Activation and up-regulation of spinal cord nitric oxide receptor, soluble guanylate cyclase, after formalin injection into the rat hind paw

PII: S 0 3 0 6 - 4 5 2 2 ( 0 2 ) 0 0 0 7 5 - 1

Neuroscience Vol. 112, No. 2, pp. 439^446, 2002 H 2002 IBRO. Published by Elsevier Science Ltd All rights reserved. Printed in Great Britain 0306-4522 / 02 $22.00+0.00

www.neuroscience-ibro.com

ACTIVATION AND UP-REGULATION OF SPINAL CORD NITRIC OXIDE RECEPTOR, SOLUBLE GUANYLATE CYCLASE, AFTER FORMALIN INJECTION INTO THE RAT HIND PAW Y.-X. TAO and R. A. JOHNS Department of Anesthesiology and Critical Care Medicine, Blalock 1415, Johns Hopkins University School of Medicine, 600 North Wolfe Street, Baltimore, MD 21287-4965, USA

Abstract7Nitric oxide synthase is expressed abundantly in the spinal cord, and nitric oxide (NO) has been shown to play important roles in the central mechanism of in£ammatory hyperalgesia. However, the expression and function of the NO receptor, soluble guanylate cyclase, is not fully understood in this processing at the spinal cord level. In the present study, we report that the soluble guanylate cyclase K1 subunit but not the L1 subunit was expressed in rat spinal cord, particularly in the dorsal horn. We showed that intrathecal administration of a selective inhibitor of soluble guanylate cyclase, 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one, produced a signi¢cant anti-nociception demonstrated by the decrease in the number of £inches and shakes in the formalin-induced in£ammatory pain model. This was accompanied by a marked reduction in formalin-induced c-fos expression in the spinal cord. During formalin-induced long-lasting in£ammation, we found that the expression of the K1 subunit of soluble guanylate cyclase was dramatically increased in the lumbar spinal cord on the second and fourth days after formalin injection into the dorsal side of a hind paw. Intraperitoneal pretreatment with an N-methyl-D-aspartate (NMDA) receptor antagonist, dizocilpine maleate (MK801), and a neuronal NO synthase inhibitor, 7-nitroindazole, not only signi¢cantly blocked formalin-induced secondary thermal hyperalgesia but also suppressed formalin-produced increase in the K1 subunit of soluble guanylate cyclase in the spinal cord. The present results indicate that peripheral in£ammation not only initially activates but also later up-regulates soluble guanylate cyclase expression via the NMDA receptor^NO signaling pathway, suggesting that soluble guanylate cyclase might be involved in the central mechanism of formalin-induced in£ammatory hyperalgesia in the spinal cord. H 2002 IBRO. Published by Elsevier Science Ltd. All rights reserved. Key words: N-methyl-D-aspartate receptor, expression, in£ammation, long-term hyperalgesia, c-fos, pain.

7-nitroindazole (7-NI) reduces nociceptive responses to formalin-produced peripheral in£ammation in mice and rats (Malmberg and Yaksh, 1993; Moore et al., 1993; Roche et al., 1996). NO donors applied intrathecally not only cause a reduction in tail £ick or paw withdrawal latency (Inoue et al., 1997; Chen et al., 2000; Tao and Johns, 2000), but also facilitate formalin-induced nociceptive behaviors in the second phase (Shibuta et al., 1996). In addition, injection of formalin into a hind paw evokes a biphasic spinal release of NO metabolites (Okuda et al., 2001). These results indicate that NO is involved in the central mechanism of in£ammatory hyperalgesia at the spinal cord level. The major action of NO is to activate the soluble form of the enzyme guanylate cyclase (GC), which has been designated as a physiological NO receptor (Furuyama et al., 1993). Soluble GC is a heme-containing protein found in the cytosolic fraction of virtually all mammalian cells. Several forms of soluble GC have been cloned and characterized, each consisting of one K (82 kDa) and one L (70 kDa) subunit (Hobbs and Ignarro, 1996). Three di¡erent isoforms of K (K1 , K2 and K3 ) and L subunits (L1 , L2 and L3 ) have been identi¢ed from rat, bovine or human tissues (Furuyama et al., 1993; Hobbs and Ignarro, 1996). Functionally, soluble GC converts

There is now a substantial body of evidence to support that nitric oxide (NO) is an intercellular messenger molecule that is thought to be involved in synaptic transmission in both the central and peripheral nervous systems (Garthwaite et al., 1988; Bredt and Snyder, 1992). In the spinal cord, NO has been demonstrated to play a role in modulation of nociceptive transmission and plasticity. The neuronal NO synthase is concentrated in the super¢cial dorsal horn of the spinal cord (Dun et al., 1993; Terenghi et al., 1993; Saito et al., 1994). Formalin-produced long-lasting in£ammation increases NO synthase expression in the spinal cord (Herdegen et al., 1994; Lam et al., 1996; Fu et al., 2001). Systemic and intrathecal administration of NO synthase inhibitors such as Ng -nitro-L-arginine methyl ester (L-NAME) or

