A Neuron-Optimized CRISPR/dCas9 Activation System for Robust and Specific Gene Regulation

CRISPRa sgRNA multiplexing for synergistic or coordinated control of gene expression. A, Illustration of pooled sgRNA multiplexing for dCas9-VPR targeting to multiple locations at a single gene (top) or simultaneous regulation of several genes (bottom). B, Single gene multiplexing at Fos (left) and Fosb (right) reveals that while individual sgRNAs are sufficient to drive gene expression, sgRNA pooling results in synergistic induction of gene expression in cultured neurons (n = 5–6, one-way ANOVA, Fos F_(4,25) = 16.17, p < 0.0001; Fosb F_(3,19) = 10.23, p = 0.0003; Tukey’s post hoc test for individual comparisons). C, CRISPRa with sgRNAs targeting Egr1, Fos, or Fosb individually results in specific and robust increases in gene expression without effects at non-targeted genes (n = 5–6, one-way ANOVA, Egr1 F_(3,16) = 56.53, p < 0.0001; Fos F_(3,16) = 17.55, p < 0.0001; Fosb F_(3,15) = 32.06, p < 0.0001; Dunnett’s post hoc test for individual comparisons). D, Pooled gRNAs result in coordinated increases in gene expression at Egr1, Fos, and Fosb (n = 6 per group). All data are expressed as mean ± SEM. Individual comparisons; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. CRISPR inactivation with the same sgRNAs as CRISPRa is shown in Extended Figure 2-1.

CRISPRa induction of Bdnf transcript variants I and IV in primary rat hippocampal neurons. A, Bdnf gene structure illustrating non-coding exons (I-IXa) and a common coding exon (IX). sgRNAs were designed upstream of exons I and IV, as indicated by the red and blue lines. B–D, Expression of Bdnf I, IV, and IX transcript variants after targeting dCas9-VPR to exons I and/or IV using sgRNAs, measured with RT-qPCR. B, Bdnf I transcript is specifically upregulated with Bdnf I sgRNA but not with Bdnf IV sgRNA (n = 8, one-way ANOVA, F_(3,28) = 15.65, p < 0.0001). C, Bdnf IV transcript is specifically upregulated with Bdnf IV sgRNA but not with Bdnf I sgRNA (n = 8, one-way ANOVA, F_(3,28) = 34.16, p < 0.0001). D, Total Bdnf IX transcript levels are upregulated with both Bdnf I and Bdnf IV sgRNAs (n = 8, one-way ANOVA, F_(3,28) = 277.7, p < 0.0001). sgRNA designed for the bacterial LacZ gene is used as a non-targeting control in panels B–D. Dunnett’s post hoc test was used for individual comparisons. All data are expressed as mean ± SEM. Individual comparisons; **p < 0.01, ***p < 0.001, ****p < 0.0001.

Transcriptome-wide selectivity of CRISPRa at Bdnf non-coding exons and the absence of off-target gene upregulation revealed by RNA-seq. A, B, Bdnf transcript variant expression (FPKM values) following dCas9-VPR targeting with Bdnf I (A) and Bdnf IV (B) sgRNAs. Bdnf I sgRNA treatment upregulated Bdnf I transcripts by 63.2x (A), while Bdnf IV sgRNA treatment upregulated Bdnf IV transcripts by 23x (B). Both Bdnf I and IV sgRNA targeted conditions increased Bdnf IX transcript expression by 4.23x and 12x, respectively. sgRNA designed for the bacterial LacZ gene is used as a non-targeting control. All data are expressed as mean ± SEM in A, B. C, D, Mirrored Manhattan plots showing degree of mRNA change across the genome for Bdnf I (C) and Bdnf IV (D) dCas9-VPR targeting. While there were no exact matches for Bdnf I or Bdnf IV sgRNA sequences elsewhere in the genome, all potential off-target sites with up to 4 nucleotide mismatches (identified with Cas-OFFinder) are shown in orange. Predicted off-target sequences for Bdnf I and IV targeting are shown in Extended Tables 4-1 and 4-2, respectively.

