Sigma-1 receptor activity in primary sensory neurons is a critical driver of neuropathic pain

Abstract

The Sigma-1 receptor (σ₁R) is highly expressed in the primary sensory neurons (PSNs) that are the critical site of initiation and maintenance of pain following peripheral nerve injury. By immunoblot and immunohistochemistry, we observed increased expression of both σ₁R and σ1R-binding immunoglobulin protein (BiP) in the lumbar (L) dorsal root ganglia (DRG) ipsilateral to painful neuropathy induced by spared nerve injury (SNI). To evaluate the therapeutic potential of PSN-targeted σ₁R inhibition at a selected segmental level, we designed a recombinant adeno-associated viral (AAV) vector expressing a small hairpin RNA (shRNA) against rat σ₁R. Injection of this vector into the L4/L5 DRGs induced downregulation of σ₁R in DRG neurons of all size groups, while expression of BiP was not affected. This was accompanied by attenuation of SNI-induced cutaneous mechanical and thermal hypersensitivity. Whole-cell current-clamp recordings of dissociated neurons showed that knockdown of σ₁R suppressed neuronal excitability, suggesting that σ₁R silencing attenuates pain by reversal of injury-induced neuronal hyperexcitability. These findings support a critical role of σ₁R in modulating PSN nociceptive functions, and that the nerve injury-induced elevated σ₁R activity in the PSNs can be a significant driver of neuropathic pain. Further understanding the role of PSN-σ₁R in pain pathology may open routes to exploit this system for DRG-targeted pain therapy.

Introduction

The σ₁R is a multi-functional, ligand-regulated molecular chaperone that is extensively expressed in the neurons of both central and peripheral nervous systems (CNS and PNS), where the σ₁R regulates multiple cellular functions, including modulating expression and activity of various receptors and ion channels important for neuronal homeostasis, synaptogenesis, neuronal plasticity, and nociception.^1–3 Mutations of σ₁R may cause juvenile amyotrophic lateral sclerosis,^4–7 distal hereditary motor neuropathy,⁸ and lower limb spasticity.⁹ Activation of σ₁R has been reported to be broadly protective in a variety of conditions, including Alzheimer disease, Parkinson disease, cancer, depression, amnesia, heart failure, and ischemic brain injury.^3,10–14

Prior studies have demonstrated a critical role of σ₁R in nociceptive function and pain hypersensitivity.^15–19 Global knockout of σ₁R in mouse and blockade by systemic administration of σ₁R small molecule antagonists attenuate pain behaviors in several pain models, while σ₁R agonism facilitates nociception and promotes sensory sensitization to mechanical stimulation.^18,20–27 These results from preclinical studies and clinical trials using σ₁R antagonists support an important role of σ₁R in modulating nociception and defined σ₁R activation as a constituent of nerve injury and inflammation-induced hypersensitization in pain pathways and provide a rationale for σ₁R antagonists for treating pain.^{15,21,25,26,28–31} As such, The σ₁R has been the focus of intense pharmaceutic research as a new analgesic target for pain treatment.

Many studies have attributed the analgesic properties of σ₁R antagonists to their central actions since σ₁R contributes to central sensitization.^{16,17,19,25,28,32–35} Specifically, increased expression of σ₁R is thought to modulate chronic pain by contributing to sensitization of CNS pathways,¹⁹ while analgesic effects of small molecule drugs result from inhibition of CNS σ₁R and reduction of central sensitization.¹⁷ However, the σ₁R is also highly expressed in the neurons of the PNS which are critical in initiation and maintenance of pain, so antinociceptive effects of systemic σ₁R antagonists could act at both central and peripheral sites. Indeed, it has been described that σ₁R expression is much higher in DRG than in CNS,³⁶ and peripheral nervous tissues show σ₁R activation in neuropathic pain and inflammatory pain models,^{18,22,37–39} local peripheral administration of σ₁R antagonists reduces inflammatory hyperalgesia,^18,40–46, and, mechanistically, σ₁R has been found to interact with and regulates the functions of nociceptive ion channels and receptors in DRG neurons.⁴⁷ These studies support that peripheral σ₁R activation may play critical roles in pain behavior probably by inducing plastic changes in the peripheral nociceptive pathways that promote nociception and the development of sensory hypersensitivity. However, direct links between DRG PSN-σ₁R and neuropathic pain has not been directly examined.

This study was designed to test the role of PSN-σ₁R in generating neuropathic pain. To interrupt σ₁R expression, we designed a dual promoter and bidirectional adeno-associated viral vector (AAV) encoding both a small hairpin RNA against σ₁R and a GFP reporter gene for in vivo optical monitor (AAV-σ₁RsiRNA). AAV vector was injected directly into the DRG for selective knockdown of σ₁R in the PSNs, sparing σ₁R expression in CNS. The GFP expression data indicates that AAV effectively delivered transgenes into DRG-PSNs; and subsequently reduced σ₁R expression was evident from immunohistochemistry. Silence of σ₁R expression selectively in PSNs by DRG-targeted AAV-encoded σ₁RsiRNA treatment produces a profound attenuation of pain behavior and concomitantly normalized hyperexcitability of nociceptive sensory neurons following spared nerve injury (SNI). These results indicate that the σ₁R activation in the PSNs after peripheral nerve injury can be a significant driver of neuropathic pain and further understanding the role of PSN-σ₁R in generation and maintenance of pain may open up a peripheral targeted σ₁R molecular therapy without CNS side-effects for those with intractable neuropathic pain.

Results

σ₁R and BiP protein levels are increased in the DRG after SNI.

