Identification of a family with HD

A. The pedigree of the family with HD. B. PCR results showed that 8 of 14 are heterogeneous; the upper band is the expanded CAGs (upper panel). Sanger sequencing revealed that the heterogeneous segment contains expanded CAG repeats (lower panel; 49 repeats; red arrows indicate the beginning and end of the CAG repeats). (See also Table S1). M, control band (CAG 35); L, Healthy local person. C. The correlations of CAG repeat with disease onset age (Spearman’s correlation).

Characteristics of polyQ assembly and polyQ nuclear deposition in neurons

A. DIC images and 3D rendered tomography showing longer polyQ assemblies form a gorge in the cytoplasm or nuclear surface of induced neurons cultured for 75 days (CAG 59 and CAG 19) (pink arrow [i, ii], the gorges; white arrow, nuclear surface dents visualized by IMARIS surface). Inner panels and 3D rendered tomography displayed the spatial relationships. B. PolyQ deposits into the nucleus and pokes the nuclear membrane of striatal neurons (white arrow, the morphology of the nucleus; yellow arrows, nucleus poked by polyQ assemblies). The lower left in right panel is an apoptotic cell (CAG, 55). C. PolyQ precipitates occupied the nucleus of GABAergic neurons cultured for 178 days (CAG, 47). D, E. HD patient striatal neurons derived from iPSCs cultured for 90 days (2D culture) (CAG 59) and 98 days (CAG 55) and 110 days (CAG 47) have spliced nucleus (white arrows, spliced nucleus). The summary table of nuclear splicing.

PolyQ assemblies of HD and healthy individuals differentially respond to stress and drugs

A. The representative polyQ assemblies in ASO-treated HD and healthy fibroblasts stained by the 3B5H10 antibody. B, C. The changes in polyQ assembly size in ASO treatments. Comparing the trends of changes in HD with healthy (B); The changes of polyQ assembly in each individual (C) (HD, n=2 [CAG 57 and 59]; Healthy, n=2; Two-way ANOVA, * <0.05; **<0.01). D. The immunostaining of Onjisaponin F-treated HD and healthy fibroblasts with LC-3B and 3B5H10 antibody. E. The changes in polyQ assembly size in Onjisaponin F treatments. The trends of changes in HD and healthy individuals (HD, n=3; Healthy, n=2; Two-way ANOVA, **<0.01). F. 3B5H10 stained polyQ assemblies under glucose starvation in the HD and healthy fibroblasts at different time points. The middle panels and the inserted image (second panel) are 3D rendering tomography in IMARIS. G, H. The size changes of polyQ assemblies under glucose starvation in the fibroblasts of healthy siblings and HD patients at different time points. The change of polyQ assembly in the fibroblast of each individual (G) (Red line, HD patients; Green line, healthy), comparing the change of polyQ in HD with healthy (H) (Two-way ANOVA, **<0.01).

