1. Introduction

N-type calcium channels play an essential role in primary afferent neurotransmission in the spinal cord dorsal horn.11,36 Indeed, these channels were first identified in dorsal root ganglion (DRG) neurons.20,42 The use of selective peptide blockers including ω-conotoxin GVIA6,46 has furthered understanding of the importance of N-type channels23,24 and their distribution.1,22 Their targeting for chronic pain therapy is also well-established.51,57 Although clinical use of ziconotide (ω-conotoxin MVIIA) is limited because of its intrathecal route of administration and side effects, nevertheless, it validates the role of N-type channels in pain pathophysiology and pharmacotherapy.45

CaV2.2 channels form a complex with auxiliary β and α2δ subunits, which are important for channel trafficking and function.18 The α2δ-1 isoform is prominent in primary afferent pathways and is up-regulated following neuropathic injury.4,34,40 Indeed, genetic ablation of either CaV2.228,47 or α2δ-144 suppresses various sensory modalities in chronic pain models.

Despite the importance of CaV2.2 channels in primary afferent transmission in nociceptive pathways, examination of their distribution and trafficking as well as altered expression following nerve injury has been hindered by the lack of reliable antibodies recognising native N-type calcium channels, which have been validated using knockout tissue. Furthermore, previous studies using antipeptide antibodies have reported conflicting results.12,29,60,61 For this reason, we previously developed CaV2.2 constructs containing exofacial epitope tags, in a position not affecting channel function.9 We then generated a knockin mouse line carrying the hemagglutinin (HA) tag in this position in the Cacna1b gene, to examine the distribution of native CaV2.2 protein in the intact nervous system.41 We previously found a dramatic effect of α2δ-1 ablation on CaV2.2_HA distribution, with the loss of cell surface CaV2.2, particularly in small peptidergic nociceptive sensory neuron somata and terminals.41

Here, we have examined the effect of partial sciatic nerve ligation (PSNL) on CaV2.2_HA distribution in sensory neurons and spinal cord and provide novel insights into cellular pathophysiological mechanisms after nerve injury. We have compared CaV2.2_HA distribution, ipsilateral and contralateral to nerve injury, both in DRG neuronal cell bodies and in their terminals in the dorsal horn, and we have then examined the effect of α2δ-1 knockout on this. We have further used several markers of different DRG subtypes, to examine their coexpression with CaV2.2_HA. This includes the glial cell line–derived neurotrophic factor (GDNF) family ligand receptor (GFRα1), which is present in certain low-threshold mechanoreceptors (LTMRs) and is up-regulated following nerve injury.5,27 Glial cell line–derived neurotrophic factor is a DRG trophic factor, which is analgesic in neuropathic pain.7

Our key finding is that CaV2.2_HA is up-regulated, ipsilateral to PSNL, in medium/large DRG neurons, where it shows increased association with GFRα1. In parallel, we observe increased CaV2.2_HA in ipsilateral medial/central deep dorsal horn, where GFRα1 is correspondingly up-regulated. The increased CaV2.2_HA in DRGs and deep dorsal horn is α2δ-1 dependent, whereas the elevation in GFRα1 is not, indicating that it represents increased CaV2.2_HA trafficking to these mechanoreceptor terminals, which may result in elevated neurotransmission.

2. Methods 2.1. Partial sciatic nerve ligation

The CaV2.2_HA mouse line was generated by Taconic Artemis on the C57BL/6 background, as described in detail previously.41 The α2δ-1−/− C57BL/6 mouse line described previously21,44 was crossed, as heterozygotes, with the Cav2.2_HA knockin mice to generate double-transgenic Cav2.2_HAKI/KI α2δ-1−/− mice. Wild-type (WT) mice were C57BL/6. Both male and female mice were used in this study. Mice were housed in groups of no more than 5 on a 12 h: 12 h light: dark cycle; food and water were available ad libitum. All experimental procedures were covered by UK Home Office license, had local ethical approval, and followed the guidelines of the International Association for the Study of Pain.62