We first performed immunohistochemistry (IHC) using a novel σ₁R antibody (Invitrogen/LF/PA5–77467) and a BiP antibody (Proteintech/11587–1-AP) to characterize the expression of these proteins in CNS and PNS, and found abundant immunopositivity of both σ₁R and BiP in lumbar DRG and spinal cord (SC). Specifically, in DRG from naïve rats, σ₁R and BiP were detected in neurons of all size groups, both displaying a ring profile in close proximity to the NKA1α-labeled neuronal plasma membrane but clearly within it (Fig. 1a–f), likely representing an interface of the mitochondrion-associated ER membrane (MAM).^48–50 We also tested the property of this σ₁R antibody by preincubation antibody solution with antigenic peptide (5 μg/ml) for 2 hr prior to IHC,⁵¹ which nearly complete elimination of immunostaining signals (Fig. S1). Similar findings were confirmed using a second σ₁R antibody (Abcam/ab53852, see below). σ₁R immunopositive profiles analyzed by these two antibodies are consistent with previous reports showing that σ₁R is highly expressed in the soma of all DRG neurons but generally absent in the sensory axon processes and afferent terminals in the dorsal horn of the spinal cord.^43,45,52 Although BiP has been predicted to be an ER luminal protein (https://www.expasy.org), BiP has also been detected in other cellular locations including the membranes and mitochondria.^53,54 We detected no immunopositivity of σ₁R or BiP in the satellite glial cell population (Fig. 1e–e2, 1f–f2). In the lumbar SC, σ₁R and BiP immunoreactivity (IR) were mostly detected in the motor neurons of the ventral and lateral horn (Fig. 1g, 1j), while immunolabeling of σ₁R and BiP were distinctly lower in the neurons and fibers of the dorsal horn (Fig. 1h–i, k–l). Double immunostaining σ₁R with CaMKII which is highly expressed in SDH neurons^55,56 verified scattered σ₁R-positive neurons in SDH (Fig. 1m–m1). Additionally, σ₁R and BiP were also extensively detected in brain neurons (Fig. S2). Based on above characterization and the availability of an antigenic peptide, the LF/PA5–77467 σ₁R antibody was used for further IHC and immunoblot experiments.

Figure 1. IHC characterization of σ₁R and BiP expression in DRG and SC of naïve rat.

Representative montage images show immunostained σ₁R expression profile (a), σ₁R co-labeled with NKAα1 (a1), BiP immunostaining (b), and BiP co-labeled with NKAα1 (b1). Amplified montage images of σ₁R (c, e), BiP (d, f), NKAα1 (c1, d1) and Vimentin (Vim, e1, f1), merged images of σ₁R with NKAα1 (c2) and BiP with NKAα1 (d2), and merged images of σ₁R with Vim (e2) and BiP with Vim (f2) show σ₁R (red arrowheads) localization in close vicinity but beneath the NKA1α-labeled neuronal plasma membrane and Vim-labeled satellite glial cell rings (green arrowheads) in DRG section from naive adult rat. In lumbar SC sections, the σ₁R (g) and BiP (j) immunoreactivity (IR, red) were mostly detected in the motor neurons reside in the ventral horn (VH) and lateral horn (IML) while immunolabeling of σ₁R (h) and BiP (k), with dorsal horn (DH) layer IIa marked by IB4 (i, l), are lower in the neurons and fibers. Dashed lines outline the SC grey matters and white matters are pseudocolored in blue. Scattered σ₁R positive cells (arrowheads) in SDH are colabeled with CaMKII (m-m2), verified their neuronal property. Scale bars: 100μm for all.

To quantify the effect nerve injury on σ₁R and BiP expression, DRG tissues were collected at 28d after SNI. The σ₁R is associated with cell organelle membranes and nuclear envelopes, where the σ₁R forms multiprotein complex assemblies.^43,48,57,58 Therefore, in the present study, we prepared separate cytosol and membrane fractions from DRG homogenates using a well-defined ProteoExtract Subcellular Extraction Kit.⁵⁹ We detected σ₁R and BiP proteins in both the cytosolic and membrane-enriched fractions from pooled L4/L5 DRG (Fig. 2a). Both σ₁R and BiP protein levels were significantly increased following SNI, compared to contralateral control samples (Fig. 2b). Additionally, the ratio of membrane to cytosol of BiP protein level was significantly higher after SNI than in control animals (2.1 ± 0.9 for contralateral, 3.6 ± 0.5 for injured side, p< 0.001, n= 9 each), possibly due to injury-induced translocation of BiP to the cell membrane components, similar to the stress-induced BiP translocation observed in other cell types.^53,54 Together, these data indicate the peripheral nerve injury induces upregulation of both the σ₁R and BiP protein levels. IHC also provided optical images consistent with an increased membrane (MAM) profile of both σ₁R and BiP in SNI DRG, showing higher immunolabeling intensity in SNI compared to control for both σ₁R and BiP, particularly in close proximity but beneath NKA1α-positive plasma membrane (Fig. 2c–c3, 2d–d3, 2e–e3, 2f–f3).

Figure 2. Increased σ₁R and BiP levels in DRG following SNI.

The NKA1α-deficient cytosolic fraction and NKA1α-enriched membrane fraction were extracted from the pooled ipsilateral (ipsi.) L4/L5 DRG tissues at 4-wk after SNI using contralateral (contra.) DRG as controls, and subjected to immunoblotting as shown in the representative immunoblots of σ₁R, BiP, Iba-1, NKA1α, and Tubb3 of cytosol (a, left panels) and membrane fractions (a, right panels), respectively. Bar charts (b) show densitometry analysis of immunoblots (***p<0.001, unpaired, two-tailed Student’s t-test). The number in each bar is the number of analyzed DRG per group. IHC of σ₁R (c), BiP (e), and each colabeled with NKA1α (d, f), respectively, are compared, from the ipsi. and the contra. DRG following 4-wk of SNI, as indicated. Scale bars: 100μm for c-f. Insets of c-f are amplified showing both σ₁R and BiP IR signals are apparently increased in the ipsi. DRG (cf. c1-c3 and d1-d3 for σ₁R, as well as e1-e3 and f1-f3 for BiP). Red arrowheads point to σ₁R/BiP and green arrowheads to NKA1α.