*Corresponding author. Tel.: +1-410-9558408; fax: +1-4109554858. E-mail address: [email protected] (R. A. Johns). Abbreviations : EDTA, ethylenediaminetetra-acetate; EGTA, ethylene glycol-bis(2-amino ethyl-ether)-N,N,NP,NP-tetraacetic acid; GC, guanylate cyclase ; GMP, guanosine-3P,5P-monophosphate ; g L-NAME, N -nitro-L-arginine methyl ester ; MK-801, dizocilpine maleate; 7-NI, 7-nitroindazole ; NMDA, N-methyl-D-aspartate ; ODQ, 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one. 439

NSC 5527 24-5-02

440

Y.-X. Tao and R. A. Johns

guanosine-5P-triphosphate to cyclic guanosine-3P,5Pmonophosphate (GMP). Cyclic GMP, as an intracellular second messenger, modi¢es several intracellular processes including activation of protein kinases, ion channels and phosphodiesterases (Knowles et al., 1989). Thus, the function of NO is mainly expressed through production of cyclic GMP, and the soluble GC plays a key role in the NO signaling pathway. Formalin injected subcutaneously into the hind paw produces typical ‘phasic’ and ‘tonic’ components of pain behaviors associated with peripheral tissue injuryinduced in£ammation (Wheeler-Aceto et al., 1990; Tjolsen et al., 1992). Recently, it was found that peripheral in£ammation after subcutaneous formalin injection produced long-term hyperalgesia (Fu et al., 2001). Considering the involvement of NO in the central mechanism of formalin-induced in£ammatory hyperalgesia, we hypothesized that soluble GC in the spinal cord might play an important role in formalin-induced in£ammatory hyperalgesia. In the present study, we ¢rst observed the expression and distribution of soluble GC subunits, K1 and L1 , in the spinal cord. Second, we tested whether inhibition of soluble GC a¡ected formalin-produced pain behaviors and formalin-induced c-fos expression as a marker of functional activity of nociceptive neurons in the spinal cord. Finally, we examined whether soluble GC expression could be up-regulated in the spinal cord during long-lasting in£ammation after formalin injection and the possible mechanisms of its up-regulation.

EXPERIMENTAL PROCEDURES

Animals Male Sprague^Dawley rats (250^300 g) were housed in separate cages on a standard 12/12-h light^dark cycle, with water and food pellets available ad libitum. The experimental procedures were approved by the Animal Care Committee at the Johns Hopkins University and were consistent with the ethical guidelines of the National Institutes of Health and the International Association for the Study of Pain. Behavioral testing Assessment of formalin-induced pain behaviors. The rats were implanted with an intrathecal catheter under pentobarbital anesthesia as described in previous work (Tao et al., 1997, 2000; Tao and Johns, 2000). In brief, a polyethylene (PE-10) catheter was inserted into the subarachnoid space at the rostral level of the spinal cord lumbar enlargement through an incision at the atlanto-occipital membrane. The animals were allowed to recover for 7^10 days before being used experimentally. Rats showing neurological de¢cits post-operatively were discarded. The agent administered intrathecally was a selective inhibitor of soluble GC, 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ, Calbiochem, USA). The drug was dissolved in dimethylsulfoxide and then dispersed in distilled water before administration. The animals were randomly assigned into four groups as follows: vehicle or distilled water (control) (n = 12); 1 Wg of ODQ (n = 6); 5 Wg of ODQ (n = 6); 10 Wg of ODQ (n = 6). The drug solution or vehicle was injected intrathecally in a volume of 10 Wl, followed by an injection of 10 Wl of distilled water to £ush the catheter. Ten minutes later, formalin (100 Wl, 4%) was injected into the plantar side of a hind paw of the rat. In addition, four rats were treated only with intrathecal ODQ (10 Wg) and another four rats received no treatment.

