Focused ultrasound widely broadens AAV-delivered Cas9 distribution and activity

Abstract

Because children have little temporal exposure to environment and aging, most pediatric neurological diseases are inherent, i.e. genetic. Since postnatal neurons and astrocytes are mostly non-replicating, gene therapy and genome editing present enormous promise in child neurology. Unlike in other organs, which are highly permissive to adeno-associated viruses (AAV), the mature blood-brain barrier (BBB) greatly limits circulating AAV distribution to the brain. Intrathecal administration improves distribution but to no more than 20% of brain cells. Focused ultrasound (FUS) opens the BBB transiently and safely. In the present work we opened the hippocampal BBB and delivered a Cas9 gene via AAV9 intrathecally. This allowed brain first-pass, and subsequent vascular circulation and re-entry through the opened BBB. The mouse model used was of Lafora disease, a neuroinflammatory disease due to accumulations of misshapen overlong-branched glycogen. Cas9 was targeted to the gene of the glycogen branch-elongating enzyme glycogen synthase. We show that FUS dramatically (2000-fold) improved hippocampal Cas9 distribution and greatly reduced the pathogenic glycogen accumulations and hippocampal inflammation. FUS is in regular clinical use for other indications. Our work shows that it has the potential to vastly broaden gene delivery or editing along with clearance of corresponding pathologic basis of brain disease.

Introduction

The mammalian brain orients the organism to food, mating and safety, and in humans is the seat of sentience. These functions require stable cellular connectivity. As a result, brain cells are for the most part and most of life non-replicating. This advantages the brain for gene therapy with non-genome-integrating plasmids. Cellular longevity necessitates enhanced protection, which is afforded by the blood-brain barrier (BBB). The BBB disadvantages the brain for gene therapy. Intravenously (IV) administered viral vector overwhelmingly transduces other organs (e.g., biodistribution in the liver is 100-fold greater than in the brain [1]) despite the brain’s comparable vascularization. Vector administration to the cerebrospinal fluid (CSF) circumvents the BBB, requires a much lower dose, affords first-pass advantage to the brain and is comparatively immune-privileged [1-4]. Notwithstanding, possibly because CSF does not contact most brain cells directly, transduction efficiency with this approach remains limited (e.g. <20% with adeno-associated virus-9; AAV9), though clearly superior to IV administration. A large portion of viral particles administered to the CSF do not enter brain cells and exit into the bloodstream [2-4], wherefrom any brain re-entry is impeded by the BBB. Focused Ultrasound (FUS) can safely and transiently open the BBB and improve gene therapy delivery to targeted brain regions [5-8]. During the open period brain tissue should be no different than other organs in virus admissibility. In the present work, we combine CSF-administered gene therapy with opening the BBB with FUS to supplement the advantages of the former with unimpeded viral re-entry through an opened BBB.

We perform the above experiment in a mouse model of Lafora disease (LD) (LKO; Epm2a^−/−). LD is caused by mutations in the EPM2A or EPM2B genes [9, 10], which respectively encode the glycogen phosphatase laforin and its interacting E3 ubiquitin ligase partner malin. Though underlying mechanisms remain unclear, a principal function of the laforin-malin complex is the regulation of glycogen chain lengths. Absence of either protein results in overlong-branched glycogen, which gradually precipitates, aggregates and accumulates into ever-enlarging inclusions (Lafora bodies; LBs) in neurons and astrocytes across the brain. This drives a neuroinflammatory and neurodegenerative process that manifests in insidious teenage onset and subsequent progressive intractable epilepsy and dementia that culminate in death usually within a decade [11]. Therapeutically, it was reasoned that since overlong glycogen branches appear to be the pathogenic basis of the disease, downregulating glycogen synthase, the enzyme that synthesizes glycogen chains, might be beneficial. Effectively, crossing Epm2a^−/− or Epm2b^−/− mice with mice lacking or with reduced glycogen synthase activity prevented formation of overlong-branched glycogen and LBs, and rescued murine LD [12-15]. This is being followed by promising therapeutic developments of antisense oligonucleotide [16], virally delivered miRNA [17] and virally delivered CRISPR-Cas9 [18] targeting of the glycogen synthase Gys1 gene. To this last approach in the present work, we add focused ultrasonic opening of the BBB and obtain a massive enhancement of LB elimination. The work serves as proof of principle for correcting the pathological basis of a neurological disease through ultrasonically enhanced gene therapy.

