| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
From the Departments of *Anesthesiology, and Neurology, Rush Medical College, Chicago, Illinois.
Address correspondence to Jeffrey S. Kroin, PhD, Department of Anesthesiology, Rush Medical College, 1653 West Congress Parkway, Chicago, IL 60612. Address e-mail to jkroin{at}rush.edu.
Abstract
BACKGROUND: Plantar hindpaw incision produces hyperalgesia, transient upregulation of cyclooxygenase-2 (COX-2) and prolonged upregulation of cyclooxygenase-1 (COX-1) in rat lumbar spinal cord. Our hypothesis in this study was that a deep thoracic incision causes COX-1 and COX-2 upregulation in the dorsal horn coincident with pain-related behavior, and that specific cell types contribute to this increase in COX expression.
METHODS: A left lateral thoracic skin incision was made in anesthetized rats, and superficial and deep muscles were incised. Postoperative pain-related behavior was quantified by recording exploratory rearing. Four and 24 h postsurgery, COX-1 and COX-2 immunohistochemistry, with co-labeling for cell type, were performed on the spinal cord.
RESULTS: Deep thoracic muscle incision produced a 42% decrease in rearing compared to sham skin-incision controls at 4 h postsurgery (P = 0.001). There was an increase in both COX-1 and COX-2 immunoreactivity in the thoracic dorsal horn at 4 h postsurgery on the ipsilateral side of surgery animals compared to the ipsilateral side of control animals, contralateral side of surgery animals or contralateral side of control animals. No surgery-induced differences were seen at the lumbar level. At 24 h postsurgery, there was no longer a decrease in rearing, and no surgery-induced differences in COX-1 or COX-2 were seen at any level. At 4 h postsurgery, 96% of COX-1 immunoreactive cells co-localized with microglia and 98% of COX-2 immunoreactive cells co-localized with neurons.
CONCLUSIONS: A unilateral deep thoracic wound produces pain-related behavior and, at the same time, ipsilateral upregulation of microglial COX-1 and neuronal COX-2 in the thoracic dorsal horn.
Plantar hindpaw incision in the rat leads to hypersensitivity and an increase in the number of cyclooxygenase-1 (COX-1) immunoreactive cells in the lumbar superficial dorsal horn that peaks at 2 days postsurgery.1 Bilateral plantar hindpaw incisions produce a transient increase in lumbar spinal cyclooxygenase-2 (COX-2) protein that peaks at 3–6 h and returns to baseline at 24 h.2 However, upregulation of COX has not been directly measured in pain models associated with deep surgical wounds. After laparotomy, rats exhibit pain-related behavior that can be attenuated by the systemic administration of the mixed COX-1/COX-2 inhibitors ketoprofen and carprofen.3,4 The mixed COX-1/COX-2 inhibitor ketorolac, given systemically or intrathecally, partially reduces laparotomy-induced pain-related behavior, and greatly enhances the effect of morphine in this model.5,6 After deep thoracic muscle incisions, both ketorolac and a COX-2 selective inhibitor, given systemically, reduce pain-related behavior and tissue and thoracic cerebrospinal fluid (CSF) prostaglandin E2 (PGE2) levels.7 Intrathecal ketorolac also reduces CSF PGE2.7 While the studies on deep surgical wounds suggest that there is spinal upregulation of COX-1 and/or COX-2, the specific cell types involved in these effects and quantification of the magnitude and timing of this COX upregulation have not been examined in these models.
Postoperative pain in patients can also be attenuated with COX inhibitors that reduce PGE2 levels. After hip replacement surgery, both pain and lumbar CSF and local tissue PGE2 can be reduced with perioperative administration of a COX-2 selective inhibitor.8 After vascular surgery, postoperative administration of a COX-2 inhibitor rapidly reduces both pain and CSF PGE2 and, to a lesser extent, a mixed COX-1/COX-2 inhibitor has the same effect.9
The hypothesis of the present study was that both COX-1 and COX-2 immunoreactivity increase in the spinal cord dorsal horn after deep thoracic surgery in the rat. Moreover, this COX upregulation is coincident with pain-related behavior, and specific cell types (neuron, microglia, astrocyte) contribute to this increased COX expression.
METHODS
Experiments were performed on 300–325 g male Sprague-Dawley rats (Charles River, Portage, MI) and were approved by the Institutional Animal Care and Use Committee.
