Efficient genetic perturbation of murine sensory neurons in vivo using CRISPR/Cas9

Abstract

Gene editing using CRISPR/Cas9 in vivo offers a powerful tool to investigate pain mechanisms. We generated a conditional knock-in mouse model where Streptococcus pyogenes CRISPR-associated protein 9 (Cas9) expression is restricted to cells that express SCN9A. Transgenic markers were detected in key tissues including the dorsal root ganglia (DRG) and sciatic nerve. To assess in vivo editing efficacy, RNA guides targeting TRPV1 were intrathecally administered. Two injections of guide RNAs resulted in a significant reduction of TRPV1 in both the DRG and sciatic nerve without triggering caspase-3-mediated apoptosis or motor deficits. Edited animals exhibited increased withdrawal latencies to heat and reduced nocifensive behaviors following capsaicin injection. Capsaicin-evoked thermal hyperalgesia and mechanical allodynia were diminished. This approach enables rapid and efficient sensory neuron-specific CRISPR/Cas9 gene perturbations for pain research in mice. We envisage that this method can be employed both for the exploration of molecular mechanisms underlying nociception and for the validation of therapeutic targets associated with pain.

Introduction

Chronic pain is pervasive. It diminishes quality of life for more than 20% of the world’s population ¹. In the vast majority of cases, pain signals originate in specialized class of sensory neuron called a nociceptor. Their cell bodies reside in the dorsal root ganglia (DRG) and their fibers innervate internal organs, the viscera, and the skin. Nociceptors detect noxious thermal, mechanical, or chemical stimuli and transmit this information to the central nervous system. In chronic pain states, nociceptors become sensitized, leading to exaggerated responses to painful stimuli (hyperalgesia) and normally innocuous inputs (allodynia) ^2,3. Identifying genes that drive chronic pain through precise genetic changes is imperative for discovering analgesic targets ⁴. Although hundreds of genes have been linked to pain in mice ^5–7 and humans ^8–10, genetic validation remains extremely challenging.

Genome engineering of rodents is a cornerstone of modern pain research. Numerous strategies including whole body mutation, conditional deletion, and temporally controlled knockouts, have enabled tremendous insights into pain signaling ^11,12. However, these approaches are time consuming, laborious, and expensive. A more rapid approach involves viral transduction. This strategy enables precise delivery of genetic cargo through intraganglionic injection of adeno-associated virus (AAV) vectors or retrograde transport of AAVs applied to the sciatic nerve ^13–15. However, viral strategies have limitations that include low and variable transduction efficiencies, transient expression, immune activation, and the limited packaging capacity ¹².

CRISPR/Cas9 genome modification provides a precise and flexible approach to gene editing. SpCas9 nuclease is recruited to genomic targets via a programable non-coding RNA ^16–18. When the guide RNA-SpCas9 complex is bound to target DNA sequence, SpCas9 induces double-strand breaks that are repaired through the non-homologous end joining (NHEJ) pathway. This results in small insertions and deletions (indels) that usually result in frameshifting and premature termination of translation. In the presence of a donor DNA template, homologous recombination (HR) can be used for genome engineering ¹⁹. Cas9 delivery via plasmid vectors or as a complex with preloaded guide RNAs has been reported in DRG sensory neurons ^{20–23 24 25}. The major goal of this work was to extend on this work by asking if highly efficient editing can be achieved through the use of stable transgenic expression of Cas9 in sensory neurons.

As a proof of concept, we targeted the transient receptor potential vanilloid 1(TRPV1). We demonstrate a procedure for rapid and facile genome perturbations within a subset of sensory neurons in mice. We envisage that our optimized strategy and genetic tool will be broadly applicable for the field.

Materials and methods

Mice

All animal procedures were conducted in accordance with approved protocols from the Institutional Animal Care and Use Committee (IACUC) at the University of Wisconsin-Madison (Protocol ID: M006582-A07). Rosa26-LSL-Cas9 knock-in (Cas9^fl/+) mice were obtained from Jackson Laboratory (strain #024857). These mice express CRISPR-associated protein 9 (Cas9) endonuclease and EGFP in a Cre recombinase-dependent manner, under the control of a CAG promoter ³⁰. Scn10a-Cre (strain 036564) was used to generate the conditional Cas9 knock-in mice ³¹. Then, Cas9^fl/fl/Scn10aCre⁺, Cas9^fl/fl and wild type C57BL/6J mice were obtained from our breeding facilities (Clinical Science Center, UW-Madison). All mice were housed under a 12-hour light/dark cycle with controlled ambient temperature (20–26°C) and humidity (30–70%). Food and water were provided ad libitum. Groups consisted of both male and female mice aged 8–12 weeks old (25–30 g) were used in this study. Mice were randomly assigned to either non-targeting control (NTC, 3 μg) or GFP (0.5–3 μg) guide RNAs, while a separate cohort was randomly assigned to NTC or TRPV1 RNA guide treatment. To minimize potential confounding factors, all animals were identified by their tag number and home cage. No group contained more than three mice from the same litter or the same cage. Cage positions within the rack were regularly rotated to avoid environmental or positional bias.

Guide RNA design

The list of equipment needed for the production and validation of guide RNAs is provided in Table 1. The guide RNA synthesis protocol described here was based on that described by Dewitt et al. with minor modifications ³². For each target gene, three guides were designed against the gene of interest to ensure sufficient depletion (Fig. 1A). Candidate target (protospacer) sequences were identified with CHOPCHOP ³³ using the GRCm38 genome assembly with the CRISPR/Cas9 and knockout settings. Protospacer sequences were selected on the basis of high predicted activity and specificity. The 20 nucleotide protospacer sequences were added to the T7FwdVar sequence (without the PAM) in place of the Xs in the T7FwdVar sequence (Table 2). The guanine upstream of the target sequence is required for efficient T7 transcription. If the target sequence begins with a guanine, the upstream guanine can be omitted.

