Background
Sensory neurons in the dorsal root ganglia (DRG) are normally not directly innervated by the sympathetic nerves. Sympathetic fibers in normal DRG are associated only with blood vessels. Sprouting of sympathetic fibers into the DRG has been described in several preclinical pain models [
1‐
5]. Anatomically, this may include formation of dramatic "basket" structures in which sympathetic fibers form a dense plexus around individual somas (particularly of large diameter cells), as well as an increase in overall sympathetic fiber density in the cellular regions of the DRG. Basket structures have also been observed in DRGs from human neuropathic pain patients [
6]. Sprouting occurs around both intact and axotomized cells [
7], and can occur rapidly even in the absence of overt axotomy, for example in locally inflamed DRG [
8], and mechanically compressed DRG [
9].
The discovery of sympathetic sprouting in the DRG provided a possible anatomical basis for clinical syndromes of sympathetically maintained pain. Clinically, many chronic pain conditions such as complex regional pain syndrome are maintained or exacerbated by sympathetic activity in some patients, especially at earlier stages, and respond to various anti-sympathetic manipulations [
10,
11]. However, studies in this field using animal models have yielded sometimes contradictory results on the behavioral importance of the sympathetic-sensory connections. Many behavioral studies have shown that various forms of sympathectomy reduce or eliminate mechanical or thermal pain behaviors in rodents, but other studies been reached the opposite conclusion (see Table
1 in ref. [
5]).
Table 1
Distribution of nociceptive markers in cells with and without sympathetic contacts
Marker
|
Condition
|
Cells with Symp Fibers
|
Cells with Symp Baskets
|
Cells w/o Symp Contact
|
Cells with Symp Fibers
|
Cells with Symp Baskets
|
Cells w/o Symp Contact
|
Overall Percent Expressing Marker
|
Percent with Symp Basket or Fibers
|
SP
| SNL | 118 | 55 | 1576 | 168 | 114 | 558 | 32.4%* | 17.6%* |
| Normal | 1 | 2 | 2637 | 2 | 0 | 622 | 19.1% | 0.2% |
TrKA
| SNL | 171 | 83 | 1267 | 109 | 68 | 237 | 21.4%* | 22.3%* |
| Normal | 1 | 0 | 1431 | 3 | 1 | 1202 | 45.7% | 0.2% |
CGRP
| SNL | 134 | 50 | 1241 | 113 | 72 | 374 | 28.2%* | 18.6%* |
| Normal | 3 | 0 | 1892 | 7 | 12 | 1429 | 43.3% | 0.7% |
Few functional studies of abnormal sympathetic-sensory neuron connections have been reported. In general, these studies used extracellular fiber recording methods to measure increased spontaneous activity after addition of alpha adrenergic agonists or preganglionic stimulation in vivo [
1,
12,
13]. These studies were not designed to detect absolute incidence of spontaneous activity, or to examine other measures of excitability such as rheobase and threshold, which require intracellular recording methods.
Recently, we demonstrated that sympathetic sprouts into the rat DRG at 3 days after spinal nerve ligation (SNL) originated primarily from the dorsal ramus of the spinal nerve [
14]. This branch of the spinal nerve diverges from the ventral ramus close to the intervertebral foramen, and innervates posterior structures. The dorsal ramus is left intact in the spinal nerve ligation model, which involves ligation and (in some cases) transection of the much larger ventral ramus of the spinal nerve, in which fibers that innervate the paw and hindlimb travel. This finding allowed electrical stimulation of the dorsal ramus in the excised rat whole DRG preparation, in which microelectrode recordings of sensory neurons with attached axon segments and intact glial cell layers could be made. We showed that repeated stimulation of the dorsal ramus had excitatory effects on large and medium diameter sensory neurons, including increased spontaneous activity and decreased rheobase, beyond the excitatory effects induced by spinal nerve ligation per se. Experiments using sympathetic blockers, or in DRG in which sprouting was prevented, provided evidence that these excitatory effects were due to stimulation of sympathetic sprouts.
In the present study, we continue the development of techniques for functional study of abnormal sympathetic-sensory connections ex vivo. We used a transgenic mouse in which enhanced green fluorescent protein (EGFP) is expressed under control of the tyrosine hydroxylase (TH) promoter, the commonly used marker for sympathetic neurons. This allowed visualization of sprouted sympathetic fibers in the excised mouse DRG preparation. Microelectrode recordings were made in sensory neurons, based on an adaptation of the technique used in rat DRG. Sensory neurons were identified during the recording as being surrounded by extensive sympathetic basket or ring formations, having nearby but less elaborate sympathetic fibers, or having no nearby fibers. We report that excitability parameters differ markedly between these three classes of cells, suggesting that the sprouted sympathetic fibers have localized, functional excitatory interactions with nearby sensory neurons.
