Research ArticleAPPLIED SCIENCES AND ENGINEERING

Recapitulating complex biological signaling environments using a multiplexed, DNA-patterning approach

See allHide authors and affiliations

Science Advances  18 Mar 2020:
Vol. 6, no. 12, eaay5696
DOI: 10.1126/sciadv.aay5696
  • Fig. 1 High-resolution surface-DNA patterning using photolithography.

    (A) Multicomponent patterns of unique 20-bp oligonucleotides instruct the spatial organization of cells and ligands through the hybridization between surface-presented oligonucleotides and complementary oligonucleotide-labeled biological components. (B) Surface DNA patterns are fabricated through the successive utilization of, first, photolithography to define regions of reactive aldehyde groups for oligonucleotide conjugation (step 1) and, second, a reductive amination step to covalently react the amine-terminated oligonucleotides to the aldehyde-functionalized glass surface (step 2). Multicomponent DNA patterns are assembled by patterning a new layer of positive photoresist and repeating steps 1 and 2 using unique oligonucleotides. (C) The use of photolithography enables the fabrication of high-resolution, spatially complex, DNA patterns across a range of length scales—from micrometers to millimeters. Patterned photoresist is used as a mask to conjugate selectively amine-terminated oligonucleotides, which can be visualized by hybridizing a complementary fluorescent oligonucleotide. A surface DNA concentration of 20 μM was used. (D) A dynamic range of surface-DNA pattern intensities (left) can be achieved by tuning the DNA solution concentration. Representative fluorescence intensity profiles (right, top) and their corresponding images of a 40-μm DNA spot array (right, bottom) are illustrated for a low (0.5 μM), medium (5 μM), and high (50 μM) DNA concentration. Error bars are SDs and n = 3. Scale bar, 100 μm. a.u., arbitrary units.

  • Fig. 2 Microfabricated DNA patterns direct the capture of NSCs.

    (A) Patterned surface oligonucleotides organize a fluorescently labeled population of NSCs with high spatial precision through Watson-Crick base pairing between the surface-conjugated DNA and the temporary lipid-modified DNA tethered to the cell membranes. (B) Cell capture efficiency of 20-μm-diameter DNA spot patterns was dependent on the concentration of DNA solution with a significant drop of efficiency occurring at a concentration below 5 μM (left). This is seen in the representative NSC-patterned images of a 10 × 10 array of 20-μm-diameter DNA spots for a range of concentrations (right). (C) The number of DNA-captured cells can be controlled by tuning the feature size of the DNA patterns. Representative images (left) demonstrate DNA spots with different diameter dimensions capturing varying numbers of fluorescently labeled NSCs. Moreover, an increase in diameter size of DNA-patterned spots (right) results in an increase in cell capture number (blue dots) that follows a similar increasing trend in spot area (yellow line). Unless noted otherwise, a surface DNA concentration of 20 μM was used for all experiments. All error bars are SDs, and n values are reported in table S2. Scale bars, 500 μm.

  • Fig. 3 Scalable, multicomponent DNA patterns organize heterogeneous cell populations.

    Characterization of multiple fabrication steps highlights the compatibility of photolithography with DNA patterning. (A) The integrity of surface DNA patterns is preserved—as indicated by the ability to hybridize with its complementary, fluorescent oligo counterpart—when subjected to repeated photolithographic fabrication steps, as would occur when patterning multiple DNA layers [i.e., removal of photoresist (PR) with acetone and patterning of a new layer]. Despite a slight initial drop upon the application of a second PR layer, the average fluorescence intensity of DNA-patterned features remains robust upon a third and fourth photolithography step. Curved black arrow indicates surface patterning of amine-terminated DNA oligonucleotides. (B) The functionality of the surface-modified aldehyde groups, which is necessary for DNA conjugation, is also preserved during successive PR layer applications. Additional photolithography steps yield surface DNA patterns with robust fluorescent intensities. Scale bars, 100 μm. All error bars are SDs and n = 3. Curved black arrows indicate surface patterning of amine-terminated DNA oligonucleotides. (C) (a) Micrometer-scale registration of three complex DNA patterns was patterned and visualized with unique complementary fluorescent oligonucleotides. Curved black arrow indicates the addition of fluorescent oligonucleotides and subsequent hybridization with the surface-patterned DNA strands. (b) To highlight their functionality, multicomponent DNA patterns assembled three distinct, fluorescently tagged NSC populations with high spatial control and specificity by labeling each population with unique complementary, lipid-modified oligos that insert into the cell membrane. (D) DNA surface patterns direct the parallel assembly of four unique cellular components to construct an in vitro breast cancer microenvironment that mimics early metastasis. An inner circular pattern of MCF-7s represents the primary tumor and is encased within an outer circular layer of nonmalignant breast epithelial MCF-10As. Vasculature is represented by HUVEC patterns, structured as vessel cross sections. White arrows highlight the varied circular pattern diameter of the final invasive cell type, MDA-MB-231, representing clusters of cancer cells that have escaped the primary tumor site and corresponding to different potential scenarios of the initial stages of metastasis. White dashed box (top right) corresponds with the day 0 and 1 zoom-in images (bottom left) of the HUVEC patterns embedded within the MCF-10A outer-patterned layer. A surface DNA concentration of 20 μM was used for all experiments. Scale bars, 500 μm.

