Neuropathic Pain Creates Systemic Ultrastructural Changes in the Nervous System Corrected by Electroacupuncture but Not by Pregabalin.

PURPOSE: It is unclear whether neuropathological structural changes in the peripheral nervous system and central nervous system can occur in the spared nerve injury model. In this study, we investigated the pathological changes in the nervous system in a model of neuropathic pain as well as the effects of electroacupuncture (EA) and pregabalin (PGB) administration as regards pain relief and tissue repair. PATIENTS AND METHODS: Forty adult male SD rats were equally and randomly divided into 4 groups: spared nerve injury group (SNI, n = 10), SNI with electroacupuncture group (EA, n = 10), SNI with pregabalin group (PGB, n =10) and sham-operated group (Sham, n=10). EA and PGB were given from postoperative day (POD) 14 to 36. EA (2 Hz and 100 Hz alternating frequencies, intensities ranging from 1-1.5-2 mA) was applied to the left "zusanli" (ST36) and "Yanglingquan" (GB34) acupoints for 30 minutes. The mechanical withdrawal thresholds (MWTs) were tested with von Frey filaments. Moreover, the organizational and structural alterations of the bilateral prefrontal cortex, hippocampus, sciatic nerves and the thoracic, lumbar spinal cords and dorsal root ganglions (DRGs) were examined via light and electron microscopy.
RESULTS: MWTs of left hind paw demonstrated a remarkable decrease in the SNI model (P < 0.05). In the SNI model, ultrastructural changes including demyelination and damaged neurons were observed at all levels of the peripheral nervous system (PNS) and central nervous system (CNS). In addition, EA improved MWTs and restored the normal structure of neurons. However, the effect was not found in the PGB treatment group.
CONCLUSION: Chronic pain can induce extensive damage to the central and peripheral nervous systems. Meanwhile, EA and PGB can both alleviate chronic pain syndromes in rats, but EA also restores the normal cellular structures, while PGB is associated with no improvement.

Introduction

Neuropathic pain (NP) is a particularly intractable and persistent type of chronic pain that arises as a direct consequence of a pathological insult to the somatosensory system.1 The pain is characterized by shooting or burning quality, hyperalgesia, allodynia, numbness, and can occur spontaneously.2 The prevalence of NP is 9.8% in the general population, and is an important public health issue.3 Moreover, if this is not treated relatively quickly, it can easily contribute to both depressive and anxiety-like complications as well as chronic cognitive changes. It also poses an incredible burden on not only individuals but also the health care system.4 Unfortunately, management of NP remains a major clinical challenge, given that the pathogenesis of NP is not yet completely elucidated.5 Moreover, regardless of the inciting cause of NP it always shares the clinical characteristics previously described.6 We speculated that a central nervous system (CNS) injury may be associated with these characteristics in patients with NP. It is therefore crucial to the further understanding of the pathophysiology of NP to investigate possible mechanisms and patterns of injury and their treatment.

Available clinical neuroimaging evidence indicates that chronic pain is considered a CNS disorder.7 Several brain regions are activated in response to nociceptive stimulation, such as the raphe and the periaqueductal gray, the primary and secondary somatosensory cortices, thalamus, insula, and the mid-cingulate cortex (MCC).8,9 Wang et al have observed reductions in the gray matter volume of many cortical and subcortical cerebral areas in trigeminal neuralgia (TN) patients.10 Previous basic research from our team has confirmed that rats with an experimental TN lesion exhibit systemic ultrastructural changes in the prefrontal cortex (PFC), hippocampus and medulla oblongata.11–13 Electro-acupuncture (EA) can restore many if not most of these ultrastructural changes, while pregabalin (PGB) does not show equivalent improvement.11 Despite significant research efforts, detailed knowledge about the pathophysiology involved and the structural and functional changes of the CNS induced by NP is still limited.

In the present study, we aimed to demonstrate the pathophysiology underlying the systemic ultrastructural changes of the nervous systems induced by neuropathic pain. In addition, the effects of EA intervention and PGB administration in pain relief and tissue repair were also assessed.

Materials and Methods

Ethical Statement

All animal experimental procedures were approved by the Animal Research Ethics Committee of Aviation General Hospital of Medical University and were performed in accordance with the ethical guidelines for the Care and Use of the International Association for the Study of Pain.14

Animals

Forty adult male Sprague-Dawley rats, weighing 250–300g, were purchased from Beijing Vital River Laboratory Animal Technology (Beijing, People’s Republic of China). Rats were housed in groups of four in plastic cages under standard laboratory conditions (22 ± 2°C, 12 h light/dark cycles and 50–70% relative humidity). Food and water were available ad libitum. Rats were divided into four groups randomly: spared nerve injury group (SNI, n=10), electro-acupuncture group (EA, n = 10), pregabalin group (PGB, n =10) and sham-operated group (Sham, n=10).

