| Summary: | 博士 === 國立臺灣大學 === 臨床醫學研究所 === 101 === Introduction
Ischemia has been reported to be important in the pathogenesis of human neuropathies caused by acute peripheral arterial occlusion and in neuropathies caused by vasculitis. It has also been implicated in metabolic disorders, e.g. diabetes mellitus and in disseminated neuropathy due to prolonged coma (critical illness polyneuropathy). The main clinical presentation of ischemic neuropathy is a mononeuropathy simplex or multiplex with asymmetric pain, sensory symptoms and motor deficits. Several clinical studies have confirmed the association between morphological and functional impairments of peripheral nerves in peripheral vascular diseases. Among the neurological deficits in these patients with peripheral vascular insufficiency or vasculopathy, neuropathic pain of variable incidence (43%-58.8%) is frequently mentioned, depending on the severity of the diseases. Skin denervation in vasculitic neuropathy has rarely been documented despite frequent manifestations of small-fiber neuropathy including reduced sensitivity and neuropathic pain.
Recently, skin biopsy has been established as a new approach to diagnose small fiber sensory neuropathy. Small-diameter sensory nerves innervating the skin are responsible for conveying noxious and thermal stimuli, and damage to these nerves is presumably related to neuropathic pain behaviors. However, the mechanisms of ischemia-induced neuropathic pain are largely not clear at present. Specifically, two issues are seldom mentioned: the relationship between the extent of ischemia and the occurrence of neuropathic pain, and the susceptibility of small-diameter sensory nerves in the skin to ischemia. Neurophysiological studies of the preferential involvement of nerve types in ischemia show wide discrepancies in several clinical investigations, ranging from predominantly sensory neuropathy to mainly involvement of motor nerves. After ischemia, there is a progressive reduction in membrane potential resulting in a transient increase in excitability followed by a failure of impulse conduction. This increased excitability may be important for the clinical symptoms of hyperesthesia, pain and numbness in patients presenting with vasculitis or diabetic neuropathy in early stages. Nerve injury elicits regenerative growth, which occurs in most peripheral neurons, and long-term hyperexcitability that appears selectively in nociceptive sensory neurons; the long-term hyperexcitability enhances inputs to nociceptive circuits in the CNS and can contribute to chronic pain after nerve injury. How information from the site of a distal lesion is communicated to the cell nucleus is not completely understood, but positive injury signals, i.e. axonal proteins that are activated at the site of the injury and retrogradely transported to the cell soma, are thought to have an import role. Several studies indicate that protein kinase G activated in axons is a limited signal that is responsive for the induction of long-term hyperexcitability.
The pathological changes of ischemic neuropathy are well described as axonal degeneration and regeneration, endoneurial edema, abnormally thin in myelinated fibers and endothelial swelling of endoneurial microvessels. The energy requirements is lower in peripheral nerves, which indicate that peripheral nerve has a greater metabolic safety factor than the central nervous system and there is a much wider window of time for therapeutic intervention in peripheral nerve than in the brain. Although restoration of blood flow to the ischemic nerve is essential to prevent irreversible damage, reperfusion can result in oxidative injury to endothelial cells with ensuing endoneurial edema, lipid peroxidation and fiber degeneration. There are various studies of neuroprotective approaches using the aforementioned models including hypothermia, hyperbaric oxygenation, and the antioxidant α-lipoic acid. However, further studies are needed to investigate the mechanisms of the treatments.
Neuropathic pain results in a series of complex but coordinated behaviors mediated by damage to large myelinated, small myelinated, and unmyelinated nerve fibers. Several groups including ours have demonstrated the rich innervation of the skin by immunohistochemistry with various neuronal markers, particularly, protein gene product (PGP) 9.5. PGP 9.5 is an ubiquitin carboxyhydrolase and probably functions as an immediate early gene for processing sensory information in the neurons. Our previous study on chronic constriction injury and laser-induced focal neuropathy indicate that PGP 9.5(+) nerve terminals in the skin were moderately depleted compared with those in completely denervated skin. These findings suggest that partial injury is another important principle for creating experimental models of neuropathic pain. Apparently, large-diameter and small-diameter nerve fibers are differentially vulnerable in models of neuropathic pain. Whether terminals of sensory nerves in the skin retain the same patterns as they have at the sciatic nerve level is an open issue. Thus, it was intriguing to investigate whether there is a relationship between nerve injury to fibers of different categories and the magnitude of neuropathic pain.