  • Fig. 4 Microfabricated DNA patterns direct the spatial organization of solid-phase ligands.

    (A) The heterobifunctional linker DBCO-PEG4-maleimide enables covalent labeling of ligands of interest with an oligonucleotide label. A free sulfhydryl group on the protein is reacted first with the maleimide moiety on the cross-linker, introducing a DBCO functional group to the ligand that then reacts via click chemistry to an azide-terminated oligonucleotide. SH, thiol group. (B) The incorporation of an oligonucleotide label having a fluorescent tag enables imaging and monitoring of DNA-directed ligand patterns. For proof of concept, eGFP with a Cy5 tag was assembled using DNA surface patterns (top). Trends in fluorescence intensity profiles for the patterned protein and the fluorescent tag closely matched one other when tuning surface DNA concentrations (2, 4, and 10 μM, from left to right), suggesting that the fluorescent oligonucleotide label can also be used as a relative readout of patterned protein concentration (bottom). (C) Multicomponent DNA surface patterns enable tunable control over each ligand concentration as evident in the DNA assembly of eGFP and mCherry. mCherry concentration was held constant as eGFP concentration was tuned as quantified by the change in eGFP fluorescence intensity (bottom) and visualized in the fluorescent composite images (top). Surface DNA concentration for mCherry was 20 μM. Surface DNA concentration for eGFP was 1, 2, and 20 μM (left to right). Scale bars, 500 μm.

  • Fig. 5 Multicomponent DNA patterns enable tight spatial control and investigation of the presentation of competing ligand cues, FGF-2 and ephrin-B2, on single-NSC behavior.

    (A) Overview of two four-layer DNA patterning schemes that direct the assembly of FGF-2, ephrin-B2 mimetic peptide, single NSCs, and PA patterns (left). (a) The top PR patterns segregate each ligand to one-half of the microisland, exposing the patterned single NSC to both solid-phase cues equally, while (b) the bottom PR patterns forces the presentation of one ligand over the other. A representative image (right) of a large-area microisland array contains both presentation strategies; the patterned ligands are visualized by their respective fluorescent oligonucleotide labels (FGF-2 in cyan and ephrin-B2 in magenta). The zoomed-in insert highlights the simultaneous assembly of three different spatial presentation configurations: half/half, FGF-2 center, and ephrin-B2 center. (B) Representative time-lapse images illustrating cell proliferation and migration for three sample microislands of the different ligand spatial configurations and their corresponding day 5 immunostaining results: (a) FGF-2 center, (b) ephrin-B2 center, (c) half/half. DAPI, 4′,6-diamidino-2-phenylindole. (C) Quantification of cell body counts within FGF-2 (cyan) and ephrin-B2 (magenta) patterns over 4-day time lapse using custom analysis script, corresponding to the same three sample microislands in (B). Black arrows indicate proliferation events, and red arrows indicate cell death. A surface DNA concentration of 20 μM was used for all patterned components. Scale bars, 100 μm.

  • Fig. 6 Cell occupancy in response to various ligand presentations of FGF-2 and ephrin-B2 and resulting end fate after 5-day differentiation.