Immediately following the formalin injection, each individual rat was placed in a transparent cage for observation of the formalin-injected paw. The pain-related behaviors, £inches and shakes, were assessed for the next 60 min by an experimenter who was unaware of the group assignment. The observational session was divided into two periods: a phasic period (0^10 min) and a tonic period (10^60 min). The mean number of £inches and shakes for each period of each treatment group was determined. Assessment of formalin-induced long-term thermal hyperalgesia. In the rats from di¡erent agent-treated groups as described below, nociceptive paw withdrawal latencies of both hind paws to radiant heat were measured prior to saline or formalin injection and on the fourth day after formalin injection. Based on previous work (Hargreaves et al., 1988), the animals were placed in transparent plastic chambers (18U29U12.5 cm) on a glass plate preheated to a constant temperature and acclimated for 5 h each day before the experiment. Radiant heat from the apparatus (Model 336 Analgesia Meter, IITC/Life Science Instruments, Woodland Hills, CA, USA) was applied from below to the planter surface of each hind paw. A cut-o¡ time latency of 20 s was used to avoid tissue damage to the hind paws. Five measurements were performed on each hind paw spaced 5 min apart to determine mean paw withdrawal latency. Western blots The animals were killed by decapitation at 0 (normal control), 2, 4, 8 and 16 h and 1, 2 and 4 days after injection of saline or formalin (4%, 100 Wl) into the dorsal side of a hind paw. Lumbar enlargement segments of the spinal cord were removed, quickly frozen in liquid nitrogen and stored at 380‡C for later use. Frozen tissues were homogenized in the homogenization bu¡er (50 mM Tris^HCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM phenylmethylsulfonyl £uoride, 1 WM leupeptin, 2 WM pepstatin A, 0.1% 2-mercaptoethanol). The crude homogenate was centrifuged at 4‡C for 15 min at 1000Ug. The supernatants (100 Wg) were heated for 5 min at 90‡C and then loaded onto 4% stacking/7.5% separating sodium dodecyl sulfate^polyacrylamide gels. The proteins were electrophoretically transferred onto nitrocellulose membrane and blocked with 2% non-fat dry milk and subsequently incubated for 1 h with polyclonal rabbit anti-K1 and L1 subunits of soluble GC antibody (1:1000, Cayman Chemical Corp, MI, USA) or with polyclonal rabbit anti-tubulin antibody (1:1000, Santa Cruz Biotechnology, Santa Cruz, CA, USA). Normal lungs were used as a positive control for soluble GC. The proteins were detected using horseradish peroxidase-conjugated anti-rabbit secondary antibodies and visualized using chemiluminescence reagents provided with the enhanced chemiluminescence kit (Amersham Pharmacia Biotech, Piscataway, NJ, USA) and exposure to ¢lm. The intensity of blots was quanti¢ed with densitometry. In some experiments, rats ¢rst were measured for baseline paw withdrawal latencies as described above, and then randomly assigned into four groups. In group 1 (as control, n = 5), 2 ml of 0.9% saline or 2 ml of peanut oil was administered intraperitoneally (i.p.) 30 min prior to the subcutaneous injection of saline (0.9%, 100 Wl) on the ¢rst day. The same doses of saline or peanut oil were given i.p. once, respectively, on the second and third days. In group 2 (n = 5), 2 ml of 0.9% saline or 2 ml of peanut oil was administered i.p. 30 min prior to the subcutaneous injection of formalin (4%, 100 Wl) on the ¢rst day. The same doses of saline or peanut oil were given i.p. once, respectively, on the second and third days. In groups 3 (n = 5) and 4 (n = 5), dizocilpine maleate [MK-801; an N-methyl-Daspartate (NMDA) receptor antagonist, 2 mg/kg in 2 ml saline, RBI, MA, USA] and 7-NI (a selective neuronal NO synthase inhibitor, Alexis Biochemicals, CA, USA) (100 mg/kg in 2 ml peanut oil) were administered i.p., respectively, 30 min prior to the subcutaneous injection of formalin (4%, 0.1 ml) on the ¢rst day. The same doses of MK-801 and 7-NI were given i.p. once, respectively, on the second and third days. The doses of MK801 and 7-NI used above were determined according to previous

NSC 5527 24-5-02

Soluble guanylate cyclase and in£ammatory hyperalgesia

441

studies (Pajewski et al., 1996; Ji and Rupp, 1997; Tao et al., 2000). On the fourth day, all rats from these four groups were killed after the measurement of paw withdrawal latencies as described above. Lumbar enlargement segments were removed and western blot analysis was employed in the same manner as described above. Speci¢city controls for anti-K1 and L1 subunits of soluble GC anti-serum included the immunoadsorption of this anti-serum with excess of K1 subunit peptide of soluble GC (Cayman Chemical) or of L1 subunit peptide of soluble GC (Cayman Chemical), the substitution of normal rabbit serum for this anti-serum and the omission of this anti-serum. Immunohistochemistry The animals were deeply anesthetized with pentobarbital sodium (60 mg/kg i.p.) and perfused with 4% paraformaldehyde in phosphate bu¡er (0.1 M, pH 7.4) 1 h after the formalin behavioral test was done as described above. The lumbar spinal cord was removed, post-¢xed in the same ¢xative solution for 4 h, cryoprotected by immersing in 30% sucrose overnight at 4‡C and frozen-sectioned at 30 Wm. Sections were processed for immunohistochemistry with the use of the conventional avidin^ biotin complex method (Hsu et al., 1980). In brief, sections were incubated in polyclonal rabbit anti-Fos serum (1:4000; Santa Cruz Biotechnology, Santa Cruz, CA, USA) diluted in 0.01 M phosphate-bu¡ered saline (pH 7.4) containing 3% normal goat serum and 0.25% Triton X-100 for 48 h at 4‡C. The speci¢city of the primary antibody has been shown in previous studies (Tao et al., 1997, 2000; Tao and Zhao, 1997). The sections were then incubated in biotinylated goat anti-rabbit IgG (1:200, Vector Laboratories, Burlingame, CA, USA) for 1 h at 37‡C followed by avidin^biotin-peroxidase complex (1:100; Vector Laboratories) for 1 h at 37‡C. The immune reaction product was visualized by catalysis of 3,3-diaminobenzidine by horseradish peroxidase in the presence of 0.01% H2 O2 . Five sections were randomly taken from the ¢fth lumbar segment of the spinal cord of each rat and were examined with a microscope linked to a computer-assisted image-processing system through a video camera. To quantify the anatomical results, the spinal cord was divided into four regions: (1) the super¢cial laminae (laminae I and II), (2) the nucleus proprius (laminae III and IV), (3) the neck of the dorsal horn (laminae V and VI), and (4) the ventral horn (laminae VII, VIII and IX) and the region around the central canal (lamina X). The number of Fos-positive neurons in the spinal cord was independently counted by two experimenters whose results were consistent within 10%. Statistical analysis The results from the immunohistochemistry, behavioral tests and western blot were statistically assessed by an analysis of variance. Intergroup di¡erences were analyzed by the Newman^Keuls test. Data were assessed as mean W S.E.M. Signi¢cance was set at P 6 0.05.

