Chemogenetic Attenuation of Acute Nociceptive Signaling Enhances Functional Outcomes Following Spinal Cord Injury

Abstract

Identifying novel therapeutic approaches to promote recovery of neurological functions following spinal cord injury (SCI) remains a great unmet need. Nociceptive signaling in the acute phase of SCI has been shown to inhibit recovery of locomotor function and promote the development of chronic neuropathic pain. We therefore hypothesized that inhibition of nociceptive signaling in the acute phase of SCI might improve long-term functional outcomes in the chronic phase of injury. To test this hypothesis, we took advantage of a selective strategy utilizing AAV6 to deliver inhibitory (hM4Di) Designer Receptors Exclusively Activated by Designer Drugs (DREADDs) to nociceptors of the L4-L6 dorsal root ganglia to evaluate the effects of transient nociceptor silencing on long-term sensory and motor functional outcomes in a rat thoracic contusion SCI model. Following hM4Di-mediated nociceptor inhibition from 0-14 days post-SCI, we conducted behavioral assessments until 70 days post-SCI, then performed histological assessments of lesion severity and axon plasticity. Our results show highly selective expression of hM4Di within small diameter nociceptors including calcitonin gene-related peptide (CGRP)⁺ and IB4-binding neurons. Expression of hM4Di in less than 25% of nociceptors was sufficient to increase hindlimb thermal withdrawal latency in naïve rats. Compared with subjects who received AAV-yellow fluorescent protein (YFP; control), subjects who received AAV-hM4Di exhibited attenuated thermal hyperalgesia, greater coordination, and improved hindlimb locomotor function. However, treatment did not impact the development of cold allodynia or mechanical hyperalgesia. Histological assessments of spinal cord tissue suggested trends toward reduced lesion volume, increased neuronal sparing and increased CGRP⁺ axon sprouting in hM4Di-treated animals. Together, these findings suggest that nociceptor silencing early after SCI may promote beneficial plasticity in the acute phase of injury that can impact long-term functional outcomes, and support previous work highlighting primary nociceptors as possible therapeutic targets for pain management after SCI.

Introduction

Spinal cord injury (SCI) is a traumatic event that frequently results in permanent neurological dysfunction. Consequences of SCI largely depend on the spinal level and severity of the lesion, and can include loss of voluntary motor function, autonomic dysfunction, and persistent neuropathic pain.^1-3 It has been reported that 50-80% of individuals living with SCI experience chronic neuropathic pain, which manifests as sharp, shooting, or burning pain sensations and can severely detract from quality of life.^1,4 The development and persistence of neuropathic pain after SCI is driven in part by changes in nociceptors of the dorsal root ganglia (DRG), which have been shown to exhibit hyperactivity, metabolic changes, and aberrant sprouting after SCI.^5–10 In addition, previous studies utilizing rodent SCI models have shown that nociceptive stimulation induces maladaptive plasticity that undermines motor functional recovery.^11–14 Hence, inhibiting nociceptive signaling in the acute phase of SCI presents an attractive therapeutic target to mitigate the development of chronic neuropathic pain and promote motor functional recovery.

Conventional pharmacological treatments for neuropathic pain management, including opioids, are only partially successful and are associated with several undesirable side effects following SCI.^15–19 In fact, opioids such as morphine that are routinely administered in the clinic for pain management impart debilitating side effects not only on sensory but also locomotor recovery, potentially by interfering with processes vital for tissue repair after injury.¹⁸ These limitations of existing pharmaceutical interventions to attenuate persistent pain signaling highlight an unmet clinical need to identify alternative targeted therapeutic strategies for attenuation of pain signaling following SCI.

One promising potential strategy to achieve this goal is through chemogenetics, the use of modified G protein-coupled receptors (Designer Receptors Exclusively Activated by Designer Drugs; DREADDs) that can be activated with exogenous ligands to induce neuronal hyperpolarization or depolarization.^20–24 The Gi-coupled inhibitory DREADD (hM4Di), developed through modification of human muscarinic receptors, has been extensively used to enable silencing of neural circuits within the CNS.^25–29 Adeno-associated viral vectors (AAVs) can be used for sustained gene expression in long-term studies, with low immunogenicity and selective infectivity through selection of AAV serotype.^30,31 Notably, Iyer and colleagues and Towne and colleagues recently described a robust strategy for selective and sustained inhibition of small-diameter nociceptors in mice using AAV6 viral vectors.^32-34 They showed that following activation of AAV6-hM4Di by the DREADD ligand clozapine-N-oxide (CNO) nociceptor activity could be noninvasively inhibited over a duration of hours. However, this strategy has not yet been utilized for nociceptor silencing in an in vivo model of SCI.

We performed this study to test the hypothesis that selectively silencing DRG nociceptors early after SCI can improve long-term sensory and motor functions. Here, we describe the use of inhibitory hM4Di DREADDs for sustained, reversible inhibition of nociceptor hyperactivity in the acute phase after a moderate thoracic (T10) contusion SCI in rats. We utilized bilateral sciatic injection of AAV6-hM4Di to achieve transduction in primary nociceptors of rat lumbar DRGs, then administered CNO starting immediately after injury up to 14 days post-injury (DPI). We assessed sensory behavioral outcomes including thermal hyperalgesia, mechanical allodynia, and cold allodynia, and locomotor outcomes through the Basso, Beattie, and Bresnahan (BBB) open field locomotor test and CatWalk gait analysis test, up to 10 weeks post-injury. We also performed histological analysis of thoracic and lumbar spinal cord sections to characterize anatomical outcomes such as lesion volume, gray matter and white matter sparing, and axonal plasticity following SCI.

Methods

Ethics statement

All animal experiments were performed in strict compliance with the National Institutes of Health Guidelines for Animal Care and Use of Laboratory Animals. All animal experiments were approved by the Texas A&M University Institutional Animal Care and Use Committee. All efforts were made to minimize pain and distress in animals.

Animals

A total of 68 adult, female Sprague Dawley rats (200-250 g; Charles River Laboratories) were used for all experiments, including pilot dosing studies. Animals were housed two per cage, in a 12-h (6:00 am - 6:00 pm) light cycle, with unlimited access to food and water, and with ambient temperature between 20-23°C and 30-70% humidity. Initial group sizes (as shown in Table 1) were n = 7 for naïve, n = 6 for Sham-yellow fluorescent protein (YFP), n = 6 for sham-hM4Di, n = 16 for SCI-YFP, and n = 16 for SCI-hM4Di. One animal in the SCI-YFP group and two animals in the SCI-hM4Di group died immediately after surgery. Two animals in the SCI-YFP group and one animal in the SCI-hM4Di group were excluded from behavioral analysis because displacement of the impactor probe was outside of an acceptable range.