AAV6-σ₁RsiRNA induces efficient σ₁R knockdown in vivo

We have developed an AAV6-encoded GFPσ₁RsiRNA (AAV6-σ₁RsiRNA, Fig. S3), and have previously verified its σ₁R silencing effect both in vitro and in vivo,^60,61 using AAV6-GFPscramble (AAV6-GFPsc) as a control (see below in Methods). AAV6 was used as the delivery vector because this serotype efficiently transduces all size groups of DRG neurons. Here, transduction rate and σ₁R knockdown effects were determined on DRG sections by IHC using GFP and σ₁R antibodies 4 weeks after AAV injection into DRG of naïve rats. Results show GFPσ₁RsiRNA expression in 40% ± 4% of neurons (915 out of 2564 total neuronal profiles positive for pan-neuronal marker β3-tubulin; full count of n=4 sections per DRG, which were selected as every fifth section from the consecutive serial sections, n = 3 DRG). Both AAV6-GFPsc and AAV6-σ₁RsiRNA transduced DRG neurons were observed in all size groups of the neuronal population that also expressed σ₁R (Fig. 3a–a2, b–b2). Effectiveness of the construct is shown by significantly lower or absent σ₁R immunolabeling in the AAV6-σ₁RsiRNA transduced neurons, compared with non-transduced neurons or those transduced by the control vector AAV6-GFPsc (Fig. 3a3, b3). Sections from the SC show that siRNA-GFP expression was restricted to the PSNs without expression in the intrinsic SC neurons, and that the pattern of σ₁R expression in the lumbar spinal cord appears not to be affected by intraganglionic AAV6-σ₁RsiRNA (Fig. 3c–d). Similar findings of σ₁R knockdown were confirmed using a second σ₁R antibody (Abcam/ab53852, Fig. S4a–a2, b–b2). BiP expression in PSNs was not affected in AAV6-σ₁RsiRNA transduced neurons (Fig. S4c–c2, d–d2). Additionally, normal mechanical and thermal behavior continued throughout the full testing period of 4 weeks after AAV-σ₁RsiRNA injection in the rats without nerve injury surgery (Fig. 4a). This indicates that σ₁R-knockdown following σ₁RsiRNA expression in PSNs had no significant effect on the normal sensory threshold for these modalities, which is in agreement with previous reports that normal sensory function is not affected by genetic σ₁R deletion^21,27 or systemic σ₁R antagonists.^26,31,62

Figure 3. Efficient knockdown of σ₁R by DRG injection of AAV6-σ₁RsiRNA.

Representative IHC montage images (a-a2) show σ₁R expression (red), AAV6-GFPsc transduced neurons (green), merged image (a2), and σ₁R intensity of transduced and nontransduced small-, medium-, and large-sized neurons (a3). No difference of σ₁R intensity was found between transduced (filled arrowheads) and nontransduced cells. Representative montage images (b-b2) show σ₁R expression (red), AAV6-GFPσ₁RsiRNA transduced neurons (green), merged image (b2), and σ₁R intensity of transduced and nontransduced small-, medium-, and large-sized neurons (b3). σ₁R intensity was significantly reduced in transduced neurons (filled arrowheads) compared to nontransduced neurons (empty arrowheads), ***p<0.001, unpaired, two-tailed Student’s t-test. The number below or above each bar of panels a2 and b2, respectively, is the number of analyzed neurons per group. Lumbar SC σ₁R pattern (c-c2, red) was not affected after intraganglionic AAV6-GFPσ₁RsiRNA that results in GFPσ₁RsiRNA (green) transported to SDH presynaptic terminals. Inset of c2 is amplified with montage images showing σ₁R (d-d1, red) scattered in GFPσ₁RsiRNA positive presynaptic terminals (green) of SDH. Dashed lines outline the SC profile (c-c2) and enumerated laminae (d). The white matters of SC are pseudocolored in blue. Scale bars: 100μm for all.

Figure 4. Attenuation of SNI-induced pain behavior by AAV6σ₁RsiRNA-mediated PSN-σ₁R inhibition.

Panel a evaluates sensory sensitivity to innocuous punctate mechanical stimulation (von Frey, left) and to heat stimulation (right) at baseline and day 28 after AAV6-σ₁RsiRNA injection in noninjured rats (n=4). Left panels of b-f show the time courses for the group averages of sensitivity to von Frey (b), hyperalgesia behavior after touch with a pin (Pin, c), dynamic brush (d), sensitivity to heat (f), and acetone stimulation (Cold, f), before SNI and after DRG injection of either AAV6-GFPsc (filled circle, n = 10 rats) or AAV6-σ₁RsiRNA (empty circles, n = 10 rats). Injection of AAV vectors into the fourth and fifth lumbar DRG was performed immediately after the procedure of SNI, denoted by arrowheads. ^#p < 0.05, ^##p < 0.01, and ^###p < 0.001 for comparison to BL and *p < 0.05, **p < 0.01, and ***p < 0.001 for comparison between groups after treatment, respectively (b and e, repeated measures two-way ANOVA and Bonferroni post-hoc; c, d and f, nonparametric analyses by Friedman’s test with Dunn’s post-hoc). Right panels of b-f show averaged area under the curve (AUC) calculated for each individual for the time period following vector injection for von Frey (b), pin (c), brush (d), heat (e), and cold (f). *p < 0.05, **p < 0.01, and ***p < 0.01 for AUC comparison between groups (unpaired, two-tailed Student’s t-tests). Representative IHC montage images (g, h) show PSN-σ₁R expression 5 weeks post SNI with AAV6-GFPsc (g) and AAV6-σ₁RsiRNA (h) treatment. Filled arrowheads point to the transduced neurons (g, h) and empty arrowheads to nontransduced neurons (h). Scale bars: 100μm for all.