RESULTS

Expression and distribution of soluble GC in the spinal cord The anti-soluble GC antibody used in the present study recognized both the K1 and L1 subunits of soluble GC. Immunoblotting analysis revealed the expression of K1 subunit of soluble GC in the spinal cord (Fig. 1A). The K1 subunit was distributed mainly in the dorsal horn of the spinal cord (Fig. 1B). In contrast, the L1 subunit of soluble GC was not detected or very weakly detected in the spinal cord (Fig. 1A). Since co-expression of K and

Fig. 1. Expression of soluble GC (sGC) in the spinal cord. (A) Only the K1 subunit of soluble GC was detected in the spinal cord, whereas both K1 and L1 subunits of soluble GC were seen in normal lung tissue (as a positive control). (B) High level of expression of K1 subunit of soluble GC was observed in the dorsal horn (DH) and low level in the ventral horn (VH). L: lung. SC: Spinal cord. Tubulin was used as a loading control.

L subunits is necessary for activity of soluble GC (Furuyama et al., 1993; Hobbs and Ignarro, 1996), at issue in the present results is whether this antibody is speci¢c against the L1 subunit of soluble GC or not. We performed three experiments. First, in parallel with the staining of spinal cord, the tissue from normal lung (as a positive control) expressed both K1 and L1 subunits of soluble GC (Fig. 1A), which is consistent with previous studies (Bloch et al., 1997; Li et al., 1999). Second, the staining for the K1 and L1 subunits in the lung and for the K1 subunit in the spinal cord was abolished when the anti-serum was immunoadsorbed with both K1 and L1 peptides. However, the staining for the L1 subunit in the lung was not abolished by immunoadsorption with the K1 peptide (data not shown). Third, the staining for the K1 and L1 subunits could not be seen when the antiserum was omitted or substituted by normal rabbit serum (data not shown). We believe that the K1 subunit might be co-expressed with other L subunits (L2 or L3 ) in the spinal cord, which remains to be further con¢rmed when a speci¢c L2 or L3 antibody becomes available. E¡ect of a selective soluble GC inhibitor on formalin-induced pain behavior Pretreatment with a selective soluble GC inhibitor, ODQ, produced signi¢cant and dose-dependent inhibition of the formalin-induced pain behavior (Table 1). Intrathecal ODQ at 10 Wg reduced the number of £inches and shakes by 48.36% (P 6 0.05) and 57.23% (P 6 0.01) in the phasic and tonic periods of the formalin test, respectively. ODQ given at a dose of 5 Wg dramatically suppressed the formalin-induced behavior by 23.75% (P 6 0.05) in the tonic period, but had no e¡ect in the phasic period. A low dose of ODQ (1 Wg) did not a¡ect the formalin response in either the phasic or tonic periods of the formalin test. The rats treated only with intrathecal ODQ (10 Wg) did not show any motor de¢ciency or abnormal locomotion.

NSC 5527 24-5-02

442

Y.-X. Tao and R. A. Johns

Fig. 2. Photomicrographs showing the distribution of Fos-labeled neurons in the dorsal horn of the ¢fth lumbar segment. (A) In intrathecally vehicle-treated animals, Fos-labeled neurons were concentrated in laminae I, II, V and VI. (B) In intrathecally ODQ (10 Wg)-treated animals, the number of Fos-labeled neurons was signi¢cantly decreased in the dorsal horn compared with (A). Scale bar = 200 Wm.

E¡ect of a selective soluble GC inhibitor on formalin-induced c-fos expression in the spinal cord Consistent with previous studies (Tao et al., 1997, 2000; Tao and Zhao, 1997), injection of formalin into a hind paw induced a high amount of c-fos expression in the ipsilateral side but not the contralateral side of the

spinal cord. Most of the Fos-labeled neurons were concentrated in the medial region of laminae I^II and in laminae V^VI (Fig. 2A). There were no Fos-labeled neurons in rats without any treatment or only with the treatment of ODQ. ODQ given intrathecally signi¢cantly attenuated formalin-induced c-fos expression in all laminae except lam-

Table 1. E¡ect of intrathecally administered ODQ (1, 5 and 10 Wg) on formalin-induced nociception in the rat

Phasic Tonic

Vehicle

1 Wg

5 Wg

10 Wg

53.25 W 7.25 503.25 W 16.76

50.75 W 6.29 430.75 W 52.03

40.25 W 6.70 383.75 W 51.65*

27.50 W 10.10* 215.25 W 87.34**

Number of £inches and shakes in the phasic and tonic periods of the formalin test was counted as described in Experimental procedures. Each value is the mean W S.E.M. *P 6 0.05, **P 6 0.01 versus vehicle-treated control group.