Table 1.

Key equipment, kits, and chemicals

Name	Manufacturer	Cat No
Thermal cycler	Thermo Scientific	4483636
Tabletop centrifuge	Eppendorf	Centrifuge 5425R
Milli-Q Reference	EMD Millipore
3Autoclave	Tuttnauer
Nanodrop OneC	Thermo Scientific
RNA Clean & Concentrator-5	Zymo Research	R1014
GoTaq Green Master Mix	Promega	M7123
HiScribe T7 High Yield RNA Synthesis Kit	New England Biolabs	E2040L
DNase I, RNase-free	Thermo Scientific	EN0521
Quick CIP	New England Biolabs	M0525S
Bleach	Clorox
RNase Away	Molecular Bioproducts	7003
DEPC	Sigma-Aldrich	40718-100ML
Ethanol	Sigma-Aldrich	1.08543.0250
Agarose	Research Products International	A20090-500.0
GeneRuler 1 kb Plus DNA Ladder	Thermo Scientific	SM1331
50x TAE Buffer	Fisher Scientific	BP13321
Ethidium Bromide	Fisher Scientific	BP102-5
TriTrack DNA Loading Dye (6x)	Thermo Scientific	R1161
DNA Oligonucleotides	IDT

Figure 1.

Overview of sgRNA synthesis and in vivo delivery. (A) Schematic of sgRNA association with genomic DNA. (B) The DNA template for sgRNA synthesis comprises a T7 promoter, the gene-specific sgRNA sequence, and elements essential for Cas9 function. For each target gene, three distinct sgRNAs were used. (C) DNA templates were synthetized by PCR using overlapping primers. sgRNAs were produced by in vitro transcription (IVT). Following DNase digestion of template and 5´-triphosphate removal, RNA-lipid complexes were generated prior to intrathecal delivery.

Table 2.

DNA oligonucleotides used in this study

Name	Sequence
T7FwdVar	GGATCCTAATACGACTCACTATAGXXXXXXXXXXXXXXXXXXXXGTTTTAGAGCTAGAA
T7RevLong	AAAAAAGCACCGACTCGGTGCCACTTTTTCAAGTTGATAACGGACTAGCCTTATTTTAACTTGCTATTTCTAGCTCTAAAAC
AmpFwd	GGATCCTAATACGACTCACTATAG
AmpRev	AAAAAAGCACCGACTCGG
GGP010_NTC	GGATCCTAATACGACTCACTATAGCTCAGTTCCAGTACGGCTCCAGTTTTAGAGCTAGAA
GGP011_trpv1_1	GGATCCTAATACGACTCACTATAGACGGGATTCGGTGGGCAACAGTTTTAGAGCTAGAA
GGP012_trpv1_2	GGATCCTAATACGACTCACTATAGCCCCAAGGCTCTATGATCGCGTTTTAGAGCTAGAA
GGP013_trpv1_3	GGATCCTAATACGACTCACTATAGGCTGCGTGCACCAACCAGCGGTTTTAGAGCTAGAA

Production of guide RNA pools

Prior to synthesis of guides, it is critical to mitigate the risk of contamination. RNase-free reagents, DEPC (diethyl pyrocarbonate)-treated water, and barrier pipette tips were used. To inactivate RNases with DEPC, 0.1% DEPC solution was prepared in ultrapure water and incubated overnight at 37°C. The solution was autoclaved with a sterilization time of at least 30 minutes. Benchtops and tube racks were decontaminated with 10% bleach for 15 minutes. The surfaces of electronic equipment were wiped down with RNase-away. Pipettes were disassembled and decontaminated with RNase-Away after soap, water, ethanol, and bleach washes.

DNA templates for in vitro transcription (IVT) were created by PCR. The T7FwdVar and T7RevLong oligos served as the template for the full-length sgRNA (Table 2, Fig. 1B). FwdAmp and RevAmp prime amplification of the full-length template during subsequent rounds of PCR. GoTaq Green was added to a final volume of 25 μl for every individual guide (Table 3, Fig. 1C). The cycling parameters and extension times are provided in Table 4. The product was a 125 bp amplicon. To determine if a guide was amplified, products were visualized on a 2% TAE agarose gel (Fig. 2A).

Table 3.

DNA template PCR conditions

Component	Concentration
GoTaq Green MM	1x
T7FwdVarV2	40 nM
T7RevLongV2	40 nM
T7FwdAmp	400 nM
T7RevAmp	400 nM

Table 4.

sgRNA Template PCR Program

Temperature (C)	Duration	Cycles
98	1m	1
98	15s
50	15s	30
72	15s
72	2m	1

Figure 2.

Validation of sgRNA synthesis. (A) Amplification of DNA templates (125 base pairs (bp)) for the targeted gene (TRPV1) or a non-targeting control (NTC) visualized on an agarose gel. (B) In vitro transcription (IVT) products corresponding to sgRNAs were visualized on a bleach-treated agarose gel. The presence of bands under ~100–200 bp indicates successful sgRNA synthesis. Bands were observed in the absence of T7FwdVar primer or in vitro transcription.

Once suitable templates were created, the next step was in vitro transcription of RNA. Reagents from the HiScribe T7 High Yield RNA Synthesis kit – specifically the 10x reaction buffer and 100 mM nucleotide stocks were placed on ice for 1 hour. Next, a 2x IVT master mix was generated through the addition of equal volumes of 10x reaction buffer and 100 mM nucleotide stocks (Table 5). Individual reactions of 50 μl were prepared by mixing 23 μl of template PCR reaction, 23 μl of 2x IVT master mix, and 4 μl of T7 RNA Polymerase mix. Transcription was conducted at 37°C overnight.

Table 5.