AAV6-σ₁RsiRNA treatment attenuates SNI-induced neuropathic pain

We next evaluated whether sustained expression of σ₁RsiRNA in DRG neurons altered pain sensation following nerve injury. In the experimental design, animals were randomized into two groups. Peripheral nerve injury was performed by SNI, which followed during the same anesthetic by injection (immediate after SNI procedure) of either AAV6-σ₁RsiRNA or AAV6-GFPsc into both the L4 and L5 DRG, which are the principal DRG containing the somata of sensory fibers in the rat sciatic nerve.⁶³ Sensitivity to cutaneous mechanical and thermal stimulation was evaluated at baseline before surgery (SNI plus AAV injection) and after surgery for 5 weeks on a weekly basis, after which tissues were harvested for IHC characterization of transgene and target gene expression, and for electrophysiological recordings (see below). Behavioral evaluations (Fig. 4b–f) showed that all rats developed hypersensitivity to the various sensory modalities within 1 week after the surgery, which persisted for the 35-day evaluating period in animals receiving the control AAV6-GFPsc vector. However, animals receiving AAV6-σ₁RsiRNA showed a pattern of recovery for all modalities, initiated by week 2 after treatment, which is consistent with the time of onset of transgene expression.^64,65 This treatment effect included normalized thresholds for withdrawal from mild mechanical stimulation (von Frey) and heat, more frequent hyperalgesic-type responses (sustained lifting, shaking, grooming) after noxious mechanical stimulation (Pin testing), more frequent withdrawals from cold (acetone), and dynamic mechanical stimulation (brush). These findings suggesting that AAV6-mediated, DRG-targeted PSN-σ₁R silencing attenuated hypersensitivity induced by nerve injury in the SNI model of neuropathic pain. Notably, sustained analgesic effectiveness was observed for several weeks during the testing course of this study and likely retain longer since AAV6-encoded transgene expression lasts at least 3-month after single injection.⁶⁴ Efficient σ₁R knockdown in the PSNs by intraganglionic AAV6-σ₁RsiRNA in SNI rats was confirmed by IHC on the DRG harvested 5 weeks after treatment (Fig. 4g–h). Together, these results indicate that sustained expression of the σ₁RsiRNA selectively in the PSNs of the L4 and L5 DRG significantly attenuates injury-induced persistent neuropathic pain.

AAV6-σ₁RsiRNA treatment normalizes SNI-induced PSN hyperexcitability.

The PSNs are a significant source of ectopic afferent discharge after peripheral nerve injury. Persistent hyperexcitability of DRG sensory neurons following peripheral nerve injury, including SNI,⁶⁶ is a common pathophysiology underlying behavior signs of neuropathic pain.^67,68 σ₁R has been reported to interact with various proteins for regulating neuronal functions and tuning neuronal excitability.^1,2,16,69 We therefore determined whether AAV6-σ₁RsiRNA treatment reverses the enhanced neuronal excitability of nociceptive DRG neurons following SNI, using whole-cell current-clamp recording. Although SNI produces DRG with co-mingled injured and uninjured axons, nerve-injury can induce increase of voltage-gated ion channel activity in both axotomized neuron and adjacent intact neurons,^70–72 as has long been recognized that some molecules related to nociception are upregulated in the intact L4 DRG neurons after L5 SNL⁷³, leading to similar electrophysiological changes and increased discharge frequency in axotomized and neighboring intact DRG neurons.^72,74,75 Consistent with this notion, a coupled activation of DRG neurons (i.e. transglial signaling of DRG sandwich synapse) following nerve injury has been recently defined.^76,77 We therefore recorded from randomly chosen small neurons in cultures from dissociated L4 and L5 DRG. GFP-expressing small neurons dissociated from DRG injected with AAV6-GFPsc and AAV6-σ₁RsiRNA were used for recording. The effect of AAV6-σ₁RsiRNA treatment on the repetitive firing properties of DRG neurons was assessed by applying a series of 500-millisecond current injections to the DRG dissociated neurons. Results showed that the frequency of AP spikes evoked by progressively greater depolarizations in small-sized DRG neurons (<30μm) from SNI rats were significantly increased, compared to sham controls (Fig. 5). SNI increased excitability, which was normalized in AAV6-σ₁RsiRNA transduced neurons, whereas AAV6-GFPsc had no effect. Interestingly, significant reduction of firing frequency of evoked APs was observed in small-sized DRG neurons from control rats injected with AAV6-σ₁RsiRNA, but this σ₁RsiRNA-induced change of neuronal excitability did not affect normal sensory thresholds (Fig. 4a). Together, these findings indicate that reversal of nerve injury-induced sensory neuron hyperexcitability by σ₁R silencing may contribute to its attenuation of neuropathic pain behaviors.

Figure 5. Current-clamp analysis of AAV6σ₁RsiRNA transduction on DRG neuron excitability.

Representative action potential (AP) traces elicited by 500ms depolarizing current inputs of 60 pA (a) and 100 pA (b) from resting membrane potentials (RMP) were recorded from DRG neurons dissociated from the rats of sham, SNI, and SNI treated with AAV6-GFPsc or AAV6-σ₁RsiRNA, as indicated. Comparison of responses (number of APs evoked by a 500ms stimulus) for the populations of DRG neurons in different groups across a range of step current injections from 10 to 100 pA (c), two-way ANOVA of main effects of groups with Bonferroni post-hoc, ***p<0.001. The bar charts show analysis of AP numbers evoked by input current at 60 pA (d, left) and 100 pA (d, right) from resting membrane potentials (RMP), respectively. The number in each bar is the number of analyzed neurons per group. *** denote p<0.001, one-way ANOVA analysis of variance with Turkey post-hoc.

