Striatal cell-type-specific molecular signatures reveal potential therapeutic targets in a model of dystonia

Abstract

Abnormal dopamine neurotransmission and striatal dysfunction is implicated in many forms of dystonia, yet the underlying molecular processes remain unknown. Here, we identified thousands of dysregulated genes within striatal spiny projection neuron (SPN) subtypes in a genetic mouse model of DOPA-responsive dystonia (DRD), which is caused by gene defects that reduce dopamine neurotransmission. Although changes in mRNA expression were unique to each SPN subtype, abnormal glutamatergic signaling was implicated in each SPN subtype. Indeed, both AMPA and NMDA receptor-mediated currents were enhanced in direct SPNs but diminished in indirect SPNs in DRD mice. The pattern of mRNA dysregulation was distinct from parkinsonism where the dopamine deficit occurs in adults, suggesting that the phenotypic outcome is dependent on both the timing of the dopaminergic deficit and the SPN-specific adaptions. By leveraging these disease-specific molecular signatures, we identified LRRK2 inhibition, among other mechanisms, as a novel therapeutic target for dystonia.

Keywords: D1 dopamine receptor; D2 dopamine receptor; MLi-2; Medium spiny neuron; Parkinson's disease; RNA-seq; Translatome.

Fig. 1.

Hundreds of genes are dysregulated in dSPNs in DRD mice. A. Robust principal component analysis outlier map. All samples (gray circles) were clustered within the cutoffs for score distance and orthogonal distance (represented by vertical and horizontal dotted lines, respectively) demonstrating no samples were outliers (n = 3 females, 5 males/genotype for panels A-F). B. Assessment of TRAP enrichment for dSPN mRNA using normalized counts of Drd1 and Drd2 from RNA-seq on striatal mRNA extracted from control and DRD mice carrying the Drd1a-TRAP transgene. Drd1 mRNA was significantly enriched compared to Drd2 mRNA in both control (t₁₄ = 16.22, p < 0.0001, one-tailed t-test) and DRD mice (t₁₄ = 18.40, p < 0.0001, one-tailed t-test). C. Volcano plot depicting protein-coding genes differentially expressed in dSPNs in DRD compared to control mice. Downregulated genes in DRD mice are represented by dark aqua dots and upregulated genes in DRD mice are represented by light aqua dots (fold change > ∣1.25∣; BH adjusted p value <0.1). All other genes are represented by light gray dots. D. The top ten downregulated (dark aqua) and top ten upregulated (light aqua) genes in dSPNs in DRD mice. E. Significantly enriched pathways in dSPNs identified by gene set enrichment analysis. All dSPN pathways were downregulated in DRD compared to control mice. The x axis represents − log₁₀ p values with a maximum cutoff of 9. F. Heatmap of individual variation of differentially expressed genes in dSPNs in the Inactivation of MAPK Activity pathway between control and DRD mice. Data are represented as z-scores, with yellow depicting higher expression and aqua depicting lower expression. Each column represents an individual mouse. G. Validation of RNA-seq results from dSPNs using mRNA extracted from whole striatum. Normalized expression (ΔΔCT) of Spred mRNAs identified in ‘F’ was measured by qPCR in an independent cohort of control and DRD mice (n = 6/genotype). Spred1 (t₁₀ = 2.414, p = 0.0182, one-tailed t-test), Spred2 (t₁₀ = 2.428, p = 0.0178, one-tailed t-test), and Spred3 (t₉ = 11.02, p < 0.0001, one-tailed t-test) mRNAs were significantly downregulated in DRD compared to control mice. For box and whisker plots, horizontal lines illustrate the minimum, 25th percentile, median, 75th percentile, and maximum with values for individual mice indicated by circles. Asterisks indicate *p < 0.05 and ****p < 0.0001.

Fig. 2.

Hundreds of genes are dysregulated in iSPNs in DRD mice. A. Robust principal component analysis outlier map. All samples (gray circles) were clustered within the cutoffs for score distance and orthogonal distance (represented by vertical and horizontal dotted lines, respectively) demonstrating no samples were outliers (n = 5 females, 3 males/genotype for panels A-F). B. Assessment of TRAP enrichment for iSPN mRNA using normalized counts of Drd2 and Drd1 from RNA-seq on striatal mRNA extracted from control and DRD mice carrying the Drd2-TRAP transgene. Drd2 mRNA was significantly enriched compared to Drd1 mRNA in both control (t₁₄ = 21.76, p < 0.0001, one-tailed t-test) and DRD mice (t₁₄ = 18.27, p < 0.0001, one-tailed t-test). C. Volcano plot depicting protein-coding genes differentially expressed in iSPNs in DRD compared to control mice. Genes significantly downregulated in DRD mice are represented by dark purple dots and genes significantly upregulated in DRD mice are represented by light purple dots (fold change > ∣1.25∣; BH adjusted p value <0.1). All other genes are represented by light gray dots. D. The top ten downregulated (dark purple) and top ten upregulated (light purple) genes in iSPNs in DRD mice. E. Significantly enriched, brain-relevant pathways in iSPNs identified by gene set enrichment analysis. All pathways identified for iSPNs were upregulated in DRD compared to control mice. F. Heatmap of individual variation of differentially expressed genes in the Response to Calcium Ion pathway in iSPN-enriched mRNAs among control and DRD mice. Data are represented as z-scores, with yellow depicting higher expression and purple depicting lower expression. Each column represents an individual mouse. G. Validation of RNA-seq results from iSPNs using western blot for ΔFOSB in striatal homogenates from control and DRD mice (n = 4/genotype). Individual values of the densitometric quantification of ΔFOSB normalized to ß-actin for control and DRD mice are superimposed over box and whisker plots. ΔFOSB protein expression was significantly increased in DRD compared to control mice (t₆ = 3.951, p = 0.0038, one-tailed t-test). Box and whisker plots illustrate the minimum, 25th percentile, median, 75th percentile, and maximum with horizontal lines with values for individual mice indicated by circles. Asterisks indicate **p < 0.01 and ****p < 0.0001.