2x In-vitro transcription master mix composition

Component	Concentration
T7 Reaction Buffer	2x
ATP	20 mM
CTP	20 mM
GTP	20 mM
UTP	20 mM

To degrade the DNA template, 2 U DNase I was added to each reaction and incubated for 15 minutes at 37°C. Next, RNA was purified using the Zymo RNA Clean and Concentrator-5 kit according to the manufacturer’s instructions with minor modifications. To reduce inflammation invoked by exogenous RNA carrying 5’-triphosphate groups, single guide RNAs (sgRNAs) were incubated with ten 10 units of calf intestinal alkaline phosphatase (Table 6, Fig. 1C) at 37°C for 3 hours. After the phosphatase reaction, sgRNAs were repurified with the Zymo RNA Clean and Concentrator kit, as described above. RNA was eluted in a total volume of 14 μl to concentrate the sample.

Table 6.

Quick CIP reaction conditions

Component	Volume (μL)
10x Cutsmart Buffer	4
sgRNA	34
Quick CIP (5U/μL)	2
Total	40

Validation of guide RNAs

RNA quantification was conducted using a NanoDrop spectrophomoteter. This protocol produces highly concentrated RNA (~5–10 μg/μL). To obtain accurate concentration measurements, samples were diluted 1:50 in DEPC-treated water before absorbance measurements were collected. To confirm the approximate size of the RNA, bleach gel electrophoresis was conducted ³⁴. The gels are made with 1x TAE buffer to 2% (w/v) agarose and 1% (v/v) bleach. Bleach is allowed to react for 15 minutes prior to melting the agarose. Ethidium bromide at 0.5 μg/mL was added immediately prior to gel casting. A total of 100–500 ng of RNA was diluted in 10 μL in 1x TriTrack loading dye in DEPC-treated water and loaded onto the gel. Because these are not denaturing conditions, this gel simply confirms the presence of small (<200 nt) RNA molecules (Fig. 2B). More precise measurements can be collected on an Agilent TapeStation system if needed.

Generation of guide RNA polymer complexes and intrathecal injection

After quality-control checks, individual sgRNAs targeting the same gene were pooled at equal concentrations. The final concentration of the sgRNA pool was 2–3 μg/μL. The remaining RNA was aliquoted and can be stored at −80°C for up to one month.

For intrathecal administration, the guide RNA was diluted with mRNA buffer to a final volume of 2 μL and mixed with 3 μL of in vivo-jetRNA^®+ reagent (101000122, Polyplus). The resulting 5 μl of complex was loaded for injection using a Hamilton syringe equipped with a 31 G needle. A new needle was prepared for each mouse to avoid RNase contamination. Syringes were cleaned and sterilized between injections involving different guide RNA complexes. Mice were anesthetized using isoflurane at a concentration of 2–3% via a nose cone. Anesthesia depth was verified by the absence of a response to the tail or hind paw pinching. The lower lumbar region was shaved and disinfected with 70% ethanol. The injection site was identified between the L4–L5 vertebrae ³⁵. The intrathecal injection was confirmed by a characteristic tail flick and the RNA complex volume was delivered slowly. Mice were returned to their home cages and monitored until full recovery from anesthesia, ensuring they were fully ambulatory and responsive. All procedures were conducted in a clean and decontaminated workspace.

Immunohistochemistry

As editing targets, we examined either a transgenic GFP or endogenous TRPV1 using immunohistochemical analysis. Lumbar DRG, sciatic nerve, spinal cord or trigeminal ganglia (TG) were obtained from mice treated with guide RNAs targeting either: GFP (0.5–3 μg, a single intrathecal injection), TRPV1 or a non-targeting control (NTC) (3 μg, a single or double intrathecal injection). Mice were briefly anesthetized with isoflurane and transcardially perfused with ice-cold PBS followed by 4% formaldehyde (in PBS, pH=7.4). Tissues were harvested and post-fixed overnight at 4°C with 4% formaldehyde. Then, tissues were cryoprotected in 30% sucrose for 3 days at 4°C. Tissues were embedded in Tissue-Tek OCT compound (Sakura Finetek, Torrance, CA) and cryosectioned at 14 μm for DRG and trigeminal ganglia, and at 20 μm for sciatic nerve and spinal cord. Tissues sections were washed twice with TBS to remove OCT compound and then, incubated with 10 mM sodium citrate buffer (10 mM sodium citrate, 0.05% Tween 20, pH 6) for 1 hour. Tissues then were blocked with 10% normal goat serum, 0.3% Triton X-100 and goat anti-mouse Fab fragment (1:1,000, 115–001-003, Jackson ImmunoResearch) in TBS for 2 hours at room temperature using a humidified chamber. Sections were incubated overnight at 4°C with chicken anti-GFP (1:1,000, GFP-1010, Aves Labs), rabbit anti-GFP (1:200, 2956s, Cell Signaling Technology), mouse anti-Nav1.8 (1:1,000, 75–166, NeuroMab), chicken anti-peripherin (1:1,000, NBP1–05423, Novus Biologicals), rabbit anti-βIII-tubulin (1:1,000, 802001, Biolegends), goat anti-TRPV1 (1:500, GP14100, Neuromics), rabbit anti-caspase-3 (1:200, 9661, Cell Signaling Technology), rabbit anti-TRPA1 (1:200, PA1–46159, Invitrogen) or rabbit anti-CGRP (1:200, ab81887, Abcam) antibodies in 1% normal goat serum and 0.3% Triton X-100 as incubation solution. Sections were washed three times with 0.1%Triton X-100 in TBS and incubated with the corresponding secondary antibody (1:1,000, in incubation solution) or isolectin GS-IB4 (1:300, I21412, Invitrogen) for 1 hour at room temperature. Sections were washed and DAPI (1:10,000 in TBS) was incubated for 15 min followed by three more washes. Finally, coverslips were mounted on the slices using ProLong^™ Glass Antifade (P36984, ThermoFisher Scientific).