Spared Nerve Injury (SNI) Sciatic Neuralgia Rat Model Establishment

The SNI model was established as previously described.15 The rats were anesthetized using 1% Pelltobarbitalum Natricum (40 mg/kg body weight intraperitoneally, AMRESCO, USA). Then, blunt dissection was performed to expose the three peripheral nerve branches of the left sciatic nerve. The common peroneal and tibial nerves were tightly ligated by 5.0 silk and 2–4 mm portion of the nerves was removed. The sural nerve remained intact and great care was taken to avoid any contact with or stretching of this nerve. In the sham operation, the animals received the same procedure but without any lesion to the sciatic nerve.

Behavioral Assessments

Animals were tested 3 days before the surgery and on 3, 6, 9, 12, 15, 20, 25, 30, and 40 days after the surgery. Testing was performed from 7 am to 7 pm. Rats were placed in transparent plastic boxes (20×20×20 cm) on an elevated metal mesh floor. A habituation of 15–30 min was allowed before the test. Then, the threshold for paw withdrawal (both ipsilateral and contralateral sides) was stimulated with calibrated von Frey monofilaments (Stoelting, Chicago, IL).16 Monofilaments were perpendicularly applied to the glabrous area of limbs with adequate force to cause filament bending. A sharp withdrawal of the paw was considered as a positive response. Behavioral testing was also done at the same time each day.

PGB Administration

PGB (75 mg, Pfizer, New York, NY) was dissolved in 7.5mL 0.9% saline and administered by oral gavage using an intragastric instrument. Previously, it was shown that PGB reduced the nociceptive response after chronic contractile injury (CCI) of the sciatic nerve in a dose-dependent manner (3–30 mg/kg). Therefore, PGB treatment (30mg/kg) started from the 14th day following surgery and continued for 23 days, from postoperative day 14 to 36.17 PGB administration occurred at the same time each day to avoid possible circadian effects.

EA Intervention

According to the procedure described by Tao et al,18 the rats were treated with EA starting on the 14th day after surgery. During the treatment, disposable intradermal acupuncture needles (gauge #32, 0.5 in length) were inserted into the left hind paw “Zusanli” (ST36) and “Yanglingquan” (GB34).19 “ Yanglingquan” (GB34) is anatomically located along the outer side of the lower leg, in the depression anterior and inferior to the fibular head.20 “Zusanli” (ST36) is located 5 mm below the capitulum fibulae, laterally and posterior to the knee joint.21 The needle was fixed with adhesive tape and the animal transferred to a transparent plastic cage. EA stimulation was subsequently performed using a Han’s Acupoint Nerve Stimulator (HANS, LH series, Peking University). The frequency of EA stimulation remained 2/100 Hz variable, and the current intensity remained at 1 mA for 10 minutes, then increased to 1.5 mA for 10 minutes, and finally increased to 2 mA for 10 minutes. EA treatment was performed once every 3 days for 30 minutes, from 14 to 36 days after surgery and conducted at the same time each day.

Morphology Studies

On day 40 after surgery, the experimental animals in the four groups (Sham, SNI, EA, and PGB group) were sacrificed and tissue samples were prepared. In brief, tissue samples of the spinal cord (T7 and L4 level) and brain were removed and fixed in 10% neutral formalin fixative for 48 h. Then, these tissue samples were embedded in paraffin wax. Paraffin-embedded tissue was cut into 5 micron thick sections and histologically examined using H&E staining and Nissl staining for general assessment of histopathological changes.22

Transmission Electron Microscopy (TEM)

The preparation of samples was described previously.11 Under deep anesthesia, animals were perfused with warm saline, solution of 4% paraformaldehyde and 2% glutaraldehyde (Sigma, St. Louis, Missouri, USA). The bilateral prefrontal cortex, hippocampus, bilateral sciatic nerves, the spinal cord and dorsal root ganglions (DRGs) at the thoracic (T7), and lumbar (L4) levels were immersed in 3% glutaraldehyde for 24 hours, and then washed 3 times with 0.1M phosphate buffer. The tissues were fixed with 1% osmium tetroxide (Sigma, St. Louis, MO, USA) for 2 hours, dehydrated, embedded in araldite, cut into 1 µm plastic sections and stained with uranyl acetate and then observed under a Hitachi H-7700 transmission electron microscope (Hitachi Ltd., Tokyo, Japan).