Table 1.

Experimental Groups in This Study

Group	AAV	Surgery	Drug	N
SCI-hM4Di	AAV6-hM4Di	SCI	CNO	16
SCI-YFP	AAV6-YFP	SCI	CNO	16
Sham-hM4Di	AAV6-hM4Di	Sham	Vehicle	6
Sham-YFP	AAV6-YFP	Sham	Vehicle	6
Naïve	None	None	Vehicle	7
			Total	51

SCI, spinal cord injury; AAV6, adeno-associated virus serotype 6; CNO, clozapine-N-oxide; YFP, yellow fluorescent protein.

Viral vector injection

pAAV-hSyn-HA-hM4D-IRES-mCitrine (Addgene #50464) and pAAV-hSyn-eYFP (gift from Dr. Karl Deisseroth) plasmids were packaged into AAV6 viral vectors by Vigene Biosciences (Rockville, MD). Rats were randomized to treatment group using a random number generator website. Animals were anesthetized with a cocktail of ketamine (50 mg/kg), xylazine (2.6 mg/kg), and acepromazine (0.5 mg/kg), and body temperature was maintained at 37°C throughout the surgery. The sciatic nerve was exposed through blunt dissection of the connective tissue between the gluteus maximus and the biceps femoris muscles. The sciatic nerve was kept taut, but not stretched, by placing a curved hemostat underneath the nerve. A small piece of wet gel foam (Ethicon) was placed on the surface of the nerve to prevent the epineurium from drying out. A total volume of 5 μL containing 8 × 10¹² genome copies/mL AAV +600 mM NaCl +0.1% w/v cholera toxin subunit B (CTB; List Labs) was injected into each sciatic nerve using a pulled glass micropipette. The virus solution was injected across three injection points on the nerve approximately 2 mm apart, over a period of approximately 30-45 min, using a Picospritzer II (General Valve, Inc., Fairfield, NJ). Following injection, the incision was closed using stainless steel wound clips. To ensure that the injections did not cause any functional impairments, sensory and locomotor behavior assessments were performed at 7 days post (DP) AAV injections and 7 days prior to SCI.

Spinal cord injury surgeries

Animals were assigned to receive thoracic (T10) contusion SCI (n = 32) or laminectomy (n = 12) surgeries. The grouping assignments were made based on their baseline thermal sensory behavior scores to ensure that group means were not statistically different prior to surgery. At 4 weeks after AAV injections, animals were anesthetized with inhaled isoflurane (4% for induction and 1.5-2.0% for maintenance) and body temperature was maintained at 37°C throughout the surgery. The lower back was shaved and disinfected with 70% ethanol and betadine solution. A 5-cm incision was made in the skin using a #15 scalpel blade, and skin was retracted. Next, two incisions were made in the muscles on either side of the vertebral column, extending about 2 cm rostral and caudal to the T9-T11 vertebral segments. The T10 dorsal spinous process was removed, and a laminectomy was performed to expose the spinal cord. A moderate spinal cord contusion injury was performed at T10 using the IH-0400 Infinite Horizon Impactor (Precision Systems and Instrumentation, Fairfax Station, VA), using a force of 150 kdynes and dwell time of 1 sec. After injury, the muscle was sutured using 4-0 polypropylene suture. The incision was closed using stainless steel wound clips. Animals in the sham group received laminectomy only, with no SCI. Following death of three animals after surgery, and exclusion of three animals due to the displacement range for SCI surgery being beyond the acceptable range, the final group sizes were n = 13 for the SCI-YFP group and n = 13 for the SCI-hM4Di group. In addition, uninjured (naïve) rats that did not receive virus injection or sham surgery (n = 7) were included as a control for behavioral experiments.

Post-operative care

Subcutaneous injections of 3 mL lactated Ringers + ampicillin (33 mg/kg) + banamine (0.3 mg/kg) were administered once daily for 3 days starting immediately after any surgical procedure. We did not administer buprenorphine or any other opioid drug in this study. After surgery, animals were placed in a post-operative cage half-on/half-off a circulating water heating pad set to 37°C, and monitored until they fully recovered from anesthesia. All SCI subjects had their bladders manually expressed twice a day until they regained bladder control, which was defined as 3 consecutive days of small/empty bladder at the time of expression. Body weight was monitored every day for the first 7-10 days, until wound clips were removed and weekly thereafter.

Drug administration

Starting immediately after SCI/laminectomy and continuing for 14 days, clozapine-N-oxide (NIDA Drug Supply Program) was systemically administered to subjects according to treatment group (Table 1). The first dose was administered via intraperitoneal injection of 3 mg/kg body weight, and thereafter animals were given CNO in drinking water. For oral administration, CNO dissolved in 0.05% dimethyl sulfoxide, or vehicle, was delivered in drinking water for a total concentration of 5 mg/kg/day. A small amount of saccharine (5 mM) was added to the water to mask the bitter taste of CNO. Water intake in all cages was carefully monitored and water bottles were changed and replaced with fresh CNO/saccharine or vehicle every day.

Behavioral assessments

For all behavioral assessments, animals were tested during the light cycle. Each behavior test was performed at approximately the same time every day.

Testing schedules for sensory assessments

Prior to AAV injections, three baseline measurements were collected per animal for each behavioral test. Another two measurements per test were collected at 7 and 21 days after AAV injections to ensure that the intrasciatic injections did not result in any significant changes in behavioral scores compared with baseline. Following SCI or sham surgery, behavioral testing continued once weekly for 10 weeks starting at 14 dpi. For each behavioral test, rats were allowed to acclimate to the testing room in their home cages for 30 min, then to the testing apparatus for 30 min on 4 consecutive days before testing began. The same acclimation schedule was also followed on each day of testing, prior to starting the trials. For all sensory behavior assessments, fruit cereal pieces were provided as treats to all subjects to create a distraction during hindlimb stimulation. All behavioral assessments were performed by experimenters blinded to treatment group.

Testing schedules for locomotor assessments

Prior to AAV injections, two baseline measurements were collected for CatWalk gait analysis and one baseline was collected for open field (BBB) locomotor testing. One measurement was collected after AAV injections on 21 days post injection. Post-SCI testing was performed at 1, 3, and 7 days post-SCI/sham for BBB locomotor test and starting at 14 days post-SCI/sham for the CatWalk gait test. Testing continued weekly thereafter for 10 weeks. All behavioral assessments were performed by experimenters blinded to treatment group.