Surgery was performed based on a method described previously.44,50 Mice were maintained under 2% vol/vol isoflurane (Baxter, Northampton, United Kingdom) anesthesia delivered in a 3:2 ratio of nitrous oxide and oxygen. Under aseptic conditions, the left sciatic nerve was exposed through blunt dissection of the biceps femoris above the trifurcation of the nerve. Approximately half of the nerve was ligated with a nonabsorbable 7-0 braided silk thread (Ethicon, VetTech, United Kingdom). The surrounding muscle and skin was closed with absorbable 6-0 vicryl sutures (Ethicon, VetTech) and topical lidocaine cream (5% wt/wt) applied to the skin. Sham surgery was performed in an identical manner, omitting the nerve ligation step. After surgery, the mice were allowed to recover. Foot posture and general behavior of the operated mice were monitored throughout the postoperative period. While blind to genotype, mechanical hypersensitivity was tested 14 days after surgery to confirm that the operated mice used for the study displayed neuropathic responses.

2.2. Immunohistochemistry

For immunohistochemistry, on days 14 or 15 after surgery, mice were deeply anaesthetized with an intraperitoneal injection of pentobarbitone (Euthatal, Merial Animal Health, Harlow, United Kingdom; 600 mg/kg), perfused transcardially with saline containing heparin, followed by perfusion with 4% paraformaldehyde in 0.1 M phosphate buffer (PB, pH 7.4) at a flow rate of 2.5 mL·min−1 for 4 minutes. Lumbar 4 DRGs and the lumbar enlargement of the spinal cord were dissected out. Following dissection, the spinal cord was postfixed for 2 hours, whereas the DRGs did not undergo extra fixation. Tissue was washed with PB, cryoprotected by incubation in PB with 20% sucrose overnight, and finally mounted in Optimal cutting temperature (OCT) compound (VWR International, Lutterworth, United Kingdom) before storing at −80°C, until sectioning. Dorsal root ganglia and spinal cord were sectioned at 15 and 20 μm, respectively, using a cryostat, placing the sections sequentially in series of 6 slides, so the distance between any section and the next on any slide is 90 or 120 µm in each case. Slides were stored at −80°C until processed.

For immunofluorescence labelling of DRGs, sections were blocked with 10% goat serum in PBS containing 0.3% Triton X-100 for more than 1 hour at room temperature (RT), followed by incubation with the unconjugated goat Fab antimouse IgG (H + L) (0.1 mg/mL in PBS, Jackson ImmunoResearch Lab, Stratech Ltd, Ely, United Kingdom, catalogue number 115-007-003) for 1 hour at RT to reduce nonspecific binding of antirat antibody to endogenous IgG in mouse tissue, washed in PBS, 0.1% Triton X-100 (PBS-T), and then incubated with rat monoclonal anti-HA antibody (Roche, catalogue number 11867423001, 1:100), for 2 to 3 days at 4°C in 5% goat serum, 0.3% Triton X-100 in PBS. Following extensive washing in PBS-T, immunolabelled samples were fixed in 4% paraformaldehyde in PBS for 30 minutes at RT, washed in PBS-T and incubated for 1 to 2 days at 4°C with the goat antirat conjugated with Alexa Fluor 488 (Invitrogen, Thermo Fisher Scientific, Oxford, United Kingdom, catalogue number A11006, 1:500). After washing, sections were treated with the nuclear stain DAPI (Molecular Probes, catalogue number D106, 0.5 μM) and mounted in VectaShield (Vector Laboratories, 2BScientific Ltd., Upper Heyford United Kingdom, catalogue number H-1000). When costaining HA with the goat antibody against GFRα1 (R&D Systems, Bio-techne, Abingdon , United Kingdom, catalogue number AF560, 1:200), the procedure was the same except that horse serum was used instead of goat serum in the blocking and antibody solution, the goat Fab antimouse was omitted, and the secondary antibodies were donkey antigoat conjugated with Alexa Fluor 488, and the donkey antirat highly cross-adsorbed antibody conjugated with Alexa Fluor 594 (Thermo Fisher Scientific, catalogue numbers A11055 and A21209 respectively, both used at 1:500).

