Huntingtin CAG expansion impairs germ layer patterning in synthetic human 2D gastruloids through polarity defects

Abstract

Huntington's disease (HD) is a fatal neurodegenerative disorder caused by an expansion of the CAG repeats in the huntingtin gene (HTT). Although HD has been shown to have a developmental component, how early during human embryogenesis the HTT-CAG expansion can cause embryonic defects remains unknown. Here, we demonstrate a specific and highly reproducible CAG length-dependent phenotypic signature in a synthetic model for human gastrulation derived from human embryonic stem cells (hESCs). Specifically, we observed a reduction in the extension of the ectodermal compartment that is associated with enhanced activin signaling. Surprisingly, rather than a cell-autonomous effect, tracking the dynamics of TGFβ signaling demonstrated that HTT-CAG expansion perturbs the spatial restriction of activin response. This is due to defects in the apicobasal polarization in the context of the polarized epithelium of the 2D gastruloid, leading to ectopic subcellular localization of TGFβ receptors. This work refines the earliest developmental window for the prodromal phase of HD to the first 2 weeks of human development, as modeled by our 2D gastruloids.

Keywords: 2D gastruloids; Epithelial polarity; Human embryonic stem cells; Huntington's disease; TGFβ signaling.

Fig. 1.

CAG length-specific reduction of SOX2⁺ ectodermal lineage in 2D gastruloid differentiation. (A) Generation of isogenic hESC collection using CRISPR/Cas9 genome engineering to model HD. The first exon of HTT was targeted and exchanged for increasing CAG lengths to create an allelic series or was deleted to generate HTT^−/−. Three independent clones were isolated for each genotype. (B) Self-organization of CAG-expanded hESCs in geometrically confined micropatterns induced with BMP4 (50 ng/ml) in conditioned medium for 48 h. Concentric rings of ectodermal SOX2⁺ (blue), mesodermal BRA⁺ (red) and extra-embryonic tissue-like CDX2⁺ (green) domains probed by immunofluorescence (top). Dot plots (bottom) represent relative SOX2 intensity (red, high expression; blue, low expression) normalized to nuclear DAPI. Each dot corresponds to a single cell. (C) Family pictures of 20CAG control (top) and CAG-expanded 56CAG (bottom) 2D gastruloids. Micropatterns were induced with BMP4 and stained for SOX2 (blue) and BRA (red). (D) Analysis of mean SOX2 intensity across colonies with different CAG lengths using immunofluorescence data. Data are mean±s.d. of n>100 colonies for each genotype using n=2 or 3 clones. (E) Quantification of SOX2⁺ domain presented as area of total 2D gastruloid size (r=250 µm). Each dot corresponds to a single colony, plotted as mean±s.d. of n>100 colonies and n=2 or 3 clones/genotype. ***P<0.001, Kruskal–Wallis followed by Dunn's test. (F) BMP4-induced self-organization of HTT^−/− and HTT^+/+ is visualized by immunostaining against SOX2⁺ (blue), BRA⁺ (red) and CDX2⁺ (green) domains (top). Spatial representation of SOX2 expression analyzed at single-cell resolution (bottom). (G) Analysis of SOX2⁺ area based on immunofluorescence data and plotted as a percentage of total 2D gastruloid size in HTT^+/+ and HTT^−/−. Data are mean±s.d. of n>100 colonies obtained using n=2 or 3 clones for each genotype. n.s. P>0.05, Mann–Whitney. (H) Strategy for generating mCitrine-SOX2 reporter lines in 20CAG, 56CAG and 72CAG genetic background using CRISPR/Cas9 technology. Coding region of SOX2 (SOX2 CDS) was labeled with mCitrine fluorescent tag (mCit) and separated from the blasticidin resistance gene (Bsd) by a P2A sequence. (I) BMP4-induced 2D gastruloid differentiation from time-lapse imaging series using mCitrine-SOX2 live reporter cell lines. Images shown before BMP4 addition (T=0 h) and every 10 h after stimulation in 20CAG, 56CAG and 72CAG. (J) Analysis of SOX2 radius as a function of time elapsed in control and HD 2D gastruloids (n=6). Images were acquired every 30 min. Individual colonies originate from a single micropatterned chip for each clone. Similar results were obtained from n=3 independent experiments. Scale bars: 100 µm.