Statistical Analysis

All statistical analyses were performed using SPSS Version 24.0 (IBM Corporation, NY, United States). All data are presented as the mean ± standard deviation (SD). Comparisons between groups were performed using one-way analysis of variance (ANOVA). Comparisons of intra-group differences across time were assessed with repeated measures analysis and p < 0.05 was considered as statistically significant.

Results

SNI Induced Behavioral Signs of Nociception in the Hind Paw

As shown in Figure 1, there were no significant differences in mechanical withdrawal thresholds (MWTs) between the 4 groups prior to surgery. However, the MWT of the SNI group showed a significant decrease after operation in comparison to the sham group. Compared with the SNI group, there were no significant differences in MWT of the non-operative hind paw between the EA and PGB groups. The MWTs on the operative hind paw of the SNI model group showed significant decreases after operation in comparison to the sham group (*p < 0.05). We observed that EA treatment starting 14 days after surgery increased the MWTs of the operative hind paw to von Frey stimulation. Significant recovery was first observed after 7 days of treatment with EA (−p = 0.015). The recovery effect continued until the end of the experimental period. Similar effects were observed in the PGB treatment group (#p = 0.021). Therefore, both EA and PGB treatment reduced the behavioral effects associated with sciatic nerve injury.

Figure 1

MWTs between four groups for hind paw at different time intervals.

Neuropathological Changes of Nervous Systems in Rat with SNI

H&E and Nissl staining demonstrated no obvious pathological change in the sham group in bilateral prefrontal cortex and hippocampus. The cell boundaries were clear and closely arranged with clear nuclei in bilateral prefrontal cortex (Figure 2A, C, a and c) and hippocampus (Figure 2E, G, e and g). In comparison, the bilateral prefrontal cortex and hippocampus of the SNI group revealed disorderly arrangement of neurons, irregularly shaped neurons with shrunken cell bodies, ruptured nuclear membranes and vanishing nucleolus (prefrontal cortex: Figure 2B, D, b and d; hippocampus: Figure 2F, H, f and h). Similarly, TEM in the sham group showed normal neurons in the bilateral prefrontal cortex (Figure 3A, C, a and c) and hippocampus (Figure 3E, G, e and g) with normal appearing endoplasmic reticulum, Golgi body, intact outer nuclear membranes and mitochondria. Compared with the sham group, the SNI group showed abnormal indentation of neurons, damaged mitochondria and Golgi apparatus, and dissolving surrounding tissues in the bilateral prefrontal cortex (Figure 3B, D, b and d) and hippocampus (Figure 3F, H, f and h).

Figure 2

Hematoxylin and eosin (H&E) stained and Nissl-stained sections from bilateral prefrontal cortex and hippocampus in sham-operated and SNI group.

Figure 3

Ultrastructural changes of bilateral prefrontal cortex and hippocampus in sham-operated and SNI group.

As for the pathological changes of different levels of the spinal cord, the rats in the SNI model group showed cell nuclei contraction in the bilateral thoracic spine (Figure 4B, D, b and d) and lumbar spinal cord (Figure 4F, H, f and h), neurons disappeared into vacuoles, and Nissl bodies were deeply stained and partially lost compared with the sham operation group (thoracic spine: Figure 4A, C, a and c; lumbar spinal cord: Figure 4E, G, e and g). TEM observations demonstrated a light demyelination in both the operative (thoracic spinal cord: Figure 5D; thoracic DRG: 5B; lumbar spinal cord: Figure 5H; lumbar DRG: 5F) and non-operative spinal cords and DRGs (thoracic spinal cord: Figure 5d; thoracic DRG: 5b; lumbar spinal cord: Figure 5h; lumbar DRG: 5f) at thoracic, and lumbar levels in SNI group rats compared with sham operation group (thoracic spinal cord: Figure 5C and c; thoracic DRGs: 5A and a; lumbar spinal cord: Figure 5G and g; lumbar DRGs: 5E and e) on POD 40. As expected, the spared nerve injury caused damage to the spinal cord DRGs at all levels examined. The myelin sheaths were swollen, twisted and showed demyelination on both the operative and non-operative sides.

Figure 4

H&E and Nissl-stained sections from thoracic and lumbar spinal cords.

Figure 5

Ultrastructural changes of thoracic and lumbar spine, DRGs and sciatic nerves in SNI group compared with sham-operated group.