NSC 5527 24-5-02

Soluble guanylate cyclase and in£ammatory hyperalgesia

443

Table 2. E¡ect of intrathecally administered ODQ (1, 5 and 10 Wg) on formalin-induced c-fos expression in the rat

Laminae Laminae Laminae Laminae

I^II III^IV V^VI VII^X

Vehicle

1 Wg

5 Wg

10 Wg

26.72 W 0.77 7.92 W 0.50 23.00 W 1.44 6.72 W 0.64

25.27 W 2.64 6.72 W 0.78 21.86 W 1.26 6.27 W 0.43

20.92 W 2.71 6.60 W 0.67 21.78 W 3.11 5.83 W 1.25

15.53 W 1.33* 5.19 W 0.77* 14.24 W 1.15* 5.51 W 0.37

Number of Fos-positive neurons in the spinal cord was counted as described in Experimental procedures. Each value is the mean W S.E.M. *P 6 0.05 for ODQ-treated groups versus vehicle-treated control group.

inae VII^X in the spinal cord (Fig. 2B and Table 2). With a higher dose of ODQ (10 Wg), the mean reduction in number of Fos-labeled neurons per section was 42% in laminae I^II, 34% in III^IV and 38% in laminae V^VI. Two lower doses of ODQ (1 and 5 Wg) failed to produce any signi¢cant change in the amount or distribution of Fos-labeled neurons compared with the control group. Up-regulation of selective soluble GC expression in the spinal cord after formalin injection As shown in Fig. 3, there is a time-dependent change in the expression of the K1 subunit of soluble GC in the spinal cord after formalin injection. Relative to normal rats (0 h), abundant K1 subunit was detected in the tissues from 2-day (n = 4) and 4-day (n = 4) formalin-treated rats (Fig. 3). In contrast, the same levels of K1 subunit protein were detectable in the tissues from 0-h (n = 4), 2-h (n = 4), 4-h (n = 4), 8-h (n = 4), 16-h (n = 4) and 1-day (n = 4) formalin-treated rats (Fig. 3). There were no signi¢cant time-dependent changes in expression of K1 subunit after saline injection (n = 4, each time point, data not shown). Quantitation showed that the K1 subunit levels were 1.55- and 1.94-fold greater in the tissues from 2- and 4-day formalin-treated groups than those in the tissues from 2- and 4-day saline-treated groups, respectively. Statistical analysis showed a signi¢cant di¡erence between the saline- and formalin-treated groups above (Table 3). Tubulin, as a loading control, was detected in the spinal cord, but no signi¢cant di¡erence of tubulin expression was seen among normal control, salinetreated and formalin-treated groups (Fig. 3). E¡ects of NMDA receptor antagonist and selective neuronal NO synthase inhibitor on formalin-induced long-term thermal hyperalgesia and on up-regulation of soluble GC expression in the spinal cord Subcutaneous formalin injection into the dorsal side of

rat hind paw produced long-lasting in£ammation and long-term thermal hyperalgesia. As shown in Table 4, on the fourth day after formalin injection, paw withdrawal latencies in ipsilateral and contralateral sides were signi¢cantly decreased by 38 and 42%, respectively, compared to those before formalin injection. Formalininduced paw edema on the injected side was still seen on the fourth day. The thickness of the injected hind paw was signi¢cantly increased compared with its thickness before formalin injection or the thickness of contralateral hind paw (data not shown). Pretreatment with MK-801 and 7-NI completely abolished reduction in paw withdrawal latencies evoked by formalin on the fourth day (Table 4). However, formalin-produced increase in paw thickness of the injected paw was not a¡ected by MK801 or 7-NI (data not shown). After i.p. pretreatment with vehicle (saline or peanut oil), the tissues from 4-day formalin-treated rats (group 2) displayed more abundant K1 subunit of sGC, when compared with those from 4-day saline-treated rats (group 1) (Fig. 4). Quantitation revealed the tissues from group 2 contained 2.1-fold more K1 -subunit protein than that from group 1. Statistical analysis showed a signi¢cant di¡erence between groups 2 and 1 (Fig. 4). However, following i.p. pretreatment with MK-801, much lower levels of K1 -subunit protein were detected in tissues from 4-day formalin-treated rats (group 3) compared to that in group 2. Quantitation showed that K1 -subunit protein in group 3 was 1.19-fold greater than that in group 1, but the increases were not statistically signi¢cant. In addition, A low level of K1 -subunit protein also was observed in tissues from 4-day formalin-treated rats after i.p. pretreatment with 7-NI (group 4) (Fig. 4). There was no signi¢cant di¡erence in the K1 -subunit proteins between groups 4 and 1, although the tissues from

Table 3. Quantitative change of K1 subunit of soluble GC in the lumbar enlargement segments of the spinal cord from control animals (0 day) and treated animals, 2 and 4 days after saline or formalin injection 0 day (control) 2 days Saline 1 Formalin 1

Fig. 3. Expression of K1 subunit of soluble GC and tubulin in the lumbar enlargement segments of the spinal cord from animals at di¡erent time points after formalin injection. Tubulin was used as a loading control for K1 subunit of soluble GC.