    (A) Average cell occupancy of cell bodies within the FGF-2 (cyan), ephrin-B2 (magenta), and spanning both (gray) protein-patterned regions was tracked over time for each of the three ligand spatial presentations: (a) FGF-2 center, (b) ephrin-B2 center, and (c) half/half. A strong spatial bias toward FGF-2 was observed. (B) Analysis of end fate through quantification of proliferation (top) and neuronal differentiation (bottom) reveals that, despite spatial preference toward FGF-2, some NSC microislands integrated both signals, generating significant heterogeneity. (C) Time-lapse snapshots (left) of a microisland exhibiting 100% neuronal differentiation (right) despite having near 100% FGF-2 occupancy throughout the 4-day culture reveal dynamic neurite processes (indicated by white arrows) occupying both FGF-2 and ephrin-B2 patterns. White, dashed boxes correspond to below zoom-in images. n = 55 for each ligand presentation. All P values were obtained from Tukey-Kramer test. ***P < 0.001. N.S., not significant. Scale bars, 100 μm.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/6/12/eaay5696/DC1

    Supplementary Materials and Methods

    Fig. S1. Characterization and optimization of DNA patterning steps.

    Fig. S2. Re-use of PR layer versus new PR layer for multicomponent DNA patterning.

    Fig. S3. Tunable multiplexed surface DNA patterns.

    Fig. S4. Stability and specificity of DNA-directed eGFP patterns.

    Fig. S5. Optimization of PA patterning using photolithography.

    Fig. S6. Microfabricated DNA and PA patterns support high-throughput clonal analysis of adult NSCs.

    Fig. S7. Characterization of adult NSC viability upon labeling with lipid-oligonucleotides.

    Fig. S8. Characterization of adult NSC fitness upon labeling with lipid-oligonucleotides.

    Fig. S9. Spatial precision imparted by photolithography provides tight control over heterogeneous intercellular communication.

    Fig. S10. DNA-based assembly of HUVEC patterns.

    Fig. S11. Multicomponent DNA patterns enable controlled, high-throughput studies of adult NSC niche solid-phase ligand cues, FGF-2 and ephrin-B2, at the single-cell level.

    Fig. S12. Soluble versus solid-phase peptide activity in adult NSCs.

    Fig. S13. Custom cell tracking pipeline using ilastik and Fiji.

    Fig. S14. Tracking changes in average cell occupancy within FGF-2 over time for each individual microisland.

    Table S1. Overview of surface-patterned DNA sequences and their complementary fluorescent, cell-labeling and ligand-labeling oligonucleotides.

    Table S2. In-depth report of experimental sample number “n”.

    Movie S1. Example time-lapse videos of single adult NSC cultures with various spatial organizations of competing solid-phase niche ligands.

  • Supplementary Materials

    The PDF file includes:

    • Supplementary Materials and Methods
    • Fig. S1. Characterization and optimization of DNA patterning steps.
    • Fig. S2. Re-use of PR layer versus new PR layer for multicomponent DNA patterning.
    • Fig. S3. Tunable multiplexed surface DNA patterns.
    • Fig. S4. Stability and specificity of DNA-directed eGFP patterns.
    • Fig. S5. Optimization of PA patterning using photolithography.
    • Fig. S6. Microfabricated DNA and PA patterns support high-throughput clonal analysis of adult NSCs.
    • Fig. S7. Characterization of adult NSC viability upon labeling with lipid-oligonucleotides.
    • Fig. S8. Characterization of adult NSC fitness upon labeling with lipid-oligonucleotides.
    • Fig. S9. Spatial precision imparted by photolithography provides tight control over heterogeneous intercellular communication.
    • Fig. S10. DNA-based assembly of HUVEC patterns.
    • Fig. S11. Multicomponent DNA patterns enable controlled, high-throughput studies of adult NSC niche solid-phase ligand cues, FGF-2 and ephrin-B2, at the single-cell level.
    • Fig. S12. Soluble versus solid-phase peptide activity in adult NSCs.
    • Fig. S13. Custom cell tracking pipeline using ilastik and Fiji.
    • Fig. S14. Tracking changes in average cell occupancy within FGF-2 over time for each individual microisland.
    • Table S1. Overview of surface-patterned DNA sequences and their complementary fluorescent, cell-labeling and ligand-labeling oligonucleotides.
    • Table S2. In-depth report of experimental sample number “n”.
    • Legend for movie S1

    Download PDF

    Other Supplementary Material for this manuscript includes the following:

    • Movie S1 (.mov format). Example time-lapse videos of single adult NSC cultures with various spatial organizations of competing solid-phase niche ligands.

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article