A critical issue in ischemia-induced neuropathic pain studies is to develop an experimental system with quantifiable degree of injury. Most of the currently available ischemic animal models tend to be less than satisfactory in their ability to simulate focal ischemic nerve damage in a reproducible and physiologically relevant manner. Models based on ligation, microembolisation or compression have been developed by Adams (1943) and Roberts (1948) and improved by Korthals and Wisniewski (1975). Nukada and Dyck (1984) obtained dose-related effects using polystyrene microspheres. Sladky (1991) and coworkers produced chronic regional nerve ischemia in rats by creating proximal arteriovenous shunts. For the ischemic models induced by arterial ligation, most studies were performed in the rats or rabbits; seldom experiments were performed in mice and did not specify whether neuropathic pain developed. There are many mice with many genetic models mimicking human disease and an ischemic neuropathy model in mice will shed light on the pathophysiology of ischemia-induced neuropathic pain.
The purposes of this study were (1) to investigate the pathologic features of cutaneous nerves in ischemic neuropathy, (2) to generate a mouse system of ischemic neuropathy with neuropathic pain behaviors and quantifiable degree of nerve injury by ligating the femoral artery with reperfusion, (3) to investigate the influence of ischemia on nerve fibers of different categories and the magnitude of neuropathic pain behaviors using multidisciplinary approaches, (4) to evaluate the susceptibility of small-diameter sensory nerves in the skin to ischemic injury by examining the skin innervation, and (5) to evaluate effects of ant anti-nociception treatment on allodynia in ischemic nerve injury. We hypothesize that the degree of ischemia has influence on categories and the magnitude of neuropathic pain.
Material and Method
I. Skin denervation in vasculitis
In the study of skin denervation in vasculitis, clinical data of patients were retrieved from the database of hospitalized patients at National Taiwan University Hospital, Taipei, from January 1, 2000, to December 31, 2003. All patients fulfilled 2 criteria: (1) clinical presentation of a mononeuropathy multiplex of acute or subacute onset and (2) pathological evidence of vasculitis on nerve biopsy specimens following the consensus criteria. All underwent detailed neurologic examinations and laboratory investigations to exclude metabolic and infectious diseases that affect the peripheral nerves, including plasma glucose level, and functional tests of the liver and kidneys. Specific autoimmune diseases, such as systemic lupus erythematosus and rheumatoid arthritis, were excluded by the following relevant laboratory tests: antinuclear antibody, antibody against double-stranded DNA, anti–Sjögren syndrome A antigen, anti-–Sjögren syndrome B antigen, anti–Smith antigen, anti–scleroderma antigen (SCL-70), anti-RNP, rheumatoid factor, C3 and C4 complement levels, cryoglobulin, hepatitis profiles (hepatitis B surface antigen, anti–hepatitis B surface antibodies, and anti–hepatitis C virus antibodies), syphilis (VDRL/rapid plasma reagent test), anti–human immunodeficiency virus antibodies, lead levels in serum, immunoelectrophoresis of proteins in the serum and cerebrospinal fluid, and tumor markers (α-fetoprotein, carcinoembryonic antigen, CA125, and CA19-9). Patients had 3-mm punch biopsy specimens taken from the distal part of the leg (without active vasculitic lesions) and a sural nerve biopsy specimen was taken in addition to detailed neurologic examinations, laboratory investigations, and nerve conduction studies. Results of nerve conduction studies, epidermal nerve fiber density studies, and immunohistochemistry were analized.
II. Patterns of nerve injury and neuropathic pain in ischemic neuropathy after ligation-reperfusion of femoral artery in mice
Effect of femoral arterial ligation on epineurial arteries and blood flow
In the study of patterns of nerve injury and neuropathic pain in ischemic neuropathy after ligation-reperfusion of femoral artery in mice, the femoral artery on one side was tied in the proximal portion with a 7-0 prolene suture using a slipknot technique for rapid release, and reperfusion was achieved by the release of the ligature on adult male 8-week-old ICR mice. Epineurial blood flow of the tibial nerve was measured with a laser Doppler flow meter before and immediately after arterial ligation, and immediately after reperfusion. Behavioral testing assessed after arterial ischemia including mechanical allodynia using a set of calibrated von Frey filaments, and cold allodynia using acetone test with modifications. Electrophysiological studies for evaluation about the motor function of the sciatic nerve after ischemia was performed by measuring the compound muscle action potential (CMAP). Immunohistochemistry of the footpads with an antibody to protein gene product (PGP) 9.5 and quantitation of epidermal innervation expressed as the epidermal nerve density (END) were performed. Pharmacological studies using gabapentin was performed for assessment about the neuropathic pain behaviors.