Thermal (Hargreaves) hyperalgesia test

The Hargreaves test is used to measure paw withdrawal latency in response to application of a radiant heat source to the plantar surface of the rat's hindpaw, and is used as a measure of hyperalgesia following SCI.^35,36 We used a plantar analgesia meter (IITC Life Science Inc., Woodland Hills, CA), which uses visible light as the heat source. Testing was performed as previously described.³⁶ During the test, rats were placed within individual acrylic chambers placed on a glass platform maintained at 30°C. A light beam of intensity of 50% was used, which was determined a priori to produce a paw withdrawal after approximately 8-12 sec of heat application in naïve animals. During testing, the light emitter was placed directly under the plantar surface of the paw, and the amount of time it took the animal to withdraw from the light stimulus was automatically recorded. Five trials were recorded for each hindlimb with an interval of 3-5 min between every trial for the same animal. For analysis, the three middle values were used to obtain the mean paw withdrawal threshold at each time-point. During the test, withdrawal response associated with supraspinal or aversive responses such as looking at and licking the tested paw, grooming, or holding the tested paw up in the air, were all recorded.

von Frey mechanical allodynia test

The von Frey test is used to assess mechanical allodynia in response to manual stimulation of the plantar surface of the animal's hindpaw with monofilaments.^36,37 Rats were placed within acrylic enclosures on a metal mesh floor (IITC Life Science Inc.). von Frey monofilaments calibrated to different grams of force (ranging from 10 g to 300 g) were applied to the plantar surface of each hindpaw of the rat with a 3-5-min interval between each application on the same animal. At a point where the filament bends, if there is a swift paw withdrawal, it is considered a positive response. Mechanical paw withdrawal thresholds in response to the monofilaments were collected using the Dixon up-down method as described previously.^38,39 If a trial produced a positive response, the next-lowest force monofilament was used for the subsequent trial; however, if there was no response the subsequent trial was performed using the next higher force monofilament. Ten such trials were taken for each hindlimb and the filament that elicited a positive response at least 50% of the times that it was tested, was considered as the paw withdrawal threshold for that testing session. During the test, withdrawal response associated with supraspinal/aversive responses such as looking at and licking of the tested paw, grooming, holding the tested paw up in the air were all recorded.

Acetone cold allodynia test

Acetone application to the plantar surface of the rat's paw produces an evaporative cooling effect resulting in aversive behaviors such as paw withdrawal, licking of paw and grooming responses, which are considered indicators of cold allodynia.³⁶ This test was performed only in a proportion of SCI rats and all the naïve rats, due to logistical constraints. Rats were placed on a metal mesh surface within a transparent acrylic box. A drop of acetone was formed at the end of a 3-mL syringe and applied to the plantar surface of each hindlimb, such that only the acetone, and not the syringe, contacted the animal's hindpaw. On each day of testing, five trials of acetone application were conducted per hindpaw with an interval of 5 min between every trial on the same animal. A swift paw withdrawal in response to the acetone application which is an indication of cold allodynia, was considered a positive response. During the test, withdrawal response associated with supraspinal/aversive responses such as looking at and licking of the tested paw, grooming, holding the tested paw up in the air were all recorded. The frequency of paw withdrawal was expressed as a percentage (100 × the number of paw withdrawals divided by the total number of trials).

Open field locomotor test

The BBB open field locomotor test was used to evaluate locomotor function after SCI.⁴⁰ Rats were allowed to acclimate to the behavior room in their home cages for 30 min and in the open field for 5 min each, on 4 consecutive days before testing began. During testing, each animal was placed in the open field and allowed to freely locomote for a total duration of 4 min while being observed by two blinded experimenters from opposite ends of the open field. Different aspects of locomotion such as movements of different joints of the hindlimb, weight support, coordination, paw placement, and toe clearance were assessed to assign a score ranging between 0-21 for each animal.⁴⁰ BBB subscores were calculated by scoring specific behavioral attributes independently and adding them together as previously described.³⁹

CatWalk gait analysis

The CatWalk XT gait analysis system (Noldus) allows unbiased, quantitative assessment of subtle aspects of gait locomotor function in rodent SCI models.^41,42 Rats were allowed to acclimate to the testing room in their home cages for 30 min, then to freely explore the CatWalk walkway either for 5 min, or until three compliant runs were observed, on 4 consecutive days before testing began. To encourage continuous and consistent movement from one end of the walkway to the other, rats were trained with pieces of fruit cereal placed at opposite ends of the walkway. A compliant run was defined by animals having continuous uninterrupted locomotion along the walkway, while having a run speed variability of less than 50%. Testing was performed as previously described.⁴¹ At each time-point, animals are placed in the walkway and allowed to freely run back and forth across the length of the walkway for a total of 10 min or until five compliant runs were obtained. Runs were then classified and gait parameters quantified using the CatWalk XT software.

Immunohistochemistry

Tissue processing

At the conclusion of the study, animals were deeply anesthetized and transcardially perfused with 200 mL of 0.1 M phosphate buffered-saline followed by 200 mL of 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer. Tissue was post-fixed in 4% PFA for 16 h at 4°C, then cryoprotected in 30% sucrose for at least 3 days at 4°C. Tissue processed for immunohistochemistry included the thoracic spinal cord (1-cm segment containing the lesion epicenter), lumbar spinal cord (L2-L5 segments), and L4-L6 DRGs for experiments validating AAV-mediated gene transduction. Tissue was harvested, embedded in optimal cutting temperature medium, frozen, and sectioned using a cryostat to a thickness of 25-30 μm. Tissue sections were mounted directly onto gelatin-coated slides and stored at -20°C. Either a 1-in-6 or a 1-in-12 tissue series was used for each immunostaining experiment.

Immunostaining

Tissue sections were washed three times in tris-buffered saline (TBS), then incubated in blocking buffer (5% goat or donkey serum in TBS +0.1% Triton-X-100) for 1 h at room temperature. Tissue was then incubated in primary antibodies (Supplementary Table S1) overnight at 4°C. Next, tissue was washed three times in TBS then incubated in AlexaFluor-conjugated secondary antibodies (Jackson ImmunoResearch) for 2.5 h at room temperature, washed three more times in TBS with the final wash containing 4',6-diamidino-2-phenylindole (5 μg/mL, Sigma-Aldrich). Sections were then rinsed with deionized water, allowed to dry, and cover-slipped with Mowiol mounting medium.

Image acquisition

A Nikon Eclipse fully motorized upright fluorescent microscope equipped with monochrome camera and Nikon NIS-Elements software was used for image acquisition. Slides stained with fluorescent probes were stored protected from light and imaged in a dark room. Images were acquired using the same acquisition settings across all samples for each experiment. Images were exported as 8-bit TIFF files for analysis.

Image analysis

All image analysis was performed in a blinded fashion using ImageJ or FIJI software.

NeuN quantification: Neuronal sparing was calculated by drawing a region of interest (ROI) around area of NeuN immunoreactivity (gray matter) for each section containing a lesion, for every subject. The final average area of neuronal sparing for each animal was calculated by averaging the area of ROI of each section.

