Digital Image analysis of 3D structures for Salivary Tissue Engineering

Abstract

Objectives: In vitro tissue engineering methods generate three-dimensional (3D) multicellular structures, within complex support scaffolds, that require new imaging methods for assessment and quantification. In the present work, we applied 3D digital image analysis to assess two related elements: morphological characteristics of multicellular salivary cell clusters, and interior pore morphology of biomanufactured hydrogel systems, for future combined use in salivary tissue engineering.

Experimental Methods: Imaris 9.2.1 was used to examine 3D confocal microscopy image stacks. For multicellular cluster analysis, single cells were clustered, encapsulated in hydrogel, stained with live/dead reagents, and imaged. ImarisCell was used to quantify the number of cells per cluster, and multicellular cluster size. Cluster diameters were approximated via either bounding-box or volume-based methods. Separately, hydrogel scaffolds were “carved” by multiphoton-based subtractive ablation, using pre-defined templates. Brightfield microscopy and FITC-dextran staining identified internal scaffold morphology. Image data were compared for fidelity against ablation templates. 3D renders were built in Imaris by examining multiple slices.

Results: The number of nuclei per cluster and diameter-adjacent measures were quantified. These demonstrated an exponential relationship, indicating changes in cell packing as cluster size increased. Non-clustered single cells were also observed in high numbers. Image stacks for carved scaffolds maintained appearances consistent with their templates, but digital reconstructions yielded overly rounded corners. FITC-Dextran produced better contrast and detail than brightfield images.

Conclusion: Many consistent clusters were achieved, with a predictable diameter relation. However, further iteration is needed to minimize non-clustered single cells. Carved scaffold analysis identified very good qualitative template preservation, but further staining and imaging revisions are needed to improve imaging accuracy. For both cells and scaffolds, 3D digital imaging enabled major improvements in feature quantification. Future work will iterate on these models to enhance both methods.

This study was supported by the UTSD Student Research Program, and NIH/NIDCR grant R03DE028988.