Discussion

The σ₁R, one of the two sigma receptors, has been the focus of intense preclinical research and clinical trials as a potential pharmacological target for pain treatment.^{15,18,21,24,26,31} However, σ₁R interacts with a large number of receptors and ion channels, and exerts a potent influence on multiple neurotransmitter systems. Thus, small molecule σ₁R inhibitors are inadequately specific for nociception. Also, systemic application cannot restrict the drugs to the neural pathways responsible for transduction and transmission of nociceptive signals, so side effects may result from σ₁R that is widely expressed in non-neural tissues, including lung, liver, gastrointestinal system, immune system, adrenal glands, and the heart,^14,60 making off-target side effects hard to avoid.^18,78 For example, decreased expression of σ₁R can facilitate hepatic tumorigenesis,⁷⁹ and removal of σ₁R can cause cardiac dysfunction^14,80, anxiety, depressive-like behavior, as well as memory alterations.^81,82 Most studies have attributed the analgesic effects of σ₁R antagonists to their central actions. However, there is also evidence suggesting that peripheral σ₁R can also participate in pain processes,¹⁷ although these only report changes in expression without confirmation of a functional role, with the exception of a study blocking σ₁R subcutaneously to produce relief of paw inflammation.²⁵ In contrast, we show that selectively targeting PSNs and inducing loss of σ₁R restricted to PSNs produces analgesia. Thus, delineation of the significance and contribution of the peripheral components of σ₁R inhibition on pain sensation could help in the development of therapeutic strategies that may achieve pain relief while avoiding CNS adverse effects and addiction.²⁶

At the DRG level, we had previously reported a decrease of σ₁R protein level in the L5 DRG proximal to L5 spinal nerve ligation.⁵² However, the σ₁R antibody used to generate these observations was subsequently called into question.⁴³ In the present study, we have used two alternative antibodies that show comparable findings of σ₁R immunolabeling pattern in DRG. We validated these antibodies by their failure to detect σ₁R in PSNs expressing an siRNA construct to knock down σ₁R. Further validation comes from their identification of σ₁R expression preferentially in the membrane-enriched fraction, consistent with their affinity for the MAMs. Although our present study uses the SNI peripheral nerve injury model that replicates clinical conditions, rather than the SNL model we use previously, the PSNs in the DRG proximal to the SNI consist of a mix of axotomized and intact neurons, so our contrasting results cannot be fully explained by the difference in these injury models and the exact reason for the discrepancy is unknown at present. We speculate that differences in preparing the sample for immunoblots may in part account for the differences. Specifically, we used a well-established commercial subcellular extraction kit that was unavailable at the time of previous study, which employs differential extraction and centrifugation to separately extract cytosolic and membrane fractions, whereas immunoblots in our earlier study was performed using soluble fraction of tissue homogenates after high-speed centrifugation, by which most of the membrane fractions enriched with σ₁R could be lost in the discarded pellet.⁵² Additionally, in comparison with the study of Mavlyutov et al.,⁴³ our present findings largely agree with theirs in showing high σ₁R in sensory neuron somata and low expression in the axonal processes and satellite glial cells. A remaining point of discordance is the claim by Mavlyutov et al.⁴³ of finding σ₁R expression in the plasmalemma, whereas our present findings by σ₁R co-labeling with an authentic plasmalemma marker support an expression pattern of σ₁R accumulation preferentially in a subplasmalemmal locations.

Increased σ₁R expression in PSNs following painful nerve injury raises the possibility that regulating PSN-σ₁R may produce analgesia. Selective targeting of sensory or nociceptive neurons in peripheral nerves is a clinically desirable goal to achieve effective analgesia while minimizing unwanted effects. Gene therapy targeting the DRG offers a potential means of long-term pain relief without off-site CNS and other systemic side effects, although potential clinical application requires an improved safe procedure. We and others have demonstrated that direct DRG delivery of AAVs encoding analgesic molecules can provide relief in chronic pain in animal models, with high transduction efficiency, flexibility for selective segmental localization, and minimal behavioral changes attributable to the surgical procedure.^65,83,84 Here, we extended the applicability of DRG-AAV strategy to the analgesic effectiveness of PSN-specific knockdown of σ₁R for neuropathic pain. Our novel findings indicate that AAV-mediated σ₁R silencing selectively in PSNs while preserving σ₁R expression in the CNS do not modify normal sensory sensitivity thresholds but exerts persist analgesic effects of SNI-induced sensitizing conditions and reduces pain behavior in rats. The results therefore support that the activation of peripheral PSN-σ₁R following nerve injury may be a critical driver in pain sensitization. Because activation of σ₁R has been found in various pain conditions,¹⁸ it will be of interest to address the analgesic efficacy of AAV-therapy for PSN-σ₁R inhibition in additional models of other pain etiologies.

A potential site of σ₁R pain pathophysiology may be elevation of sensory neuron membrane excitability. Our initial electrophysiological observations show that silencing of PSN-σ₁R is associated with reduction in the hyperexcitable state of sensory neuron membrane following peripheral nerve injury. These results support the conclusion of prior works that σ₁R can modulate neuronal excitability.^69,85 Injury-induced ectopic hyperactivity of PSNs causes hypersensitization in multiple sites of peripheral sensory nervous system, including augmented pain perception in the peripheral terminals, enhanced nociceptive signal transduction in PSN soma, and increased neurotransmission in spinal dorsal horn.⁶⁸ Our experiments did not investigate the anatomical sites at which PSN-σ₁R silencing could alter pain generation processes, but this could include peripheral sensory transduction at the neuronal termini in the receptive field, transmission of impulse trains through the T-junction, generation of ectopic activity at the injury site or neuronal soma, or in regulation of neurotransmitter release in the dorsal horn. Additionally, it is known that σ₁R interacts with various ion channels types,⁴⁷ and new findings also point to an important role of PSN-σ₁R in neuron-macrophage/monocyte communication in the DRG of painful neuropathy.⁸⁶ Therefore, the precise site and mechanism by which PSN-σ₁R knockdown produces analgesia remain to be determined.

In conclusion, we have shown that activation of σ₁R following SNI is associated with neuropathic pain and enhanced excitability of DRG neurons, and that PSN-specific knockdown of σ₁R alleviates pain hypersensitivity. Our study therefore presents support that PSN-σ₁R activation is a critical driver of neuropathic pain, and raises the possibility that AAV-mediated PSN-σ₁R inhibition could be a useful treatment of chronic neuropathic pain.

