Optogenetic control of Wnt signaling models cell-intrinsic embryogenic patterning using 2D human pluripotent stem cell culture

ABSTRACT In embryonic stem cell (ESC) models for early development, spatially and temporally varying patterns of signaling and cell types emerge spontaneously. However, mechanistic insight into this dynamic self-organization is limited by a lack of methods for spatiotemporal control of signaling, and the relevance of signal dynamics and cell-to-cell variability to pattern emergence remains unknown. Here, we combine optogenetic stimulation, imaging and transcriptomic approaches to study self-organization of human ESCs (hESC) in two-dimensional (2D) culture. Morphogen dynamics were controlled via optogenetic activation of canonical Wnt/β-catenin signaling (optoWnt), which drove broad transcriptional changes and mesendoderm differentiation at high efficiency (>99% cells). When activated within cell subpopulations, optoWnt induced cell self-organization into distinct epithelial and mesenchymal domains, mediated by changes in cell migration, an epithelial to mesenchymal-like transition and TGFβ signaling. Furthermore, we demonstrate that such optogenetic control of cell subpopulations can be used to uncover signaling feedback mechanisms between neighboring cell types. These findings reveal that cell-to-cell variability in Wnt signaling is sufficient to generate tissue-scale patterning and establish a hESC model system for investigating feedback mechanisms relevant to early human embryogenesis.

diodes (LEDs) illuminate a tissue culture (TC) plate placed onto LAVA board. i) Image of LAVA boards kept inside a TC incubator. j) Immunostaining for LRP6 in optoWnt hESCs in the dark (left) and after 1 hr of 0.8 µW mm -2 illumination (right). Scale bar 25 µm. k) Representative images of immunostaining for bcatenin in optoWnt hESCs in the dark (left) and after 6 hr illumination (right). Nuclear outline from DAPI stain overlaid in blue. Scale bar 25 µm. l) Quantification of b-catenin nuclear intensity shown in (k). Graph shows pooled analysis of 14 fields of view per biological replicate (n = 3), each point representing a single cell. Unpaired two-samples Wilcoxon test (p < 10 -16 ). m) Schematic of eGFP knock-in strategy to make BRA/T reporter line using CRISPR-mediated modification of endogenous BRA/T locus. n) FACS analysis of optoWnt hESCs (mCherry+) modified with an eGFP reporter at the endogenous BRA/T gene locus ("BRA/T reporter"), kept in the dark (top) or illuminated for 24 hrs (bottom). Graph shows pooled data from 3 biological replicates, ~30,000 cells per condition. o) FACS quantification of percent eGFP-positive cells under indicated conditions, treated for 24 hrs, e.g. Wnt3a recombinant protein (250 ng/mL), CHIR (5 µM), or light stimulation (0.8 µW mm -2 ). ANOVA followed by Tukey test (p < 10 -12 ). Graph shows mean ± 1 s.d., n = 3 to 6 biological replicates.  Figure 1C. Graph shows analysis of biological replicates (rep 1-3) after 48 hrs illumination or CHIR99021 (5 µM) treatment. Each point represents a single cell. Threshold for BRA+ classification indicated with red arrow. b) Representative images of immunostaining for BRA in optoWnt iPSCs in the dark (top) or after 48 hrs illumination (bottom). Scale bar 100 µm. c) qPCR timecourse of T mRNA expression in optoWnt hESCs at indicated durations of illumination. Graph shows mean log fold change in mRNA expression (log2FC) relative to dark (0 hr) condition ± 1 S.E.M, n = 3 biological replicates. d-f) Quantification of mean SOX2 (d), OCT4 (e), and NANOG (f) nuclear intensity from immunostaining shown in Figure 1E. Graphs show analysis of biological replicates (rep 1-3) after 48 hrs illumination or CHIR99021 (5 µM) treatment. Each point represents a single cell. Threshold for SOX2+, OCT4+, or NANOG+ classification indicated with red arrow. g) Quantification of percent positive cells for pluripotency markers based on immunostaining shown in (d)-(f) and Figure 1E. Graph shows percent positive cells ± 1 s.d., n = 3 biological replicates. ANOVA followed by Tukey test.