Confocal imaging and image analysis

Images were captured on a Leica SP8 laser scanning confocal microscope with LAS X Life Science software. Lumbar spinal cord sections were imaged using a 4x objective. Lumbar DRG, sciatic nerve and trigeminal ganglion sections were imaged using 20X objective or 2.5x zoom magnification. Imaging was performed with a pinhole of 1 Airy and scan speed of 400 Hz, using sequential acquisition to minimize spectral overlap. Images are representative projections of Z-stacks obtained from samples collected from 3–4 mice per experimental condition. Fluorescence intensity between treatments was quantified using ImageJ software by calculating corrected total cell fluorescence (CTCF) values. For DRG analysis, regions of interest (ROIs) were drawn to measure the TRPV1 GFP, or TRPA1 fluoresce intensity specifically in GFP or Nav1.8 positive cells, respectively. For sciatic nerve, ROIs were created to measure the area of fibers in each fluorescence channel. Colocalization analysis was then performed to determine the area of overlap between markers in the sciatic nerve. For caspase-3 analysis, the number caspase-3-positive puncta was divided by the total number of GFP-positive cells and multiplied by 100 to calculate the percentage of positive cells. Results were calculated as either the mean of CTCF values or percent of at least 200 positive cells.

Capsaicin-induced inflammatory pain model and behavioral tests

Intraplantar injection of capsaicin (5 μg/paw) elicits a rapid nocifensive response, characterized by flinching or licking of the injected paw. Therefore, we used capsaicin-induced inflammatory pain model to determine if delivery of TRPV1 guide RNAs affects nocifensive behaviors ^36–38. Mice were habituated to clear acrylic behavioral chambers for 1 hour per day over 2 days prior to the experimental days. Then, 20 μL of capsaicin stock solution (1 mg/mL, diluted in a solution containing 10% ethanol, 5% Tween-80 and 85% saline) were injected into the plantar surface of the left paw. The number of flinches was recorded during the first 10 minutes post-injection.

Forward looking infrared (FLIR) imaging was used to visualize thermal changes of the injected paw with capsaicin, using a FLIRT T31030sc thermal imaging camera (FLIR Systems, Wilsonville, OR) ^38,39. Mice were placed in acrylic boxes with wire mesh floors and allowed to acclimate 2 hours daily during the two days prior to testing, and for 30 min on the test day before the experiment. Mice were intraplantarly injected with 5 μg of capsaicin and placed into clear acrylic chambers boxes with wire mesh floor. The temperature of the medial plantar surface was registered twice, 1 h after capsaicin injection. The average temperature was then calculated and plotted.

To assess heat sensitivity, the hot plate test and Hargreaves test were employed ^40,41. Mice were habituated to the room for 1 hour daily on the two days prior to testing, and for 30 min before the start of the experiment. Then, mice were placed on a hot plate maintained at 50°C and the time taken to observe nocifensive behavior was recorded. Nocifensive behaviors were considered when mice showed hind paw withdrawal or licking, stamping, leaning posture, or jumping. The Hargreaves test was performed to measure paw withdrawal latency in response to a focused radiant heat source. Mice were placed on a glass floor within an enclosed chamber, and a high-intensity infrared light beam was directed at the plantar surface of the left hind paw. The light intensity was set at 25–30% of maximum output, as per the manufacturer’s guidelines (model 390, IITC). Paw withdrawal latency was recorded in triplicate, with a minimum interval of 10 minutes between trials to minimize experimental variability. The cutoff time was 20 seconds for both thermal tests. The average latency was then calculated and plotted.

Mechanical hypersensitivity was assessed using calibrated von Frey filaments (Stoelting, Wood Dale, IL, USA) to determine the paw withdrawal threshold in response to mechanical stimuli. Mice were placed in acrylic boxes with wire mesh floors and allowed to acclimate two hours daily during the two days prior to testing, and for one hour on the test day before the experiment. This procedure aimed to reduce stress-induced variability and obtain more homogeneous responses. Subsequently, von Frey filaments were applied to the plantar surface of the left hind paw for 2 seconds, and the up–down method ⁴² was performed to calculate the mechanical withdrawal threshold in grams (g) ⁴³.

Motor coordination was assessed using rotarod test (Panlab, model LE8205) before and after i.t. injection of guide RNAs. All the mice were habituated to the room conditions for at least 30 minutes before training. Mice previously received an adaptive training three consecutive days before the test started. The initial speed was 4 rpm and then, adjusted up to 10 rpm for 3 minutes. The time for each mouse to fall from the rod was recorded in each test as latency to fall.

During behavioral data acquisition, mice were randomly assigned to chambers for mechanical and thermal testing, or lanes for motor coordination assessment. At the end of each evaluation, the experimenter recorded the mouse’s tag number and home cage. The experimenter remained blinded to the guide RNAs and treatments administered to the mice up to the final analysis.

TRPA1 functionality in disrupted TRPV1 Cas9 knock-in mice

TRPV1 is an ion channel that plays a critical role in the processing of nociceptive stimuli. However, it is known that TRPV1 can regulate other ion channels such as TRPA1, which also have an important function in the transduction of inflammatory nociceptive stimuli. Activation of the TRPA1 receptor induces inflammatory pain and mechanical hypersensitivity ⁴⁴. To determine whether TRPA1 function is maintained despite the genetic editing of TRPV1 through guide RNA delivery, we injected ASP7663 (1 μg/paw), a selective agonist of TRPA1 ⁴⁵. Mice were habituated to clear acrylic behavioral chambers for 1 hour per day over 2 days prior to the experimental sessions. Then, 20 μL of the ASP7663 stock solution (50 μg/mL, diluted in a solution containing 10% DMSO and 90% saline) were injected into the plantar surface of the left paw. Then, mechanical hypersensitivity was assessed as we described before.