Fig. 3.

Unique cell-type–specific adaptations in dSPNs and iSPNs. A. Venn diagram of dysregulated genes in dSPNs and iSPNs. 106 genes were upregulated in both dSPNs (light aqua) and iSPNs (light purple) while 88 genes were downregulated in both dSPNs (dark aqua) and iSPNs (dark purple) in DRD compared to control mice (fold change > ∣1.25∣, BH adjusted p value <0.1). B. Significantly enriched pathways identified by gene set enrichment analysis using genes that were regulated in opposite directions in dSPNs and iSPNs of DRD mice. C. AMPA currents were increased in dSPNs (t₃₀ = 2.346, p = 0.0256, two-tailed t-test) but decreased in iSPNs (t₃₂ = 2.179, p = 0.0368, two-tailed t-test) of DRD mice compared to controls in response to electrically evoked EPSCs from corticostriatal synapses. D. NMDA currents were increased in dSPNs (t₃₀ = 2.560, p = 0.0158, two-tailed t-test) but decreased in iSPNs (t₃₂ = 3.303, p = 0.0024, two-tailed t-test) of DRD mice compared to controls. E. The paired pulse ratio (PPR) did not change in dSPNs (t₃₆ = 1.339, p = 0.1888, two-tailed t-test) but decreased in iSPNs (t₄₁ = 2.623, p = 0.0122, two-tailed t-test) of DRD mice compared to controls. For C, D and E, number of cells appear in parenthesis (6–10 mice/group). Box and whisker plots illustrate the minimum, 25th percentile, median, 75th percentile, and maximum with horizontal lines. Asterisks indicate *p < 0.05 and **p < 0.01 compared to control. F, G. Representative traces for AMPA and NMDA components of EPSCs when SPNs were held at −80 and + 40 mV, respectively, in dSPNs (F) and iSPNs (G) of control and DRD mice.

Fig. 4.

Distinct molecular adaptations in dystonia and parkinsonism. A. Turning asymmetry in control (solid circles) and parkinsonian (PD; open circles) mice carrying the Drd1a-TRAP transgene for isolation of translating mRNA from dSPNs. Spontaneous 360° rotations were recorded over 5 min and ipsilateral rotations toward the 6-OHDA or sham-injected hemisphere were calculated as a percent of the total (n = 14 mice/treatment for all panels after exclusion of outliers in B). B. Robust principal component analysis outlier map. Most samples (gray circles) were clustered within the cutoffs for score distance and orthogonal distance (represented by vertical and horizontal dotted lines, respectively) while three samples (red circles) were outliers and excluded from subsequent analyses (sham, n = 15 mice; 6-OHDA-lesioned, n = 16 mice). C. Assessment of TRAP enrichment for dSPN mRNA using normalized counts of Drd1 and Drd2 from RNA-seq on striatal mRNA extracted from control and 6-OHDA-lesioned mice carrying the Drd1a-TRAP transgene. Drd1 mRNA was significantly enriched compared to Drd2 mRNA in both sham-treated (t₂₆ = 26.76, p < 0.0001, one-tailed t-test) and 6-OHDA-treated mice (t₂₆ = 24.22, p < 0.0001, one-tailed t-test). D. Volcano plot depicting protein-coding genes differentially expressed in dSPNs in 6-OHDA-treated compared to control mice. Genes significantly downregulated in 6-OHDA-treated mice are represented by dark blue dots and genes significantly upregulated in 6-OHDA-treated mice are represented by light blue dots (fold change > ∣1.25∣; BH adjusted p value <0.1). All other protein-coding genes are represented by light gray dots. E. Venn diagrams of upregulated and downregulated genes in dSPNs in DRD and parkinsonian mice. Thirty-eight genes were upregulated in both conditions and 80 genes were downregulated in both conditions. Box and whisker plots illustrate the minimum, 25th percentile, median, 75th percentile, and maximum with horizontal lines with values for individual mice indicated by circles. Asterisks indicate ****p < 0.0001.

Fig. 5.

Identification of potential therapeutic targets for dystonia. A. Mechanisms of action predicted to correct both dSPN and iSPN gene expression in DRD mice were identified using CMAP. Negative connectivity scores of <−1.0 are predicted to reverse the gene expression abnormalities. B. Normalized expression of Lrrk2 translating mRNA in dSPNs and iSPNs of control and DRD mice (n = 8/genotype for each cell type from data described in Figs. 1 & 2). Lrrk2 expression was significantly reduced in dSPNs of DRD mice compared to control mice (padj <0.0001) but significantly increased in iSPNs of DRD mice compared to control mice (padj = 0.00027). C. The LRRK2 inhibitor MLi-2 attenuates dystonia in DRD mice. Abnormal movements and locomotor activity were assessed simultaneously after administration of vehicle or MLi-2 in a crossover design (n = 5 females, 3 males). Compared to vehicle, MLi-2 significantly reduced abnormal movements in DRD mice (t₇ = 2.728, p = 0.0147, paired one-tailed t-test). Compared to vehicle, MLi-2 did not affect locomotor activity (p = 0.3125, two-tailed Wilcoxon test). Data from individual mice are plotted as open circles with lines indicating the individual response to vehicle and MLi-2. Box and whisker plots illustrate the minimum, 25th percentile, median, 75th percentile, and maximum with horizontal lines with values for individual mice indicated by circles. *p < 0.05, ***p < 0.001, ****p < 0. 0001.