PolyQ assemblies circle the large Golgi apparatus

A. GM130 and 3B5H10 antibodies and phalloidin staining in healthy sibling and HD fibroblasts showed that polyQ assemblies circle the Golgi apparatus (z-stack intervals, 0.2µm). The rendering tomography of Z-stacked images revealed the spatial relationship of polyQ assemblies with the Golgi stack; polyQ assemblies circle the Golgi stack, and a narrowing of GM130-stained Golgi exists in the site where polyQ encircles the Golgi (yellow arrows, isthmus; blue arrows, the arm of polyQ; the boxed region, the magnified part in the right panel). B. The size of the GM130-stained Golgi apparatus in HD and healthy individuals (HD, n=3; Healthy, n=3. Student t-test; **<0.01). C. The representative images of clathrin and 3B5H10 antibody immunostaining in the fibroblasts of an HD patient and healthy fibroblasts. The left is a large polyQ assembly. The inner inserts on the left show angled and rendered views of the spatial relationship between the polyQ assembly and clathrin+ puncta. The right is the interacting pattern of polyQ with the nucleus and attaching patterns of clathrin to polyQ assemblies that interact with the nucleus (inner inserts, rendered view [lower] and the conjugation pattern of clathrin with polyQ assembly [upper]). D. The number of Clathrin+ vesicles overlapped with polyQ assemblies. The object-object statistics in IMARIS were used to measure overlapping. HD, n=3; Healthy, n=3. Data, mean± SD. Student t-test; *<0.05. E. HTT C-terminal antibody and polyQ antibody immunostaining revealed that C-terminal antibody signals colocalize with cytoplasmic polyQ assemblies. F. Coimmunostaining of 3B5H10 antibody with HAP40 antibody revealed that polyQ assemblies were attached by HAP40. The lefts are sectional views, and the rights are 3D renderings of the middle panel by IMARIS surface. G. The count of HAP40 puncta attached to polyQ assemblies in healthy and HD fibroblasts. The distances between HAP40 puncta and polyQ assemblies were measured by IMARIS. HD, n=2; Healthy, n=2. Student t-test; Data, mean ± SD. Student t-test; ns, p>0.05 H. The 3B5H10 and GM130 antibodies immunostaining images in the mitotic fibroblasts, including prometaphase, anaphase, and telophase, were obtained by SIM microscopy. The inserted DAPI images (gray) showed the nuclear chromatin structure. The right panels are a sectional view of the boxed region (anaphase and telophase). I. ARF1 and 3B5H10 antibody immunostaining in the fibroblasts revealed that ARF1 preferentially interacts with polyQ assemblies. J. The measurement of polyQ assembly-conjugated AFR1 puncta in the healthy and HD fibroblasts. Student t-test; **<0.001 K. GM130 and 3B5H10 antibody immunostaining in BFA-treated fibroblasts revealed that most fragmented Golgi liberate from polyQ. The boxed region in the upper panel is the magnified region. The lower insert is a 3D rendering of a tomographic image. L. The volume of polyQ assemblies/puncta in fibroblasts after 2 hours and 7 hours of treatment by BFA. The measurements were performed using IMARIS software. (Images, n: 7-12. Sample size: n=2 in each group. One-way ANOVA, **<0.01). M. The count of 3B5H10-stained assemblies/puncta in the fibroblasts (Images, n: 7-11. Sample size: n=2 in each group. One-way ANOVA, **<0.01). N. Image sequences of Golgi labelled by GOLGI ID Green assay kit in the living fibroblasts of healthy siblings and HD patients (CAG 59, 49, and CAG 49) after glucose starvation (The displays, the inverted channel of the original image). O. The fragmentation ratio of the Golgi under glucose starvation at different time points in healthy and HD fibroblasts. Fragmentation ratio = number of Golgi of a cell at each time point/ number of Golgi at 0 min (Two-way ANOVA. **<0.01. The number of images in the cell line: n>5).

The characteristics of polyQ assembly in the induced striatal and cortical neurons