Surgical Model
Animals were anesthetized with 1.5% isoflurane in oxygen and, under sterile conditions, a skin incision was made over the left lateral thoracic region. In the deep muscle surgery animals, both superficial and deep chest wall muscles were incised by creating 3-cm long lateral cuts over the 3rd, 5th, and 7th ribs.7 The intercostal muscles were spared. The muscle wounds were closed with silk sutures and the skin incision closed with nylon sutures. This model is different from another one from our group examining long-term neuropathic pain after pleura opening and 60-min rib retraction.10 In sham-surgery control rats, only a skin incision was performed, and closed with nylon sutures. After surgery, which lasted 15 min, animals were returned to their cages.
Pain-Related Behavior
We have previously assessed exploratory locomotive activity as a measure of pain, and the effect of COX inhibitors, after thoracic muscle surgery,7 using a method similar to one used to evaluate postoperative pain in a rat laparotomy model.5,6 Briefly, animals were tested in the morning in a lighted room, in fresh clear vivarium plastic cages surrounded by a cage rack Photobeam Activity System (San Diego Instruments, San Diego, CA). The animals’ 24-h light/dark cycle was not altered.
Beam interruptions were automatically recorded. A set of horizontal beams 11 cm above ground measured rearing (beam brakes in the vertical direction) over a 60-min testing period.
In preliminary experiments, we determined that mechanical allodynia of the thoracic skin was not present at 2 or 4 h postsurgery in this muscle incision model. Von Frey filaments, over the range 20–282 mN, were applied in ascending steps in three tests, with the lowest force of the three tests recorded as the withdrawal force threshold.11 The median withdrawal force was at 282 mN pre-surgery, and remained at 282 mN at 2 and 4 h postsurgery (n = 6). Therefore, behavioral testing was used as the primary indicator of pain.
Rats (n = 48) underwent deep thoracic muscle surgery or sham thoracic skin incision surgery. Rearing was then recorded at 4 h after surgery (n = 12 per group), and at 24 h postsurgery in another set of animals (n = 12 per group). Behavioral data between groups were analyzed (SPSS 11.5 statistical package, Chicago, IL) using Bonferroni-corrected t-tests (P < 0.05). In addition, rearing at 4 h after sham thoracic skin incision surgery was compared (t-test) with additional sham animals that only had the thoracic region shaved, all rats receiving the same anesthesia (n = 24).
Immunohistochemistry
At 4 h (n = 10) and 24 h (n = 10) after surgery, equal numbers of deep muscle surgery and sham skin-incision control animals were perfused intracardially with saline followed by cold 4% paraformaldehyde in phosphate buffered saline (PBS). After overnight immersion of the spinal column in fixative at 4°C, 8 mm lengths of thoracic (T1–3) and lumbar (L4–6) spinal cord were removed, and placed in PBS containing 30% sucrose for at 3 days. Transverse sections were cut on a freezing sliding microtome (40 µm) and stored at –20°C in cryoprotectant. For COX-1 and COX-2 protein immunostaining, one series of sections from each rat was washed sufficiently in PBS, and then incubated in 0.3% H2O2 in Tris buffered saline (TBS) for 20 min to eliminate endogenous peroxide activity. After TBS rinse, the sections were incubated in a blocking solution (5% normal goat serum and 2% bovine serum albumin in TBS) for 1 h. Sections were then incubated overnight with primary antibody (murine polyclonal COX-1, 1:1000 or COX-2, 1:1000; Cayman Chemical, Ann Arbor, MI) in the above blocking solution. In Western blots, the COX-1 antibody (#160109) reacts with rat COX-1 protein, but does not react with rat COX-2; the COX-2 antibody (#160126) reacts with rat COX-2 protein, but does not react with rat COX-1 (manufacturer's data sheet). The COX-1 antibody shows no immunohistochemical staining in COX-1 knockout mice,12,13 and the COX-2 antibody shows no immunohistochemical staining in COX-2 knockout mice.12 In our study, omission of primary antibodies resulted in negative staining in all tested sections. Sections were washed with TBS (thrice for 10 min) and incubated for 1 h with a biotinylated secondary antibody (goat anti-rabbit IgG; 1:200; Vector Labs, Burlingame, CA). After TBS wash (thrice for 10 min), sections were incubated for 75 min in an avidin-biotin complex (Elite ABC kit, 1:500; Vector, Burlingame, CA) in TBS, separated by three washes in TBS. The chromogen solution that completed the reaction consisted of 0.05% 3'3-diaminobenzidine, 0.005% H2O2, and 2.5% nickel II sulfate. Sections were mounted on gelatin-coated slides, dehydrated through graded alcohol (50, 70, 95, and 99%), cleared in xylene, and coverslipped with Cytoseal. Similar processing was performed for the animals used to compare rearing at 4 h after sham thoracic skin incision surgery to animals that only were shaved.