Results

FUS Enhances Intrathecally administered AAV-delivered Cas9 distribution in the Hippocampus

The viral vector utilized in this study (AAV-SaCas9) has been described previously. Its transgene, packaged in AAV9, comprises the genes for Staphylococcus aureus Cas9 driven by the JetI promoter and a murine Gys1-targeting guide RNA driven by the U6 promoter (Fig. 1a) [18].

Fig. 1.

Viral vector structure; FUS system and procedures. a. Organization of vector transgene components [18]. b. Focused Ultrasound (FUS) system. c. Mice are fixed to a stereotactic FUS system and lambda and bregma are identified after opening the scalp to expose the skull. d. System software registers the Allen Brain atlas to the skull coordinates. The locating pointer is then replaced with the FUS transducer. Brain regions to be targeted can then be selected in the software (white arrow). After an intravenous infusion of microbubbles, FUS is delivered across an intact skull onto a specific brain region. e. BBB opening is confirmed by enhancement seen from gadolinium leakage on T1 weighted MRI images (white arrow).

One of the major challenges in brain-targeted gene therapy is achieving efficient transduction. AAVs are currently the leading vectors for gene therapy; however, even with direct intra-CSF administration, their transduction efficiency remains limited. To overcome this limitation, we utilized FUS to transiently open the hippocampal BBB. To evaluate the effect of FUS on the extent of viral transduction, we injected Epm2a^−/− mice intrathecally in the lumbar region with the above AAV-SaCas9 vector at postnatal day 21, followed by FUS directed to the hippocampal BBB in the right hemisphere (Fig. 1b-e). The left hippocampi were not treated with FUS and served as controls. Mice were sacrificed at 3 months of age and SaCas9 mRNA distribution in each of the hippocampal regions was measured following in situ hybridization (RNAscope) with a probe targeting the Cas9 mRNA. In all mice, the FUS-treated right hippocampi showed massively, on average approximately 1200-fold, increased Cas9 mRNA expression and distribution compared to the FUS-untreated left hippocampi (Fig. 2).

Fig. 2.

Massive increase in SaCas9 expression and distribution in FUS targeted hippocampi. a. Representative Epm2a^−/− mouse brain sections from both right (FUS treated) and left (control) hemispheres stained with a SaCas9 RNA probe using RNAscope. Small red puncta represent SaCas9 mRNA. Insets show enlarged regions for clarity. Scale bar is 2 mm. b. Quantification of SaCas9 mRNA positive area of hippocampal region (n = 4). Data is presented as mean ± SEM. ** p < 0.01

FUS enhances Cas9-Mediated Reduction of GYS1 mRNA and Protein

To determine the effect of FUS-complemented Gys1-targeting of SaCas9 on Gys1 mRNA, we measured the latter using RNAscope with a murine Gys1 specific probe. Gys1 mRNA signal in the FUS-targeted right hippocampi was fourfold lower than in the FUS-untreated left side (Fig. 3a, b). To quantify the effect on GYS1 protein, we performed immunofluorescence microscopy using a GYS1-specific antibody. GYS1 signal in puncta in the FUS-targeted right hippocampi was likewise approximately fourfold lower than in the FUS-untreated left side (Fig. 3c, d).

Fig. 3.

Reduction of glycogen synthase mRNA and protein expression with combined SaCAS9 and FUS. a. Representative Epm2a^−/− mouse brain sections from both right (FUS treated) and left (control) hemispheres subjected to RNAscope analysis using a specific Gys1 mRNA probe. Insets show enlarged regions for better clarity. Black arrows point to light red puncta indicating Gys1 mRNA. b. The numbers of red puncta (Gys1 mRNA signals) per hippocampal region are quantified (n = 3). c. Representative images of IF stained Epm2a^−/− mouse brain sections for GYS1 protein. Insets show enlargement of the boxed regions for better clarity. d. Quantification of GYS1 signals. All Scale bars are 2 mm. Data is presented as mean ± SEM. * p < 0.05.

FUS enhances Cas9-Mediated Reduction in LB Formation

The primary pathology in LD is the formation of overlong-branched insoluble glycogen precipitating and accumulating into ever-enlarging LBs [19]. In the present experiment, Gys1, the gene encoding the enzyme that elongates glycogen branches is targeted by the Cas9 enzyme.

In the LD mouse models, brain LBs are already present and widespread at 3 months of age. To measure the effect of FUS on the reduction of LBs by AAV-SaCas9, we stained sections from paraffin-embedded hemispheres with the LB stain periodic acid-Schiff following diastase pretreatment (PASD) and quantified PASD signal per hippocampal area. FUS-applied right hippocampi showed 50% less LB accumulations compared to the left FUS-untreated hippocampi (Fig. 4).