Fig. 2.

CAG expansion leads to enhanced late pSMAD1 and SMAD2/3 response. (A) Short-term BMP4 response (5 ng/ml) of 20CAG (left), 56CAG (middle) and 72CAG (right) in conditioned medium. Samples were fixed after 1 h of stimulation and immunostained for pSMAD1 (red); nuclei are visualized with DAPI (gray). (B) Mean radial pSMAD1 profile analyzed for the early response based on immunofluorescence data. Data are mean±s.d. of n>20 colonies/genotype. (C) Quantification of pSMAD1 activation from colony edge. The width of pSMAD1⁺ ring is determined based on radial intensity profile in early BMP4 signaling. Each data point corresponds to a single colony. Data are mean±s.d. of n>20 colonies/genotype. n.s. P>0.05, Kruskal–Wallis followed by Dunn's test. (D) Immunofluorescence data demonstrating activation of both TGFβ signaling branches after 24 h of BMP4 treatment (50 ng/ml) probed against the effectors pSMAD1 (red) and SMAD2/3 (green) in 20CAG (top), 56CAG (middle) and 72CAG (bottom). DNA is stained with DAPI (gray). (E) pSMAD1⁺ ring width in expanded CAG-length cell lines (56CAG and 72CAG) compared with 20CAG in late BMP4 signaling. Values are calculated from the radial intensity profile of individual colonies. Scatter plot of n=10 colonies/genotype. Data are mean±s.d. *P<0.05, **P<0.01, Kruskal–Wallis followed by Dunn's test. (F) Radial profiling of pSMAD1 intensity at 24 h. Data are mean±s.d. n>10 colonies/genotype. (G) 56CAG and 72CAG exhibit expanded nuclear SMAD2/3⁺ ring width compared with 20CAG. Width values of n=10 colonies/genotype. Data are mean±s.d. **P<0.01, ***P<0.001, Kruskal–Wallis followed by Dunn's test. Individual colonies originate from a single micropatterned chip for each clone. Similar results were obtained from n=3 independent experiments. Scale bars: 100 µm.

Fig. 3.

Unraveling the effect of HTT-CAG expansion on the signaling events governing 2D gastruloid patterning. (A) Schematic illustration of signaling hierarchy driving 2D gastruloid differentiation. BMP4 induces WNT, which in turn stimulates NODAL. Through a positive-feedback mechanism, BMP4 and NODAL also induce themselves. IWP2 is a pharmacological inhibitor of WNT secretion downstream of BMP4. CHIR 99021 is a chemical compound stimulating canonical WNT signaling cell-intrinsically. SB431542 selectively inhibits the activin/NODAL branch of TGFβ signaling. Activin acts through the same receptors and activates the same effectors as NODAL. (B) Dissection of 2D gastruloid signaling in 20CAG (top) and 56CAG (bottom). Three germ layers induced by BMP4 (50 ng/ml). IWP2 (2 µM) in combination with BMP4 (50 ng/ml) leads to abolished mesendoderm differentiation. In the absence of activin/NODAL signaling [blocked by SB431542 (10 µM)], WNT3a (100 ng/ml) induces mesoderm from the periphery. Activin (100 ng/ml) in combination with CHIR 99021 (6 µM) promotes endoderm differentiation in a radially symmetrical manner. Immunostaining was performed against SOX2 (blue), BRA (red), GATA3 (green) and SOX17 (gray) after 48 h of differentiation in conditioned medium. (C) The size of the central domain is determined as a relative change to the mean central domain area of 20CAG colonies as a percentage. Scatter plot represents normalized data points of n>250 colonies obtained from n=2 or 3 independent clones/genotype. Data are mean±s.d. in all respective conditions. Relative differences between the genotypes were statistically compared between each condition. n.s. P>0.05, ***P<0.001, ANOVA. (D) Fate acquisition in HTT^−/− compared with HTT^+/+ in CHIR 99021 and activin-induced micropatterns probed by immunofluorescence against SOX2 (blue), BRA (red) and SOX17 (gray). (E) Quantification of the percentage change in central domain area relative to the mean central domain area of HTT^+/+. Scatter plot represents normalized data points from n≥200 colonies/genotype. Data are mean±s.d. ***P<0.001, Mann–Whitney. Individual colonies were obtained from a single micropattern chip for each clone. Similar results were obtained from n=3 independent experiments in each condition. Scale bars: 100 µm.