Quantification of COX-1 and COX-2 Immunoreactive Cells in Dorsal Horn
Quantification of the immunohistochemical signals was performed on an Olympus microscope coupled to a computer-assisted morphometry system (Image 1200; NIH). The person performing the analysis was blinded to the animal groups. The optical density (OD) of COX-1 and COX-2 immunoreactive complexes in laminae I and II of each section was measured automatically by the computer using the National Institute of Health Image software. To overcome any differences in background staining among sections, five background regions, in areas with few immunoreactive cell bodies, on each section were measured, and the mean of those five OD's was subtracted from each OD in laminae I and II of each section. Data are expressed as number of pixels per square millimeter in the laminae I–II area containing immunoreactive cells, and are presented as mean ± sem. At least six slides were analyzed at each spinal level. For statistical analysis at a spinal level and at a postsurgery time, pixels were compared with one-way ANOVA with post hoc Scheffé test (P < 0.05 for significance). An unbiased, stereological cell counting software (optical dissector method; MicroBrightField, Colchester, VT) was used.
Identification of Cell Type Expressing COX-1 and COX-2 in Dorsal Horn
Double labeling immunofluorescence was performed to identify the specific type of cell expressing either COX-1 or COX-2. Cells staining for COX-1 or COX-2 were colabeled with either an neuronal marker for neuron nuclei (NeuN), an astrocyte marker for glial fibrillary acidic protein (GFAP), or a microglia marker (OX-42). Briefly, after several washes in the TBS solution containing 0.05% Triton X-100 (TBS-Tx), sections were incubated in 5% normal goat serum in the TBS-Tx for 1 h at room temperature, followed by incubation with primary antibodies in a 0.1M PBS solution containing 0.5% Triton X-100 (PBS-Tx) overnight at room temperature. The following primary antibody pairs were used: COX-1 (rabbit IgG, 1:150, Cayman Chemical, Ann Arbor, MI) or COX-2 (rabbit IgG; 1:150, Cayman Chemical), with NeuN (mouse IgG, 1:200, Chemicon, Temecula, CA) or GFAP (mouse IgG, 1:200, Chemicon) or OX-42 (mouse IgG, 1:1000, Serotec, Raleigh, NC). The sections were rinsed in TBS-TX (3 times for 10 min) and then incubated with secondary antibodies in TBS-TX containing 3% normal goat serum for 1 h. The following fluorochrome-conjugated secondary antibodies were used: Cy2-conjugated goat anti-rabbit IgG (1:400; Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) and Cy3-conjugated goat anti-mouse IgG (1:500; Jackson ImmunoResearch Laboratories, Inc., West Grove, PA). After several washes in TBS, sections were mounted on gelatin-coated glass slides, dehydrated in ethyl alcohol, cleared in xylene, and coverslipped with DPX mounting media. Sections were evaluated by a Fluoview laser confocal system equipped with an Olympus microscope and an argon/krypton laser. For this analysis, three thoracic muscle surgery rats and three sham control rats were analyzed at 4 h postsurgery. For each colabeling combination, seven sections were analyzed from each animal.
RESULTS
Pain-Related Behavior
Thoracic muscle incision produced a 42% decrease in rearing compared with sham skin incision controls at 4 h postsurgery (P < 0.001) (Fig. 1A). By 24 h after surgery, there were no significant differences in rearing between the 2 groups of animals (P = 0.617). There was no difference in rearing at 4 h postsurgery between sham skin incision control animals and rats that only had the thoracic region shaved and were maintained under anesthesia for an equal amount of time (P = 0.688) (Fig. 1B). For the rest of the study, all sham control animals were those with sham skin incisions.
|
COX-1 Immunoreactivity in the Dorsal Horn
In sham-surgery control rats, COX-1 immunoreactivity was expressed in the spinal cord with relatively higher levels in dorsal horn laminae I–III. The higher levels were due to both increased grain number and higher intensity of each grain in those regions, compared with other regions of the spinal cord. Moderate COX-1 immunostaining was seen in the lateral part of lamina V, the meningeal layer surrounding the central canal, and in motor neurons. Sparse COX-1 staining was observed throughout the remaining gray matter and the white matter.
