. 1a)
    • “derivatization of the PEG macromers with short peptidic substrates for FXIIIa allows site-specific enzymatically mediated amide bond formation between the PEG chain termini under physiological conditions”
    • “single mESC encapsulated within these PEG-based hydrogels show a very good viability (89.1±7.3% s.d.) that is not significantly different (P 1⁄4 0.44, two- sample unpaired t-test, unequal variance, two-tailed distribution) from standard culture conditions on gelatin-coated plastic (92.5±3.6% s.d.)”
      • Note, I’m not really sure what a whole lot of this means, I should look it up.
    • Cells are physically separated from each other in the gel-capsules
    • “To allow for efficient growth in 3D, the gels can be rendered susceptible to cell-secreted proteolytic remodelling by designing PEG–peptide conjugates bearing an additional protease substrate site for gel degradation (Fig. 1a)”
    • Biologically active molecules (oligopeptides or proteins) can be tether to gel networks during the cross-linking stage (Fig. 1a) and thus allows varied 3D microenvironments

Experimental setup

  • To demenstrate this system, varied 5 microenvironment signaling types (Fig. 1b)
    • matrix mechanical properties (MP)
    • proteolytic degradability (DG)
    • Extracellular matrix proteins (EC)
    • cell-cell interaction (CC) proteins
    • soluble factor (SF)
  • Young’s moduli (E) of the gels were specified between ca. 300 and 5400 Pa by varying the polymer content (Fig. 1b)
  • Gels were rendered differentially degradable by matrix metalloproteinases via the incorporation of peptides of different sensitivities ($k_cat / K_m $) to cell-secreted MMPs.
    • “adjusted the precursor content of gels having different susceptibilities to MMP-mediated degradation in order to perfectly match their mechanical properties”
      • This allowed independent control of key mechanical and biochemical properties of this synthetic gel system.
    • To modulate cell signaling properties of the matrices
      • Chose set of ECM and recombinant growth factor proteins
        • These are implicated in regulating ECS pluripotency
      • Enzymatically tethered to the gels (Fig. 1b).
      • While ESC interaction with laminin, fibronecting, and collagen IV had previously been associated with a loss of pluripotency in 2D, ligation of their integrin homdimers by peptide analgoues in a 3D culture system indicated that they could instead promote maintenance of self-renewal
        • I’m assuming this means that you can do something to make it helpful in 3D where it is counter-productive in 2D
      • “added the soluble ESC regulatory factors leukemia inhibitory factor (LIF), bone morphogenetic protein 4 (BMP4) and fibroblast growth factor 4 (FGF4)14,20 to a serum-free medium formulation such that they could reach the encapsulated cells via the diffusive properties of the gel network”
      • Specified 4 levels of modulation for each of the 5 categories
    • Currently this level of complexity can only be reduced to practice by miniaturizing the sample volumes and adapting automated methods for gel synthesis (Fig. 1c) and cell fate analysis (Fig. 1d-f)
      • Robot synthesized the stuff in an automated manner
      • All wells were isolated physically from its neighbors
  • To obtain quantitative information on ESC self-renewal and differentiation in response to various 3D environments
    • Used Oct4-GFP reporter cell line, in conjunction with automated imaging and image analysis (Fig. 1e,f).
      • The transcription factor Oct4 is widely considered as a marker of embryonic stem cell pluripotency (I’m assuming hence why it was used)
    • Cells embedded in PEG hydrogels were imaged the day after seeding to get initial counts.
      • In the most permissive conditions, formed spherical colonies within 3 days and kept proliferating until fixation 5 days after gel embedding (Fig. 1d)
        • High-resolution confocal microscopy confirmed that colonies growing in these conditions were in a 3D space of B500mm in thickness (Fig. 1e)
        • soluble molecules in the medium could rapidly diffuse homogeneously throughout the thickness of the gel
    • Automated analysis procedure
      • colony area as a measure of proliferation, and GFP intensity as a measure of ESC pluripotency (Fig. 1f)

Fig. 2

  • For every combination of the 5 environmental characteristics, GFP intensity and colony area were averaged across 3 replicates and organized by input conditions (Fig. 2a)
    • soluble factor modulation was found to be the strongest predictor of heat map intensity, with the LIF condition clearly leading to both high proliferation and self-renewal, with the opposite effect for FGF4 and BMP4, and an intermediate regime appearing for conditions with no soluble factors
    • The top ten self- renewing and proliferating conditions were all within the LIF condition, and most tended to be in conditions of low mechanical properties and absence of cell–cell interaction proteins.
    • LIF is known to be a potent mediator of ESC self-renewal via phosphorylation of the transcription factor STAT3
      • STAT3 is a signal transducer and activator of transcription 3
    • Thus this validates the model? somehow?
    • distinct area versus GFP relationships emerged out of such a representation
      • suggesting that the correlation between self- renewal and proliferation is linked to a particular soluble factor regime
    • “Interestingly, the most self-renewing and proliferating (i.e., high GFP and high area) matrix characteristics in all subpopulations were found in conditions of low matrix stiffness (Fig. 2c,d).”


Figure 3:

  • used generalized linear models (GLMs) to quantify the self-renewal data (Fig. 3)
  • This allowed explanation for more than 70% of the variability in the system
    • Decoupling of subtle effects from the more dominant ones allowed quantification of the various factors (Fig 3a)
    • also investigated the role of individual factors within these categories (Fig. 3b).
  • Degradable matrices favoured self-renewal, but not necessarily proliferation. Thus, the physical parameters of the matrix may not only dictate colony growth but also coordinate stemness.
  • Synergistic or antagonistic effects between factors that were statistically significant were represented by a network interaction map (Fig. 3c) and as clustered heat maps showing significant pair-wise interactions (Fig. 3d–g).
  • There’s more, but basically they found various interactions that you can see in the figures and paper.

Figure 4:

And Now, This

Written on October 6, 2014