Discussion
Because we could visualize sympathetic fibers in live tissue, using the transgenic mouse allowed us to efficiently identify and record from basket cells which are relatively rare. The present results validate the use of pTH-EGFP mice to investigate sympathetic-sensory coupling under pathological conditions. The rapid appearance of sprouting sympathetic fibers (by POD 3) after spinal nerve ligation that we observed in our whole mount preparations is consistent with results obtained with higher resolution, thin section methods in rats [
27]. The rapid increase of spontaneous activity (especially, bursting activity) in large diameter cells with high conduction velocities, and their relatively narrow size distribution, is also similar to that previously reported in rats [
14,
26,
28,
29]. It is worth noting that cells with high conduction velocities but relatively small diameters have been described as belonging to the small group of nociceptive Aβ-fiber neurons in guinea pig neurons [
30]. Cells with similar properties were also observed to express Substance P only after chronic constriction injury in mice [
31] though the N values were too small to be significant. The great reduction in sprouting at POD3 if the dorsal ramus is cut along with the ventral ramus is also similar to observations made in rat [
14]. It is of interest that cutting the grey ramus to the L5 DRG, which also reduces sprouting within the DRG, can greatly reduce SNL-induced pain behaviors in the rat [
14,
32]. Co-localization of sympathetic sprouts with spontaneously active neurons was demonstrated using the rat sciatic nerve transection model to induce sprouting [
33]. These comparisons suggest that studies of sympathetic sprouting give similar results in mice and rats, enhancing the usefulness of transgenic mice in studying sympathetic sprouting.
Sympathetic basket formations around sensory neurons after peripheral nerve injury or inflammation, as first observed by McLachlan et al.[
1], initially generated much interest because they provided a possible anatomical basis for sympathetically enhanced or maintained pain states. However, their relevance to pain models has been questioned in part because of conflicting behavioral studies (see discussion in ref [
14]), and in part because basket formations are relatively rare. The data presented in this study provide evidence for important functional roles for these sympathetic sprouts. We find that at a time early after injury (POD 3) sensory neurons with basket formations are much more excitable than neighboring neurons without basket formations; and, in particular, the basket cells have a very high incidence of spontaneous activity (71% showed spontaneous activity, primarily of the bursting pattern). In the data from DRG sections stained for TH, the average incidence of basket formations was 6.8% (Table
1). For comparison, the incidence of spontaneous activity in all DRG cells at this time point was 10% (Figure
3, after allowing for the fact that small cells were excluded from the electrophysiological data). This suggests that over half of the overall spontaneous activity may be attributed to basket cells, despite their overall rarity. In turn, spontaneous activity has been shown to play key roles in pain models, including roles in satellite glia activation, neurotrophin release, spinal cord sensitization, and establishment of pain behaviors[
28,
34‐
37]. Additional evidence for functional relevance of sympathetic sprouting comes from our observations that cells that have nearby sympathetic fibers, but not a full-blown basket or ring formation, have an intermediate electrophysiological phenotype between that of basket cells and cells lacking any nearby sympathetic fibers. Though studies based on microscopy have tended to focus on and quantitate only the basket formations, the cells with the lower density nearby fibers are more numerous (11.5% vs. 6.8%; Table
1) and insofar as the presence of these nearby fibers is also associated with a more excitable phenotype, it seems that these structures may also contribute to a functional role for sympathetic sprouting.
The data presented here show an association between spontaneous activity in vitro and the presence of sympathetic fibers around the neuron. This suggests a highly localized interaction between the sympathetic fibers and nearby sensory neurons - neighboring cells lacking nearby fibers have different properties. This may reflect long-lasting, pro-excitatory interactions in vivo that are preserved in our acutely isolated in vitro preparation, and/or may indicate ongoing release of excitatory sympathetic transmitters in our in vitro preparation. Some evidence for the latter possibility was obtained in rat DRG, in medium but not large diameter cells [
14]. Do spontaneously active neurons attract sympathetic fibers, or do nearby sympathetic fibers enhance excitability and help drive spontaneous activity? Or could both processes occur in a positive-feedback cycle? This study provides evidence for interactions going in both directions. On the one hand, spontaneous activity is highest on day 1, before much sympathetic sprouting has occurred, suggesting spontaneous activity cannot be initiated by sympathetic inputs. However, we also found that simply cutting the small dorsal ramus in addition to the larger ventral ramus that is usually ligated in the SNL model, reduced sprouting on POD3 and also dramatically reduced spontaneous activity to levels seen in control animals. Insofar as spontaneous activity is initiated by axonal ligation, cutting both dorsal and ventral ramus would have been expected to increase spontaneous activity, not decrease it, though such an effect would have been quite small given the very small number of lumbar DRG neurons with projections in the dorsal ramus (less than 3% in rat [
14]). This suggests that it is more likely the large reduction in sympathetic sprouting that explains the observation that cutting the dorsal ramus greatly reduces spontaneous activity.