Fig. 4.

The effect of AAV-SaCas9 on LB levels is highly enhanced by FUS. a. Representative images from PASD-stained hippocampi from right (FUS treated) and left (control) hemispheres are shown. Scale bar is 0.1 mm. b. LB quantification in the hippocampal region (n = 12). Data presented as mean ± SEM. *** p < 0.001

In healthy mice GYS1 protein is diffusely distributed in tissues and usually not detectable with standard immunofluorescence. In LD mice GYS1 concentrates at LBs, where it is distinctively observable and quantifiable [20]. Thus, the GYS1 punctal protein quantification in the preceding section is itself also a sign of the much greater reduction in LBs on the FUS-treated versus untreated side.

The AAV-SaCas9 - FUS Combination Induces Modest Microgliosis

One of the chief concerns with Cas9 based therapeutics is the potential immune response to the foreign bacterial protein. To assess for effects of FUS-based enhancement of AAV-SaCas9 delivery on neuroinflammation, we quantified the two most commonly used neuroinflammatory markers, GFAP for astrocytes and Iba1 for microglia, by immunohistochemistry. There was no increased astrogliosis with FUS-enhanced Cas9 delivery, but there was modest microgliosis (Fig. 5).

Fig. 5.

Combined AAV-SaCas9 and FUS applications mildly increase microgliosis. a. Representative IHC images for GFAP and d. Iba1 in the hippocampal region are shown. Scale bar is 0.5 mm. b, c. Quantification of GFAP and e, f. Iba1 signals (n = 12). Data is presented as mean ± SEM. * p < 0.05. ns denotes non-significant.

FUS-Mediated BBB Opening Reduces Neuroinflammation

In the above experiments, it was not possible to track the main effector of the observed reduction in LB formation. In other words, because FUS and AAV-SaCas9 were combined, we could not rule out that it was the FUS, and not the SaCas9, that resulted in the reduced LBs. We therefore conducted a second study where FUS was applied to the right hemispheres with no intrathecal AAV-SaCas9 administration. No difference in hippocampal LBs was observed between the FUS-applied right versus FUS-untreated left hemispheres (Fig. 6a-c), indicating that FUS alone has no effect on LB formation.

Fig. 6.

FUS application alone does not impact LB quantities but improves neuroinflammation. a. Representative images from PASD stained brain sections. Scale bar is 0.1 mm. b, c. LB quantification in the hippocampal region (n = 6). d. Representative IHC images for GFAP and g. Iba1. Scale bar is 0.5 mm. e, f. Quantification of GFAP and h, i. IBA1 signals (n = 6). Data is presented as mean ± SEM. p < 0.01. ns denotes non-significant.

Interestingly, there was a statistically significant reduction in astrocytosis and a trend toward the same in microgliosis in the FUS-applied right hippocampi (Fig. 6d-i). These results suggest that FUS alone reduces LB-associated brain neuroinflammation. They also suggest that the increased microgliosis with the combination of FUS and AAV-SaCas9 was likely related to Cas9.

Discussion

The brain utilizes the lion’s share of the genome’s genes for its development and function. Accordingly, there are several thousand monogenic neurological diseases, individually rare but collectively exceeding the commonest diseases. At the same time, its chief cells being mostly non-replicating, the brain is prime candidate organ for ‘one-and-done’ gene replacement or editing therapies. A major obstacle, however, is the protective barrier unique to the brain (BBB), which prevents taking full advantage of the circulatory capillary system that reaches every individual cell. AAV vectors are able to cross the BBB but with low efficiency, greatly limiting the number of cells transduced through an intravenous delivery route. Delivering the viral vector through the CSF circumvents the BBB, but does not provide direct access to individual brain cells. This approach has several advantages, including requiring much lower doses, and affording a first pass to the target organ, but improves the extent of vector product distribution only modestly. Overall, brain gene therapy at the present time, through any route, cannot access more than 20% of brain cells.

In the future, gene editing, where applicable, is expected to supersede gene replacement, because correction through editing happens in the target gene’s natural genomic milieu, retaining the regulated aspects of its expression. However, delivery of genome editing machinery, such as Cas9, faces even greater challenges than disease gene delivery. Among these is the size of these enzymes. For standard gene replacement therapy, packaging the gene in the virus in self-complementary arrangement (essentially two copies of the gene in tandem) greatly enhances efficiency [21]. Cas9 enzymes are too large for self-complementary packaging in AAV viruses. SaCas9 is among the smallest Cas9’s, and can be packaged, along with a guide RNA, in non-self-complementary fashion.

