Cerulenin partially corrects the disrupted developmental transcriptomic signature in Huntington's disease striatal medium spiny neurons

Abstract

Huntington's disease (HD) is a neurodegenerative disorder caused by an expansion of CAG repeats in exon 1 of the huntingtin (HTT) gene, resulting in a mutant HTT (mHTT) protein. Although mHTT is expressed in all tissues, it significantly affects medium spiny neurons (MSNs) in the striatum, resulting in their loss and the subsequent motor function impairment in HD. While HD symptoms typically emerge in midlife, disrupted MSN neurodevelopment is important. To explore the effects of mHTT on MSN development, we differentiated HD-induced pluripotent stem cells (iPSCs) and isogenic controls into neuronal stem cells, and then generated a developing MSN population encompassing early, intermediate progenitors, and nascent MSNs. Single-cell RNA sequencing revealed that the developmental trajectory of MSNs in our model closely emulated the trajectory of human fetal striatal neurons. However, in the HD MSN cultures, several crucial genes required for proper MSN maturation were downregulated, including members of the DLX family of transcription factors. Our analysis also uncovered a progressive dysregulation of multiple HD-related pathways as MSNs developed, including the NRF2-mediated oxidative stress response and mitogen-activated protein kinase signaling. Using the transcriptional profile of developing HD MSNs, we searched the L1000 dataset for small molecules that induce the opposite gene expression pattern. We pinpointed numerous small molecules with known benefits in HD models and previously untested novel molecules. A top candidate, Cerulenin, partially restored the DARPP-32 levels and electrical activity in HD MSNs, and also modulated genes involved in multiple HD-related pathways.

Keywords: Huntington’s disease; development; induced pluripotent stem cells; medium spiny neurons; neurodegeneration; single-cell RNAseq.

Graphical Abstract

Figure 1.

Differentiation of developing MSNs. (A) Diagram of differentiation process for MSNs. (B) C116 and HD72 NSCs immunolabeled with Nestin and SOX2. (C) DARPP-32 staining on developing MSNs. Scale bar: 200 μm. (D) RNA expression of multiple markers of MSNs in C116 NSCs and developing MSNs determined with qPCR. P values were calculated using t-tests followed by Benjamini-Hochberg correction for multiple tests. ^#<.1, *<.05, **<.01.

Figure 2.

Transcriptional characterization of HD72 NSCs and developing MSNs. (A, B) PCA plot of NSCs (A) and developing MSNs (B). A HD72 NSC outlier has been removed only for visualization purposes. (C, D) Volcano plots illustrating DEGs when comparing HD72 vs C116 NSCs (C) and HD72 MSNs vs C116 MSNs (D). (E, F) IPA of HD72 NSCs (E) and HD72 MSNs (F). Gray dots indicate an undetermined direction for the pathway. (G, H) BioPlanet enrichment analysis of HD72 NSCs (G) and HD72 MSNs (H). The transcriptomics from HD72 NSCs from Ring et al. were used.

Figure 3.

Single-cell transcriptional characterization of HD72 and C116 developing MSNs. (A) UMAP plots of C116 and HD72 MSNs. (B) Heatmap showing levels of expression of various markers of early progenitors, intermediate progenitors, and nascent MSNs in combined C116 and HD72 cells. (C) Heatmap showing level of activation of pathways that become more activated in HD72 over C116 with MSN maturation (HD72/C116). (D) Heatmap showing level of activation of pathways that become less activated in HD72 over C116 with MSN maturation (HD72/C116). (E) Violin plots showing expression of the top upregulated and downregulated genes in HD72 early progenitors, NMB and CCND2. (F) Violin plots showing expression of the top upregulated and downregulated genes in HD72 intermediate progenitors, DLX6-AS1 and HOXB5 . (G) Violin plots showing expression of the top upregulated and downregulated genes in HD72 nascent MSNs, CRYBA2 and TCEAL7.

Figure 4.

Comparison of iPSC-derived C116 and HD72 MSNs with human fetal LGE. (A, B) UMAP plot of fetal LGE (A) C116 and HD72 developing MSNs (B). (C) Expression of NES, TCF7L1, GSX2, ASCL1, DLX1, DCX, MEIS2, and FOXP1. (D) Proportion of apical progenitors, basal progenitors, and MSNs in LGE. (E) Proportion of early progenitors, intermediate progenitors, and nascent MSNs in C116 and HD72 MSNs.

Figure 5.

Developmental dysregulation in developing HD72 MSNs. (A) Pseudotime in fetal LGE, C116 MSNs, and HD72 MSNs. Expression of neuronal maturation markers, NES (B), ASCL1 (C), and DCX (D). Expression of transcription factors DLX1 (E), DLX2 (F), DLX5 (G), DLX6 (H), DLX6-AS1 (I), and GAD1 (J) in human LGE and developing C116 and HD72 MSNs. (K) Comparison of DLX2 targets at transcriptional start sites with DEGs in HD72 MSNs. P values were calculated using Fisher’s exact tests followed by Benjamini-Hochberg correction for multiple tests. (L) Expression of DLX5, DLX6, GAD1, and GAD2 in iSPNs and dSPNs from caudate and putamen of HD patients and R6/2.

Figure 6.

Small molecules predicted to reverse HD dysregulation. (A) List of small molecules predicted to reverse gene dysregulation in developing HD72 MSNs. Each dot represents an individual signature obtained from the treatment of a cell line with the small molecule. (B) Canonical targets for predicted small molecules. (C) tSNE plot based on scores of predicted mechanisms of action for each small molecule. Molecules with more than a 40% confidence are color-coded to its top mechanism of action.

Figure 7.

Effects of Cerulenin treatment on HD72 MSNs. (A) Quantification of DARPP-32 levels from immunostaining of Cerulenin-treated HD72 developing MSNs. Each transparent data point represents individual cell intensities, and the darker points indicate well averages. Cells and wells are color-matched. P values were calculated using Dun’s test followed by Benjamini-Hochberg correction for the comparisons to HD72 DMSO. (B) Immunostaining of HD72 and C116 MSNs treated with 250 nM Cerulenin. Scale bar: 200 μm. (C) Quantification of DARPP-32 in C116 and HD72 MSNs treated with 250 nM Cerulenin. Each transparent data point represents individual cell intensities, and the darker points indicate well averages. Cells and wells are color-matched. P values were calculated using pairwise Wilcoxon’s tests followed by Benjamini-Hochberg correction. (D) Active electrodes/minute in cultures of C116, and HD72 developing MSNs treated with DMSO or 250 nM Cerulenin. P values were calculated using pairwise Wilcoxon’s tests followed by Benjamini-Hochberg correction. (E) Log2 fold changes of genes significantly changed between HD72 and C116 developing MSNs (x-axis) and by treatment of HD72 developing MSNs with Cerulenin (y-axis). (F) Bar plot showing pathways altered by Cerulenin treatment in HD72 MSNs. (G) Images of BCL11B staining in HD72 developing MSNs; blue outline shows outline of DAPI-stained nuclei. White outline shows outline of neuronal processes. (H) Quantification of BCL11B foci per nuclei in the HD MSNs. P values were calculated using Dunnett’s test for the comparisons to HD72 DMSO.