The literature describes a number of possible mechanisms for the postulated two-way interaction between spontaneous activity and sympathetic sprouting. Spontaneously active cells may release neurotrophic factors that attract sympathetic fibers, either directly or via activation of surrounding satellite glia cells that then synthesize neurotrophic factors (reviewed in [
36]). In the rat sciatic nerve transection model, blocking nerve activity early after injury reduces sympathetic sprouting, and increasing activity increases sprouting[
33]. Conversely, sensory neurons (and their associated satellite glia) also express receptors for sympathetic neurotransmitters norepinephrine and ATP that have excitatory effects on neurons, particularly after injury or inflammation [
38‐
40], and sympathectomy reduces SNL-induced spontaneous activity recorded in rat dorsal root fibers [
41].
The membrane capacitance distributions suggested that spontaneously active cells as well as basket cells were in the size range near the boundary between medium and large diameter cells. The very largest cells were not generally spontaneously active and did not have basket formations. However, in experiments in which conduction velocity was measured, basket cells always had conduction velocities above 9 m/s; though the higher velocities were not accurately measured due to the shortness of the stimulated dorsal root and the physiological temperature used. These velocities were clearly distinct from the much slower, accurately measured values seen in a subset of cells including small cells. The conduction velocities and size distribution suggested that spontaneously active cells and cells with basket formations might overlap with myelinated nociceptors. Though most nociceptors are small cells with nonmyelinated axons, a significant minority of cells conducting in the Aαβ range show nociceptive (high) thresholds and project into lamina I/II of the spinal cord [
42,
43]. In rat it is estimated that 20% of Aαβ cells are myelinated nociceptors [
42]. Consistent with this possibility, we observed that several commonly used nociceptive markers - TrkA, CGRP, and substance P - were markedly enriched in basket cells compared to cells lacking nearby sympathetic fibers. TrkA in particular has been proposed as a marker for the myelinated nociceptors [
42], and an association of TrkA with sprouting induced by NGF overexpression in mice has been reported [
44]. However, there are also numerous mechanisms by which Aαβ neurons that are not normally nociceptive can contribute to abnormal pain states [
45]. Further studies are needed to more precisely define the physiological roles and projections of the basket cells and spontaneously active cells.
We found that substance P was largely confined to small and medium diameter cells (Additional file1, Figure S1), while CGRP and TrkA had a somewhat broader distribution, similar to previous studies in both rats [
46‐
49] and mice [
50,
51]. Some studies, primarily in rat, have shown upregulation of substance P and/or CGRP specifically in medium or large diameter neurons in various pain models including SNL [
50‐
54]. Our immunohistochemical data did not provide clear evidence for this phenomenon, possibly because we studied a much earlier time point than most of the above studies or because the time course may differ in mice. Comparison of cells with upregulated substance P or CGRP in previous studies, with the basket cells described in this study, is confounded by the fact that the basket cells described here (especially the spontaneously active ones) would have been defined as having Aβ fibers in functional studies but probably would be classified as "medium" diameter cells in many microscopy studies; the categorization of cells as small, medium, or large diameter does not correlate very precisely with C, Aδ, and Aβ conduction velocities respectively [
31]. The present study was not designed to specifically address the question of selective upregulation of substance P or CGRP in medium and large diameter cells, but does show that these nociceptive markers are enriched in cells with nearby sprouting sympathetic fibers.
Methods
Animals
Transgenic mice expressing EGFP under the TH promoter [
55,
56] were obtained from the Mutant Mouse Regional Resource Centers and bred via hemizygous x wildtype (ND4 Swiss Webster) crossings in the local animal facility. Genotyping was performed using DNA isolated from tail clips. Mice were housed four per cage under a controlled diurnal cycle of 12 h light and 12 h dark with free access to water and food. The ambient environment was maintained at constant temperature (22 ± 0.5°C) and relative humidity (60-70%). All the surgical procedures and the experimental protocol were approved by the Institutional Animal Care and Use Committee of the University of Cincinnati (Cincinnati, OH).