In the present work we asked whether transiently and focally opening the BBB enhances gene delivery. We chose the LD model, because it permits demonstration of a step beyond enhanced delivery, namely a functional impact on causative pathology of the disease, namely on LBs. We chose to deliver Cas9 (as opposed to another protein) as a higher bar than gene replacement, toward proof of principle of potential utility of BBB opening in the armamentarium of future genome editing therapies.

FUS dramatically enhanced delivery of the Cas9 gene to the BBB-opened hippocampus, as evidenced by the RNAscope results, and Cas9 function as shown by reductions in the expression and translation of the target gene (Gys1) and in LB formation.

The causation of neurodegeneration in LD is multifactorial, one important component being inflammatory, evidenced by expanding astrogliosis and microgliosis that tracks with the progressively widening field of LBs in the course of the disease [19, 22]. Cas9 itself is immunogenic [23] as is FUS at unduly high pressures [24]. In the present work we applied FUS at low pressures, which, combined with circulating microbubbles, transiently disrupts the BBB without tissue damage or persisting inflammation [7, 24]. FUS itself did not lead to any astrogliosis or microgliosis. There was a small degree of microgliosis when FUS was combined with AAV-SaCas9. Most likely, this microgliosis was a response to the relatively very large number of virus particles and their cargo of bacterial protein gene delivered to the FUS-opened side.

Our and other transcranial-focused ultrasound laboratories are continuously refining the safe and effective dosing of FUS, while at the same time developing methods to open the BBB briefly, transiently and consecutively, region after region, to cover ever broader volumes of the brain. Current automated methods in development are beginning to allow the consecutive opening of the BBB across most of the brain over periods of time [5] during which AAV9 circulates in blood following infusion (around 24 hours) [25]. It is hoped that the technology will continue to improve, making consecutive brain regions transiently and briefly barrierless for broad corrections with gene replacement or editing.

Transcranial FUS is in clinical use for the treatment of essential tremor and Parkinson’s disease through thermal ablation of target white matter tracts [26-29]. Investigational studies are ongoing to establish the safety and effectiveness of transcranial FUS for transient opening of the BBB to both enhance the delivery of molecular agents into the brain and to also release brain tissue components into the blood for remote sampling [30-33].

The primary mechanism for increased AAV delivery is through enhanced delivery to the brain. The stimulation of intravascular microbubbles with ultrasound mechanically stimulates the endothelium of brain microvasculature leading to transient opening of the BBB. During this period of a few hours, molecules up to the size of viral vectors can diffuse at a much higher rate than in unexposed brain. Other than some indication of a molecular effect on the endothelial cells of the brain, it is assumed there is no direct molecular interaction with cells in the brain parenchyma since the microbubbles are too large to extravasate into this region [34].

Advantages of transcranial FUS for molecular delivery to the brain include non-invasiveness, focal targeting, large volume coverage, and the ability to perform repeat exposures. In regards to gene therapy, an important advantage is that it does not require new delivery vehicle development, but rather simply facilitates delivery of a vector (AAV) with existing long track record of safety and efficacy. Other methods to deliver genetic cargo broadly to the brain include development of new viruses that distribute more widely than AAV9. Multiple laboratories are working along those lines, usually to mutate or force the evolution of the AAV capsid genes to generate virions with greater transduction efficiency [35-37]. Others are including epitopes in the viral capsids that interact with receptors (e.g. transferrin receptor) at the BBB endothelium to translocate virus across the barrier [38, 39]. Finally, forgoing viruses altogether, large efforts are underway to artificially engineer transport vehicles (e.g. liposomes), with or without endothelial cell receptor interacting components [40-42].

The transient opening of the BBB application is under investigation for Alzheimer’s disease with which LD has historical and neuropathological (brain accumulations, neuroinflammation, neurodegeneration) associations. When Alzheimer turned down the United States government’s effort to recruit him, he sent his two Spanish trainees Achúcarro and Lafora instead who, working in the then ‘Government Hospital for the Insane’ near Washington D.C., described the first, third and fifth American cases of Alzheimer’s disease [43]. Lafora described what would become his eponymous disease during that same stay [44]. For many years the brain accumulations were thought to be amyloid in both diseases (of course in Alzheimer’s they subsequently proved to be protein), and LD was considered a juvenile variant of Alzheimer’s. Fast forward to the present, amyloid plaques of Alzheimer’s disease are being treated with intravenous antibodies, and recent inclusion of FUS opening of the BBB is dramatically enhancing amyloid clearance (32 versus 6 %) [45]. Possibly, emerging gene [46] and antibody-based [47] therapies for LD will likewise benefit from inclusion of FUS BBB opening.