Experimental designs and statistical analyses
In the first phase of the experiment, we investigated the duration of the ischemic effect of inducing neuropathic pain. The criterion of successful induction was the presence of sustained mechanical allodynia in the interval between postoperative day (POD) 7 and 14. We studied the effects of different ischemic durations on the development of neuropathic pain behaviors. Mice were randomly divided into four groups and respectively received 3, 4, 5, and 6 h of ischemia, followed by reperfusion. Neuropathic pain was determined by the presence of sustained mechanical allodynia. There were six mice at each time point for each laboratory procedure. Behavioral and laboratory data are presented as the mean ± SEM at different time points after ischemia. For the statistical analysis of values obtained from nerve-conduction studies and ENDs over the experimental period, analysis of variance (ANOVA) and a post-hoc analysis were used. Differences in the CMAP amplitude and END between the control and operated sides were tested using a paired Student''s t-test. Any difference with p < 0.05 was considered statistically significant.
Results
I. Skin denervation in vasculitis
Clinical presentations and laboratory examinations
All patients had motor and sensory impairments with acute or subacute onset. Motor and sensory symptoms usually began focally or asymmetrically in one leg or thigh and then progressed to affect the contralateral side and the upper limbs. Three patients had simultaneous symptoms in the upper and lower limbs as the initial manifestation. Pain was noted in most patients (5 of 6). Elevated protein levels in the cerebrospinal fluid were noted in 3 patients. Nerve conduction studies showed a pattern of axonal neuropathy or mononeuropathy multiplex. After immunotherapy with corticosteroids, plasma exchange, or cyclophosphamide, there were significant improvements in muscle strength (at least 1 grade in the weakest limb according to the grading of the Medical Research Council) and reduction of neuropathic pain in all patients.
Pathological demonstration of vasculitis
Inflammatory cells surrounding vessels were demonstrated in the sural nerves and dermal vessels of the skin, and cells were immunoreactive for markers of macrophages (CD68) and T cells (CD3), but not for B cells (CD20). In patients with vasculitis, the quantity of epidermal nerves was markedly reduced. The epidermis of some patients had even become completely denervated. We quantified epidermal innervation by measuring the epidermal nerve density. All patients had significantly reduced epidermal nerve fiber density (0-3.18 fibers per millimeter) compared with the aforementioned normative values from our laboratory. In addition, dermal nerves showed a pattern of degeneration; sweat glands became denervated in some patients.
II. Patterns of nerve injury and neuropathic pain in ischemic neuropathy after ligation-reperfusion of femoral artery in mice
Effect of femoral arterial ligation on epineurial arteries and blood flow
After femoral arterial ligation, the blood flow on the operated side (flow velocity of post-operated/baseline level) was significantly reduced to < 40% of the baseline value (flow ratio 37.5% ± 9.2%, p = 0.002) within 10 min. The decrease in epineurial blood flow persisted through the entire period of arterial ligation. After reperfusion, the blood flow returned to the pre-ligation level within 1 h after releasing the ligature.
Behavioral observations after arterial ligation with subsequent reperfusion
Ischemic injury resulted in distinct neuropathic behaviors in the mice, showing significant changes in behaviors, gait, and stance, showing guarding of the affected paw from touching the floor which was most pronounced during PODs 4~14, and these gradually decreased after POD 21. In the first week after the operation, mice showed spontaneous pain behaviors, for example, elevating the hind paw with shaking movements, particularly when walking on uneven surfaces. When walking, the mice always elevated the hind paw of the affected side to avoid touching the floor with the footpads.
Effects of the ischemic duration on neuropathic pain behaviors: mechanical allodynia and acetone-induced cold allodynia
Femoral arterial ligation could induce mechanical allodynia during the first postoperative week after various ischemic duration of 3, 4, 5, or 6 h. In the 6-h ischemic group, there was a transient increase in the mechanical threshold on the operated side compared to that on the contralateral side from PODs 1 to 2 (POD 1: 0.31 ± 0.14 vs. 0.13 ± 0.06 g, p = 0.039; POD 2: 0.29 ± 0.18 vs. 0.14 ± 0.02 g, p = 0.037, Fig. 2). In the 5-h ischemic group, the withdrawal thresholds to mechanical stimuli on the operated sides were significantly reduced compared to those on the control sides from PODs 4 to 56 (POD 4: 0.04 ± 0.06 vs. 0.1 ± 0.03 g, p < 0.001; POD 56: 0.01 ± 0.008 vs. 0.06 ± 0.03 g, p = 0.017, Fig. 3). In preliminary studies, a shorter ischemic duration (3 h) produced less-sustained mechanical allodynia lasting for < 2 weeks, compared to 4 weeks with 4-h ischemic injury, and > 2 months with 5- and 6-h ischemia. The presence of longer initial hypoesthesia in the 6-h ischemic group made it less suitable for assessing acute neuropathic pain behaviors after ischemic injury, compared to the 5-h ischemic group. Therefore we determined that femoral arterial ligation for 5 h was an optimal ischemic duration and used this setting to induce neuropathic pain in animals in subsequent studies.