0.97 W 0.11 1.51 W 0.06*#

4 days 1.03 W 0.11 2.00 W 0.20**##

The density of the normal control group was set at 100%. The relative density value for each saline- or formalin-treated group was determined by dividing the optical density value from the treated group by one from the normal control group. Each value is the mean W S.E.M. # P 6 0.05, ## P 6 0.01 versus control group. *P 6 0.05, **P 6 0.01 versus corresponding saline-treated groups.

NSC 5527 24-5-02

444

Y.-X. Tao and R. A. Johns Table 4. E¡ects of intraperitoneally administered MK-801 and 7-NI on formalin-induced long-term thermal hyperalgesia in the rat Formalin

Ipsilateral Contralateral

MK-801+formalin

7-NI+formalin

baseline

4 days

baseline

4 days

baseline

4 days

10.19 W 0.91 10.34 W 0.67

6.27 W 0.74** 6.04 W 0.47**

10.40 W 1.27 9.34 W 1.28

11.31 W 0.92 11.04 W 0.58

11.07 W 1.52 9.74 W 1.60

11.97 W 0.81 11.38 W 0.70

Withdrawal latencies (s) of both hind paws to radiant heat were measured as described in Experimental procedures. Formalin (Group 2); MK-801+formalin (Group 3); 7-NI+formalin (Group 4). Each value is the mean W S.E.M. **P 6 0.01 versus corresponding baseline values (before formalin injection).

group 4 contained 1.22-fold more K1 -subunit protein than that from group 1.

DISCUSSION

The regional expression of soluble GC in the mammalian brain and peripheral tissues has been extensively investigated using a variety of experimental approaches (Furuyama et al., 1993). To our knowledge, however, its expression and distribution in the spinal cord are not fully examined. With the use of immunoassay, a high level of cGMP (the main product of soluble GC) was detected in the dorsal horn of lumbosacral segments of spinal cord, whereas the level of cGMP was much lower in the ventral horn (Pavel et al., 2000). Furthermore, with the use of immunocytochemistry, NO-mediated cGMP synthesis was observed in neuronal cells and ¢bers in all spinal cord laminae, particularly, in the dorsal horn of spinal cord (Vles et al., 2000). The indirect evidence above suggests that soluble GC may exist in the neurons of the spinal cord and may be expressed mainly in the dorsal horn. In the present study, we provided direct evidence that showed that the K1 subunit of soluble GC was richly expressed in the spinal cord, while L1 subunit was absent or present at extremely lower levels. More important, K1 subunit was found to be distributed

Fig. 4. E¡ects of intraperitoneally administered MK-801 and 7-NI on the expression of the K1 subunit of soluble GC (sGC) in the lumbar enlargement segments of the spinal cord from 4-day formalin-treated animals. The upper panel depicts a representative western blot. Lane I: group 1; lane II: group 2; lane III: group 3; lane IV: group 4. The treatment regimen for each group is described in Experimental procedures. The lower panel is the statistical summary of the densitometric analysis expressed relative to normal control groups. The data were presented as mean W S.E.M. **P 6 0.01 for group 2 versus group 1.

mainly in the dorsal horn of spinal cord. Although the detailed neuronal structural localization of soluble GC remains unclear (since the soluble GC antibody we used is not recommended for immunostaining), speci¢c expression and distribution of soluble GC in the spinal cord demonstrated in the present study suggest that soluble GC may be involved in the processing of nociceptive information at the spinal cord level. Several lines of previous evidence have demonstrated a role of soluble GC in excitatory amino acid-induced acute thermal hyperalgesia. Intrathecal injection of methylene blue (an inhibitor of soluble GC) signi¢cantly blocked reduction in tail-£ick latencies produced by NMDA (Meller et al., 1992, 1996). Similarly, ODQ (a potent and selective inhibitor of soluble GC) co-administered with glutamate dose-dependently antagonized the glutamate-induced hyperalgesia (Ferreira et al., 1999). These data indicate the involvement of soluble GC in hyperalgesia triggered via NMDA receptor activation. The present behavioral experiments showed that pain responses evoked by formalin in the tonic period, which is driven largely by tissue injury-induced in£ammation and is generally considered as a model of hyperalgesia (Wheeler-Aceto et al., 1990; Tjolsen et al., 1992), were decreased signi¢cantly and dose-dependently by intrathecal administration of ODQ. This ¢rstly demonstrates that soluble GC in the spinal cord can be activated in in£ammatory hyperalgesia and is involved in mediating the behavioral pain response to peripheral in£ammatory stimuli. Interestingly, ODQ at the highest dose (10 Wg) also attenuated pain response in the phasic period. The phasic period of the formalin test is believed to be due to direct activation of small diameter primary a¡erents (Wheeler-Aceto et al., 1990; Tjolsen et al., 1992). Previous work has demonstrated the presence of soluble GC in the dorsal root ganglion (Kummer et al., 1996). It is very possible that soluble GC in the dorsal root ganglion is related to the modulation of noxious stimulation transmission. ODQ at the highest dose injected intrathecally might di¡use into the dorsal root ganglion and directly inhibit formalin-induced activity in a¡erent ¢bers. In our current study, the reduction of pain behaviors following the administration of ODQ was accompanied by a dramatic decrease in the number of Fos-positive neurons in the dorsal horn. C-fos, a proto-oncogene, has been widely used as a marker of functional activity of neurons, especially in the study of the response of spinal cord neurons to noxious stimulation (Hunt et al., 1987; Bullitt, 1990). The present study indirectly demonstrates that noxious stimulation-