Similar phenomenon has been observed in sciatic nerve. On postoperative day 40, complete and clearly visible myelin sheaths were observed on the operative (Figure 5I) and non-operative (Figure 5i) sciatic nerve in the sham group. In the SNI group, demyelination and Schwann cell degeneration were noted in the operative sciatic nerve (Figure 5J). Similarly, damage was observed in the contralateral sciatic nerve (Figure 5j).

EA but Not PGB Treatment Restored Normal Structures of the Nervous Systems

After EA treatment, cell atrophy and nucleolar loss were reduced in the bilateral prefrontal cortex (Figure 6A, a, C and c) and hippocampus (Figure 6E, e, G and g). However, in PGB group, many cytoplasmic inclusions and nuclear shrinkage appeared in the bilateral hippocampus (Figure 6F, f, H and h) and prefrontal cortex (Figure 6B, b, D and d). Cells exhibited a disorderly arrangement, and neuronal eosinophilia was present. TEM observation showed intact and normal neuronal cells in the bilateral prefrontal cortex (Figure 7A, a, C and c) and hippocampus (Figure 7E, e, G and g), as well as intact endoplasmic reticulum and mitochondrial membranes in EA group. However, in the PGB treatment group, damage in the bilateral hippocampal (Figure 7F, f, H and h) and prefrontal cortex (Figure 7B, b, D and d) did not show similar improvement. Partial cytoplasmic lysis and changes in neuronal cell bodies and mitochondria were still present.

Figure 6

Effects of EA and PGB in neuropathological alterations of bilateral prefrontal cortex and hippocampus.

Figure 7

Ultrastructural changes of bilateral prefrontal cortex and hippocampus in EA and PGB group.

Similarly, H&E and Nissl staining demonstrated that the spinal cord tissue of the EA group showed no obvious neuronal damage in bilateral thoracic spine (Figure 8A, a, C and c) and bilateral lumbar spinal cord (Figure 8E, e, G and g), while those in the PGB group (thoracic spine: Figure 8B, b, D and d; lumbar spinal cord: 8F, f, H and h) showed no obvious improvement in neuronal damage compared to the rats in the SNI group. In addition, TEM observations also showed that myelin sheaths on the operative (thoracic spinal cord: Figure 9C; thoracic DRG: 9A; lumbar spinal cord: Figure 9G; lumbar DRG: 9E) and non-operative (thoracic spinal cord: Figure 9c; thoracic DRG: 9a; lumbar spinal cord: Figure 9g; lumbar DRG: 9e) thoracic, lumbar spinal cords levels and DRGs were intact and normal in EA treated rats. However, in PGB-treated rats, operative (thoracic spinal cord: Figure 9D; thoracic DRG: 9B; lumbar spinal cord: Figure 9H; lumbar DRG: 9F) and non-operative (thoracic spinal cord: Figure 9d; thoracic DRG: 9b; lumbar spinal cord: Figure 9h; lumbar DRG: 9f) side DRGs showed partial myeloid dissolution and onion-like demyelinating changes. As for bilateral sciatic nerve, slight demyelination was observed in EA group (Figure 9I and i) compared to SNI group. However, severely swollen myelin sheaths were shown in the PGB group (Figure 9J and j).

Figure 8

Effects of EA and PGB in neuropathological alterations of bilateral thoracic and lumbar spinal cords.

Figure 9

Ultrastructural changes of thoracic and lumbar spine, DRGs and sciatic nerves in EA and PGB group.

Discussion

Generally, a better rodent animal model is the premise and foundation for elucidating the mechanism of NP and evaluating potential treatment. The current mainstream models for NP include CCI,23 spinal nerve ligation model (SNL),24 and SNI.15 Of which, the SNI rat model is one of the most widely used in the study of NP. In recent years, many studies have used the SNI model to explore the mechanism of NP and innovative therapies.25–27

In this study, we found that the systemic ultrastructural alterations were observed at different levels of the nervous system in the SNI rat model. That is, the micro-structure of the bilateral PFC, hippocampus, spinal cord (T7 and L4 level), DRGs, and sciatic nerves were damaged in SNI group. Specifically, the shrunken cell bodies, ruptured nuclear membranes and vanishing nucleolus were observed bilaterally in the prefrontal cortex and hippocampus. Meanwhile, the neuropathological changes (mainly demyelination) were also observed at different levels of the spinal cord, bilateral DRGs, and bilateral sciatic nerves. This phenomenon was similar to that seen in our previous studies.17 We surmise that one of the main causes of this phenomenon is neuroinflammation.