A. The characteristics of polyQ assembly in iPS -derived striatal GABAergic neurons at day 90 and 120 (i, the nucleus without nuclear accumulation; ii, the nucleus with nuclear polyQ accumulation). The inner inserts show the 3D tomographic rendering, which revealed the spatial relationship between polyQassemblies and the nucleus. B. The size of polyQ assemblies in iPSCs-derived GABAergic neurons (HD, n=4 [CAG 59, 55, 49, and 47]; Healthy, n=3 [CAG 19, CAG 19, and H9]). Data, mean ± SD. t-test; **<0.01. C. The percentage of nuclear polyQ volume/total polyQ volume in iPS-derived GABAergic neuronal nucleus cultured over 90 days (HD, n=4 [CAG 59, 55, 49, and 47]; Healthy, n=3 [CAG 19, CAG 19, and H9]). t-test; **<0.01. D. The representative images of polyQ assemblies and Golgi stained by 3B5H10 and GM130 antibodies in hCO (Taken by SIM microscopy; Z-stack interval, 0.2µm). The right panels show the rendered tomography of the Golgi by Imaris 9.8 surface (white arrows indicate the Golgi with an irregular surface). E. The representative images of polyQ assemblies and Golgi stained by 3B5H10 and GM130 antibodies in hStrO (taken by SIM microscopy; Z-stack interval, 0.2 µm). The middle panels are a sectional view of the rectangle. The lower panels are 3D rendering tomography by Imaris 9.8 surface. F. Comparing the overlapped volume of the Golgi with polyQ assemblies in HD striatal neurons with that of healthy striatal neurons. The object-object statistic function in Imaris 9.8 measures the overlapped volume (Student t-test; **, p <0.01. HD hStrO: CAG 55 and CAG 47; Healthy hStrO: H9, CAG 19. Student t-test, *<0.05). G. The images of polyQ assemblies stained by 3B5H10 and clathrin antibodies in hStrOs were taken by SIM microscopy (The upper panel, the lateral view of the boxed region; the inner insert in the rendered image, a larger view of the original image). The renderings are done by Imaris surface. H. Comparing clathrin+ vesicle volume overlapped with 3B5H10-stained polyQ assemblies of striatal neurons in the healthy sibling with an HD patient. t-test, *, p<0.05. HD hStrO: CAG 55 and CAG 47; Healthy hStrO: H9, CAG 19. Student t-test; * <0.05. I. 3B5H10 and GM130 antibodies immunostaining images and their sectional views in the neurons of hStrO of HD patients and healthy siblings revealed the spatial relationships between polyQ assemblies and Golgi (white arrows, the Golgi in the inner hole of the ring; pink arrows, the scaffolding Golgi; the boxed region, the right or left panel). The right panel is a 3B5H10-stained assembly in a human striatal neuron with a long projection. J. 3B5H10 and clathrin antibodies staining images of the neurons in hStrO of HD patients and healthy siblings revealed the spatial relationships between polyQ assemblies and clathrin-stained vesicles. The left panel is a sectional view between two arrows (hStrO at day 80; two white arrows, the segment in the right panel) K. The raster plots and network activity of the cortical neurons in HD and healthy hCOs. L. Quantification of network features at the burst level in HD and healthy hCOs, including the number of spikes per burst, the interburst intervals (IBI), and IBI CV (CTR group: n=3 hCOs; HD group: n=3 hCOs; The mean ± SD, Welch’s t-test, * < 0.05; ** < 0.01).

GABAergic neurons from HD patients reduced Golgi- or vesicle-related activities and intensified nucleus-related activities

A, B. Uniform manifold approximation and projection (UMAP) plot of all cells colored by their sample origin (A) and by cell types(B). C. Dot plot displaying an expression of example marker genes. D. Venn diagram summarizing the number of overlapped genes differentially expressed in CAG59 vs. CAG19/H9 GABAergic neurons. E. GO enrichment analysis of down-regulated (top) and up-regulated (bottom) genes in CAG59 vs. CAG19/H9 GABAergic neurons.

Onjisaponin F did not mitigate polyQ nuclear accumulation, and ASO corrected the aberrant HD striatal neuronal firing

A. The schematics of treatment (upper) and the representative images of polyQ assemblies stained by 3B5H10 antibodies in HD and healthy striatal neurons at 63 days treated by one dose of 20 µM onjisaponin F for 72 h.