Fig. 4.

HTT-CAG expansion selectively affects cell-autonomous SMAD2 signaling dynamics without altering transcriptional output. (A) BMP4 (5 ng/ml) response of RFP-SMAD1 reporter single hESCs obtained from live imaging data in 20CAG pre and 0.5, 1, 2, 4 and 8 h post-ligand presentation in TeSR-E7 medium lacking TGFβ ligands. (B) Quantification of nuclear RFP-SMAD1 signal intensity in 20CAG, 56CAG and untreated 20CAG control single cells over time. Mean intensity values±s.d. are plotted for n>200 cells at each time-point in both genotypes. Cells were induced with BMP4 at T=0 h and imaged every 10 min. Similar results were obtained from n=2 independent experiments. n.s., P>0.05, unpaired t-test. (C) Response to activin (1 ng/ml) of 20CAG (left) and 56CAG (right) mCitrine-SMAD2 reporter single hESCs in defined TeSR-E7 medium. Snapshots from live imaging shown prior to and 1, 2, 4 and 12 h following stimulation. (D) Quantitative analysis of nuclear-to-cytoplasmic ratio (N:C) of mCitrine-SMAD2 intensities in single cells as a function of time. Data are mean signal intensities±s.d. of n>200 cells at each time-point are shown. Activin was added at T=0 h and images were acquired every 10 min in n=3 independent experiments. *P=0.0164, unpaired t-test. (E) Time-course RT-qPCR analysis of key pluripotency associated genes induced by SMAD2 upon activin (10 ng/ml) stimulation in both 20CAG (gray) and 56CAG (red) genotypes. (F) Expression of the SMAD2 target mesendoderm differentiation gene EOMES for 12 h following activin treatment. (G) SMAD2 target mesendoderm differentiation genes induced for over 12 h when stimulated with activin in combination with WNT3a (100 ng/ml). RT-qPCR analyses were performed in single cells using n=3 or 4 independent clones/genotype in defined TeSR-E7 medium and are plotted as mean±s.d. Data points obtained from n=4 technical replicates for each clone were normalized to internal GAPDH expression, then to the pre-stimulus (T=0 h) levels of each gene in 20CAG. n.s., P>0.05, unpaired t-test. Similar results were obtained from n=2 or 3 independent experiments in each condition. Scale bars: 25 µm in A; 50 µm in C. a.u. arbitrary units.

Fig. 5.

HTT-CAG expansion and loss of HTT disrupt edge restriction in response to activin signaling. (A) Early nuclear SMAD2/3 (red) accumulation probed by immunofluorescence with antibody recognizing total protein levels in response to activin (100 ng/ml) in conditioned medium. DAPI (gray) was used as a DNA stain. Samples were analyzed 1 h after induction. (B) Mean radial intensity profile of nuclear SMAD2/3 in 20CAG, 56CAG, 72CAG and HTT^−/−. Error bars represent ±s.d. of n=15 colonies obtained from each genotype, all within a similar range of cell density (top). Mean nuclear SMAD2/3 intensity displayed in colony center versus colony edge in each genotype (bottom). ***P<0.001, Kruskal–Wallis followed by Dunn's test. (C) Temporal analysis of 20CAG and 56CAG mCitrine-SMAD2 reporter colonies before and 1, 2, 4 and 12 h after activin (10 ng/ml) treatment in defined TeSR-E7 medium. Cells were stimulated with activin at T=0 h. Dot plots represent mCitrine-SMAD2 intensity analyzed at a single-cell resolution. (D) Quantification of nuclear-to-cytoplasmic ratio (N:C) of mCitrine-SMAD2 intensities as a function of colony radius in 56CAG (red) compared with 20CAG (gray). Mean response±s.d. is calculated for n=6 (r=250 µm) colonies at 1, 2, 4 and 12 h after treatment with activin. n.s. P>0.05, **P<0.01, ***P<0.001, unpaired t-test. Individual colonies originate from a single micropatterned chip corresponding to each clone. Similar results were obtained from n=3 independent experiments. Scale bar: 100 µm. a.u. arbitrary units.