Lesion volume/glial fibrillary acidic protein (GFAP) quantification

To quantify lesion volume, an ROI was drawn along the area bordering GFAP immunoreactivity for every section containing a lesion. Total area of the ROI was calculated for each section using imageJ or FIJI software using the following formula:

One subject in the SCI-hM4Di group was excluded from analysis because the tissue sections were poor quality.

B-tubulin quantification

We quantified white matter sparing through beta-tubulin immunofluorescence at the thoracic spinal level. For each subject, considering the epicenter as the section with the largest lesion cavity, six rostral and six caudal sections were imaged for quantification (13 total sections). A region of interest was drawn around the white matter in each transverse section and the average intensity of B-tubulin expression per pixel area was quantified. The total volume containing spared white matter was also quantified using the formula above.

CGRP and IB4 quantification

To quantify CGRP and IB4 immunoreactivity in both thoracic and lumbar spinal cord transverse sections, ROIs were drawn around bands of CGRP/IB4 immunoreactivity for both right and left dorsal horns of each subject. For the thoracic spinal level: considering the epicenter as the section with the largest lesion cavity, six rostral and six caudal sections were imaged for quantification (13 total sections). For the lumbar spinal level: all sections from L2-L6 levels were imaged and considered for quantification. Once ROIs were drawn, the average intensity of CGRP/IB4 expression per pixel area was calculated. To calculate CGRP immunoreactivity in the lumbar deep dorsal horn, we performed analysis as previously described.⁴³ Briefly, we drew a ROI around the dorsal horn band of PKCγ immunoreactivity in laminae IIi, and defined the dorsoventral extent of this box as distance “x”. Then, we drew a second, rectangular ROI with dimensions 4 × (mediolateral) by 2 × (dorsoventral) and placed this new ROI immediately ventral to the PKCγ ROI. CGRP immunoreactivity was thresholded using Phansalkar's auto local threshold function in ImageJ,⁴⁴ then the number of thresholded pixels per unit area within this region was quantified.

Statistical analysis

GraphPad Prism 8 was used for all statistical analyses. Details of all statistical tests can be found in Supplementary Table S2. All data are presented as mean ± standard error of the mean (SEM). Statistical significance was defined as p < 0.05.

Results

Validation of AAV-mediated gene expression in rat lumbar dorsal root ganglion nociceptors

Previous studies have demonstrated that intrasciatic injection of AAV6 is sufficient to achieve selective transduction of small-diameter nociceptors in the DRG of mice and non-human primates.^32,33,45 We first sought to determine whether this strategy can also enable selective nociceptor transduction in adult rats. We delivered AAV6-hSyn-HA-hM4D-IRES-mCitrine (“AAV-hM4Di”) into bilateral sciatic nerves of adult, naïve Sprague-Dawley rats and allowed 4 weeks for transgene expression, then sacrificed for immunohistochemical analysis (Fig. 1A). We analyzed L4-L6 DRGs and observed bright mCitrine fluorescence in hM4Di⁺ cells. Cholera toxin subunit B (CTB) was used as a control for accuracy of intrasciatic injection; however, CTB immunoreactivity appeared to be mostly restricted to larger-diameter DRG neurons and did not colocalize with hM4Di+ cells (Fig. 1B). mCitrine colocalized with the neuronal marker, NeuN, and appeared restricted to small-diameter DRG neurons (Fig. 1C). Indeed, quantification revealed that 99.3% of mCitrine⁺ neurons had a diameter of less than or equal to 30 μm (Fig. 1D). A Mann-Whitney test revealed a statistically significant difference in the cell body diameters of hM4Di+ neurons compared with all NeuN⁺ neurons (p = 0.0244). This is consistent with previous findings that AAV6-hM4Di selectively transduces small-diameter DRG neurons following sciatic nerve injection in mice.³² We next examined colocalization of the mCitrine reporter with neuronal subtype-specific markers in order to verify whether expression was restricted to nociceptors. hM4Di was found to be expressed in neurons that were immunoreactive for calcitonin gene-related peptide (CGRP) and neurons that were IB4-binding (Fig. 1E).

FIG. 1.

Intrasciatic AAV6 injection allows for selective transduction of hM4Di in rat lumbar dorsal root ganglion nociceptors. (A) Experimental design. AAV6-hSyn-HA-hM4D-IRES-mCitrine (“AAV6-hM4Di”) was co-injected with cholera toxin subunit B into bilateral sciatic nerves of adult rats, and animals were sacrificed 4 weeks later for immunohistochemical assessments. A subset of n = 4 animals was subjected to hindlimb thermal withdrawal latency testing before and after administration of clozapine-N-oxide (CNO) at 4 weeks post-AAV injection. Illustration created with Biorender.com. (B) Image of L4 dorsal root ganglia (DRG) containing neurons expressing hM4Di and the mCitrine reporter (arrowheads). Cholera toxin subunit B (CTB immunoreactivity confirms accuracy of intrasciatic injection. (C) Image of L4 DRG immunolabeled with neuronal marker NeuN. hM4Di expression is restricted to neurons. (D) Cell body size distribution of all NeuN⁺ neurons (purple), and all NeuN⁺/hM4Di⁺ neurons (yellow) in a total of 7 L4-L6 DRGs. Data is from n = 3 individual animals (2164 total neurons). (E) Colocalization of hM4Di⁺ neurons (arrowheads) with DRG neuronal subtype markers calcitonin gene-related peptide (CGRP) and IB4. (F) Correlation of the percentage of L4-L6 neurons that are hM4Di⁺ versus the fold change in hindlimb thermal withdrawal latency after CNO delivery. Data was analyzed using linear regression analysis. Scale bars = 100 μm (B), 50 μm (C), 20 μm (E).

We performed three different pilot studies to optimize AAV transduction, and found that co-injection of AAV6 with 600 mM NaCl yielded the greatest extent of transgene expression, with the mCitrine reporter expressed in 6.77 ± 1.56% of DRG neurons (Supplementary Fig. S1). We then sought to determine whether this low transduction rate was sufficient to observe behavioral effects in hM4Di-mediating silencing experiments. Animals were subjected to hindlimb thermal withdrawal latency testing at 4 weeks post-AAV injection, both before and 30 min after systemic administration of the DREADD agonist, clozapine-N-oxide (CNO). We observed a significant positive correlation between the percentage of hM4Di+ neurons in the L4-L6 DRG and the change in hindlimb withdrawal latency post-CNO delivery (p = 0.0106; Fig. 1F). Indeed, one animal with 3.86% of DRG neurons transduced with hM4Di had a 2.57-fold increase in hindlimb withdrawal latency after CNO delivery. Hence, a modest amount of transduction in the DRG is sufficient to produce significant changes in behavior upon hM4Di-mediated nociceptor silencing. It is important to emphasize that even with optimal transduction conditions, only between 5-10% of all DRG neurons express the mCitrine reporter (Supplementary Fig. S1). Because small-diameter nociceptors comprise approximately 70% of DRG neurons,⁴⁶ this indicates that a maximum of approximately 15% of nociceptors are actually expressing the hM4Di transgene in this study. Hence, not all nociceptors are subjected to silencing using our experimental approach.