Western blotting

DRG, TG, spinal cord, thalamus and cerebral cortex were harvested from Cas9^fl/fl/Scn10aCre⁺ mice. Tissues were homogenized in ice-cold RIPA buffer supplemented with Pierce^™ EDTA-free protease inhibitor (ThermoFisher, Waltham, MA, USA). Lysates were centrifugate at 4°C for 20 minutes at 12,000 xg. Protein concentration was quantified by Pierce^™ BCA protein assay (Invitrogen, Waltham, MA, USA). Total protein (30 μg) was loaded onto an 8% SDS-polyacrylamide gel and subjected to gel electrophoresis. SDS-PAGE gel was then transferred to polyvinylidene difluoride membranes, followed by blocking with 5% non-fat dry milk in TBS plus 0.1% Tween-20 (TBS-T) for one hour. Membrane was then incubated overnight at 4°C in TBS-T with mouse anti-Cas9 (1:1,000; 14697, Cell Signaling Technology) or rabbit anti-GAPDH (1:20,000, 10494IAP, Proteintech). Membranes were then washed three times with 1% non-fat milk in TBS-T. Following washes, they were incubated with horseradish peroxidase (HRP)-conjugated anti-rabbit (1:10,000; 111–036-144, Jacskon Immunoresearch) or anti-mouse (1:5,000; 115–036-062, Jacskon Immunoresearch) secondary antibodies for one hour at room temperature. The protein signal was detected using the Immobilion ECL Ultra Western HRP substrate as chemiluminescence system (WBULS0100, Millipore) according to the manufactureŕs instructions. GAPDH was used as an internal loading control.

Paclitaxel-induced cell death

To validate whether RNA guide treatment induced cell damage, we used paclitaxel-induced peripheral neuropathy as positive control of neurotoxicity and cellular cell damage ⁴⁶. Mice were injected via intraperitoneal with paclitaxel (Cayman Chemical Company, 10461, Ann Arbor, MI) at dose of 4 mg/kg, administered every other day (a total of three injections). DRGs were obtained one day after last intraperitoneal injection.

Statistical analysis of results

All data were analyzed using GraphPad Prism software. For comparisons between two groups, unpaired t-test was applied. For comparisons involving two variables, one or two-way ANOVA analysis followed by Dunnett or Bonferroni’s post hoc tests were used. Results were presented as mean ± SEM. Statistical significance was set at P < 0.05. Sample sizes were determined based on our prior work, which have consistently reported biologically and statistically meaningful results. Data were considered normally distributed if P > 0.05 and α= 0.05 by the Shapiro-Wilk test. For immunofluorescence analysis, tissue samples were obtained from 3 mice per condition. During sample collection, tissue processing, data acquisition, and quantitative analysis, all samples and data were recorded according to the mouse’s tag number. The experimenter remained blinded to the guide RNAs and treatments administered to the mice until the final analysis. Data points were excluded if values were identified as outliers, defined as more than two standard deviations from the mean. The experimenter remained blinded to the genotype or guide RNAs administered to the mice up to the final analysis.

For in vivo experiments, behavioral analyses were conducted using a sample size of 7 mice per group. Mice were included if they demonstrated consistent baselines responses to mechanical and thermal stimulation, as well as normal paw temperature and latency to fall, prior to the injection of guide RNAs or drugs. After this point, no behavioral data were excluded from the analysis.

Patient involvement

This was not a factor in either the design or execution of this project.

Results

Characterization of conditional Cas9 knock-in mice in sensory neurons

We generated a transgenic mouse where Cas9 is selectively expressed in sensory neurons. Cas9^fl/fl mice were crossed with mice that express Cre recombinase under the control of the Scn10a promoter (Nav1.8-Cre) ^30,31. Scn10a encodes the voltage-gate sodium channel 1.8 (Nav1.8) which is primarily expressed in sensory neurons. We performed immunostaining to confirm that Cas9 expression is restricted to Nav1.8-positive neurons. We examined GFP as a transgene reporter and peripherin which is co-expressed with Nav1.8 in over 90% of DRG sensory neurons (Fig. 3)^47,48. In conditional knock-in mice (Cas9^fl/fl/Scn10a^Cre/+), neurons from lumbar DRGs (Fig. 3A) and TG (Suppl. Fig.1A and 1B) expressed GFP in Nav1.8 positive cells. In contrast, GFP signal was not detected in Cas9^fl/fl mice. Furthermore, GFP-positive neurons colocalized with peripherin ⁴⁹. GFP expression was also detected in sciatic nerve fibers (Fig. 3B), where it colocalized with Nav1.8 and βIII-tubulin. In the spinal cord, GFP signal was localized to laminae I and II of the spinal dorsal horn, where it colocalized with CGRP and IB4, markers of nociceptive primary inputs ^50,51. Ad an additional validation, we performed immunoblots of Cas9 in pain pathway-related tissues obtained from conditional knock-in mice. We observed that Cas9 was expressed only in the DRG and TG, but not in the spinal cord, thalamus and cerebral cortex (Suppl. Fig.1C). These findings confirm that Cas9 expression is present in both DRG and TG sensory neurons in the Cas9^fl/fl/Scn10aCre⁺ transgenic model.

Figure 3.

GFP expression in sensory neurons and fibers of Cas9 conditional knock-in mice. (A) Lumbar dorsal root ganglia (DRG) and (B) sciatic nerve sections from Cas9^fl/fl/Scn10aCre⁺ mice, in which expression of Cas9 endonuclease and GFP was induced in a Cre recombinase-dependent manner. GFP expression was not observed in the absence of Cre recombinase (Cas9^fl/fl). Cre expression was driven by the Scn10a (Nav1.8) promoter. Immunofluorescence demonstrates colocalization of GFP with peripherin and βIII-tubulin, confirming selective Cas9 expression in small-diameter sensory neurons and neuronal fibers, respectively. Images are representative of three replicates.