Four hours after thoracic muscle surgery, there was an increase in COX-1 immunoreactivity in laminae I and II at the thoracic (T1–3) spinal level on the side ipsilateral to the thoracic muscle incisions (Fig. 2A). Measurement of pixels per square millimeter of COX-1 immunoreactive cells (Fig. 2B) demonstrated a higher value in the ipsilateral dorsal horn of muscle incision animals (4541 ± 163) compared with the ipsilateral dorsal horn of the sham-operated control animals (3251 ± 48), the contralateral dorsal horn of muscle incision animals (3337 ± 109), and the contralateral dorsal horn of control animals (3084 ± 80) (F = 37.7; df = 3,20; P < 0.001). In contrast, in the lumbar spinal cord at 4 h postsurgery, there was no difference in pixels per square millimeter between ipsilateral dorsal horn of surgery animals (3532 ± 259), ipsilateral side of control animals (3362 ± 187), contralateral side of surgery animals (3423 ± 258), and contralateral side of control animals (3273 ± 257) (F = 0.202; df = 3,20; P = 0.894). At 24 h after surgery, there was no longer any increase in COX-1 immunoreactive pixels in laminae I and II at the thoracic spinal level ipsilateral to the thoracic muscle incisions: ipsilateral side, surgery (5004 ± 169), ipsilateral side, control (4988 ± 124), contralateral side, surgery (4949 ± 189), and contralateral side, control (5224 ± 106) (F = 0.675; df = 3,20; P = 0.580). At the lumbar level at 24 h, there were no pixel differences: ipsilateral side, surgery (5118 ± 199), ipsilateral side, control (5405 ± 131), contralateral side, surgery (5179 ± 160), and contralateral side, control (5477 ± 216) (F = 0.931; df = 3,20; P = 0.448). In the control experiments at 4 h comparing sham skin incision to a skin shave, there was no difference in pixel density between groups (F = 0.455; df = 3,20; P = 0.717).
|
COX-2 Immunoreactivity in the Dorsal Horn
In sham-surgery control rats, COX-2 immunoreactivity was expressed in the same regions as the COX-1 staining described in the previous section. COX-2 immunoreactivity in laminae I and II at the thoracic spinal level on the side ipsilateral to the thoracic muscle incisions increased at 4 h postsurgery (Fig. 3A). Pixels per square millimeter of COX-2 immunoreactive cells (Fig. 3B) demonstrated a larger value in the ipsilateral dorsal horn of muscle incision animals (3737 ± 36) compared with the ipsilateral dorsal horn of the sham-operated control animals (2945 ± 123), the contralateral dorsal horn of muscle incision animals (2734 ± 68), and the contralateral dorsal horn of control animals (2945 ± 59) (F = 31.9; df = 3,20; P < 0.001). In contrast, at 4 h postsurgery in the lumbar spinal cord, there was no difference in pixels per square millimeter between ipsilateral dorsal horn of surgery animals (3313 ± 132), ipsilateral side of control animals (3101 ± 253), contralateral side of surgery animals (3195 ± 169), and contralateral side of control animals (3056 ± 227) (F = 0.320; df = 3,20; P = 0.811). At 24 h postsurgery, there was no longer any increase in COX-2 immunoreactive pixels in laminae I and II at the thoracic spinal level ipsilateral to the thoracic muscle incisions: ipsilateral side, surgery (3857 ± 51), ipsilateral side, control (3916 ± 157), contralateral side, surgery (3776 ± 172), and contralateral side, control (3848 ± 111) (F = 0.188; df = 3,20; P = 0.903). At 24 h after surgery at the lumbar level, there were no pixel differences: ipsilateral side, surgery (3855 ± 162), ipsilateral side, control (3673 ± 80), contralateral side, surgery (4072 ± 111), and contralateral side, control (3889 ± 204) (F = 1.23; df = 3,20; P = 0.330). In the control experiments at 4 h comparing sham skin incision to a skin shave, there was no difference in pixel density between groups (F = 0.384; df = 3,20; P = 0.765).