Methods

Mice and Viral vector and administration

The Epm2a^−/− LD mouse model was described previously [48]. All procedures were carried out according to NIH guidelines and the Institutional Animal Care and Use Committee regulations at the University of Texas Southwestern Medical Center. 21 day-old mice were injected with 7×10¹¹ vector genomes of AAV-SaCas9 in 5μl PBS (or PBS alone) intrathecally in the lumbar region prior to FUS application. Construction and production of AAV-SaCas9 (Fig. 1a) were described previously [18].

Focused ultrasound

FUS was performed as described previously using the stereotactic guided RK-50 system (FUS Instruments Inc.) [7]. Briefly, mice were sedated with isoflurane, and a 26-gauge intravenous catheter (SuperCath 5, 26G Safety IV catheter; ICU Medical, Inc.) assembly with a dead volume of 69 μl was placed in the tail vein for bubble administration and affixed with tissue glue. A physiologic monitoring system (PhysioSuite, Kent Scientific Corp.) was used to monitor respiration and rectal temperature throughout the experiment, while a heating pad under the animal was used to maintain body temperature. Hair over the cranial surface of the skull was removed using an animal clipper and depilatory cream VEET sensitive formula, Reckitt Benckiser, Parsippany, NJ, USA). Carbrofen (0.1 ml/25g) and buprenorphine SR (0.5-1.0 mg/kg) were administered. A 1-2 cm incision was made over the skull to allow for visualization of the bregma and lambda sutures, and the two points were registered on the stereotaxic module in the RK-50 software. Once the skull landmarks were established, ultrasound gel was added to the animal’s skull for acoustic coupling, taking care to avoid air bubbles. A tank filled with deionized and degassed water was lowered onto the ultrasound gel to a fixed point coincidental with the skull, and the transducer was lowered into the tank. The mouse was sedated with 3 L/min 70% Nitrous Oxide mixed with 30% Oxygen with 1.5-2% isoflurane. Microbubbles (DEFINITY^®) were activated using the shaking amalgamator provided by the manufacturer for 45 sec for IV infusion and prepared in preservative-free saline per the manufacturers at 10⁹ microbubbles/ml and were infused at 50ul/min using infusion pump (Nanojet, Chemyx Inc, Stafford, TX, USA). After accounting for the dead volume of the catheter for the microbubbles infusion, the sonication was performed. The sonication target was the hippocampus. Three exposures were performed (see Fig. 1d). Each exposure consisted of 60 bursts with a pulse length of 10 ms and a repetition period of 1000 ms with 0.335 MPa of pressure. After ultrasound sonication, gadolinium-based MRI was used to confirm opening of the BBB. The animals were placed into a custom-built mouse bed. The bed was placed in a small animal 7.0 T (16-cm horizontal bore) magnetic resonance scanner (Bruker Biospec) with a 38 mm volume RF coil run on ParaVision software version 6.0.1 (Bruker Bio- Spin MRI GmbH). T1-weighted MR images were acquired before and after injection of gadolinium, Gadobutrol (Gadovist, Bayer Healthcare Pharmaceuticals Inc.) administered via tail catheter (1.0 mmol/kg) followed by a 70 μL saline flush. All images were acquired in the axial anatomical plane to capture lateral cross-sections in the focal zone of the FUS beam where BBB opening was targeted. After confirming the BBB was opened, the ultrasound gel was cleaned and the skin over the skull was approximated and closed with tissue glue.

RNA in situ hybridization (RNAScope)

RNAscope red chromogenic assay was performed by the UT Southwestern Metabolic Phenotyping Core to detect the transcript for SaCas9 in 5 μ thickness paraffin-embedded brain sections. Slides were deparaffinized and pretreated following the manufacturer’s instructions (Advanced Cell Diagnotics, Inc (ACD), California). Briefly, after boiling slides in a Target Retrieval solution, slides were rinsed in distilled water and dehydrated in 100% ethanol. Tissue was enzymatically digested at 40 °C for 15mins. Hybridization itself was performed following the recommended ACD procedure and reagents from the RNAscope^® 2.5 HD Detection Kit (RED; cat# 322360). The probes (Mm-SaCas9 cat# 501621) were applied at 40 °C for 2hrs. Amplification steps were done following the manufacturer's instructions. Lastly, chromogenic signal detection (red) was achieved using a mix of Fast RED-B and Fast RED-A in a ratio of 1:60 at room temperature for about 10 min. Slides were washed in distilled water and counterstained with hematoxylin before applying mounting medium (EcoMount, BioCare Medical) and a coverslip over the tissue section.