Spinal nerve ligation
Female transgenic mice (pTH-EGFP) weighing 20-30 g at the time of surgery were anesthetized with isoflurane. An incision was made on the back between L2 and S1. The L4 and L5 spinal nerves were exposed and tightly ligated with 6-0 silk and cut about 2 mm distal to the ligature. The procedure was based on the original version of this pain model in rat [
57] in which the L5 and L6 spinal nerves were ligated; except that as many subsequent researchers have done, the spinal nerves were also cut after being ligated; in addition, cutting L4 and L5 spinal nerves in the mouse has been proposed on an anatomical basis to most closely match cutting the L5 and L6 spinal nerves in the rat [
58].
Recording from whole excised DRG preparation
Intracellular recording was performed on sensory neurons with or without surrounding sympathetic fibers in whole DRG preparations isolated from normal or SNL mice at POD 1, 3, 5 or 7 as indicated. The ipsilateral L5 or L4 DRG was dissected out of the mouse under barbiturate anesthesia, placed in the recording chamber and mounted on the stage of an up-right microscope (BX70-WI, Olympus). A U-shaped stainless steel rod with 3 pieces of fine nylon filaments crossing from one side to the other was used to gently hold the ganglion in place within the recording chamber. Reducing the size of this holder was the primary change required to adapt the recording method used in rat DRG [
14,
24] to the smaller mouse DRG. The DRG was continuously perfused with artificial cerebrospinal fluid (ACSF) bubbled with 95%O
2 5%CO
2 containing (in mM): NaCl 130, KCl 3.5, NaH
2PO
4 1.25, NaHCO
3 24, Dextrose 10, MgCl
2 1.2, CaCl
2 1.2 (pH = 7.3). The temperature was maintained at 36 ± 1°C. Intracellular, electrophysiological recordings were made with microelectrodes filled with 3 M KCl,10 mM HEPES, 10 mM EGTA (pH = 7.2) coated with Sigmacote (Sigma, St. Louis, USA). Satisfactory recordings were obtained with electrodes of 30-50 MW from large- (> 40 mm) and medium-sized (< 40 mm) neurons, classified by the diameter of the soma. Small diameter neurons as defined by capacitance and the presence of a broad inflected action potential, approximately 33% of the cells that could be impaled from the DRG surface in our preparation, were observed to rarely receive sympathetic sprouts under the conditions used and rarely manifested spontaneous activity. Hence data obtained in recordings from these cells are not presented in this study, which focuses on medium and large diameter cells. The electrophysiological data were collected with the use of single-electrode continuous current-clamp (Axoclamp 2B, Molecular Devices, Sunnyvale, CA, USA) and analyzed with pClamp 9 software (Molecular Devices). In experiments to determine the incidence of spontaneous activity, individual DRG neurons were first impaled with a recording electrode. If spontaneous activity was absent during the first 60 sec of the impaling, incremental currents (up to 5 nA) were then injected to ensure that action potentials could be evoked indicating a healthy cell. If any spontaneous activity was present, then we waited for 3 min to ensure that the activity was not caused by penetrating the somata with the sharp electrode. Next, the following parameters were measured using a series of current pulse injections: the threshold current (rheobase), action potential (AP) threshold and duration (measured at threshold), resting membrane potential (Vm) and input resistance (Rin, measured from the size of the voltage response to hyperpolarizing current injections), and capacitance (measured by fitting a single exponential function to the early part of the response to hyperpolarizing current injection) of the recorded DRG cell. In some experiments, the dorsal root (8 - 13 mm) was pulled into a suction electrode for applying single extracellular current stimuli to evoke action potentials for measurement of conduction velocities. All electrophysiological data presented are from transgenic mice, though the green fluorescence was not observed in every experiment.
For experiments in which cells were identified before recording as having basket formations or nearby sympathetic fibers, green fluorescent protein was imaged using a spinning disk confocal, so as to observe the live tissue with relatively low light intensity to reduce any possible tissue damage caused by the laser. The surface of the DRG was observed for fluorescence before recording. Any neuron identified as being surrounded by sympathetic basket or having nearby sympathetic fibers was localized for recording as quickly as possible (usually within 1 min). The light source was turned off during recording and turned on again to take a picture of the cell just after recording to document the status of sympathetic fiber sprouting. In addition to recording neurons with nearby sympathetic fibers, some neurons not having sympathetic fiber baskets or nearby fibers were randomly picked for recording after examining fluorescence, to eliminate the possibility that fluorescent light exposure per se might evoke spontaneous activity in DRG neurons. Because in these experiments cells were chosen based on their sympathetic fiber status, the relative numbers of cells in each group will not reflect the overall incidence of each cell type.
