(22h) DNA-Caged Polymer Nanocomposites for Erasable Fluorescence Imaging

Imaging pathology samples is often difficult because of limited colorimetric dye options and tissue autofluorescence resulting from structural complexity. If fluorescent labels are used, they are limited to color channels with minimal overlap with tissue autofluorescence, which is not ideal for image multiplexing. Sequential erasable labeling with fluorescent antibodies has been proposed as a potential solution, often with the use of harsh chemicals to remove the antibodies. These chemicals can damage tissue, which may lead to loss of structural information or misidentification of important structures. Newer processes featuring single stranded DNA (ssDNA) modified with fluorophores as reporters may address these limitations. DNA labels can be easily erased by strand displacement reactions. However, unprotected fluorophores are easily bleached and lack stability. Here, we report the creation of DNA caged-micelles (i.e., DNA cages) for use as an erasable fluorescent labeling system. DNA cages are DNA nanostructures (modified from Kurokawa et al.¹) that form a network on spherical surfaces (i.e., micelles). The proposed labeling system is based on gentle DNA dehybridization interactions and labels are protected by micelle coatings, which would greatly improve current technology. Additionally, the dyes used in cell and tissue labeling can include colorimetric, fluorescent, and quantum dot labels.

Proof of cage formation was measured using DNA tiles modified with fluorescent dyes and/or fluorescence quencher FRET pairs, as well as via TEM images. Three different fluorescent curves were generated based on interaction of tiles with the micelle surface and/or each other. Absorption of quenching DNA tiles to a fluorescent micelle surface indicated binding of DNA cages to the micelle surface, with complete quenching at a molar ratio of 13.5:1 (DNA:PEG (micelle)). This result was validated by TEM images showing formation of a new structure not observed in the absence of DNA or polymer (micelles). Additionally, cage saturation on the micelle surface was observed at a molar ratio of 18:1 and interlocking of the DNA tiles (based on a FRET quencher pair) was observed at a molar ratio of 9:1.

Erasable labeling was demonstrated in solution by binding the targeting strand on DNA cages to ssDNA labeled with fluorophores that was then erased in sequential cycles via strand displacement reactions. In solution, an erase depth of greater than 90% was achieved for multiple cycles. DNA cages were then loaded with fluorescent dyes and used as labels in fixed cells. U87 glioblastoma cells were labeled with primary antibodies targeting either integrin β1 (an extracellular target) or actin (an intracellular target). Then, secondary antibodies conjugated to ssDNA were used to bind primary antibodies. Finally, cages containing complementary ssDNA to that on the secondary antibodies were introduced for target labeling. Labeling of both targets was observed, consistent to that of gold standard controls. Next, DNA with greater complementarity to secondary antibodies was introduced, erasing DNA cages. An erase depth of up to 85% was achieved. DNA cages could also penetrate up to 100 μm in tissue sections with no damage observed over the course of treatment and washing. These results suggest that DNA cages show great promise for use in multiplexed cell labeling.

Reference

1 Kurokawa, C. et al. DNA cytoskeleton for stabilizing artificial cells. Proc Natl Acad Sci U S A 114, 7228-7233, doi:10.1073/pnas.1702208114 (2017).