For spinal cord immunohistochemistry, sections were incubated with rat monoclonal anti-HA antibody (as above) and costained with rabbit anti-calcitonin gene-related peptide (CGRP), IB4 conjugated with FITC (Sigma, catalogue numbers C8198 and L2895), or with goat anti-GFRα1. Some sections were labelled for α2δ-1, as described previously,44 with the following modifications: after heat-induced epitope retrieval (10 mM citrate buffer, pH 6.0, 0.05% Tween 20, 95°C for 10 minutes), the sections were washed, blocked with 10% goat serum in PBS containing 0.3% Triton and treated with the unconjugated goat Fab anti-mouse IgG (H + L) (0.1 mg/mL in PBS) for 1 hour at RT. Mouse monoclonal antidihydropyridine receptor (α2-1 subunit) antibody (Sigma, D219, 1:100) was applied for 2 or 3 days at 4°C. After extensive washes, the samples were incubated with biotin-conjugated goat antimouse Fab fragment (1:500, Jackson Immuno Research Lab, catalogue number 115-067-003), overnight at 4°C, followed by washes and Streptavidin-AlexaFluor-488 overnight at 4°C (1:500, Invitrogen, catalogue number S32354).

2.3. Confocal image acquisition and analysis

Immunostaining was visualized using an LSM 780 (Carl Zeiss UK Ltd., Cambourne, United Kingdom) confocal microscope. Images were acquired with constant settings in each experiment from at least 3 sections per DRG or 7 per spinal cord from at least 3 mice unless otherwise stated. Only intact tissue sections, unfolded and with uniform staining, were selected for imaging and analysis. For DRG sections, multiple images were acquired with a 63× 1.4NA objective (0.7 μm optical section) covering the whole area of the section containing neurons using the tiling mode with a 5% overlap and stitched with Zen software (Zeiss). For analysis, using ImageJ software (Schneider et al., 2012), in every intact DRG neuron with a visible nucleus, we selected 2 different types of regions of interest (ROI). Using images with temporarily enhanced brightness and contrast, solely to aid visualization of the circumference of even dimly stained cells, first we drew a 10 pixel-wide line (0.9 µm) following the perimeter of the cell from which we recorded the length as an estimation of the size of the cell (small <61 µm, medium 61-94 µm, or large >94 µm) and the mean membrane intensity. Next, we selected an ROI for the area inside the first ROI, excluding the plasma membrane and the nucleus, to record the mean intracellular intensity. Regions of interest outside each section were used as background and deducted from sample measurements.

To determine the proportion of DRG neurons expressing CaV2.2_HA, GFRα1 or both, cells with staining above a threshold (3× the SD of the contralateral side from each animal) were selected using the multipoint tool of ImageJ and merged with the list of ROIs around the perimeter of each section, as described above, to quantify the different populations according to staining and size.

Spinal cord images (at least 7 per animal, unless otherwise stated) at low magnification were acquired using a 20× 0.8 NA objective (5-µm optical section) covering the whole dorsal half of each section also in tiling mode and were stitched with Zen software as for DRG sections. For analysis, using the same software, the mean intensity was recorded from a profile scan of a rectangular ROI of 50 × 300 µm placed across the superficial layers of the medial, central, and lateral part of the ipsilateral and contralateral dorsal horn of each section. The ratio of ipsilateral to contralateral per ROI was calculated to determine the relative level of change in fluorescence and then averaged per each animal. To quantify that data, the mean of the superficial (0-80 μm) or deep (140 and 300 μm) regions was extracted. The HA, CGRP, IB4, and GFRα1 data from different experiments were pooled according to genotype and presented as the mean ± SEM.

For high magnification examination of the spinal cord sections, a 63× 1.4 NA objective in Airyscan mode (0.2-µm optical sections) for HA, CGRP, IB4, and GFRα1 was used in tiling mode to generate multiple images covering the ipsilateral or contralateral dorsal horn of each section; superresolution images then underwent Airyscan processing and were stitched using Zen software. To quantify density of CaV2.2_HA immunoreactivity and that of the other markers, a square ROI (70 × 70 µm) in the deeper layers of the medial region or the superficial layers of the medial, central, and lateral dorsal horn was analysed with FIJI software 6. Each ROI was split into 2 or 3 channels (depending on number of markers used) and thresholded (4 × the SD of the deeper ROI or the combined 3 superficial ROIs from the contralateral side from each section) to create a mask per channel with the all the clusters between 110 and 2800 pixels (0.2-5 μm2) selected using the particle analyzer command. All the clusters were saved as list of ROIs and used in the original image to record their size and mean intensity. The corresponding masks for HA and the other marker/s were merged, and the overlapping clusters (>1%) were extracted using the plugin Binary Feature Extractor from the BioVoxxel Toolbox (http://www.biovoxxel.de), to obtain the associated clusters. Overall, 126 superficial and 42 deeper ROIs were analysed, from a total of 21 sections with 2 or 3 sections per mouse stained for HA, CGRP and IB4, or HA and GFRα1, from 2 CaV2.2_HAKI/KI -α2δ-1+/+ and 2 CaV2.2_HAKI/KI -α2δ-1−/− mice per experiment.

