Pathological pain induced by peripheral nerve trauma and inflammation is regarded as neuropathic pain (Zimmermann,2001). Current therapies for this kind of pain are only partially effective, mainly because we do not have the complete knowledge concerning the induction and maintenance of neuropathic pain. Previous studies have shown that spinal astrocytes works as a key player in the pathological pain states (Watkins and Maier,2003), but the underlying mechanisms remain unclear. GLT-1 is one of the most abundant glutamate transporters (Rothstein et al.,1996; Danbolt,2001; Mao et al.,2002), and it locates on the cell membrane of astrocytes, takes up the bulk of synaptically released glutamate, and diminishes the glutamate spilled over to other synapses (Rothstein et al.,1994). Astrocytes are capable of metabolizing incorporated glutamate into glutamine by the enzyme glutamine synthetase. As both glutamate and astrocytes are closely related to neuropathic pain, GLT-1, which is exclusively expressed on the astrocytes, might be crucial in the induction and/or maintenance of neuropathic pain (Schlag et al.,1998; Sung et al.,2003; Weng et al.,2006). Previous researches had revealed the involvement of GLT-1 in chronic nerve constriction injury (CCI; Sung et al.,2003), but the detailed functions of GLT-1 in nociception and the exact relationship to its host astrocytes have not been studied. In this study, we attempted to clarify: (1) whether GLT-1 is specifically involved in the induction and/or maintenance of neuropathic pain induced by the widely used neuropathic pain model, spinal nerve ligation (SNL); and (2) the relative temporal changes of astrocyte activation and GLT-1 expression under the neuropathic pain states.
Astrocyte activation is involved in the neuropathic pain. As a glutamate scavenger, the glutamate transporter-1 (GLT-1) is exclusively expressed on the astrocytes and probably correlates with astrocyte activation. In the present study, we attempted to clarify the temporal changing courses of astrocyte activation and GLT-1 expression, as well as their correlations induced by a neuropathic pain model, namely, spinal nerve ligation (SNL) in which rapidly appearing (<3 days) and persistent (>21 days) mechanical allodynia and thermal hyperalgesia were presented. Immunofluorescent staining showed that GLT-1 was expressed exclusively in most (not all) of the astrocytes, even when the GLT-1 expression reached its peak. The expression of GLT-1 displayed an interesting biphasic change, with an initial up-regulation followed by a down-regulation after SNL. Our results also demonstrated that SNL induced a marked and long-term (>21 days) activation of astrocytes in the ipsilateral spinal dorsal horn. These results suggest that astrocyte activation, the change of GLT-1 expression and the potential relationship between them might play key roles in the induction and/or maintenance of neuropathic pain. The present results provide novel clues in understanding the mechanisms underlying the involvement of astrocytes and GLT-1 in the neuropathic pain. Anat Rec, 291:513–518, 2008. © 2008 Wiley-Liss, Inc.
MATERIALS AND METHODS
Spinal Nerve Ligation
Adult male Sprague-Dawley rats (body weight 250–300 g) were housed in plastic cages, three rats in each cage, and maintained on a 12:12 hr light/dark cycle under conditions of 22–25°C ambient temperature with food and water available. All experimental procedures agreed with the Animal Use and Care Committee for Research and Education of the Fourth Military Medical University (Xi'an, P. R. China) and also the ethical guidelines to investigate experimental pain in conscious animals (Zimmermann,1983). All efforts were made to minimize animal's suffering and to reduce the number of animals used.
To produce an SNL model, under pentobarbital anesthesia (40∼50 mg/kg, IP), the left L5 transverse process was removed to expose the L4 and L5 spinal nerves. The L5 spinal nerve was then carefully isolated and tightly ligated with 6-0 silk thread (Kim and Chung,1992). The surgical procedure for the sham group was identical to that of SNL group, except that spinal nerves were not ligated.