Chronic pain is associated with neuroinflammation.28 Several lines of evidence suggest that the mechanical allodynia induced by SNI is secondary to a neuroinflammatory response in the CNS, that is characterized by aberrant microglial activation in the spinal cord.29,30 Gordh et al also reported that patients with severe peripheral neuropathic pain demonstrated upregulation of several chemokines in their cerebrospinal fluid.31 Neuroinflammation plays an essential role in the generation of central sensitization which is a major contributing factor to the development and maintenance of chronic pain.32 Neuroinflammation occurred in the peripheral nervous system (PNS) and CNS and was characterized by the infiltration of leukocytes as well as increased production of inflammatory mediators. These inflammatory mediators could lead to demyelination and by neuronal injury by exerting cytotoxic effects on oligodendrocytes and neurons.33

Demyelination in nervous systems is a key mechanism in the development of plasticity in neuropathic pain.34 This may also explain why ultrastructural alterations were observed bilaterally in tissues in the CNS. As for pathological changes in the PNS, we identified both demyelination and Schwann cell degeneration bilaterally in the sciatic nerve that was not in line with our expectations.

It was easy to understand why the left sciatic nerve (operative side) showed demyelination and Schwann cell degeneration that is expected with the peripheral mechanisms of neuropathic pain.34 However, a similar damage pattern appeared in the right sciatic nerve (non-operative side). This phenomenon is in line with previous studies35–38 that suggest a systemic inflammatory response appears to be associated with this phenomenon. Kleinschnitz35 et al demonstrated interleukin-10 (IL-10), monocyte chemoattractant protein-1 (MCP-1), and IL-1 beta in the contralateral sciatic nerve are significantly increased following unilateral sciatic nerve injury. Another two possible mechanisms have been proposed by Koltzenburg et al.36 The first is that the contralateral effects are mediated by circulating factors. After unilateral nerve injury, the breakdown products from the denervated tissue and damaged nerve might circulate and induce alterations in the contralateral neuronal populations. The second possibility is that the contralateral effects are mediated by transmedian sprouting. That is, unilateral nerve injury could result in collateral sprouting. Such sprouting could cause the changes in contralateral neurons.

Another interesting finding in our study is that both EA and PGB were able to relieve painful hypersensitivity in SNI rats. However, EA interventions could also restore the normal structures, while PGB treatment did not show any improvement. This was also consistent with our previous studies. These results imply that neuromodulation techniques have additional advantages when compared to a pharmacologic approach in the treatment of chronic pain. EA, a therapeutic means based on acupuncture combined with electrical stimulation,39 has proven effective in pain relief40 and multiple organ protection.41 Parmen et al42 showed that EA can facilitate an improvement in tissue mechanical stress and enhance the wound healing process without detrimental side effects. In addition, EA can also promote tissue repair, ultrastructural reconstruction and pain relief by reducing neuronal apoptosis,43 inhibiting the spinal glial cell activation40 and neuroinflammatory response,44 decreasing cytokines and mediating the release of mesenchymal stem cells.45 Conversely, PGB administration did not affect the ultrastructural changes, although it did show a beneficial effect on pain relief in the SNI rat model. Unfortunately, it is unclear what the possible mechanism is for this interesting phenomenon even though we carefully consulted the literature.

This study has several flaws and limitations. Firstly, the systemic ultrastructural alterations were observed at different levels of the nervous system in SNI rat. However, we are not explicitly aware of the detailed timeline of the damage process. For example, we cannot establish with certainty if the spinal cord damage preceded or followed the onset of the brain changes. Secondly, the morphological study was a non-quantitative and observational study. A quantitative study is the direction for our future work. Thirdly, there were inadequate numbers of dorsal horn neuron images suitable for direct comparison. Fourthly, we do not know the precise molecular and biological mechanism of the micro-structural changes induced by our chronic pain model. Fifthly, although we found several meaningful outcome changes, the significance behind these phenomena is unknown. We can only surmise that many of the mental health disorders induced by chronic pain are partially explained by the systemic ultrastructural alterations of the nervous systems. Future research should also aim to specifically address these questions. Sixthly, we have no information as to whether the pathological changes we observed in the present study were due to tearing of the tissue or vacuolization during preparation of the tissues. Finally, only the mechanical withdrawal thresholds (MWTs) were assessed in this study. Thermal and cold hyperalgesia were not assessed among groups.

Conclusion

Chronic pain can induce extensive damage to the central and peripheral nervous systems. Meanwhile, both EA and PGB can alleviate chronic pain syndrome, but EA also restores the normal structures, while PGB is associated with no improvement.