|
Localization of COX Immunoreactivity in Dorsal Horn Neurons and Microglia
Almost all (96.0 ± 0.9%) COX-1 immunoreactive cells in the gray matter were colocalized with the microglia marker OX-42 (based on counts of 768 cells: red, green, or colabeled) (Fig. 4). COX-1 immunoreactive cells were not colocalized with NeuN positive cells or GFAP positive cells. This selectivity was the same for both sham control animals (Figs. 4A–C) and thoracic muscle surgery animals (Figs. 4D–F).
|
Almost all (98.0 ± 0.4%) COX-2 immunoreactive cells in the gray matter were colabeled with the neuronal marker NeuN (based on counts of 274 cells) (Fig. 5). COX-2 immunoreactivity was not colocalized with GFAP positive cells, but was present in a small number (4.2 ± 0.1%) of OX-42 positive cells (based on counts of 1063 cells). This selectivity was the same for both sham control animals (Figs. 5A–C) and thoracic muscle surgery animals (Figs. 5D–F). Since the experiments did not show any dorsal gray matter colabeling with COX-1 and NeuN or GFAP, or with COX-2 and GFAP, this argues against fluorescence break-through having produced the COX-1/OX-42 or COX-2/NeuN or COX-2/OX-42 colabeling, since all colabeling experiments used the same Cy2/Cy3 fluorochromes.
|
In the dorsal white matter, COX-1 immunoreactivity was localized mainly on microglia, with only an occasional GFAP positive cell showing COX-1 staining. COX-2 in white matter was only seen on a small number of the GFAP or OX-42 positive cells.
DISCUSSION
The importance of our findings is that they extend the understanding of spinal COX-1 and COX-2 upregulation in a deep surgical incision model by identifying the cellular location as well as magnitude and timing of these effects. Our experiments show that a unilateral thoracic muscle incision induces upregulation of both COX-1 and COX-2 immunoreactivity in the ipsilateral thoracic spinal cord of the rat at 4 h postincision, along with an increase in pain-related behavior (decrease in exploratory rearing). COX upregulation and pain are related, since our earlier study had demonstrated that a systemic nonsteroidal antiinflammatory drug (ketorolac) or a COX-2 selective inhibitor (rofecoxib) can reverse these postoperative rearing deficits after thoracic muscle surgery.7 Postoperative decreases in rearing have been well characterized after laparotomy with intestinal manipulation in rats.3,5,6,14 By 24 h, there was no longer any difference in either COX-1 or COX-2 immunostaining between thoracic muscle incision rats and skin incision controls, nor was there a difference in pain-related behavior. Since we do not have data at intermediate time points (6, 8, or 12 h), we do not know if there is a complete temporal correlation between spinal COX expression and pain. This COX-2 time course is similar to the upregulation of COX-2 protein in the lumbar spinal cord after bilateral foot incision in the rat, in which COX-2 levels are elevated at 3 and 6 h after incision but return to control values by 12 h.2 The COX-1 time course after thoracic muscle incision differs from the COX-1 upregulation after a unilateral planter foot incision, which has a more prolonged increase in COX-1 immunoreactivity that peaks in the ipsilateral lumbar dorsal horn at day 2 postsurgery, is still elevated at day 5 and returns to normal by day 7.1 In addition, the continuing mechanical hypersensitivity of the plantar hindpaw seen in that model is also prolonged: paw withdrawal thresholds are greatly decreased at 1–3 days postincision and do not return to baseline until 7 days after incision.1 One possible reason why a deep thoracic muscle wound could have a shorter duration of spinal COX-1 upregulation and pain than a seemingly more superficial paw incision is that innervation density of muscle is less than that of skin (in a sensitive region).15 In addition, input to dorsal horn neurons from muscle nociceptors has a stronger descending inhibition than the input from cutaneous nociceptors.16 In summary, the present results, in combination with previous studies,1,2 lead to a consensus that in the spinal cord both COX-1 and COX-2 upregulate in surgical pain models.
In the plantar incision model, intrathecal ketorolac or a selective COX-1 inhibitor reduced mechanical hypersensitivity,1 but intrathecal COX-2 selective inhibitors did not.1,17 Intrathecal injection of ketorolac blocked spinal CSF PGE2 upregulation after thoracic muscle surgery, but an intrathecal COX-2 selective inhibitor did not reduce CSF PGE2.7 These results suggest that spinal COX-1 upregulation, not COX-2, contributes most to postoperative pain in surgical pain models.