Acute nociceptor silencing following spinal cord injury improves chronic thermal sensitivity

Because nociceptor hyperactivity or noxious stimulation during the acute phase of spinal cord injury is thought to contribute to long-term maladaptive outcomes,^{7,9–13,47-49} we hypothesized that acute nociceptor silencing following SCI would mitigate these negative outcomes. To test this hypothesis, we conducted a chronic study in which rats were subjected to a battery of behavioral assessments over 10 weeks following moderate thoracic (T10) contusion SCI, with or without sustained hM4Di-mediated nociceptor silencing for the first 14 days post-injury (Fig. 2A). Spinal cord-injured and sham animals received bilateral sciatic nerve injections of either AAV-hM4Di (SCI-hM4Di) or the control vector AAV6-hSyn-eYFP (SCI-YFP) four weeks prior to SCI, to allow for peak expression of transgene product at the time of surgery. Control animals received laminectomy only, without SCI (sham), or received no interventions at all (naïve). Experimental groups are described in Table 1. We did not identify any significant differences in the contusion displacement values between the SCI-YFP and SCI-hM4Di groups (Fig. 2B). Moreover, we did not identify any significant differences in baseline behavioral scores (pre-AAV injections) and post-AAV injection scores (pre-SCI) on any behavioral outcomes. This demonstrates that there were no adverse effects of intrasciatic AAV injections on behavior.

FIG. 2.

Nociceptor silencing in the acute phase of spinal cord injury (SCI) reduces the incidence of thermal sensitivity following spinal cord injury. (A) Experimental design. Timeline is relative to day of SCI surgery (Day 0). Baseline scores for all behavioral assessments were recorded on Day-35. Either AAV6-hM4Di or AAV6-yellow fluorescent protein (YFP) was injected into bilateral sciatic nerves on Day-28. Moderate thoracic (T10) contusion (SCI) was administered on Day 0. Immediately following SCI, either clozapine-N-oxide (CNO) or vehicle was administered in drinking water until Day 14. Behavioral assessments were conducted until Day 70, when rats were sacrificed for histological analysis. Illustration created with Biorender.com. (B) Contusion displacement values for individual animals in the study. (C) Hindlimb thermal withdrawal latency (Hargreaves test) scores, normalized to baseline (pre-injury) scores. “Baseline” represents an average of 3 independent pre-injury baseline scores per animal. *p < 0.05, **p < 0.01, ***p < 0.001 for SCI-YFP vs. SCI-hM4Di, #p < 0.05, ##p < 0.01 for Naïve vs. SCI-YFP by two-way repeated measures analysis of variance + Tukey's multiple comparisons test. Naïve (n = 7), SCI-YFP (n = 13), SCI-hM4Di (n = 13). (D) Frequency of thermal hypersensitivity for the left and right hindpaws exhibited by individual animals from 14-70 days post-injury (DPI). Each row represents an individual animal's response on the indicated day. Hypersensitivity (pink) is defined as an animal scoring lower than 1 standard deviation below the mean of the baseline (pre-injury) scores. (E) Frequency of positive responses (e.g., withdrawal or paw licking) to acetone application to the hindpaws following SCI, normalized to baseline scores. “Baseline” represents an average of three independent pre-injury baseline scores per animal. Naïve (n = 7), SCI-YFP (n = 9), SCI-hM4Di (n = 8). All data are mean ± standard error of the mean.

We first examined the effects of nociceptor silencing on thermal sensory function following SCI. At the first time-point tested following SCI (14 DPI), animals in the SCI-yellow fluorescent protein (YFP) group showed significantly reduced withdrawal latencies to a thermal stimulus relative to naïve controls, and this continued until 35 DPI (Fig. 2C). Withdrawal scores for SCI-YFP animals were also significantly less than SCI-hM4Di animals until 42 DPI. SCI-hM4Di animals were never significantly different than naïve animals at any time-point. To better understand how individual animals were affected by treatment, we collected withdrawal latencies prior to injury for all rats, and determined the “normal range” of paw withdrawal for each animal as equal to the mean paw withdrawal threshold ±1 standard deviation, similar to what has been previously described.⁵⁰ Rats were determined to exhibit thermal sensitivity if they scored below their normal range at a given time-point. This allowed us to generate a frequency plot of sensitivity for the left and right hindlimbs for all animals (Fig. 2D). From this data, it is evident that animals in the SCI-YFP group exhibited hindpaw sensitivity more frequently than SCI-hM4Di animals. We also measured cold allodynia using the acetone test^51,52; however, we did not detect any statistically significant differences between groups at any point post-SCI (Fig. 2E). Together, these data demonstrate that acute nociceptor silencing prevented the development of thermal hyperalgesia but did not affect cold allodynia.

We also performed mechanical sensitivity (von Frey) testing to determine whether there was an effect of treatment on mechanical allodynia, which has previously been shown to develop following thoracic contusion SCI.^8,38,53,54 However, we did not observe the development of mechanical allodynia in any group (Supplementary Fig. S2). Rather, beginning at 14 DPI, SCI-hM4Di animals exhibited significantly increased withdrawal thresholds versus shams, and with the exception of 21 DPI scores, remained significantly elevated throughout the remainder of the study. In contrast, scores of SCI-YFP animals were not significant different than sham groups at any time-point. These data suggest that hM4Di activation in nociceptors from 0-14 DPI impart a lasting mechanical desensitization effect that is sustained for weeks after treatment, up to at least 10 weeks post-injury (Supplementary Fig. S2).