2.4. Statistical analysis

Data were analysed with GraphPad Prism 7 or 9 (GraphPad software, San Diego, CA) or Origin-Pro 2021 (OriginLab Corporation, Northampton, MA). Where error bars are shown they are SEM, “N” refers to number of mice or clusters, unless indicated otherwise. Statistical significance between 2 groups was assessed by Student t test or paired t test, as stated. Repeated-measures 2-way ANOVA followed by Šídák's multiple comparisons test was used to analyse the number of clusters according to cluster size (or density) and the side of the spinal cord from multiple sections. Details of statistical test results are given in Figure legends and in Supplementary Information (available at https://links.lww.com/PAIN/B762).

3. Results 3.1. Effect of partial sciatic nerve ligation on distribution of CaV2.2_HA and α2δ-1 in dorsal root ganglion neurons in vivo

The use of CaV2.2_HA knockin mice has revealed that CaV2.2_HA is present both intracellularly and on the cell surface of DRG neuronal cell bodies.41 We confirmed this distribution in our current experiments using sections of dorsal root ganglia from 12- to 16-week-old CaV2.2_HAKI/KI mice that have undergone unilateral PSNL (Fig. 1A, i). Immunoreactivity for HA was absent from CaV2.2WT/WT mice (Fig. 1A, ii). In agreement with our previous quantification,41 the level of both cell surface and intracellular CaV2.2_HA was highest in small DRG cell bodies (Fig. 1B, i and ii), and it was very low in large DRG neuronal somata contralateral to the PSNL injury (Fig. 1B, i and ii).

Figure 1.: Effect of PSNL and α2δ-1 knockout on distribution of CaV2.2_HA in dorsal root ganglia from CaV2.2_HAKI/KI mice. (A) Immunostaining for CaV2.2_HA (green) with nuclear marker DAPI (blue) in DRGs from CaV2.2_HAKI/KI, α2δ-1+/+ mice. (i) Contralateral (contra, left) and ipsilateral (ipsi, right) to PSNL. (ii) Lack of CaV2.2_HA immunoreactivity shown in wild-type naive DRGs. Scale bars are 10 µm. (B) Membrane (i) and intracellular (ii) HA staining for CaV2.2_HA, quantified with respect to cell size (small 94-μm perimeter). Data were analysed for 6 mice. For each mouse, all DRG neurons from at least 3 sections from the ipsilateral and contralateral L4 DRG were analysed and normalised to the mean of the contralateral side for each size group (each colour corresponds to mean data from one mouse). Black lines indicate mean ± SEM. Statistical analysis compares ipsilateral and contralateral for each DRG size group, *Pt test. Individual P values for small, medium, and large DRGs = 0.4169, 0.0496 and 0.0123 in i; and 0.7651, 0.3030 and 0.0376 in ii, respectively. Data from individual DRGs, and for sham-operated controls are in Supplementary Figs. 1A and B, available at https://links.lww.com/PAIN/B762. (C) Immunostaining for CaV2.2_HA (green) with nuclear marker DAPI (blue) in DRGs from CaV2.2_HAKI/KI, α2δ-1−/− mice contra (left) and ipsilateral (right) to PSNL. Scale bars are 10 µm. (D) Intracellular HA staining for DRGs from CaV2.2_HAKI/KI, α2δ-1−/− mice quantified with respect to cell size, from 6 mice, exactly as in (B). Paired t test: individual P values for small, medium, and large DRGs = 0.8546, 0.4386, and 0.6847, respectively. There was no discernible cell surface staining, and therefore, this was not quantified. Data from individual DRGs are in Supplementary Fig. 1C, available at https://links.lww.com/PAIN/B762. DRG, dorsal root ganglion; HA, hemagglutinin; PSNL, partial sciatic nerve ligation.