The animals were habituated to the testing environment for 3 days before baseline testing. For testing mechanical sensitivity, animals were put under inverted plastic boxes (11 × 13 × 24 cm3) on an elevated mesh floor and allowed to habituate for 30 min before the threshold testing. Mechanical allodynia was tested using von Frey filaments (Stoelting, Kiel, WI) in a blinded manner. The paw was pressed with one of a series of von Frey filaments with increasing stiffness (2, 4, 6, 8, 10, 15, and 26 g), presented to the plantar surface (5–6 sec for each filament). The paw withdrawal threshold was determined using Dixon's up-down method (Chaplan et al.,1994). Thresholds for behavioral response to heat stimuli applied to each hindpaw were assessed using the Hargreaves test (Hargreaves et al.,1998). Paw withdrawal thresholds were calculated from an average of three consecutive withdrawal latencies of the ipsilateral hindpaws measured between 15-min intervals. A cutoff time of 40 sec was imposed to avoid tissue damage.
Animals were killed at different time points (1, 3, 5, 7, 14, 21 days) after SNL. After being anesthetized with pentobarbital (60 mg/kg, IP), the rats were perfused through the ascending aorta with 100 ml of 0.9% saline followed by 500 ml of 0.1 M phosphate buffer (PB, pH 7.3) containing 4% paraformaldehyde and 2% picric acid. Subsequently, the fifth lumbar cord segment (L5) was removed and post-fixed in the same fixative for 2–4 hr and then cryoprotected for 24 hr at 4°C in 0.1 M PB containing 30% sucrose.
Transverse spinal sections (30 μm thickness) were cut in a cryostat, collected serially into three dishes, each of which contained a complete serial section, and processed for immunofluorescent staining. All of the sections in the first dish were rinsed in 0.01 M phosphate buffered saline (PBS, pH 7.3) for three times (10 min each), blocked with 2% goat serum in 0.01 M PBS containing 0.3% Triton X-100 for 1 hr at the room temperature (RT), and then used for the double immunofluorescent staining. The sections were incubated overnight at 4°C with a mixture of two primary antibodies: mouse anti–glial fibrillary acidic protein (GFAP; 1:5,000; Chemicon, Temecula, CA) and rabbit anti-GLT-1 (1:2,000; Chemicon). The sections were washed for three times in 0.01 M PBS (10 min each) and then incubated for 2 hr at RT with the mixture of secondary antibodies: fluorescein isothiocyanate–conjugated horse anti-mouse IgG (1:200; Vector, Burlingame, CA) and Alexa 594-conjugated donkey anti-rabbit IgG (1:800; Molecular Probes, Rockford, IL).
The specificities of the staining were tested on the sections in another dish by omission of the primary specific antibodies. No immunoreactive products were found on the sections as predicted.
Confocal images were obtained under a confocal laser scanning microscope (FV1000; Olympus, Tokyo, Japan). Digital images were captured with Fluoview 1000 (Olympus). Five nonadjacent sections from the L5 segments were selected randomly. For semiquantification, the fluorescent brightness value of GFAP- and GLT-1–like immunoreactivities were detected on the same areas of the dorsal horn by using software under IX-70 confocal microscope.
After the images were captured, optical density of the same areas of the ipsilateral superficial dorsal horn (lamina I and II, Fig. 2A) was calculated and averaged across the five spinal sections. All data were showed as mean ± SEM (n = 5) were analyzed for statistical significance with the Student's t-test between the sham/SNL rats and ipsilateral/contralateral dorsal horn. P-values of less than 0.05 were considered as significant.
SNL Significantly Induced Pain Behavioral Changes
Paw withdrawal threshold (PWT) scores were decreased in animals with SNL compared with sham group animals (Fig. 1A), and the Hargreaves test showed that SNL induced a remarkable reduction of thermal response latency (Fig. 1B). SNL produced rapidly appearing (<3 days) and persistent (>3 weeks) mechanical allodynia and thermal hyperalgesia. In addition, PWT and thermal response latency of contralateral hindpaw of SNL-treated rats decreased from baseline compared with the sham group (Fig. 1) as a sign of mirror image pain (Twining et al.,2004).