Fig. S3. Validation of RNA-seq study of optoWnt-induced hESC differentiation a)
RNA-seq results of indicated genes that mark WT (eGFP+, Neomycin-resistant) and optoWnt (Cry-LRP6c-2A-mCherry+, Puromycin-resistant) cells. Graphs show mean expression (read count + 1) ± 1 s.d., n = 3 biological replicates. b) Top upregulated (red) and downregulated (grey) genes in illuminated optoWnt vs. unilluminated optoWnt hESCs. Graph shows mean log2 fold change for each indicated gene. c) Heat map of mRNA expression (read count + 1) of indicated lineage markers. Biological replicates displayed for each condition, with undetected genes (read count < 150) shown in grey. d) Heat map of mean log2 fold change in lineage markers normalized to WT expression level from qPCR validation of RNA-seq results with indicated lineage markers and comparison to CHIR (3 µM) treatment for 48 hrs. Heatmap shows mean fold change of n = 3 biological replicates. Scale bar 500 µm. c) Representative image of optoWnt/WT co-cultures seeded at lower cell density (20k cell cm -2 ) show single-cell scattering and morphology changes in optoWnt (mCh+) cells after 48 hrs illumination. Scale bar 100 µm. d) Representative fluorescence images of optoWnt/WT co-cultures after indicated durations of illumination, stained for OCT and BRA/T. Scale bar 100 µm. e) Confocal images of optoWnt monocultures (left panels) and WT monocultures (right panels) in the dark or after 48 hrs illumination, stained for OCT4 and BRA/T. OptoWnt cells labelled with mCh. Scale bar 100 µm, YZ and XZ axial cross-sections shown through indicated slices (white lines), 63 µm in height. f) Sample images of cell neighbor analysis in CellProfiler of illuminated (left panel) and dark (right panel) co-cultures. Nuclear outline of mCh+ nuclei (green) overlaid on mCh channel image (left column). DAPI channel image (right column) shown with overlay of mCh+ nuclei colored by mCh+ positive neighbor count. Scale bar 100 µm. g) Histogram of total cell neighbor counts (bottom) and mCh+ cell neighbor counts (top) across all analyzed cells show comparable cell densities between light and dark conditions.  a) Representative brightfield images of optoWnt/WT co-cultures after 48 hrs illumination in E8, E6, and APEL2 media. Scale bar 250 µm. b) Confocal images of optoWnt/WT co-cultures in indicated media conditions, stained for OCT4 and BRA/T. OptoWnt cells labelled with mCh. Scale bar 100 µm. c) Cell neighbor analysis of optoWnt (mCh+) cells in co-culture. Graph shows count of total cell neighbors vs. count of mCh+ cell neighbors across total population of analyzed mCh+ cells (25,742 cells analyzed, pooled analysis from n=3 biological replicates for each condition). Area and color of points is proportional to the fraction of total population. Constant ratios of mCh+ to total neighbors are highlighted with dotted lines as indicated. d) Quantification of fraction of optoWnt (mCh+) cells whose neighbors are all mCh+. Each point represents an analyzed field of view (16 fields of view analyzed per condition, n = 3 biological replicates). Unpaired two-samples Wilcoxon test (pE8 = 0.011 ; pE6 = 3.3 x 10 -4 ; pAPEL2 = 3.9 x 10 -4 ). e) Stitched images of optoWnt/WT co-cultures in APEL2 media show large-scale pattern of cell selforganization. Scale bar 500 µm. Development: doi:10.1242/dev.201386: Supplementary information Fig. S7. EMT and increased cell proliferation upon optoWnt stimulation a-b) qPCR for EMT markers in WT and optoWnt hESC monocultures cultured in APEL2 media. Graphs show heatmap of mean log2 fold change (a) and mean fold change (b) over WT hESCs ± 1 s.d., n = 3 biological replicates. ANOVA followed by Tukey test. c) Representative images of immunostaining for Ecadherin (ECAD) and N-cadherin (NCAD) in unilluminated (top) and illuminated (bottom) optoWnt cells. Scale bar 100 µm. d) No observed clonal expansion of optoWnt cells when optoWnt-mVenus-NLS cells were dosed into optoWnt-mCh/WT co-cultures. Scale bar 100 µm. e-f) EdU stain of optoWnt monocultures after indicated illumination duration, with representative images of EdU staining (e) and quantification of percent EdU+ cells (f). Graph shows mean ± 1 s.d., n=3 replicates. ANOVA followed by Tukey test. Scale bar 100 µm.

Fig. S8. Characterization of optoWnt-mVenus-NLS hESC line. a)
Schematic diagram of knock-in strategy at the AAVS1 locus. b) Representative images of immunostaining for OCT4 and mVenus in optoWnt-mVenus-NLS hESCs in routine cell culture, kept in the dark. Scale bar 100 µm. c) Light-induced BRA expression of optoWnt-mVenus-NLS line is comparable to optoWnt-mCherry line. Scale bars 100 µm.