Animal models of peripheral inflammation, with hindlimb injections of either carrageenan or complete Freund's adjuvant, show an upregulation of COX-2 protein in the lumbar spinal cord, but not of COX-1 protein.18–20 Injury to the L5–6 spinal nerves, which produces a neuropathy in rats, was shown to cause an increase of COX-2 protein in the lumbar dorsal spinal cord 1 day after surgery that returned to normal by day 3,21 and a localized increase in COX-1 protein in the ipsilateral superficial dorsal horn that appeared by day 4.22 However, in another neuropathic pain model, involving lesioning of the peroneal and tibial nerves, but sparing of the sural nerve, only a small increase in COX-2 mRNA was seen in the deep dorsal horn at 12 h postsurgery.23 Therefore, it is not possible to generalize whether COX-1 or COX-2 is the dominant isoform in response to peripheral injury.
In our study, unilateral thoracic muscle incision produced increases in COX-1 and COX-2 immunoreactivity in the ipsilateral thoracic dorsal horn, but not the contralateral dorsal horn, or in the lumbar spinal cord. With hindpaw incision, the upregulation of COX-1 immunoreactivity is also seen in the corresponding ipsilateral cord, with no contralateral change.1 This suggests a COX-related functional connection between afferent nerves at the thoracic injury site and their destination in the dorsal horn. However, sciatic nerve block did not eliminate COX-2 mRNA upregulation in the lumbar spinal cord after hindpaw inflammation,20 nor did it prevent COX-2 protein increase in the lumbar spinal cord after plantar hindpaw incisions.2 Hence, it is possible that afferent nerve firing is not the only mechanism by which a peripheral tissue injury can cause regional upregulation in the spinal cord, and retrograde transport in afferent nerves may play some role.24
COX-1 immunoreactivity in the dorsal horn was almost exclusively (96%) localized in microglia in both surgery and sham animals. This suggests that the increase in COX-1 staining in the unilateral thoracic dorsal gray matter in surgery animals was due to upregulation of COX-1 in microglia. Since the remaining 4% of the COX-1 immunostained cells were not costained with neuronal or astrocyte markers, we assume that these COX-1 cells were oligodendrocytes, endothelial cells, undifferentiated cells, or residual blood cells. The COX-1 co-localization with microglia in our animals is consistent with a study 2 days after plantar foot incision in which COX-1 immunoreactivity was seen in microglia, but not astrocytes, in the dorsal horn.25 Microglial activation has been associated with the development of behavioral hypersensitivity in a rat model of neuropathic pain26; however, a preliminary study of the microglial inhibitor, minocycline, given before plantar hindpaw incision did not suggest that microglia had a role in the development of postoperative hypersensitivity.27
COX-2 immunoreactivity in the dorsal horn was almost exclusively (98%) localized in neurons, in both surgery and sham animals. This suggests that the increase in COX-2 staining in the unilateral thoracic dorsal gray matter in surgery animals is due to upregulation of COX-2 in neurons. A small number of COX-2 cells (4%) were colabeled with microglia. The probable reason why the total number of COX-2 co-labeled cells is not exactly 100% is that the COX-2/OX-42 sections are not the exact same sections used for COX-2/NeuN counting. The distribution of COX-2 immunostaining in our animals is consistent with previous studies.28–30 Similar results to our COX-2 findings were seen after a hindpaw crush using a hemostat, with upregulation of COX-2 protein in the ipsilateral lumbar dorsal horn, and colocalization of COX-2 immunoreactivity with cells stained with the neuronal marker NeuN.28 However, in the hindpaw crush model, mechanical hyperalgesia has a relatively short duration (<2 h) and has resolved by the time spinal COX-2 protein begins to increase, implying that this early hyperalgesia is due to constitutive, not induced, COX-2.28 Overall, our results with a deep wound model are consistent with the consensus opinion from other surgical and trauma models that dorsal gray matter COX-1 is expressed primarily in microglia25 and COX-2 in neurons.28
In addition to identifying the specific cell types involved in spinal COX-1 and COX-2 upregulation, our results (based on the magnitude and time course of these COX responses) suggest that future research strategies to reduce spinal prostaglandin production after deep tissue surgery should address both COX-1 and COX-2 inhibition.
Footnotes
Accepted for publication November 20, 2007.
Supported by University Anesthesiologists S.C.
Reprints will not be available from the author.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