In vivo evaluation of RNA guide efficiency

Approximately 50% of guide RNAs do not effectively generate double-stranded breaks in vivo ^52,53. Accurate prediction of DNA accessibility remains a major challenge in guide RNA design. Additionally, in cell lines the use of more than three RNA guides simultaneously results in genotoxic stress ^54,55. To balance these two considerations, we limited the number of guide RNAs to three. Three guides should have a success rate of approximately 90%. As an initial approach to quantify editing efficacy, we examined the effect of guide RNA dose on editing efficiency. The guides targeted GFP or a non-targeting control (NTC) control guide pool. The NTC pool is introduced in an identical genetic background as the target pool. We performed immunofluorescence analysis of DRG and sciatic nerve samples one week after RNA guide administration. We found that only the 3 μg dose significantly reduced GFP expression (Suppl. Fig. 2 A–B) (One-way ANOVA, F_(4,10) = 10.05, Dunnett’s test **P = 0.0015). Based on these findings, we selected the 3 μg dose for subsequent in vivo experiments.

To establish proof of concept, we focused on TRPV1 as a target. TRPV1 is an integral membrane protein that has been implicated in the transduction of noxious thermal stimuli and inflammatory pain ^27,56,57. We evaluated the efficacy of RNA guides directed to diminish TRPV1 expression in DRG neurons. Injection of 3 μg RNA guide pools two times resulted in a significant decrease in TRPV1 signal of approximately 65% in the DRG (Fig. 4A) (Unpaired t-test, t₍₄₎ = 3.27, *P = 0.03 ) and 55% in sciatic nerve (Fig. 4B) (Unpaired t-test, t₍₄₎ = 3.88, *P = 0.02 ).

Figure 4.

Intrathecal delivery of TRPV1 guides reduces TRPV1 expression in vivo. (A) DRG and (B) sciatic nerve sections showing immunohistochemical staining for TRPV1 expression following two intrathecal injections of TRPV1 or NTC guides. TRPV1 fluorescence intensity was normalized to GFP signal. Quantification revealed a significant reduction in TRPV1 expression in TRPV1 guide-treated mice compared to NTC. Bars represent the mean ± SEM of relative intensity values from n = 3 mice per group. *P < 0.05 was determined by an unpaired t-test. Scale bar: 100 μm.

Intrathecal TRPV1 guides does not induce caspase-3 activation

Caspase-3 is a critical effector of apoptosis, a form of programmed cell death. Upon activation, caspase-3 can degrade intracellular structural and functional proteins, ultimately leading to cell death ⁵⁸. The CRISPR/Cas9 system induces double strand breaks in DNA, which can trigger DNA damage responses and, under certain conditions, initiate apoptotic signaling pathways ^59,60. To determine whether TRPV1 or NTC guide RNA pools induce DNA damage and apoptosis in DRG neurons, we assessed cleaved caspase-3 as a marker of apoptosis. We found that a minority of GFP positive cells showed cleaved caspase-3 puncta in either group, suggesting a low incidence of apoptosis (Unpaired t-test, t₍₄₎ = 0.1, P = 0.92). As a positive control for apoptosis, we examined DRG cells from mice treated with paclitaxel (4 mg/kg, administered every other day for one week) ⁴⁶. Immunohistochemical analysis revealed that approximately 37% of DRG cells in paclitaxel-treated mice displayed immunoreactivity for caspase-3, however, no significant increase was detected with either TRPV1 and NTC guides (Fig. 5 A–B).

Figure 5.

Perturbation of TRPV1 does not induce apoptosis. (A) Representative images of DRGs from mice treated with TRPV1 or NTC guides, stained for cleaved caspase-3 (apoptosis marker), GFP (Cas9 expression) and DAPI (nuclei). DRGs from mice treated with paclitaxel (4 mg/kg, administered every other day for one week) were included as a positive control of neurotoxicity and cellular damage. Scale bar: 50 μm. (B) Quantification of the percentage of cleaved caspase-3 positive cells among GFP positive neurons. Bars represent the mean ± SEM from n = 3 mice per group. No significant difference (ns) was determined by an unpaired t-test.

TRPV1 depletion diminishes heat-evoked behavioral responses

TRPV1 plays a crucial role in thermoregulation and is activated by temperatures above 42°C. It is also associated with the sensation of noxious heat and pain ^26,27,56. We asked if disruption of TRPV1 using conditional Cas9 knock-in mice with a single injection of RNA guides impacts behavioral responses the noxious heat. We found that a single injection did not alter on capsaicin-evoked thermal hyperalgesia, mechanical allodynia, or paw temperature (Suppl. Fig. 3 A–C) (Two-way ANOVA, for thermal hyperalgesia: F_(4,50) = 0.37, P = 0.83; for mechanical allodynia: F_(8,75) = 1.36, P = 0.22; for paw temperature: unpaired t-test, t₍₁₂₎ = 0.90, P = 0.38). However, mice that received two TRPV1 guide injections exhibited a delay in response latency of 11 seconds, compared to 6 seconds in the NTC group, when exposed to a 50°C heated plate (Fig. 6A) (Unpaired t-test, t₍₁₂₎ = 5.02, ***P < 0.001), Similarly, mice treated with TRPV1 guides showed increased response latency to radiant heat applied to the hind paw in the Hargreaves test, compared to the NTC group (Fig. 6B) (Unpaired t-test, t₍₁₂₎ = 2.61, *P < 0.05). These findings suggest that intrathecal delivery of TRPV1 guides can modulate heat-evoked behavioral responses.

Figure 6.