Materials and methods

Animals

Adult male Sprague Dawley (SD) rats weighing 100–125g body weight (Charles River Laboratories, Wilmington, MA) were used. All animal experiments were performed with the approval of the Zablocki VA Medical Center Animal Studies Subcommittee and the Medical College of Wisconsin Institutional Animal Care and Use Committee (Permit number: 3690–03) in accordance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals. Animals were housed individually in a room maintained at constant temperature (22±0.5°C) and relative humidity (60±15%) with an alternating 12h light-dark cycle. Animals were access to water and food ad libitum throughout the experiment, and all efforts were made to minimize suffering. The estimated numbers of animals were derived from our previous experience with similar experiments and the number of experiments needed to achieve statistically significant deviation (20% difference at p<0.05) based on a power analysis.^87,88

AAV constructs

A dual promoter and bidirectional plasmid expressing a U6 promoter-driven σ₁RshRNA with a CMV promoter-driven EGFP (referred to hereafter as GFP) and a plasmid coding U6 promoter-driven shRNA-scramble control (sc) with CMV promoter-driven GFP, both kindly provided by Dr. Tsung-Ping Su, IRP/NIDA/NIH), were used to construct single-strand AAVs coding either the U6 promoter-driven σ₁RshRNA or the U6-promoter-driven scramble. Efficacy of the σ₁R knockdown by these constructs has previously been demonstrated,^60,61 and also verified in our study using a neuronal NG18–15 cells (Not shown). To construct our AAV vectors, DNA fragments of either the U6 promoter-driven σ₁RshRNA or the scramble from above plasmids were released by Mlu I and subcloned into the Mlu I sites of a single-strand AAV expressing plasmid pAAV-CMV-GFP (Cell Biolabs, San Diego, CA). This generated pAAV-U6-σ₁RshRNA-CMV-GFP (Fig. S3) and pAAV-U6-scramble-CMV-GFP that encode either the σ₁RsiRNA or the scramble driven by U6 promoter and the GFP downstream a chimeric intron for enhancing transcription driven by the CMV promoter. These plasmids were used to pack AAV2/6- U6-σ₁RshRNA-CMV-GFP and AAV2/6- U6-scramble-CMV-GFP as a control (subsequently referred to as AAV6-σ₁RsiRNA and AAV6-GFPsc, respectively) for in vivo injection. AAV vectors were produced and purified in our laboratory by previously described methods.⁶⁴ This included AAV particle purification by optiprep ultracentrifugation and concentration by use of Centricon Plus-20 (Regenerated Cellulose 100,000 MWCO, Millipore, Billerica, MA). AAV titer was determined by PicoGreen (life technologies, Carlsbad, CA) assay, and final aliquots were kept in 1x phosphate buffered saline (PBS) containing 5% sorbitol (Sigma-Aldrich, St. Louis, MO) and stored at −80°C. The titers of AAV6-σ₁RsiRNA and AAV6-GFPsc vectors were 2.45 ×10¹³ GC/ml and 1.36 × 10¹³ GC/ml, respectively. The same lots of viral preparations were used for all in vivo experiments.

Microinjection of AAV vectors into DRG

AAV vector solution was microinjected into right L4 and L5 DRGs using previously described techniques.⁸⁹ Briefly, the surgically exposed intervertebral foramen was minimally enlarged by removal of laminar bone. Injection was performed through a micropipette that was advanced ~100 μm into the ganglion. Rats received L4 and L5 DRG injections of either AAV6-σ₁RsiRNA or AAV6-GFPsc (one vector per rat), consisting of 2 μl with adjusted titers containing a total of 2.0 ×10¹⁰ genome viral particles. Injection was performed over a 5-min period using a microprocessor-controlled injector (Nanoliter 2000, World Precision Instruments, Sarasota, FL, USA). Removal of the pipette was delayed for an additional 5 min to minimize the extrusion of the injectate. Following the injection and closure of overlying muscle and skin, the animals were returned to the animal house where they remained as the designed experiments required.

Neuropathic pain by spared nerve injury (SNI)

To model clinical traumatic painful peripheral neuropathy, we performed a SNI, an established model of peripheral nerve injury-induced chronic neuropathic pain.⁸⁸ Animals were anesthetized using isoflurane at 4% induction and 2% maintenance. Under anesthesia, the right sciatic nerve was exposed under aseptic surgical conditions by blunt dissection of the femoral biceps muscle. The sciatic nerve and its three branches (sural, common peroneal, and tibial nerves) were identified. The tibial and common peroneal nerves were then tightly ligated and transected distal to the ligation. The overlying muscle and skin were then sutured following surgery. Sham-operated rats were subjected to all preceding procedures without nerve ligation and transection.