CRISPRa targeted induction of Bdnf I and IV transcript variants causes coordinated upregulation of genes involved in neuronal activation and synaptic function. A, B, RNA-seq volcano plots showing DEGs detected by DESeq2 in LacZ versus Bdnf I sgRNA (A) and LacZ versus Bdnf IV sgRNA (B) targeted conditions. Standard cutoff point is represented by the horizontal dotted line (adjusted p < 0.05). Upregulated (red or blue) and downregulated (orange or green) genes are indicated for each comparison. Bdnf is the top upregulated gene in both conditions. C, D, Heat maps representing all DEGs comparing LacZ versus Bdnf I sgRNA (C) and LacZ versus Bdnf IV sgRNA (D) targeted conditions for three biological replicates. Values in each row represent LacZ-normalized counts for each DEG (adjusted p < 0.05). Log₂ fold change increases (red or blue) or decreases (orange or green) in gene expression are presented relative to the LacZ mean (white). E, Venn diagram representing 664 DEGs after Bdnf I sgRNA targeting (red) and 2842 DEGs after Bdnf IV sgRNA targeting (blue), with 259 overlapping genes. F, Scatter plot representing all shared 259 DEGs in Bdnf I versus Bdnf IV sgRNA targeted conditions. Genes upregulated in both groups (141), downregulated in both groups (97), upregulated after Bdnf I and downregulated after Bdnf IV sgRNA targeting (11), downregulated after Bdnf I and upregulated after Bdnf IV sgRNA targeting (10) are indicated. Select upregulated IEGs are specified. G, Top significant GO terms for 141 co-upregulated and 97 co-downregulated genes in Bdnf I and Bdnf IV sgRNA targeted conditions.

CRISPRa induction of Bdnf mRNA increases spike and burst frequency in hippocampal neurons cultured on microelectrode arrays (MEAs). A, CRISPRa induction of Bdnf I and IV increases Bdnf protein quantified by immunoblotting (n = 6 per group; Mann–Whitney U test, U = 0, p = 0.0022). B, Primary hippocampal neurons grown on MEAs and transduced with dCas9-VPR and LacZ (top) or Bdnf I and IV (bottom) sgRNAs. mCherry signal indicates successful transduction of sgRNAs in live cultures (right). Scale bar = 100 μm. C, Experimental timeline for viral transduction and MEA recordings. Representative traces (D) and raster plots (E) from 10 units after LacZ (top) or Bdnf I and IV (bottom) targeting. F, The number of active units per well does not change between LacZ and Bdnf I and IV targeted conditions (n = 10–12, unpaired Student’s t test; p = 0.1783). G, Action potential frequency across DIV7–DIV11 showing an increase of mean frequency after Bdnf I and IV sgRNA treatment by DIV11, as compared to LacZ sgRNA (n = 57–98 neurons, two-way ANOVA with main effect of sgRNA, F_(1,493) = 8.561, p = 0.0036, Sidak’s post hoc test for multiple comparison). H, Spike frequency at DIV11 for all units ranked from highest to lowest mean frequency showing an increase in activity for the top 1/3 most active units in Bdnf I and IV versus LacZ targeted conditions. I, Burst frequency at DIV11 is increased after Bdnf I and IV versus LacZ targeting (n = 98, unpaired Student’s t test; p = 0.0392). All data are expressed as mean ± SEM; *p < 0.05 and **p < 0.01. The physiological consequences of CRISPRa induction at another gene target (Reln) is shown in Extended Figure 6-1.

CRISPRa-mediated induction of Fosb in hippocampal, striatal, and cortical neurons in vivo. A–C, Lentiviral infusions were bilaterally targeted to the brain region of interest (Paxinos and Watson, 2009) in adult male rats (n = 4 rats/region). Two weeks following stereotaxic viral infusions, animals were transcardially perfused and IHC was performed to measure Fosb upregulation. IHC reveals high transduction efficiency of the guide RNA (expressing mCherry, signal not amplified) bilaterally in (A) the CA1 region of the dorsal hippocampus, (B) the nucleus accumbens core (NAc), and (C) the medial PFC. Fosb protein is enhanced in the hemisphere that was infused with the Fosb-targeting sgRNA (right) compared to the hemisphere that received a sgRNA targeting the bacterial LacZ gene (left). Cell nuclei were stained with DAPI. Scale bar = 500 μm. Schematics of target regions are adapted from Paxinos and Watson. D–F, dCas9-VPR increases the number of Fosb+ cells in the CA1, NAc, and PFC, compared to a non-targeting control (LacZ; n = 4, ratio paired t test; CA1: t₍₃₎ = 8.73, p = 0.003, R ² = 0.96; NAc: t₍₃₎ = 4.62, p = 0.019, R ² = 0.87; PFC: t₍₃₎ = 3.43, p = 0.041, R ² = 0.79). All data are expressed as mean ± SEM. Individual comparisons; *p < 0.05 and **p < 0.01. Or: oriens layer, Py: pyramidal cell layer, Rad: radiatum layer, LMol: lacunosum moleculare, DG: dentate gyrus, ac: anterior commissure, LV: lateral ventricle.