Disruption of TRPV1 prevents capsaicin-induced hypersensitivity in Cas9 knock-in mice. Mice received two intrathecal injections of TRPV1 or non-target control guide RNAs (3 μg per injection) over a period of 2 weeks. One week after the last intrathecal injection, behavioral tests were conducted. Thermal hyperalgesia was assessed using (A) either a 50°C hot plate or (B) the Hargreaves test. Bars represent mean ± SEM of withdrawal latency (s) from n = 7 mice per group. *P < 0.05 and ***P < 0.001 was determined by an unpaired t-test. (C) Motor coordination was evaluated using the rotarod test at baseline (BL), and one week after the first (1) and second (2) intrathecal injection. Bars represent mean ± SEM of latency to fall (s) from n = 7 mice per group. No significant differences (ns) were observed across time points and RNA guide treatments by one-way repeated measures ANOVA. (D) Spontaneous pain was assessed by counting the number of flinches during the first 10 minutes following intraplantar injection of capsaicin (5 μg). Bars represent the mean ± SEM of number of flinches from n = 7 mice per group. ***P < 0.001 was determined by an unpaired t-test. (E) Representative images of the left paw injected with capsaicin (dotted line) from TRPV1 or NTC guide-treated mice. Forward looking infrared (FLIR) imaging was used to measure paw temperature 1 h after capsaicin injection. Bars represent the mean ± SEM of temperature (°C) from n = 7 mice per group. ***P < 0.001 was determined by an unpaired t-test. (F) Thermal hyperalgesia and (G) mechanical allodynia were evaluated before intrathecal injection of TRPV1 or NTC guides, as baseline (BL), and one week after second guide delivery at 0, 2, 4, 24 and 48 hours after capsaicin injection. Mice treated with TRPV1 guides displayed significantly reduced thermal and mechanical allodynia. Lines represent the mean ± a shaded area indicating SEM of withdrawal latency (s) or threshold (g) from n = 7 mice per group. ***P < 0.001 and ****P < 0.0001 were determined by two-way ANOVA followed by Bonferroni’s post hoc test. Arrows indicate intrathecal injections of RNA guides (i.t. gRNA).

To assess if delivery of guide RNAs impacted motor neuron function, we conducted a rotarod test. Motor coordination and balance were evaluated by measuring the latency to fall from the rotating rod at baseline (prior to injection), or after intrathecal administration of TRPV1 or NTC guides. Our analysis revealed no significant differences in motor performance either within treated groups over time or between groups at any time point (Fig. 6C) (Two-way ANOVA, F_{(2.09, 12.59)} = 1.76, P = 0.21). These results suggest that intrathecal injection of guide RNAs did not diminish motor function.

Intrathecal TRPV1 RNA guides reduced capsaicin-induced pain behaviors

To determine if intrathecal delivery of TRPV1 guides impacts spontaneous pain associated behavior, we administered capsaicin (5 μg) into the paw. Capsaicin activates the TRPV1 receptor ⁶¹. Intraplantar injection of capsaicin elicits a rapid nocifensive response, characterized by flinching or licking of the injected paw. We examined this behavior immediately after capsaicin injection over a 10-minute period. We observed that mice treated with TRPV1 guides displayed significantly fewer flinches than the NTC group (Fig. 6D) (Unpaired t-test, t₍₁₂₎ = 5.42, ***P < 0.001). As an additional test, we measured paw temperature using FLIR imaging. Heat is a cardinal hallmark of inflammation. We found capsaicin injection increased paw temperature on the ipsilateral (injected) paw in the NTC group but not in the TRPV1 edited group (Fig. 6E) (Unpaired t-test, t₍₁₂₎ = 4.54, ***P < 0.001). There was no effect on the contralateral paw (Unpaired t-test, t₍₁₂₎ = 0.75, P = 0.46) (Suppl. Fig. 4A). These results suggest that depletion of TRPV1 with Cas9 perturbation impacts behavioral responses evoked by capsaicin injection.

Next, we asked if TRPV1 depletion impacts thermal hyperalgesia or mechanical allodynia following capsaicin injection. Capsaicin produced a reduction in thermal latency and mechanical withdrawal threshold in both the TRPV1 and NTC groups two hours after injection. However, mice treated with TRPV1 guides recovered more rapidly than mice that received NTC guides at 4, 24 and 48 h after capsaicin injection (Fig. 6F and 6G) (Two-way ANOVA, for thermal hyperalgesia: F_(5,60) = 4.98, Bonferroni’s multiple comparisons. ****P < 0.0001; for mechanical allodynia: F_(5,60) = 6.74, Bonferroni’s multiple comparisons. ***P < 0.001). Thermal and mechanical withdrawal responses in the contralateral paw were not impacted (Suppl. Fig. 4 B–C) (Two-way ANOVA; for thermal hyperalgesia: F_(4,60) = 0.82, P = 0.51; for mechanical allodynia: F_(4,50) = 0.45, P = 0.77). These results suggest that intrathecal delivery of TRPV1 guides can regulate pain and thermal responses evoked by capsaicin injection. Based on these observations, we conclude that TRPV1 perturbation with CRISPR/Cas9 blunts capsaicin-induced evoked pain behaviors in mice.

Editing does not diminish TRPA1 function

Next, we asked if TRPV1 perturbation diminished the function co-expressed channels. We focused on TRPA1 given their extensive co-localization in sensory neurons ⁴⁴. Knockdown TRPV1 signal did not impact TRPA1 expression in the DRG (Suppl. Fig. 5A) (Unpaired t-test, t_(3.8) = 0.52, P = 0.63 ). Furthermore, behavioral responses to ASP7663 (1 μg intraplantarly injected), a selective TRPA1 agonist, were not impacted by TRPV1 depletion (Suppl. Fig. 5B) (Two-way ANOVA, for ipsilateral F_(5,60) = 1.72, P = 0.14, for contralateral F_(5,60) = 0.96, P = 0.44). These observations suggest that our strategy enables selective disruption of TRPV1.

Discussion

Our approach leverages existing tools for genome editing to selective disrupt factors of present in a subpopulation of sensory neurons. We contextualize the major technical challenges and conceptual barriers with this strategy in the following section.