Following PSNL, the CaV2.2_HA signal was significantly increased on the cell surface, ipsilateral to the nerve injury, compared with the contralateral side, particularly in large and medium DRG neurons (by 34.6% and 13.8%, respectively) but not in small DRG neurons (Fig. 1A, arrow, Fig. 1B, i). Furthermore, analysis of intracellular CaV2.2_HA showed that although intracellular CaV2.2_HA density was highest in small DRG neurons (Fig. 1A, 1B, ii), it was increased ipsilateral to PSNL relative to the contralateral side, only in large DRG neurons (by 17.1%, Fig. 1B, ii). There was no increase in CaV2.2_HA in sham-operated animals (Supplementary Fig. 1B compared with A, which also shows the data from individual DRG neurons for these experiments, available at https://links.lww.com/PAIN/B762). Together, these results indicate that PSNL produces an increase of CaV2.2_HA in large and medium DRG neurons ipsilateral to the injury, particularly on their cell surface.

3.2. Genetic ablation of α2δ-1 prevents the increase of CaV2.2_HA in dorsal root ganglion neurons ipsilateral to partial sciatic nerve ligation

It has been found in several studies that α2δ-1 is up-regulated following sensory nerve injury and is important for the development of neuropathic allodynia and mechanical hypersensitivity.4,30,40,44 We therefore examined the effect of α2δ-1 knockout on CaV2.2_HA distribution following PSNL. We observed that the CaV2.2_HA signal at the cell surface of DRG neurons was almost abolished in α2δ-1−/− mice (Fig. 1C), as we had found previously.41 Following PSNL in these mice, there was no appearance of CaV2.2_HA on the cell surface of the DRG neurons (Fig. 1C, ipsilateral) or any increase in intracellular CaV2.2_HA signal ipsilateral to PSNL (Figs. 1C and D). These results highlight the importance of α2δ-1 in the elevation of CaV2.2 in large and medium DRGs that we observed ipsilateral to PSNL.

3.3. CaV2.2_HA is decreased in patches of superficial dorsal horn ipsilateral to partial sciatic nerve ligation, in parallel with loss of IB4 and CGRP

Next, we examined the effect of PSNL on the distribution of CaV2.2_HA, in the dorsal horn of the spinal cord, in parallel with other DRG subtype markers, CGRP and IB4 (Fig. 2). As we previously described,41 there is strong immunoreactivity for CaV2.2_HA in the superficial laminae I and II of the dorsal horn (Fig. 2A, i), and there was no signal in naive WT mice (Fig. 2B). This localization shares topographic distribution with the presynaptic marker CGRP, which is present in peptidergic nonmyelinated primary afferent C-fiber terminals in laminae I and II-outer (Fig. 2A, ii), and also with IB4, which is present in nonpeptidergic primary afferent C-fiber terminals, mainly in lamina II-inner (Fig. 2A, iii).

Figure 2.: Comparison of the effect of PSNL and α2δ-1 knockout on CaV2.2_HA, IB4, and CGRP distribution in superficial dorsal horn. (A) Representative images of dorsal horn sections following PSNL, in CaV2.2_HAKI/KI, α2δ-1+/+ mice. Images are always oriented as contralateral (left) and ipsilateral (right) to PSNL, and stained for CaV2.2_HA (i; green), CGRP (ii; magenta), and IB4 (iii; red). Panel (iv) shows merged images. Solid arrows, areas of decreased staining in superficial dorsal horn. open arrows, area of increased staining of CaV2.2_HA in medial deep dorsal horn. Scale bar: 200 μm. (B) Lack of CaV2.2_HA immunoreactivity in wild-type naive dorsal horn. Scale bar: 200 µm. (C) Diagram of spinal cord dorsal horn showing the position of the 3 ROIs (medial, central, and lateral; 300 × 50 µm) placed in dorsal horn sections, ipsilateral (red) and contralateral (black) to side of PSNL, to measure the fluorescence intensity of CaV2.2_HA and other markers. (D) Immunofluorescence profiles for CaV2.2_HA (green), IB4 (red), and CGRP (magenta) in the medial ROI shown in (C), contralateral (black-filled symbols) and ipsilateral (color-filled symbols) to PSNL in CaV2.2_HAKI/KI, α2δ-1+/+ dorsal horn. Data are the mean ± SEM for N = 5 mice (7 sections/mouse). (E) Ratio ROI profiles (ipsilateral/contralateral) for data shown in D, for CaV2.2_HA (green), IB4 (red), and CGRP (magenta). Data are the mean ± SEM for N = 5 mice (7 sections/mouse). Quantification of mean average of superficial and deep laminae for CGRP and IB4 from this data and from CaV2.2_HAKI/KI, α2δ-1−/− are shown in Supplementary Fig. 2, available at https://links.lww.com/PAIN/B762. HA, hemagglutinin; PSNL, partial sciatic nerve ligation; ROI, regions of interest.