Sensory behavioral testing

Behavioral tests were conducted between 9:00 AM and 12:00 AM. Animals were habituated in individual test compartments for at least one hour before each testing. Behavior tests carried out as previously described⁸⁹ were performed by personnel blind to AAV6-GFPsc (control) or AAV6-σ₁RsiRNA treatments. 1) Dynamic mechanical stimulation (Brush): animals were placed on a wire mesh platform, and a camel hair brush with a width of 5mm (Ted Pella Inc., Redding, CA) was passed along the bottom of the hindpaw from front to back in a smooth motion at a rate of approximately 2 cm/s. Each hindpaw was tested 3 times in alternating fashion. The frequency of withdrawal from the stimulus was recorded. 2) Cold stimulation: acetone was applied from a syringe attached to PE220 tubing to make a meniscus that was touched to the plantar surface of hindpaw, such that the drop spread out on the plantar surface of the paw without contact of the tubing to the skin. Each hindpaw was tested 3 times in alternating fashion. Any withdrawal was considered a positive response. The frequency of withdrawal from the stimulus was recorded. 3) Mild mechanical stimulation (von Frey): withdrawal threshold was determined using calibrated monofilaments (Patterson Medical, Bolingbrook, Illinois) with forces of 0.3, 0.5, 0.8, 1.0, 2.8, 5, 9, 14, and 24g, applied in an up-down fashion, allowing calculation of the 50% withdrawal threshold.⁹⁰ Beginning with the 2.8g filament, filaments were applied to the plantar skin with just enough force to bend the fiber and held for 1 s. If a response was observed, the next smaller filament was applied, and if no response was observed, the next larger was applied, until a reversal occurred, defined as a withdrawal after a previous lack of withdrawal, or vice versa. Following a reversal event, four more stimulations were performed following the same pattern. The forces of the filaments before and after the reversal, and the four filaments applied following the reversal, were used to calculate the von Frey threshold. Rats not responding to any filament were assigned a score of 25g. 4) Heat stimulation: this was performed using a device designed for the purpose of identifying heat sensitivity (Paw Thermal Stimulator System, University Anesthesia Research & Development Group, San Diego, CA). Rats were placed on a temperature-regulated glass platform heated to 30°C, and the lateral plantar surface of hindpaws stimulated with a radiant heat source (50W halogen bulb) directed through an aperture. The time elapsed from initiation of the stimulus until withdrawal (withdrawal latency) as detected by a series of photocells was measured. Each hindpaw was tested four times, and the withdrawal latency values averaged. 5) noxious mechanical stimulation (Pin): a point of 22g spinal anesthesia needle was gently applied 5 times to the plantar surface of hindpaw with enough force to indent but not puncture the skin. Five applications were separated by at least 10s, which was repeated after 2 min, making a total of 10 touches. For each application, this evokes either a simple withdrawal response with immediate return of the foot to the cage floor, or a response characterized by sustained elevation with grooming (e.g. licking or chewing the toes) and possibly shaking, lasting at least 1 s. This latter behavior was referred to as hyperalgesia behavior. This hyperalgesic response has been associated specifically with an aversive experience.⁹¹

Immunohistochemistry and imaging

Staining of paraffin-embedded DRG, spinal cord, and brain sections was performed by a standard fluorescent protocol, as previously described.⁹² In brief, 5μm-thick sections were de-paraffinized in xylene and rehydrated through graded alcohols, and treated by heat-induced epitope retrieval in 10mM citrate buffer, pH 6.0~7.0 (depending on the antibody used). Sections were first immunolabeled with the selected primary antibodies overnight at 4°C (Table 1). All antibodies were diluted in 1 x PBS, containing 0.05% Triton X-100 and 3% bovine serum albumin (BSA). Normal immunoglobulin G (IgG from same species as the first antibody) (Table 1) was replaced for the first antibody as the negative controls. The appropriate fluorophore-conjugated (Alexa 488 or Alexa 594, 1:2000) secondary antibodies (Jackson ImmunoResearch, West Grove, PA) were used to reveal immune complexes. The sections were washed three times 5 min each with PBS containing 0.05% tween-20 between incubations. To stain nuclei, 1.0μg/ml Hoechst33342 (Hoechst, ThermoFisher, Pittsburgh, PA) was added to the secondary antibody mixture. The sections were examined, and images acquired on a Nikon TE2000-S fluorescence microscope (El Segundo, CA), equipped with an Optronics QuantiFire digital camera and acquisition software (Ontario, NY), as well as filters suitable for selectively detecting the green, red and blue fluorescence. We also tested the specificity of σ₁R antibody (LF/PA5–77467) by preincubating the antibody solution with the specific σ₁R antigenic peptides (SRGHSGRYWAEISD, 5 μg/ml) for 2 hr prior to immunostaining, as described previously.⁵¹ For each comparative experiment, tissues from compared groups were prepared and stained concurrently and all images were acquired with identical settings for detector gain and under a 10x objective (0.5 numerical aperture at 2048 × 2048 pixel resolution) or 20x objective (0.3 numerical aperture at 1024 × 1024 pixel resolution). Some IHC images were captured under a Nikon C1 digital eclipse confocal microscope. For double label colocalization, images from the same section but showing different antigen signals were overlaid.

Table 1.

Primary antibodies and IgG controls used in this study

Antibodya	Host	Supplier/Cat#/RRIDc	Dilution
EGFP	Mouse monoclonal	SCB/sc-9996	1:100 (IHC), 1:400 (Wb)
IB4		LF/I21413	1.0μg/ml
σ1Rb	Rabbit polyclonal	LF/PA5-77467	1:500 (IHC), 1:1000 (Wb)
σ1Rb1	Rabbit polyclonal	Abcam/ab53852/AB881796 Cat#ab53852, RRID: AB_881796	1:200 (IHC), 1:800 (Wb)
BiPb2	Rabbit polyclonal	Proteintech/11587-1-AP	1:200 (IHC), 1:800 (Wb)
CGRP	Mouse monoclonal	SCB/sc-57053	1:500 (IHC)
NKA1α	Mouse monoclonal	SCB/sc48345	1:400 (IHC and Wb)
VIM	Goat polyclonal	SCB/sc-7557	1:200 (IHC)
Tubb3	Mouse monoclonal	SCB/sc-80016	1:400 (IHC), 1:1000 (Wb)
CaMKII	Mouse	SCB/sc-13141	1:200 (IHC)
Iba1	Rabbit polyclonal	Wako/019-19741	1:1000 (Wb)
IgG control	Mouse	LF/31903	1:100~400
IgG control	Rabbit	LF/MA5-16384	1:100~1000
IgG control	Goat	LF/02-6202	1:200 (IHC)

^aAntibody abbreviations: EGFP, enhanced green fluorescent protein; IB4, isolectin IB4; σ1R, Sigma-1 receptor; BiP, σ1R–binding immunoglobulin protein; CGRP, Calcitonin gene related peptide; NKA1α, Sodium/potassium ATPase 1 alpha; Tubb3, β3-Tubulin; Iba1, Ionized calcium-binding adapter molecule 1.

^b-b2Synthetic peptide: ^b(C)SRGHSGRYWAEISD, residues 113–126 of human σ1R; ^b1peptide corresponding to rat σ1R C terminal; and ^b2 residues 351–654 of human BIP-GST fusion.