Characterization of polyQ assemblies and Golgi activities

A. LSB11011 (C-terminal of HTT) and 3B5H10 (N-terminal) antibody and phalloidin immunostaining revealed that actin filaments are enriched on the edge of large polyQ assemblies (i) or some small polyQ assembly near the cellular membrane (ii) (the boxed region indicated the sectional view in the right panel; white arrows, the polyQ assemblies interact with actin; the inner panels in ii are the spliced channels; white arrows, the interacting patterns of polyQ with F-actin). B, C. LSB11011 (C-terminal of HTT) and 3B5H10 (N-terminal) antibodies, and phalloidin immunostaining revealed a linear assembly of polyQ on the ruffling membrane. The right panel shows a 3D image (B) revealing the alignment pattern of polyQ on membrane curvature and ruffling membrane. The image of sectional views (C) revealed the interaction pattern of polyQ with actin filaments. D. Image sequences of the Golgi in the fibroblasts of the HD family under oxidative stress at the indicated time points. E. The fragmentation ratio of the Golgi under H2O2 stimulation at different time points. Fragmentation ratio = number of Golgi of one cell at each time point/ number of Golgi at 0 min. Two-way ANOVA. *<0.05; **<0.01.D F. The images of 3B5H10 and GM130 antibody immunostaining in the fibroblasts after 48 h glucose starvation, taken by SIM microscopy. The 3D-rendered tomography showed the spatial relationship between polyQ assemblies/puncta and the Golgi. The boxed regions are the magnified part or rendered region. G, H. The changes of Golgi fragments after 48 h glucose starvation in the fibroblasts of healthy siblings (two siblings, CAG 19) and HD patients (CAG 55 and CAG 49). HD/healthy ctr, the fibroblasts cultured in high glucose medium; HD/healthy starvation, the fibroblasts cultured in low glucose medium for 48h. Two-way ANOVA; **0.01. Cell count in each group: n>120. I. The changes of GM130-stained Golgi sizes after 48 h glucose starvation in the fibroblasts of healthy siblings (two, CAG 19) and HD patients (CAG 55 and CAG 49). HD/healthy ctrl, the fibroblasts cultured in high glucose medium; HD/healthy starvation, the fibroblasts cultured in low glucose medium for 48h. One-way ANOVA; **0.01. The number of images in the cell line: n>20. Cell count in each group: n>400.

Characterization of endogenous polyQ aggregate structure

A. 3B5H10 antibody and phalloidin staining showed polyQ distribution patterns in the fibroblasts of healthy (CAG 19) and patient (CAG 59). 3D-reconstructed tomography by IMARIS revealed the spatial relationships between polyQ assemblies and the nuclear surface and F-actin filaments (white arrows, polyQ puncta within the nucleus; yellow arrows, nuclear interactions). The yellow arrows in the DAPI-rendered image indicated the nuclear surface. B. The size of polyQ assemblies in the fibroblasts of healthy siblings (n=4; cells, n=306) and HD patients (n=4; CAGs: 59, 55, 49, and 47; cells, n=783). Fiji ImageJ measured the size. C. Average size of polyQ assembly in HD patients and healthy siblings. Student t-test; ns, p >0.05. D. Mito-tracker, phalloidin, and 3B5H10 antibody immunostaining in the fibroblasts of HD patients (CAGs, 49) and healthy siblings. 3D rendering tomography in the right two panels revealed the spatial relationships of polyQ assemblies with mitochondria and actin filaments. E. 3B5H10 (three rights) or 3B5H10 (N-terminal) and LSB11011 (C-terminal) (left) stained polyQ puncta/patches in the mitotic cells. F. The representative image of 3B5H10 antibodies immunostaining in HD patient (CAG 44) and healthy sibling fibroblasts by two different protocols (immediate staining is stained 15-30 minutes after 4% PFA fixation; delayed staining is stained 48 h after 4% PFA fixation). Left panels show freshly fixed and immediately stained groups; right panels show freshly fixed and later stained groups. Images were taken by SIM microscopy with Z-stacks.

PolyQ assemblies in human fetal and child brains and human brain organoids with necrotic cells

A. Typical appearance of intact and continuous polyQ assemblies in the neurons of hStrO. B. PolyQ assemblies stained by 3B5H10 antibody in fresh surgical brain tissues collected from a 4-year-old glioblastoma patient. Three representative appearances of polyQ assemblies in human brain cells. The middle panel (image-i) is a sectional view that shows a ring-like structure; the left corner insert in image-ii is a larger view; the rendered tomography shows the nuclear surface. C. 3B5H10 and ZO-1 antibodies immunostained intact polyQ assemblies in the pseudostratified neuroepithelium of the neural tube in the human fetal brain at 7.5 gestation weeks, and hCOs from HD patients, H9, and healthy siblings showed that the polyQ assemblies in the neuroepithelial cells are intact and long and formed by a small ring-like unit, and paralleled the apical-basal axis of cells. D. Fragmented/discontinuous polyQ assemblies in freshly surged human brain samples from cortex (tissues from 4-year-old and 2-year-old glioblastoma patients). E. The fragmentation of polyQ assemblies depends on the necrotic cells in the surrounding regions (green arrows, necrotic cells).