Acute nociceptor silencing following spinal cord injury improves locomotor functional outcomes

We also assessed locomotor function in the same animals. A repeated measures analysis of variance (ANOVA) identified a significant effect of treatment on open field locomotor (Basso, Beattie, and Bresnahan [BBB]) scores (p = 0.0489; Fig. 3A), indicating that acute nociceptor silencing improved long-term locomotor recovery. We next converted raw BBB data to BBB subscores, as previously described.³⁹ The subscore is a metric that is weighted by paw position and toe clearance, aspects of locomotion for animals that score in the middle range of the scale. We found that overall, there was a significant effect of treatment on BBB subscores (Fig. 3B). We further analyzed BBB scores at 70 DPI and found that only hM4Di-treated animals had scores of 13 and above, indicating frequent to consistent stepping and frequent coordination (Fig. 3C). We also evaluated locomotion using CatWalk gait analysis from 14-70 DPI. We examined several measures of gait output from the CatWalk gait analysis program; however, we did not observe any significant differences in individual outcomes between SCI-YFP and SCI-hM4Di treated rats during recovery (Supplementary Fig. S3). We therefore performed probabilistic linear discriminant analysis (pLDA) of CatWalk data. This is a linear discriminant analysis, combining several aspects of gait recovery into a single score of gait index, which was previously shown to reveal differences in gait following SCI in rats.⁴¹ We found that compared with the naïve and sham groups, both SCI groups exhibited significantly lower pLDA scores at all time-points post-SCI (p < 0.0002 in all cases); however, the SCI groups were not significantly different to each other (Fig. 3D). We also performed a separate two-way repeated ANOVA analysis to compare only the SCI groups and found a near-significant effect of time × treatment [F (8, 179) = 1.877, p = 0.0662]. Taken together, these data demonstrate an effect of early nociceptor silencing on locomotor recovery into the chronic phase of SCI.

FIG. 3.

Acute nociceptor silencing improves locomotor functional recovery following spinal cord injury (SCI). (A) Basso, Beattie, and Bresnahan (BBB) open field locomotor scores for SCI-yellow fluorescent protein (YFP) and SCI-hM4Di animals from 1-70 days post-injury (DPI). BBB scores for sham and naïve animals are not included in the analysis. *p = 0.0489 effect of treatment by two-way repeated measures analysis of variance (ANOVA). SCI-YFP (n = 13), SCI-hM4Di (n = 13). (B) BBB subscores for SCI-YFP and SCI-hM4Di animals. *p = 0.0214 effect of treatment by two-way repeated measures ANOVA. SCI-YFP (n = 13), SCI-hM4Di (n = 13). (C) Number of animals in each group with BBB scores of 10 (occasional stepping), 11 (frequent to consistent stepping, no coordination), 12 (frequent to consistent stepping, occasional coordination), and 13 (frequent to consistent stepping, frequent coordination) at 70 days post-SCI. SCI-YFP (n = 13), SCI-hM4Di (n = 1 3). (D) Probabilistic linear discriminant analysis scores for animals in all groups. Naïve (n = 7); Sham-YFP (n = 6); Sham-hM4Di (n = 6); SCI-YFP (n = 13); SCI-hM4Di (n = 13). Scores for both SCI-YFP and SCI-hM4Di are significantly lower than the naïve, Sham-YFP, and Sham-hM4Di groups at all time-points (see Supplementary Table S2 for details). All data are mean ± standard error of the mean.

In a pilot study, we observed that the percentage of h4MDi-transduced DRG neurons in uninjured rats positively correlated with thermal withdrawal scores (Fig. 1F). Because we observed that there was some variability in behavioral outcomes of injured animals (i.e., some animals exhibited more frequent sensitivity than others or had differences between left and right sides; Fig. 2D), we attempted to correlate these behavioral outcomes with the proportion of transduced nociceptors in each animal. However, we found that AAV transgene expression was undetectable in animals transduced with AAV-hM4Di, but not control AAV-YFP, at 14 weeks post-AAV injection (10 weeks post-SCI; Supplementary Fig. S4). This suggests that the hM4Di receptor, along with the mCitrine reporter, may be downregulated in DRG neurons at some point between 4 and 14 weeks.

Lesion severity does not significantly differ between treatment groups

To better understand the mechanisms of how treatment impacted behavioral outcomes, we performed histological analyses on spinal cord tissue. We first analyzed some commonly assessed measures of lesion severity in order to determine whether nociceptor silencing impacted the size or extent of the spinal cord lesion. Lesion volume, calculated by drawing regions of interest around GFAP⁺ reactive glial cell layers immediately surrounding the lesion cavity, was not significantly different between SCI-YFP and SCI-hM4Di subjects; however, there was a great deal of variability in the lesion volume from animal to animal (Fig. 4A, 4B). We next quantified the amount of spared gray matter as an alternative indicator of lesion size, as previously described.⁵⁰ Again, we did not detect any significant differences between treatment groups, with significant variability from animal to animal (Fig. 4C, 4D).

FIG. 4.

Histological measures of lesion severity do not differ between spinal cord injury (SCI) groups. (A) Images of glial fibrillary acidic protein (GFAP) immunoreactivity at the lesion epicenter for SCI-yellow fluorescent protein (YFP) and SCI-hM4Di subjects. (B) Quantification of lesion volume based on GFAP immunoreactivity. SCI-YFP (n = 13), SCI-hM4Di (n = 13). (C) Images of NeuN immunoreactivity at the lesion epicenter. Regions of spared gray matter containing NeuN⁺ neurons are outlined with dotted lines. (D) Quantification of the average volume of spared gray matter per 180-μm segment of spinal cord tissue (6 consecutive tissue sections). SCI-YFP (n = 13), SCI-hM4Di (n = 12). (E) Quantification of the total amount of white matter volume in a 2.34-mm length of tissue centered around the lesion epicenter. ****p < 0.0001 by one-way analysis of variance (ANOVA) + Tukey's multiple comparisons test. Naïve (n = 5); SCI-yellow fluorescent protein (YFP) (n = 13); SCI-hM4Di (n = 13). (F) Image of beta-tubulin (Tuj1) immunoreactivity. Inset: beta-tubulin immunoreactivity in the ventral spinal cord white matter. (G) Quantification of beta-tubulin intensity in the spared spinal cord white matter. *p < 0.05, **p < 0.01 by one-way ANOVA + Tukey's multiple comparisons test. Naïve (n = 7); SCI-YFP (n = 13); SCI-hM4Di (n = 13). All data are mean ± standard error of the mean. Scale bars = 250 μm.

We next analyzed the total area of white matter based on beta-tubulin immunoreactivity, and even though both SCI groups had significant white matter loss compared with intact controls, SCI-YFP was not significantly different than SCI-hM4Di (Fig. 4E). Finally, we also assessed beta-tubulin immunoreactivity as a measure of the axon density in the remaining white matter (Fig. 4F). Similar to our observations with white matter sparing, we observed a significant reduction in beta-tubulin immunoreactivity in both SCI groups compared with naïve controls but no significant differences between SCI groups (Fig. 4G). Thus, there were no apparent changes in histological measures of lesion anatomy between SCI-YFP and SCI-hM4Di groups.