SNL Induced a Persistent Increase of Astrocyte Activation and a Biphasic Change of GLT-1 Expression
Compared with the sham group, SNL induced a marked expression of GFAP in the ipsilateral compared with the contralateral spinal cord (Fig. 2B). A time course study showed that GFAP up-regulation was not evident on day 1, when symptoms of neuropathic pain began to develop. The rise in GFAP levels was significant on day 3, reached a peak on day 7, and remained at high levels on day 21 after SNL (Fig. 3), when the neuropathic pain response was still very predominant, and previous studies indicated that this nociceptive response could last more than 2 months (Kim and Chung,1992; Zhuang et al.,2006).
SNL also induced a significant change in the expression of GLT-1 within the ipsilateral spinal dorsal horn. A time course study exhibited the expression displayed an interesting biphasic change, with an initial up-regulation followed by a down-regulation after SNL (Fig. 3). The intensity of GLT-1 immunoreactivity increased during the development of neuropathic pain (<5 days), but decreased in the maintenance phase (>7 days), and almost disappeared on day 21 after SNL (Fig. 3).
The results of the double immunofluorescent staining revealed that GLT-1 immunoreactivity was completely colocalized, with GFAP immunoreactivity in the spinal cord during neuropathic pain development (Fig. 2D–F). But the density of GLT-1 was higher in the soma than in the processes, indicating that GLT-1 was mostly expressed on the cell bodies of the astrocytes. In addition, not all the astrocytes expressed GLT-1, even when the GLT-1 expression reached its peak on day 5.
The results of the present study demonstrated that SNL induced a persistent increase of astrocyte activation, whereas a biphasic change of GLT-1 expression appeared during the induction and maintenance of the neuropathic pain. In the early stage of the nerve ligation injury, GLT-1 up-regulation may be regarded as a sort of feedback regulation, because it probably prevents accumulation of glutamate in the synaptic cleft and protects neurons from ongoing glutamate receptors activation and, thus, central sensitization (Sung et al.,2003). In the late stage of the SNL, ligation may cause a retrograde degeneration of the dorsal root ganglion neurons, which produce a loss of primary afferents, the same as the changes in the CCI model (Sung et al.,2003). These changes can reduce the release of neurotransmitters (including glutamate; Kajander and Xu,1995), which indirectly leads to down-regulation of the GLT-1. Previous studies have also demonstrated that culturing of astrocytes without neuron down-regulates GLT-1 expression (Schlag et al.,1998); transection of the facial nerve leads to a down-regulation of GLT-1 protein (Lopez-Redondo et al.,2000). It seems that down-regulation of GLT-1 could result from neuron degeneration induced, which may be a mechanism of GLT-1 reduction in the late phase of neuropathic pain. As demonstrated in the previous study, GLT-1 down-regulation may contribute to neuropathic pain and inhibition of glutamate uptake in the rat spinal cord results in mechanical allodynia and hyperalgesia (Weng et al.,2006).
In the spinal cord, glial cells play an important role in modulation of neuropathic pain (Watkins et al.,2001), and GLT-1 is exclusively expressed in the astrocytes (Rothstein et al.,1994) so that the relationship between GLT-1 expression and astrocyte activation is worth revealing. In the present study, GLT-1 was up-regulated just 1 day after SNL. This was earlier than the astrocyte activation. These results suggest that GLT-1 up-regulation functions as an initiator of astrocyte activation, because astrocytes could respond to glutamate, which is transported intracellularly by GLT-1 (Lalo et al.,2006). Then astrocyte activation contributes to the neuropathic pain, and in this way, GLT-1 indirectly functions as a promoter of neuropathic pain, but further evidence needs to be elucidated. The results of the present study also suggest that, during the late stage, GLT-1 functions differently as in the early stage, whereas astrocytes still keep their activation at a high level and contribute to the maintenance of the neuropathic pain state, which is consistent with the previous report (Zhuang et al.,2006).
Further studies can be performed to investigate the underlying molecular mechanisms for the astrocyte activation and GLT-1 expression induced by SNL, by which the involvement of astrocytes in neuropathic pain can hopefully be understood. This finding might supply a new avenue for treatment of neuropathic pain.