NSC 5527 24-5-02

Soluble guanylate cyclase and in£ammatory hyperalgesia

induced neuronal activity in the spinal cord can be reduced by ODQ, an outcome which is consistent with the behavioral results above. Taken together, our results suggest that soluble GC may play an important role in the processing of in£ammatory pain at the spinal cord level. Recently, it was found that subcutaneous formalin injection into the rat’s hind paw not only produces the typical two phases of nociceptive behavior but also results in long-lasting in£ammation and long-term hyperalgesia. Fu et al. (2001) reported that the magnitude of paw edema increased rapidly over the ¢rst hour after formalin injection, remained elevated for 10 days and returned to a preinjection appearance approximately 4 weeks after injection. Importantly, they found that hyperalgesic responses to thermal and mechanical stimulation were induced on the opposite surface of the injected hind paw as well as in the contralateral noninjected hind paw. These hyperalgesic responses were observed 2 h after formalin injection, enhanced 1 to 3 days after injection and lasted 3 to 4 weeks. The results indicated that formalin-produced long-lasting in£ammation might result in long-term central sensitization in the spinal cord. However, possible mechanisms for these changes are not clear. Neuronal NO synthase expression was up-regulated in the spinal dorsal horn 1 to 4 days after formalin injection (Herdegen et al., 1994; Lam et al., 1996). In the present study, the expression of K1 subunit of soluble GC in the spinal cord was signi¢cantly increased 2^4 days after formalin injection. Consistent with previous work (Fu et al., 2001), our behavioral study also showed that paw withdrawal latencies to radiant heat were markedly decreased on the ipsilateral and contralateral sides on the fourth day compared with those before formalin injection into the dorsal side of the hind paw. The parallel changes between pain behaviors and the expression in neuronal NO synthase and soluble GC suggest a contribution of NO and its receptor to the secondary hyperalgesia caused by formalin. Obviously, the increase in K1 subunit of soluble GC in the present study is delayed with respect to that in neuronal NO synthase, suggesting that the activation and upregulation of soluble GC expression might be due to the increase in NO rather than a direct e¡ect of formalin. Indeed, we pretreated the animals with a selective neuronal NO synthase inhibitor, 7-NI, before formalin injection. It was found that 7-NI not only completely blocked formalin-induced reduction in paw withdrawal latencies but also suppressed formalin-evoked up-regulation of the

445

K1 subunit of soluble GC expression in the spinal cord. Moreover, an NMDA receptor antagonist, MK-801, had similar e¡ects. This indicates that formalin-produced long-lasting in£ammation may activate and up-regulate soluble GC expression via an NMDA receptor-NO signaling pathway in the spinal cord. It is well known that prolonged noxious stimulation, such as chronic in£ammation and nerve injury, produces central reorganization of the nociceptive processing system (Malmberg et al., 1997; Go¡ et al., 1998; Honore et al., 2000). The present results indicate that formalin-produced long-lasting in£ammation may also produce similar neurochemical changes in the spinal cord. Our recent behavioral results revealed that inhibition of soluble GC abolished formalin-induced secondary thermal hyperalgesia (data not shown). Thus, it is very likely that the activation and upregulation of soluble GC in the spinal cord have important implications in the central mechanisms of formalinproduced in£ammatory hyperalgesia.

CONCLUSIONS

In the present study, we have found abundant expression of soluble GC K1 subunit but not L1 subunit in the dorsal horn of spinal cord. The inhibition of soluble GC signi¢cantly blocked formalin-induced pain behaviors and c-fos expression as a marker of nociceptive neuronal activity in the spinal cord. Furthermore, we demonstrated that formalin-produced long-lasting in£ammation up-regulated the K1 subunit of soluble GC in the spinal cord. Pre-treatment with NMDA receptor antagonist and neuronal NO synthase inhibitor not only abolished the formalin-induced secondary thermal hyperalgesia but also suppressed formalin-produced increase of K1 subunit of soluble GC in the spinal cord. Our data indicate that peripheral in£ammation not only initially activates but also later up-regulates soluble GC expression via the NMDA receptor-NO signaling pathway, suggesting that soluble GC may play an important role in the central sensitization induced by formalin-produced peripheral in£ammation.

Acknowledgements,This work is supported by NIH Grants RO1 GM49111 and RO1 HL39706. The authors thank Mrs. Claire F. Levine for her assistance in western blot analysis and Mrs. Fengying Wang for her help in Fos immunohistochemistry.