Following PSNL, we observed a patchy loss of staining for IB4, and to a lesser extent CGRP, ipsilateral to the nerve injury, which was paralleled by a loss of CaV2.2_HA (Fig. 2A, ipsilateral on right side of each section, i-iii, and merged image in iv; closed arrows). This irregular loss of IB4 and CGRP staining has previously been described in many studies, and it is believed to be due to deafferentation of neurotrophin-dependent terminals following nerve injury.2,38,55 For the first time, we can now see that CaV2.2_HA present in those terminals is also reduced. To quantify the observed signals, and the effect of PSNL, we took ROIs perpendicular to the pial surface in medial, central, and lateral regions of the dorsal horn (Fig. 2C) and quantified the fluorescence intensity profiles through the different laminae (Fig. 2D, data shown for the medial ROI), as described previously.41 We then determined the ratio of ROI intensity (ipsilateral/contralateral for each section) with respect to the PSNL injury (Fig. 2E). For both IB4 and CGRP, taking the mean intensity for each animal for the superficial layers (Supplementary Fig. 2B, C, available at https://links.lww.com/PAIN/B762), although there is an obvious patchy reduction in most cases, no significant overall reduction was observed, presumably because of the irregular and variable nature of the loss in this nerve injury model. The CaV2.2_HA signal showed a similar patchy loss of staining in the same areas as found for IB4 and CGRP (Fig. 2A, iv, solid arrow in merged image), suggesting that it may be present in the same terminals, as we previously concluded using dorsal rhizotomy and high-resolution microscopy.41

3.4. Superresolution analysis of distribution of CaV2.2-HA in superficial dorsal horn following partial sciatic nerve ligation: comparison with distribution of IB4 and CGRP

We then analysed superresolution Airyscan images taken from ROIs from the medial, central, and lateral regions of the superficial dorsal horn sections contralateral and ipsilateral to PSNL (ROI for medial region shown in Figs. 3A and B) and examined the size and intensity of CaV2.2_HA, CGRP, and IB4 clusters and their association (mask and data for medial ROI shown in Figs. 3C and D). CaV2.2_HA, together with either IB4 or CGRP, are found in rosette-like glomerular clusters in the superficial dorsal horn (images A and B, below Fig. 3B), as previously observed.41 Quantification of all the clusters from the combined ROIs contralateral to PSNL in superficial dorsal horn shows that 35.2% of IB4 and 34.1% of CGRP clusters were associated with CaV2.2_HA. Similarly, 22.5% and 19.1% of CaV2.2_HA clusters were associated with IB4 and CGRP, respectively (data determined from 15 ROIs, 3 from each side, in 5 sections from 2 mice; 2607 CaV2.2 clusters, 1657 IB4 clusters, and 1442 CGRP clusters). We also examined the effect of PSNL on the area and intensity of CaV2.2_HA, CGRP, and IB4 clusters with respect to their density of distribution. We found that only the density of CaV2.2_HA clusters (number of clusters/ROI) was decreased ipsilateral to nerve injury (Figs. 3C and 3D, i-ii), with no clear change in size profile (Fig. 3D, i) or intensity distribution (Fig. 3D, ii). Similarly, the density of both IB4 clusters (Fig. 3D, iii-iv) and CGRP clusters (Fig. 3D, v-vi) decreased ipsilateral to PSNL, again with no change in size profile (Fig. 3D, iii, v) or intensity distribution (Fig. 3D, iv, vi).