Fig. 6.

HTT-CAG expansion disrupts the polarized activin response by affecting receptor relocalization. (A) 20CAG, 56CAG and HTT^−/− hESCs grown on transwell inserts at high density are used to measure the apical versus basal BMP4 (10 ng/ml) response in conditioned medium. Immunofluorescence data for pSMAD1 (red) after ligand delivery to top and bottom compartments (top). Activin (10 ng/ml) induced nuclear SMAD2/3 (red) accumulation in 20CAG, 56CAG and HTT^−/− genotypes in defined TeSR-E6 medium probed by immunofluorescence (bottom). DNA visualized using DAPI (gray). Samples were analyzed 1 h after stimulation. (B) Percentage of pSMAD1⁺ cells (top) and nuclear SMAD2/3⁺ cells (bottom) across different HTT genotypes. Samples were analyzed at a single-cell resolution. The fractions of activated cells are presented as mean±s.d. of n≥5 images of apically and basally stimulated transwell filters. n.s. P>0.05, ***P<0.001, ANOVA. (C) Tight junction integrity was assessed in high-density hESC cultures on transwell filters by immunostaining against tight junction protein ZO-1 (cyan) at apical SMAD2/3 (red) activation sites in 20CAG, 56CAG and HTT^−/− followed by activin (10 ng/ml) stimulation in defined TeSR-E6 medium. Cell nuclei are visualized using DAPI (gray). Samples were analyzed 1 h after stimulation. (D) TEER was measured to determine epithelial integrity in 20CAG, 56CAG and HTT^−/−. Data are mean±s.d. of n=3 independent measurements. n.s. P>0.05, ***P=0.001, ANOVA. Similar results were obtained from n=3 independent experiments. Scale bars: 50 µm.

Fig. 7.

Ectopic SMAD2/3 signaling is mediated by ACVR2B receptor mislocalization. (A) Immunostaining of ZO-1 (gray) along with transiently expressed ACVR2B-HA (red, left) and BMPR2-HA (red, right) following doxycycline induction in 20CAG, 56CAG and HTT^−/−. DNA is visualized using DAPI (blue). (B) Distance to apical plane defined by ZO-1 and measured for DAPI and ACVR2B (left) or BMPR2 (right) in 20CAG, 56CAG and HTT^−/−. Histograms are represented as violin plots with each of their medians shown in black; dotted lines are 1st and 3rd quartiles. n.s. P>0.05, **P<0.01, ***P<0.001, Kruskal–Wallis followed by Dunn's test. (C) Immunostaining for pSMAD1 (magenta) and the H2B nuclear marker of doxycycline-induced NOG-expressing cells (green) after BMP4 treatment in 20CAG, 56CAG and HTT^−/−. DAPI (gray) is used for DNA staining. Samples were fixed 16 h post-induction. Double-headed arrows indicate the distance between NOG-expressing and pSMAD1⁺ cells. (D) pSMAD1 intensity (a.u.) measured proximal to NOG-secreting cells in 20CAG, 56CAG and HTT^−/−. Data are mean±s.d. for n≥15 NOG-secreting cells across n=8 images. (E) Schematic of the model by which CAG expansion disrupts polarity selective activin signaling and loss of HTT protein further compromises cellular polarity by impairing epithelial integrity in hESCs. Similar results were obtained from n=2 independent experiments. Scale bars: 10 µm in A; 100 µm in C. a.u.: arbitrary units.