Immunostaining of TH-positive sympathetic nerve fibers in whole-mount DRG
Mice were anesthetized with pentobarbital sodium and fixed by perfusing with 20-30 ml of Zamboni's fixative through the left ventricle of the heart. Whole DRGs were first incubated with rabbit anti-TH (1:1000) (Pel-Freeze, Rogers, AR) 48 h at 4°C followed by a reaction with Alexa Fluor 594 conjugated goat anti-rabbit secondary antibodies (1:1000) (Invitrogen, Carlsbad, CA). Using SlideBook 4 Digital Microscopy Software (Intelligent Imaging Innovation, Denver, CO), confocal images were captured and stored in a computer for later analysis. The microscope was the same as used for recording, and the confocal was used to reduce background fluorescence. For quantification, the summed intensity of TH signal in the cellular regions of the whole mount DRG was measured via using Slidebook 4.1 imaging acquisition software (Intelligent Imaging Innovation, Denver, CO).
Immunostaining of TH-positive sympathetic nerve fibers and nociceptor markers in sectioned DRG
On postoperative day 3 or in normal mice as indicated, mice were anesthetized with pentobarbital sodium (40 mg/kg, i.p.) and were fixed by perfusing 50-100 ml of Zamboni's fixative (4% paraformaldehyde in 0.1 M phosphate buffer, pH = 7.4) through the left ventricle of the heart.
Ipsilateral DRGs (L4 and L5) were removed. Tissue was post-fixed in the perfusion fixative for 2 hours at room temperature. The ganglia were horizontally sectioned with a Cryostat at thicknesses of 10 μm. DRG sections were incubated in sheep anti-TH (Abcam, Cambridge, MA) at a dilution of 1:500 and rabbit anti TrkA, or CGRP, or SP all at a dilution of 1:500 (Abcam) overnight at 4°C, followed by reaction with secondary antibodies, donkey anti rabbit conjugated to Alexa Fluor 594 and donkey anti sheep conjugated to Alex Fluor 488 (1:1000, Invitrogen, Carlsbad, CA) for 1 hour at room temperature. After drying, the sections were mounted on coverslips with Vector Hard Set mounting medium (Vector Laboratories Inc., Burlingame, CA, USA).
Slides from control and experimental groups were labeled with numbers so that the person performing the image analysis was blinded as to the experimental group. In addition, all images were captured and analyzed by an investigator other than the one who performed immunohistostaining to avoid possible bias. Images from ~10 sections of each DRG were captured under a confocal microscope using Slidebook 4.1 imaging acquisition software (Intelligent Imaging Innovation, Denver, CO). The number of neurons surrounded or innervated by sympathetic fibers was counted and then normalized by the total number of neurons in the analyzed image area to give a percentage of neurons with basket formations or nearby sympathetic fibers. In double staining experiments, neurons expressing TrkA, CGRP or SP were counted and classified as having basket formations, nearby sympathetic fibers, or no nearby fibers.
Data analysis
Data are expressed as the mean ± the standard error of the mean. ANOVA or Kruskal-Wallis One Way ANOVA on ranks, was used compare multiple groups; t-test or Mann-Whitney sum test was used to compare two groups; the nonparametric test was chosen whenever the data were not normally distributed. The test used in each case is indicated in the text or figure legend. Fisher's exact test was used to test the significance of differences in 2 × 2 tables except for tables with larger N values for which the Chi Square test was used. For electrophysiological data, all experimental groups included cells from at least 3 different animals. Quantification of immunohistochemical data was also done using sections or samples from at least 3 different animals per experimental group. Significance was ascribed for p < 0.05. Level of significance is indicated by the number of symbols, e.g., *, p = 0.01 to < 0.05; **, p = 0.001 to 0.01; ***, p < 0.001.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
WX conducted electrophysiological and microscopy experiments, data analysis, experimental design, and manuscript revision. JAS contributed to experimental design, data analysis, and drafting of the manuscript. JM contributed to immunohistochemistry experiments and maintained the transgenic mouse line. J-MZ contributed to experimental design, manuscript revision, and data analysis. All authors read and approved the final manuscript.