The impaired nuclear activities in HD striatal organoids

A. The validation of hCO and hStrOs by immunostaining with GSH2/MASH1 or DARPP321/CTIP2 antibodies. B. GO analysis of differentially expressed proteins in HD striatal neurons derived from iPSCs and healthy GABAergic neurons derived from iPSCs at 60 100 days (CAG 55 vs. H9 and CAG 19). C. The validation of hCO by immunostaining with PAX6/TBR2 /HOPX or TBR1/CTIP2/SATB2 antibodies.

ARF1 levels were reduced in HD neurons

A. The immunostaining results of ARF1, CTIP2, and 3B5H10 antibodies in the neurons of hStrOs (CAG 49 and CAG 19). The middle panels showed ARF1 signals. B, C. The overlapped volume of ARF1 with polyQ in the HD and healthy neurons of hStrOs (G). The ratio of ARF1 volume/polyQ volume in CTIP2 + regions. The surface function of IMARIS was used to calibrate volume. Object-object statistics in IMARIS measure the overlapped volume. t-test, *<0.05; **<0.01. D. The immunostaining results of ARF1, CTIP2, and GM130 antibodies in the neurons of hStrOs (CAG 49 and CAG 19). The middle panels showed ARF1 signals. E, F. The overlapped volume of ARF1 with the Golgi in the HD and healthy neurons of hStrOs. The ratio of ARF1 volume/Golgi volume in CTIP2 + regions. The surface function of IMARIS was used to calibrate volume. Object-object statistics in IMARIS measure the overlapped volume. t-test, **, p<0.01. G. Bulk RNAseq data of HD hStrOs (CAG 55 and 59) and healthy hStrOs (CAG 19 and H9) at 60 days revealed ARF1 transcription levels. Student t-test; *<0.05; **<0.01. H. The bulk RNAseq data of HD hStrOs (CAG 55 and 59) and healthy hStrOs (CAG 19 and H9) at 80 days revealed ARF1. Student t-test; **, p<0.01. CTR, healthy.

Characterizing and sorting the cellular types from scRNA-seq data of hStrOs and GO analyses of different cellular clusters in scRNA-seq data of hStrOs

A. Expression of known markers of progenitor, neuron, astrocyte and oligodendrocyte. B. Stacked bar plots and a pie chart representing the proportion of cell types. C, D. Expression of known markers of LGE progenitors and cortical progenitors(C), GABAergic neurons and glutamatergic neurons (D). E. Dot plot displaying expression of GABAergic neurons marker genes. F. UMAP plot of neurons colored by GABAergic neurons (red) and other neurons (grey). G. GO enrichment analysis of down-regulated (top) and up-regulated (bottom) genes in CAG59 VS CAG19/H9 progenitor, other neuron, astrocyte, and oligodendrocyte.

ASO treatments did not alter Golgi morgphology and ARF1 scaffolding in polyQ assembly

A. The representative super-resolution images of polyQ assemblies stained by 3B5H10 antibody and Golgi stained by GM130 antibody in ASO-treated and control h hCOs. The middle and right panels show the spatial relationships between polyQ assemblies and the Golgi, and the tomography of the Golgi, respectively. The renderings are done by IMARIS surface. B. The representative super-resolution images of polyQ assemblies stained by 3B5H10 antibody and ARF1 puncta in ASO-treated and control (CTR) hCOs. The images are displayed as MIP. The rights are the magnified region in the rectangle on the left. C. The density of ARF1+ puncta in polyQ assemblies ASO-treated and control (CTR) hCOs (HD, n=2 [CAG 47 and CAG 59]; hCOs, n=6; t-test, ns>0.05). Puncta are counted with the IMARIS surface. D. The overlapped volume of ARF1 puncta with polyQ assemblies in ASO-treated and control (CTR) hCOs. (HD, n=2 [CAG 47 and CAG 59]; hCOs, n=6; t-test, ns>0.05). polyQ assembly and puncta volume are measured by the IMARIS surface.

The baseline characteritics of HD family members.