Fig. S9. Single-cell tracking of optoWnt cells in optoWnt/WT co-cultures shows increased cell migration and no change in cell persistence a) Representative images of single-cell trajectories of mVenus-NLS+ optoWnt cells. Track color distinguishes different cells in field of view. b) Average cell velocities at indicated time intervals after onset
of light stimulation. Graph shows mean (>1,000 tracks over 5 fields of view) ± 95% confidence interval. c) Average cell speed at indicated time intervals after onset of light stimulation. Graph shows mean (>1,000 tracks over 5 fields of view) ± 95% confidence interval. d) Average ratio of cell displacement over distance at indicated time intervals after onset of light stimulation. Graph shows mean (>1,000 tracks over 5 fields of view) ± 95% confidence interval. e) Distribution of median cell velocities across all cell tracks. Tracks were binned into three equal groups by median velocity (low: 0-0.15 µm/min; medium: 0.15-0.18 µm/min; high: 0.18-0.6 µm/min) as indicated. f) Number of tracks in each median velocity bin over time. g) Tracks in each velocity bin over time, shown as percentage of total tracks. h) Average cell velocity over time, binned by median cell velocity (low, medium, high). Graph shows mean ± 95% confidence interval. i) Average percent of cells undergoing persistent migration at indicated time intervals. Graph shows mean percentage ± 95% confidence interval. j) Average percent of cells undergoing persistent migration over time, binned by median cell velocity. Graph shows mean percentage ± 95% confidence interval.

Fig. S10. Gene knockdown of EMT regulator SNAI1 in optoWnt cells shows decreased cell selforganization in co-cultures a)
Western blot (WB) image (left) for SLUG protein levels in indicated WT and optoWnt cell lines in response to shRNA knockdown. Cell lines labelled +FACS designates that lines were sorted for GFP+ expression, which marks cells infected with shRNA construct. WB quantification (right), normalized to b-actin loading control. The optoWnt line expressing shRNA snai2 #2 (+FACS) was used for subsequent experiments. b) Western blot image (left) for SNAIL protein levels in indicated WT and optoWnt cell lines in response to shRNA knockdown. WB quantification (right), normalized to b-actin loading control. shRNA snai1 #1 (+FACS) was used for subsequent experiments. c) Representative images of optoWnt/WT co-cultures (left column) in the dark and after 48 hrs illumination. OptoWnt cells express scrambled shRNA (shSCRAM, top panel) or SNAI1 shRNA (shSNAI1, bottom panel). mCherry fluorescence marks optoWnt cells, while GFP nuclear fluorescence marks WT cells. Sample images from cell neighbor analysis in CellProfiler (right column) show nuclear outline of mCh+ cells (green) overlaid on mCh channel image. Scale bar 100 µm. d) Cell neighbor analysis of optoWnt (mCh+) cells in optoWnt-shSCRAM/WT or optoWnt-shSNAI1/WT cocultures kept the dark or illuminated for 48 hrs. Graph shows the count of total cell neighbors vs. count of mCh+ cell neighbors across total population of analyzed mCh+ cells (72,338 cells analyzed, pooled analysis from n=3 biological replicates). Area and color of points is proportional to the fraction of total population. Constant ratios of mCh+ to total neighbors are highlighted with dotted lines as indicated. e) Quantification of fraction of optoWnt (mCh+) cells whose neighbors are all mCh+ in optoWnt-shSCRAM/WT and optoWnt-shSNAI1/WT co-cultures. Each point represents an analyzed field of view (72 fields of view analyzed, n=3 biological replicates). Unpaired two-samples Wilcoxon test (p = 0.0033). f) Histogram of total cell neighbor counts (right) and mCh+ cell neighbor counts (left) across all analyzed cells show comparable cell densities between light and dark conditions, as well as between optoWnt-SCRAM/WT and optoWnt-SNAI1/WT cocultures. Gating strategy for sorted cells collected for RNA seq analysis. c) RNA-seq results of indicated proliferation markers in WT (monoculture, illuminated) and WT (co-culture, illuminated) cells. Graphs show mean expression (read count + 1) ± 1 s.d., n = 3 biological replicates. d) KEGG enrichment analysis for TGF-b signaling pathway components in WT cocult + light over WT light (monoculture) cells. Each gene of TGF-b pathway is color-coded by its enrichment score, n = 3 biological replicates. e) Gene set enrichment analysis (GSEA) using the TGF-b hallmark gene set in WT cocult + light over WT light (monoculture) cells, n = 3 biological replicates. f) Representative immunofluorescence images of optoWnt/WT co-cultures and that are untreated or under SB431542 inhibitor treatment and stained for proliferation marker Ki67. Scale bar 50 µm. g) Cell count of WT cells in optoWnt/WT co-cultures normalized per unit area. WT cells are defined as mCherry-negative and cell count is quantified from fluorescence imaging data. Each point represents a field of view, and graph shows mean ± 1 s.d. (one-way ANOVA, p = 0.941).