Axon growth is not influenced by treatment

Because the AAV-hM4Di treatment acts by silencing nociceptors in the acute phase of injury, we next evaluated whether nociceptive axon sprouting was altered by treatment. We evaluated density of CGRP⁺ and IB4-binding axons in the superficial dorsal horn of the spinal cord, both at the lesion epicenter (T10) and in the lumbar spinal cord (L4). We found that both spinal levels, density of both CGRP⁺ and IB4-binding axons did not differ significantly between groups (Fig. 5A-H). Notably, neither SCI group were significantly different than the naïve group in any of these analyses, suggesting that SCI did not alter the density of nociceptive axons in the dorsal horn. Maladaptive plasticity of nociceptors has previously been shown to result in sprouting of CGRP⁺ axons into the deep dorsal horn after SCI.^43,55,56 We therefore analyzed CGRP immunoreactivity in laminae III/IV of the lumbar (L4-L6) dorsal horn; however, we did not detect any significant differences among treatment groups (Supplementary Fig. S5). Finally, we also analyzed the density of presynaptic punctae on ChAT⁺ motor neurons in the L4 spinal cord. We quantified the average number of synaptophysin⁺ punctae that were present on the ChAT⁺ cell bodies in the spinal cord ventral horn (Fig. 5I). However, we failed to detect any differences between naïve, SCI-YFP, and SCI-hM4Di (Fig. 5J).

FIG. 5.

Axon density is not influenced by treatment. (A, C) Immunoreactivity of calcitonin gene-related peptide (CGRP)⁺ nociceptive axons (A) at the lesion epicenter in the T10 thoracic (Th) cord, and (C) in the L4 lumbar (Lb) spinal cord. (B) Density of CGRP⁺ axons in the T10 spinal cord. Naïve (n = 7); SCI-yellow fluorescent protein (YFP; n = 13); SCI-hM4Di (n = 12). (D) Density of CGRP⁺ axons in the L4 spinal cord. Naïve (n = 7); SCI-YFP (n = 12); SCI-hM4Di (n = 11). (E, G) Immunoreactivity of IB4-binding nociceptive axons in the (E) T10 and (G) L4 spinal cord. (F) Density of IB4-binding axons in the T10 spinal cord. Naïve (n = 7); SCI-YFP (n = 13); SCI-hM4Di (n = 13). (H) Density of IB4-binding axons in the L4 spinal cord. Naïve (n = 7); SCI-YFP (n = 10); SCI-hM4Di (n = 10). (I) Density of synaptophysin⁺ (Syp) punctae on ChAT⁺ motor neurons in the L4 spinal cord. (J) Quantification of the average number of Syp⁺ punctae per motor neuron in the L4 spinal cord. Naïve (n = 7); SCI-YFP (n = 12); SCI-hM4Di (n = 10). All data are mean ± standard error of the mean; all data were analyzed by one-way analysis of variance + Tukey's multiple comparisons test. Scale bars = 100 μm (A, C, E, G), 25 μm (I).

Discussion

In this study, we selectively delivered the hM4Di chemogenetic construct to lumbar DRG nociceptors through intrasciatic delivery of AAV6. We confirmed that AAV6-mediated transgene expression is restricted to small diameter nociceptors, confirming previous findings in mouse and non-human primate studies.^32-34 Using this approach, we have shown that acute chemogenetic nociceptor inhibition significantly attenuated thermal hyperalgesia and improved hindlimb locomotor recovery. These findings highlight the therapeutic efficacy of a nociceptor silencing strategy to yield dual beneficial effects on motor dysfunction as well as pain-associated behavior. However, it is important to emphasize that the effects of treatment we have observed on locomotor recovery are modest. While statistically significant effects of treatment were observed on BBB scores, the biological degrees of improvement are subtle and it is up for debate whether this constitutes a biologically meaningful degree of functional recovery. Although we did not identify the mechanistic basis underlying these effects, it is possible that our strategy prevents or inhibits maladaptive plasticity within nociceptive circuitry following SCI. Future work is needed to investigate changes in the electrophysiological properties of DRG neurons in response to treatment; for example, to characterize levels of spontaneous activity during and after hM4Di-mediated inhibition.^7–10

It is interesting to note that acute nociceptor silencing for 14 days after SCI imparted differential effects with regard to distinct modalities of sensation. SCI resulted in a significant reduction in thermal paw withdrawal latency in the YFP-treated subjects, but this effect was rescued with hM4Di-mediated nociceptor inhibition. It is interesting to note that SCI-YFP subjects exhibited significantly lower thermal paw withdrawal scores than controls until 35 DPI, but from 42-70 DPI their group mean scores increased such that they were not significantly different from control group mean scores. This suggests that thermal hyperalgesia resolved over time, at least in a subset of injured animals in this study. This is not likely to be attributed to damage to the sciatic nerve resulting from AAV injection, because we did not observe signs of inflammatory or axotomy at the injection site in any animals (data not shown). Graphs illustrating the frequency of hindlimb sensitivity confirm this and show that a few animals in each SCI group remain hypersensitive at the end of the study (Fig. 2D). One possible explanation for some animals showing reduced hyperalgesia over time while others do not could be variability in lesion anatomy; for example, neuronal sparing in the dorsal horn, as we have previously reported.⁵⁰

In contrast, we failed to observe mechanical allodynia in either SCI group, regardless of treatment (Supplementary Fig. S2). This is in contrast to several previously published studies, in which rats developed mechanical allodynia after moderate thoracic contusion.^{9,38,53,54,57,58} It is unclear why this discrepancy exists. It is possible that the severity and location of injury may vary slightly between labs or even surgeons, resulting in different extents of damage along the dorsoventral axis. We previously reported that dorsal gray matter sparing was positively correlated with the development of mechanical sensitivity in a mouse model of cervical hemicontusion.⁵⁰ However, we observed dorsal horn gray matter sparing in many of the animals of this study, and we rarely observed mechanical sensitivity. Future work is needed to uncover the anatomical basis of mechanical allodynia in the current model and to understand how differences in surgical techniques between labs might contribute to these differences in outcomes. While mechanical paw withdrawal thresholds of the SCI-YFP group did not significantly differ from controls (except at 14 DPI, at which time-point scores of all SCI groups were increased), withdrawal thresholds of SCI-hM4Di subjects remained significantly elevated compared with all other groups from 28 DPI throughout the duration of the study. It is unclear why these subjects became “desensitized” to mechanical stimulation following treatment, but not to thermal stimulation. Because different modalities of sensation and nociception are conveyed through distinct neural circuits,⁵⁹ it is plausible that the hM4Di-silencing strategy used here acts to prevent maladaptive plasticity in thermal sensory pathways while permanently attenuating activity of mechanical sensory pathways. More mechanistic work is needed to characterize how transient hM4Di-mediated nociceptor silencing affects the long-term physiology of distinct populations of sensory neurons.