PASD, Immunohistochemistry (IHC), and Immunofluorescence (IF) staining

Brain tissues were fixed in 10% neutral-buffered formalin overnight and sagittally-grossed hemispheres of mouse brain were paraffin processed, embedded, and sectioned by members of UT Southwestern’s Histo Pathology Core according to standard procedures [49, 50]. Paraffin sections were concomitantly prepared and checked by low-magnification dark-field illumination [51] to ensure comparative parasagittal planes of anatomy were achieved at 1.08 mm and 1.56 mm lateral to midline according to Paxinos & Franklin [52]. The serial sections were stained by regressive hematoxylin & eosin (H&E) [49, 50], Periodic-Acid-Schiff with Diastase (PASD) [53], as well as IHC for GFAP and Iba1 using their commercially available specific antibodies (anti-GFAP, Biogenex, Fremont, CA, Cat# MU020-UC, 1:800, anti-Iba1, Fujiilm Wako Chemicals USA, Richmond, VA, Cat# 019-19741, 1:400). In brief, slides for GFAP and Iba1 were deparaffinized and run to water, followed by GFAP slides proceeding without antigen retrieval, and Iba1 slides proceeding with antigen retrieval in pH 6.0 citrate-buffer at 60°C for 20-hours. Sections for GFAP were blocked against endogenous mouse IgG and secondary antibody host-serum affinity utilizing commercially available blocking reagents (Vector Impress Mouse on Mouse “MOM” Polymer Kit, Vector Laboratories, Burlingam CA, Cat# MP-2400) and incubated overnight with primary antibody at 4°C. Sections for Iba1 were blocked against secondary antibody host-serum affinity utilizing commercially available blocking reagents (Vector Impress Horse anti-Rabbit Polymer Kit, Vector Laboratories, Cat# MP-7401). Subsequent detection of bound primaries was carried out with reagents and instructions from MOM and Horse anti-Rabbit kits. Finally, brown diaminobenzidine-chromagen (ImmPACT DAB EqV Substrate Kit, Vector Laboratories) was revealed at the location of bound primaries, nuclei were counterstained lightly with hematoxylin, the slides were then dehydrated, cleared and covered by cover slip and using synthetic mounting media. For GYS1 immunofluorescence staining, paraffin-embedded tissue sections were de-paraffinated and rehydrated by processing through xylenes and decreasing concentrations of ethanol in water. Then, sections were subjected to antigen retrieval using citrate buffer pH 6.0 (Sigma, Cat# C9999). After blocking with 5% normal donkey serum, sections were incubated with rabbit anti-Gys1 antibody (Abcam, Cat# ab40810) diluted in blocking solution (1:400) overnight at 4°C. Slides were then washed with PBS and successively incubated with Alexa Fluor 594 donkey anti-rabbit secondary antibody (ThermoFisher, Cat# A21207, 1:500) and DAPI.

Microscopy and Image analysis

RNAscope, PASD, and IHC stained slides were scanned using the Hamamatsu Nanozoomer 2.0 HT digital slide scanner using its 40x objective and IF stained slides were scanned by Zeiss Axioscan.Z1 digital slide scanner at 20x resolution. For SaCas9 mRNA (RNAscope), PASD and IHC stained slides, after defining positive signals based on pixel color, the % area of hippocampus covered with positive signal was measured and reported [16]. For IF stained sections, ImageJ was used; quantitation was same. For Gys1 mRNA (RNAscope), the signal was too faint for the Nanozoomer and the signal was counted by eye. In this case, values are expressed as object frequency per unit area.

Statistical analysis

Student’s unpaired t test was used to compare single means. All data analysis and graph preparations were done using the GraphPad Prism software (v. 8.0.2; GraphPad Software). Statistical significance was set at p < 0.05. Asterisks denote level of significance based on p value: *p < 0.05, **p < 0.01, ***p < 0.001, and ****p <0.0001.

Data availability

All raw data used to prepare this manuscript is available upon request from the corresponding authors.