^cSCB, Santa Cruz Biotechnology, Santa Cruz, CA; LF, Life Technologies, Carlsbad, CA; Proteintech, Rosemont, IL; Wako, Richmond, VA

Cellular σ₁R immunofluorescence in AAV vector injected DRG was measured, in a blinded fashion, individually for all neurons in a field with a nuclear profile. Neurons were considered σ₁R positive or to be transduced by AAV vectors if σ₁R or GFP immunofluorescence exceeded a background standard from naive neurons with first antibody absent negative stain by >2 SDs. To test the effect of GFPsc or GFP-σ₁RsiRNA expression on σ₁R fluorescence levels, cellular σ₁R immunofluorescence intensity of transduced neurons was normalized against the average fluorescence of nontransduced neurons in the same field analyzed by use of Adobe photoshop CC software. Neuronal size influences σ₁R fluorescence levels, so small (<700μm²), medium (700–1500μm²), and large neurons (>1500μm²) were evaluated, respectively. Data were derived from L4 and L5 DRG of 2 animals and 4–7 fields per vector.

Immunoblots

The lysates from tissues and cultured cells were extracted using 1x RIPA buffer (20 mm Tris-HCl pH 7.4, 150 mm NaCl, 1% Nonidet P-40, 1% sodium deoxycholate, 0.1% SDS, with 0.1% Triton X100 and protease inhibitor cocktail). To examine the subcellular localization of the target protein, DRG tissues were fractionated to obtain sodium/potassium-ATPase (NKA1α) enriched membrane and NKA1α-eliminated cytosol fractions using the ProteoExtract Subcellular Proteome Extraction Kit (Millipore), which contains extraction buffers with ultra-pure chemicals to ensure high reproducibility, protease inhibitor cocktail to prevent protein degradation and benzonase nuclease to remove contaminating nucleic acids, according to the manufacturer’s instructions. Protein concentration was determined by using the Pierce BCA kit (ThermoFisher). Equivalent protein samples were size separated using 10% or 4–20% SDS-PAGE gels (Bio-Rad Laboratories, Des Plaines, IL), transferred to Immun-Blot PVDF membranes (Bio-Rad), and blocked for 1 hr in 5% skim milk. The blots were cut into high (>70KDa), medium (50–70KDa), and low (>50KDa) protein-size strips or two halves along protein size around 70KDa,⁹³ then subsequently incubated overnight at 4°C with appropriate antibodies. Immunoreactive proteins were detected by Pierce enhanced chemiluminescence (ThermoFisher) after incubation for 1 hr with HRP-conjugated second antibodies (1:5000, Bio-Rad). Densitometry of bands of interests was analyzed using ImageJ v.1.46. Ratios of the band density of the target proteins to the sum of Tubb3 and NKA1α band density were calculated and the percentage changes of target proteins in the experimental samples compared with those from the control samples.^94,95

Excitability of DRG neurons.

Whole-cell current-clamp recording of dissociated DRG neurons was performed with an Axopatch 700B amplifier (Molecular Devices, Downingtown, PA) as described previously,^96–98 to determine the effects of AAV-mediated σ1R knockdown on the excitability of DRG neurons. Dissociated small-sized DRG neurons (<30μm) from sham animals and rats with SNI only, and dissociated small-sized DRG neurons with clear GFP expression from SNI rats injected with AAV6-GFPsc or AAV6-σ₁RsiRNA at 5-week after SNI and vector injection were used for recording (n=4 rats per group). Dissociated cells were studied no later than 6 h after harvest. For whole-cell current-clamp, patch electrodes had a resistance of 0.7–1.5 MΩ when filled with the pipette solution, which contained the following (in mM): 140 K-gluconate, 5 KCl, 2 MgCl_2, 0.2 EGTA, 10 HEPES, 4 Mg-ATP, and 0.3 Na2+-GTP, 10 Na2-phosphocreatinepH 7.2 with KOH and osmolarity of 296 to 300 mOsm. The extracellular solution contained the following (in mM): 140 NaCl, 4 KCl, 2 CaCl₂, 2 MgCl₂, 10 D-glucose, 10 HEPES at pH of 7.4 with NaOH and an osmolarity of 300 mOsm. Whole-cell configuration was obtained in voltage-clamp mode before proceeding to the current-clamp recording mode. The membrane input resistance was calculated by dividing the end amplitude of steady-state hyperpolarizing voltage deflection by the injected current.⁹⁹ The rheobase current and current threshold were determined by the first action potential (AP) elicited by a series of 200-millisecond (ms) depolarizing current injections through the recording pipette with increases in 5pA increments at 10 second intervals. Given the knowledge that nerve injury induces high RMP and low rheobase in the small-diameter DRG neurons,^66,100 the neurons with stable resting membrane potentials (RMP) more negative than −40 mV and overshooting APs (>80 mV RMP to peak) were used for additional data collection. AP frequency was determined by quantifying the number of APs elicited in response to depolarizing current injections (500 ms). Current-clamp recordings on the DRG neurons from AAV6-GFPsc or AAV6-σ₁R-siRNA treatment were performed in a blind manner where the electrophysiologist was not aware of the treatment. All experiments were performed at room temperature (22°C to 25°C).

Statistics

Statistical analysis was performed with GraphPad PRISM 6 (GraphPad Software, San Diego, CA). Behavioral changes over treatment baseline and between groups for von Frey and heat measurements were generated using repeated measures two-way ANOVA and post-hoc analysis with Bonferroni test. Pin, brush and cold test results in discrete numerical data without normal distribution so conservative nonparametric analyses were performed by Friedman’s test for analysis of variance and Dunn’s test for post hoc analysis. The differences of the targeted gene expression by immunoblots analysis and electrophysiological experiments were compared with one-way ANOVA, two-tailed unpaired t-test, or Mann-Whitney test, where appropriate. Results are reported as mean and standard deviation of mean (SEM). P<0.05 were considered statistically significant.