One caveat of this study is that we were not able to compare viral transduction efficiency with behavioral outcomes because transgene expression was diminished at chronic time-points. Notably, this downregulated expression was observed only in hM4Di-treated animals, but not control YFP-treated animals. It is not clear exactly when receptor downregulation occurs; a future time course study will be needed to explore this. Downregulation of DREADD receptors has been suggested as a possibility for chemogenetics studies, due to the normal internalization/downregulation response of G-protein coupled receptors to repeated activation.²⁰ However, we are not aware of any reports of chronic hM4Di downregulation. Regardless of the mechanism by which this downregulation occurs, our therapeutic strategy only requires hM4Di expression through the subacute phase of SCI, as the receptor-mediated silencing only took place for the first 14 days post-injury.

One limitation of our study is that we only used female rats. We chose to use only females for logistical reasons, in order to avoid any confounding factors of testing males and females in the same testing environment. Previous studies have highlighted sex differences in pain mechanisms between males and females, reporting that both in rodent models and human patients, females have shown a heightened sensitivity to pain compared with males.^60–66 Therefore, it will be important to repeat this study using male rats in order to determine whether there are sex-dependent differences in outcomes following chemogenetic nociceptor silencing. Another limitation of our study is that we failed to detect any histological correlates of improved functional outcomes. Histological assessments of thoracic spinal cord tissue suggested statistical trends toward reduced lesion volume, increased neuronal sparing and increased CGRP+ axon sprouting in hM4Di-reated animals. However, we did not observe significant differences in lesion severity, gray matter sparing, white matter sparing, or the density of presynaptic 5-HT⁺ punctae onto ChAT⁺ motor neurons in the lumbar spinal cord.

Further, we did not detect sprouting of CGRP⁺ fibers into deeper dorsal laminae, in contrast to previous studies that demonstrated this phenomenon in the acute to chronic periods after thoracic SCI.^43,55,56 It is unclear why we failed to replicate these previous findings, but perhaps this can be attributed to slight differences in the surgical model or spinal levels assessed. We did not use opioid drugs such as buprenorphine or morphine in this study; rather, we used the anti-inflammatory drug banamine for analgesia. Our failure to detect histological differences suggest that treatment works through a different mechanism of action that may be activity-based, and not necessarily reflected in fixed tissue sections that were collected at the study endpoint. Previous studies have shown that the onset and maintenance of neuropathic pain requires an ongoing spontaneous activity within the soma of nociceptors, which further requires continual activity of signaling molecules like Adenylyl Cyclase, cAMP, and protein kinase-A.^7,9,10 Previous work has also shown that hM4Di expression in peripheral nociceptor neurons results in ligand-dependent and ligand-independent changes in ion channel activity and secondary messenger signaling.⁶⁷ Therefore, in the context of our study we speculate that the transient activation of inhibitory hM4Di early after SCI prevents the setting in of maladaptive spontaneous hyperactivity within the DRG nociceptors, potentially through downregulation of cAMP, adenylyl cyclase and permanent hyperpolarization.

Our findings suggest that the approach of hM4Di-mediated nociceptor silencing can be applied to other experimental models of SCI-induced neuropathic pain. For instance, several studies have showed that a unilateral cervical hemicontusion injury results in the development of at-level mechanical allodynia.^50,68-70 In this case, this chemogenetics approach could be applied to study changes in forelimb pain-associated behavior. Further, the gene delivery technique we have used to transduce DRG nociceptors with hM4Di, has potential for clinical translation. Several AAV-based gene therapy approaches have successfully been tested in clinical trials.^69–72 Overall, our findings not only underscore the role of DRG hyperactivity in the development of neuropathic pain after spinal cord injury, but also suggest a new potential therapeutic approach to attenuate sensory hyperalgesia and promote locomotor recovery.

Transparency, Rigor, and Reproducibility Summary

The study design was registered following completion of the study at the Open Data Commons for Spinal Cord Injury (https://odc-sci.org/data/851). The analysis plan was not formally pre-registered. A sample size of 16 rats per group for the SCI groups, six rats per groups for the sham groups, and seven rats for the naïve group was planned based on a power analysis with anticipated effect size of 1.5 for the primary outcome measure (BBB scores) calculated to yield 80% power to detect a statistically significant difference in SCI groups using a two-way repeated measures ANOVA and p value <0.05, and taking into account attrition due to mortalities during surgery. Rats were randomly assigned to treatment group using a random number generator. Surgeries, tissue processing, and data analysis were performed by investigators blinded to treatment group. All rats were tested between 0900 and 1400 in a fed, well-rested state. AAVs were stored at -80°C for 1 month prior to use. Tissue sections were stored at 4°C for 3-6 months prior to use. Experimental materials were analyzed in a single batch for each histological stain performed. The specificity of antibodies used for immunohistochemistry was confirmed using negative controls (no primary antibody) and comparison to historical results. No replication or external validation studies have been performed or are planned/ongoing at this time to our knowledge. Data from this study are available in a FAIR data repository (odc-sci.org) and can be accessed at http://doi.org/10.34945/F5730S. There is no analytic code associated with this study. No future use of these samples is possible because insufficient quantities remain.

Data Availability

All data generated or analyzed during this study are included in this published article and the supplementary files. Raw data will be made available upon request, within 2 weeks of the request. Additionally, data from this study are available in a FAIR data repository, the Open Data Commons for Spinal Cord Injury (odc-sci.org) under http://doi.org/10.34945/F5730S, and can be accessed at https://odc-sci.org/data/851.

Authors' Contributions

P.A.K. designed the study, performed experiments, analyzed data, and wrote the manuscript.

J.S. performed BBB testing, animal perfusions, DRG dissections, cryosectioning, immunohistochemistry, image acquisition and data analysis.

Y.K. performed immunohistochemistry and image analysis

J.H. performed immunohistochemistry, image acquisition, and image analysis.

A.L. performed CatWalk testing, animal perfusions, and immunohistochemistry.

E.K. performed BBB testing, animal perfusions, cryosectioning, immunohistochemistry, and image acquisition.

B.N. performed CatWalk testing, AAV injections, and assisted with surgeries.

R.L.A. performed immunohistochemistry and data analysis.

A.B. calculated BBB subscores.

S.M. performed Hargreaves testing, immunohistochemistry, and data analysis.

S.L. cryosectioned DRG tissue and conducted image acquisition.

M.A.H. contributed to statistical analysis.

J.N.D. designed the study, performed experiments, analyzed data, and wrote the manuscript.

Funding Information

We gratefully acknowledge funding from Mission Connect, a program of TIRR Foundation (#018-001 to J.N.D.), National Institutes of Health (R01NS116404 to J.N.D.), and the Craig H. Neilsen Foundation (#546639 to J.N.D.).

Author Disclosure Statement

No competing financial interests exist.