Editing efficiency is governed by multiple factors. Cas9 dose influences editing rates when delivered by viruses ^62,63 or as nanoparticles ⁶⁴. We made use of transgenic expression as a means of holding Cas9 levels constant throughout adulthood. This enables perturbation of multiple genetic loci in a single animal, a challenging problem with standard approaches. While the Scn10a promoter offers the benefit of restricting Cas9 expression to a defined subset of sensory neurons, its expression is not strictly confined to nociceptors ⁶⁵. While other strains such as Advillin-Cre and Pirt-Cre are used to drive Cre expression in sensory neurons, each has its own limitations regarding specificity ^66,67. Nevertheless, the general protocol we report here is likely compatible with other Cre lines that enable perturbation in different populations of sensory neurons.

A second variable is significant variation in the efficiency of guide RNAs. Approximately half of sgRNAs fail to induce double-strand breaks ^53,54. To mitigate this risk, we employed guide pools, which are known to increase editing efficiency ^68–70. Our pilot experiments demonstrated poor editing efficiency with single guide injections. We reasoned that a pool of three guides should yield success in approximately 90% of cases. However, increasing the number of guides also raises the risk of off-target effects and cumulative DNA damage ^54,55. To balance efficacy and specificty, we limited each pool to three guides selected for high predicted activity and low off-target potential. A single 3 μg intrathecal dose of GFP guides reduced GFP expression in vivo, supporting the advantage of using multiple guides. This was further substantiated by editing of TRPV1, which achieved approximately 66% depletion after two rounds of guide RNA delivery.

A concern with the use of guide RNA pools is genotoxic stress. To determine whether the guide RNA pools we injected resulted in substantial DNA damage, we measured the expression of cleaved caspase-3. We did not observe a significant increase in apoptosis following guide RNA administration. Our approach is consistent with previous reports demonstrating that multi-guide delivery enhances editing efficiency without significantly increasing off-target risks through systemic delivery ^69–71. Additionally, behavioral testing revealed no deficits in motor coordination, suggesting that our approach is well tolerated in vivo. Therefore, our approach aligns with current best practices for achieving efficient and specific editing in target tissues. A problem with injection of exogenous RNAs is activation of the RIG-I pathway and induction of a type I interferon response. To circumvent this problem, we removed the 5’-triphosphate from our synthetic RNA guides ⁷². Our behavioral data suggest that the non-targeting guide RNA pools appeared to have minimal effects on behavior at baseline, further supporting the safety and tolerability of our approach.

We applied an optimized strategy to target the capsaicin receptor TRPV1 ^26,27,56. While traditional TRPV1 knockout mice show impaired thermal nociception ^26,73, our initial results with delivery of a single TRPV1 guide failed to reproduce this phenotype. We found that mice remained sensitive to both thermal and mechanical capsaicin-evoked responses. This might be due to incomplete knockdown of TRPV1. Possible reasons for this include the inherent instability of RNA guides, which are rapidly degraded by endogenous RNases despite protective complex formation, limiting their bioavailability before reaching target tissues ⁷⁴. Moreover, the efficiency of each guide is not uniform and may vary, even when computational predictions suggest high efficiency and specificity. To address this, we administered a second round of guide RNAs and validated knockdown efficacy through immunostaining of TRPV1. We observed a significant reduction of TRPV1 expression in DRGs and sciatic nerve of approximately 65% and 55%, respectively. Similar results have been reported in human DRG neurons using CRISPR/Cas9 plasmid constructs, achieving a 57% reduction in TRPV1 protein levels ²². Our behavioral measurements are consistent with a reduction in TRPV1 function. Given these observations, it is likely that primary DRG sensory neurons after editing would display insensitivity to capsaicin in vitro. A limitation of this study is we did not probe if edited neurons displayed a reduction in TRPV1 function on sensory neuronal physiology.

Why is a high dose of RNA required for editing? siRNAs are extremely potent and can work with as little as a single microgram ⁷⁵. We found that triple that quantity of guide RNA was required. There are three parsimonious explanations. First, the complex between RNA targets and AGO is exceptionally stable resulting in half-lives that last from hours to days ^76–79. Second, sgRNAs must transit to the nucleus whereas siRNAs act in the cytoplasm. Numerous ribonucleases are present in the cytoplasm. It is likely that some of the guide RNA is degraded prior to import into the nucleus. Third and finally, the half-life of our guide RNAs is likely lower than what can be achieved through chemical modifications. These alterations can substantially enhance the stability of synthetic RNAs and are commonly employed in siRNAs.

A major advantage of having stable Cas9 expression is the capacity for sequential rounds of genome editing in the same organism. This is potentially critical for understanding the function of paralogous genes. A notable example of this are voltage gated sodium channels. Considerable debate exists surrounding the necessity of Nav1.7 and Nav1.8 for pain ^80,81. Genetic deletion of Nav1.8 results in a loss of pain ⁸². The importance of Nav1.7 is more complicated. Embryonic loss of Nav1.7 in mouse or human sensory neurons results in profound analgesia ^83,84. Antagonists of either factor reduce pain in rodents ^85–87. However, simultaneous deletion of Nav1.7 and Nav1.8 results in deficits in inflammatory pain but no effect on the development of neuropathic pain ⁸³. This implies additional redundancy in voltage gated sodium channel function. One way to identify these factors is through the use of multi-locus genome perturbations. The strategy we report provides a means to this critical end but is only one example of how this highly plastic tool can be leveraged for the field.

Highlights.

• Stable genetic perturbation of sensory neurons in vivo remains challenging and resource intensive

• SpCas9 provides an easily programable approach for genome perturbation

• A novel model is reported where SpCas9 is expressed in a subset of sensory neurons

• Disruption of TRPV1 diminishes capsaicin sensitivity with minimal toxicity

Perspective.

There are tremendous opportunities afforded by facile multi-locus genome perturbation of sensory neurons in vivo. This model and approach enables rapid and low-cost genetic depletion experiments in mice.

DATA AND REAGENT AVAILABILITY

The animal model and data are available upon request to the corresponding author.