Mice with ischemic neuropathy after 5-h ligation of femoral artery also showed significant acetone-induced allodynia. The response to cold stimuli (application of acetone to the plantar skin) on the operated side became accentuated compared with that on the control side from PODs 1 to 56 (POD 1: 2.81 ± 0.97 vs. 1.33 ± 0.47, p < 0.001; POD 56: 2.57 ± 0.53 vs. 1.71 ± 0.7, p = 0.003).
Changes in nerve conduction studies after femoral arterial ligation with reperfusion
On POD 7, ischemia for 3, 4, 5, or 6 h resulted in significant reductions in CMAP amplitudes on the operated side compared to that on the control side, providing evidence of large-fiber deficits. There was a dose effect of ischemic duration, i.e., the longer the ischemic duration, the smaller the CMAP amplitude was (p = 0.009 on ANOVA and post-hoc analysis). In the 5-h ischemic group, the reduction in CMAP amplitudes on the operated side persisted through POD 56 without improvement (3.38 ± 1.83 mV on POD 7 vs. 4.00 ± 1.68 mV on POD 56, p = 0.559).
Skin innervation in the vicinity of the sciatic nerve after femoral-arterial ligation
On POD 7, the abundance of epidermal nerves was significantly reduced as validated by lower ENDs on the operated side than that on the control side, indicating the effect of ischemia on small-diameter sensory nerve terminals in the skin. The effect of skin denervation was related to the ischemic duration, and a longer ischemic duration caused much-lower ENDs compared to those with a shorter ischemic duration (p = 0.027 on ANOVA and post-hoc analysis), for example, 6.67 ± 1.65 fibers/mm for the 6-h ischemic group vs. 13.52 ± 3.55 fibers/mm for the 3-h ischemic group, p = 0.032). In the 5-h ischemic group, the skin remained denervated on POD 56 with no sign of recovery in skin innervation (8.34 ± 1.62 fibers/mm on POD 7 vs. 7.89 ± 1.44 fibers/mm on POD 56, p = 0.84).
Vulnerability of large and small fibers to ischemic injury
The aforementioned results indicated there was damage to both large- and small-diameter nerve fibers after ischemic injury as quantified by reductions in CMAP amplitudes and ENDs, respectively. The damage to small fibers was more robust than that to large fibers as determined by comparing the ratios (operated/control sides) betweens ENDs and CMAP amplitudes on POD 56 (27.5% ± 6.1% vs. 50.5% ± 24.0%, respectively, p = 0.026).
Effects of an intraperitoneal injection of gabapentin on allodynia with ischemic nerve injury
To understand whether the above behaviors could be alleviated by pharmacological treatment, we administrated gabapentin (25, 50, and 100 mg/kg) intraperitoneally on POD 7. Mechanical allodynia was significantly reduced by the use of gabapentin (50 and 100 mg/kg) between 0.5 and 3 h after the intraperitoneal injection, with a maximal effect at 1 h. Gabapentin (50 and 100 mg/kg, ip) also significantly reduced cold allodynia at 0.5~3 h with a maximal effect at 1 h. Gabapentin alleviated mechanical allodynia in a dose-dependent manner. Gabapentin at 25 mg/kg did not significantly affect the neuropathic pain behaviors.
Discussion
The main contribution of this study included revealing small-diameter sensory nerves are affected in vasculitis in addition to the well-known effect of vasculitis on large-diameter nerves; the denervation of cutaneous small-fiber may correlate to the neuropathic pain. Significant inflammatory vasculopathy is present in the skin despite the absence of clinically active vasculitic lesions. We generate a new mouse system of ischemic neuropathy after ligation-reperfusion of the femoral artery. Direct vascular injury of the femoral artery was remote from the sciatic nerve and thus eliminated any concerns about direct mechanical compressive injury of the sciatic nerve. This system of ischemia-reperfusion injury replicated two major clinical manifestations: neuropathic pain and nerve degeneration. For neuropathic pain, these included spontaneous pain behaviors and evoked pain behaviors (mechanical allodynia and cold allodynia) which developed within 1 week after nerve injury and persisted through POD 56 of the experimental period. The neuropathic pain can be alleviated by gabapentin. The reduction of ENDs in the current model provides an index to assess the ischemic effect on small-diameter sensory nerve terminals in the skin, and this model replicates what was observed in human vasculitic neuropathy. This system produced by a simple procedure provides an opportunity to investigate mechanisms and further treatments of ischemic neuropathy on genetically engineered mice.
|
|---|