Figure 3.: High-resolution analysis of CaV2.2_HA, IB4, and CGRP clusters in superficial dorsal horn following PSNL. (A and B) Representative Airyscan tiled images of contralateral (left) and ipsilateral (right) dorsal horn from the same section following PSNL stained for CaV2.2_HA (green), IB4 (red), and CGRP (magenta). ROIs (70 × 70 μm) were placed in the medial, central, and lateral regions of the superficial layer for quantification of clusters. The medial ROI is indicated with a square in (A) and shown enlarged in (B). Representative CGRP (a) and IB4 (b) positive glomeruli are indicated by small ROIs (2 × 2 μm) in the contralateral side of (B) and enlarged underneath. Scale bars in (A and B): 100 μm and 10 μm, respectively. (C) Composite mask from the 3 channels (CaV2.2_HA [green], IB4 [red], and CGRP [magenta]) from ROIs as shown in (B), for particles between 0.2 and 5 μm2, with signal above threshold (see Methods). (D) Size and intensity distribution of clusters positive for CaV2.2_HA, IB4, and CGRP on the ipsilateral (red) and contralateral (black) sides in the medial superficial ROI. N =545, 126, and 166 (ipsi) and 1058, 500, and 470 (contra) clusters for CaV2.2_HA-, IB4-, and CGRP-positive clusters, respectively. Data from 10 superficial medial ROIs (5 contra and 5 ipsi, from 2 or 3 sections per each of 2 mice). Statistical significances are given by *PPPPhttps://links.lww.com/PAIN/B762. ANOVA, analysis of variance; CGRP; calcitonin gene-related peptide; HA, hemagglutinin; PSNL, partial sciatic nerve ligation; ROI, regions of interest.

Taken together, these results show that the patchy reduction observed at low magnifications for CaV2.2_HA, IB4, and CGRP in the superficial dorsal horn, ipsilateral to PSNL, corresponds to a decreased number of glomerular clusters rather than a change in their size or intensity in response to the nerve injury.

3.5. Effect of α2δ-1 knockout on the CaV2.2_HA distribution in the superficial dorsal horn

We next examined the effect of genetic ablation of α2δ-1 on the changes in CaV2.2_HA distribution in the dorsal horn following PSNL. We have previously shown that the signal for CaV2.2_HA in the dorsal horn was markedly reduced in CaV2.2_HAKI/KI/α2δ-1−/− mice, particularly in the superficial laminae,41 and this is confirmed here (Fig. 4A ii, compared with i). However, we found here that the patchy reduction in CaV2.2_HA following PSNL (Fig. 4A, i) in the superficial layers of the dorsal horn was still evident in α2δ-1−/− mice (Fig. 4A, ii). We further quantified the PSNL-mediated reduction of CaV2.2_HA in laminae I and II of the dorsal horn, to examine the effect of α2δ-1- knockout (KO), by means of the ROI ipsilateral:contralateral ratio profiles (Fig. 4B, α2δ-1+/+ [red] compared with α2δ-1−/− [black] intensity profiles), which show that the reduction in laminae I-II remains evident in all ROIs of α2δ-1−/− dorsal horn (Fig. 4B, i-iii). This is confirmed from the mean intensity ratios for laminae I-II, which showed a statistically significant reduction in the central and lateral ROIs of α2δ-1−/− mice (Fig. 4C). There were no evident differences between the results in male and female mice for either genotype (Fig. 4C and Supplementary Table 1, available at https://links.lww.com/PAIN/B762).

Figure 4.:

Quantification of the effect of α2δ-1 knockout on distribution of CaV2.2_HA in the superficial and deep dorsal horn of CaV2.2_HAKI/KI mice following PSNL. (A) Representative images for CaV2.2_HA immunostaining (depicted as Rainbow look-up table, LUT), from CaV2.2_HAKI/KI, α2δ-1+/+ (i) and CaV2.2_HAKI/KI, α2δ-1−/− (ii) mice following PSNL. Sections are oriented as contra (left) and ipsi (right) to PSNL. Closed arrows indicate patchy loss of CaV2.2_HA in superficial dorsal horn, and open arrow in (i) indicates increase in CaV2.2_HA in deep dorsal horn, ipsilateral to PSNL. Scale bars 200 μm. (B) Plots of the change in CaV2.2_HA intensity following PSNL (expressed as a ratio of ipsi to contra) in the medial (i), central (ii), and lateral (iii) ROIs from CaV2.2_HAKI/KI, α2δ-1+/+ (α2δ-1 WT, N = 12 mice, red) and CaV2.2_HAKI/KI, α2δ-1−/− (α2δ-1 KO, N = 7 mice, black). A total of 7 sections were averaged per mouse. Data are mean ± SEM of the number of mice stated. Statistical significance between α2δ-1+/+ and α2δ-1−/− in the region indicated was determined by unpaired t test (*P < 0.05, **P < 0.01). The individual P values for LI-II, LIII and deeper layers are, respectively, 0.6077, 0.0611, and 0.0331 in (i); 0.2115, 0.0864, and 0.0096 in (ii): and 0.9912, 0.4349, and 0.0748 in (iii). (C, D) Quantification of the change in immunofluorescence of CaV2.2_HA (expressed as a ratio of ipsi/contra) in (C) superficial layer (laminae I and II, from pial surface to 80 μm) and (D) deeper layers (laminae IV and part of V, from 140 to 300 μm) of the dorsal horn. Data are given for α2δ-1+/+ mice (WT, N = 12) or α2δ-1−/− mice (KO, N = 7). Each coloured symbol represents the mean ratio for each animal (squares represent male and circles female mice), and the black line represents mean ± SEM. Statistical significance of ipsilateral/contralateral data determined by one sample t test with 1 as reference value (*P < 0.05, **P < 0.01, ***P < 0.001). The individual P values for medial, central, and lateral regions are 0.1870, 0.0261, and 0.0883 for α2δ-1+/+ and 0.1786, 0.0019, and 0.0396 for α2δ-1−/−, respectively in (C) and 0.0002, 0.0014, and 0.0309 for α2δ-1+/+ and 0.0391, 0.9829, and 0.7527 for α2δ-1−/− in (D), respectively. DH, dorsal horn; HA, hemagglutinin; PSNL, partial sciatic nerve ligation; ROI, regions of interest.

In parallel, the patchy reduction in IB4 and CGRP signal, ipsilateral to PSNL injury in the superficial ROIs also remained evident in α2δ-1−/− dorsal horn (Supplementary Fig, 2A, vi-vii, arrows, available at https://links.lww.com/PAIN/B762), although it was only statistically significant for IB4 (Supplementary Fig. 2B, C, available at https://links.lww.com/PAIN/B762), indicating that these PSNL-induced reductions were not α2δ-1 dependent. In Airyscan analysis of sections from α2δ-1−/− dorsal horn (Supplementary Fig. 3A-D, available at https://links.lww.com/PAIN/B762), we found the reductions in CaV2.2_HA, IB4, and CGRP cluster density ipsilateral to PSNL were still present in superficial dorsal horn (Supplementary Fig. 3D i-vi, available at https://links.lww.com/PAIN/B762), with no change in size (Supplementary Fig. 3D i, iii, v, available at https://links.lww.com/PAIN/B762) or intensity distribution (Supplementary Fig. 3D ii, iv, vi, available at https://links.lww.com/PAIN/B762).

Together, these results show that CaV2.2_HA is present in the glomerular nerve terminals in the superficial dorsal horn that undergo nerve injury–dependent deafferentation ipsilateral to PSNL, and this partial loss is not α2δ-1 dependent.

3.6. CaV2.2_HA is increased in deep dorsal horn of spinal cord ipsilateral to partial sciatic nerve ligation

Surprisingly, we observed a consistent and marked increase in CaV2.2_HA ipsilateral to PSNL, in the deep layers of the dorsal horn (layers IV-V, Fig. 2A, i; Fig. 4A, i, open arrows), particularly in the medial and central ROIs (Fig. 2E green profile; Fig. 4B, orange profiles, i, ii), which was less evident in the lateral ROI (Fig. 4B, iii). Quantification shows a statistically significant increase in the deep dorsal horn for all 3 regions (Fig. 4D). There were no evident differences between the results in male and female mice for either genotype (Fig. 4D and Supplementary Table 1, available at https://links.lww.com/PAIN/B762). By contrast, there was no parallel increase in CGRP or IB4 in the deep do