REFERENCES

Bloch, K.D., Filippov, G., Sanchez, L.S., Nakane, M., de la Monte, S.M., 1997. Pulmonary soluble guanylate cyclase, a nitric oxide receptor, is increased during the perinatal period. Am. J. Physiol. 272, L400^L406. Bredt, D.S., Snyder, S.H., 1992. Nitric oxide, a novel neuronal messenger. Neuron 8, 3^11. Bullitt, E., 1990. Expression of c-fos-like protein as a marker for neuronal activity following noxious stimulation in the rat. J. Comp. Neurol. 296, 517^530. Chen, S.R., Eisenach, J.C., Pan, H.L., 2000. Intrathecal S-nitroso-N-acetylpenicillamine and l-cysteine attenuate nerve injury-induced allodynia through noradrenergic activation in rats. Neuroscience 101, 759^765. Dun, N.J., Dun, S.L., Wu, S.Y., Forstermann, U., Schmidt, H.H., Tseng, L.F., 1993. Nitric oxide synthase immunoreactivity in the rat, mouse, cat and squirrel monkey spinal cord. Neuroscience 54, 845^857.

NSC 5527 24-5-02

446

Y.-X. Tao and R. A. Johns

Ferreira, J., Santos, A.R.S., Calixto, J.B., 1999. The role of systemic, spinal and supraspinal L-arginine-nitric oxide-cGMP pathway in thermal hyperalgesia caused by intrathecal injection of glutamate in mice. Neuropharmacology 38, 835^842. Fu, K.-Y., Light, A.R., Maixner, W., 2001. Long-lasting in£ammation and long-term hyperalgesia after subcutaneous formalin injection into the rat hind paw. J. Pain 2, 2^11. Furuyama, T., Inagaki, S., Takagi, H., 1993. Localizations of K1 and L1 subunits of soluble guanylate cyclase in the rat brain. Mol. Brain Res. 20, 335^344. Garthwaite, J., Charles, S.L., Chess-Williams, R., 1988. Endothelium-derived relaxing factor release on activation of NMDA receptors suggests role as intercellular messenger in the brain. Nature 336, 385^388. Go¡, J.R., Burkey, A.R., Go¡, D.J., Jasmin, L., 1998. Reorganization of the spinal dorsal horn in models of chronic pain: correlation with behaviour. Neuroscience 82, 559^574. Hargreaves, K., Dubner, N., Brown, F., Flores, C., Joris, J., 1988. A new and sensitive method for measuring thermal nociception in cutaneous hyperalgesia. Pain 32, 77^88. Herdegen, T., Rudiger, S., Mayer, B., Bravo, R., Zimmermann, M., 1994. Expression of nitric oxide synthase and colocalisation with Jun, Fos and Krox transcription factors in spinal cord neurons following noxious stimulation of the rat hindpaw. Mol. Brain Res. 22, 245^258. Hobbs, A.J., Ignarro, L.J., 1996. Nitric oxide-cyclic GMP signal transduction system. Methods Enzymol. 269, 134^148. Honore, P., Rogers, S.D., Schwei, M.J., Salak-Johnson, J.L., Luger, N.M., Sabino, M.C., Clohisy, D.R., Mantyh, P.W., 2000. Murine models of in£ammatory, neuropathic and cancer pain each generates a unique set of neurochemical changes in the spinal cord and sensory neurons. Neuroscience 98, 585^598. Hsu, S.-M., Raine, L., Fanger, H.A., 1980. The use of avidin^biotin complex (ABC) in immunoperoxidase techniques : a comparison between ABC and unlabeled antibody (PAP) procedures. J. Histochem. Cytochem. 29, 577^580. Hunt, S.P., Pini, A., Evan, G., 1987. Induction of c-fos-like protein in spinal cord neurons following sensory stimulation. Nature 328, 632^634. Inoue, T., Mashimo, T., Shibuta, S., Yoshiya, I., 1997. Intrathecal administration of a new nitric oxide donor, NOC-18, produces acute thermal hyperalgesia in the rat. J. Neurol. Sci. 153, 1^7. Ji, R.R., Rupp, F., 1997. Phosphorylation of transcription factor CREB in rat spinal cord after formalin-induced hyperalgesia: relationship to c-fos induction. J. Neurosci. 17, 1776^1785. Knowles, R.G., Palacios, M., Palmar, R.M.J., Moncada, S., 1989. Formation of nitric oxide from L-arginine in the central nervous system: a transduction mechanism for stimulation of the soluble guanylate cyclase. Proc. Natl. Acad. Sci. USA 86, 5159^5162. Kummer, W., Behrends, S., Schwarzlmuller, T., Fischer, A., Koesling, D., 1996. Subunits of soluble guanylyl cyclase in rat and guinea-pig sensory ganglion. Brain Res. 721, 191^195. Lam, H.H., Hanley, D.F., Trapp, B.D., Saito, S., Raja, S., Dawson, T.M., Yamaguchi, H., 1996. Induction of spinal cord neuronal nitric oxide synthase (NOS) after formalin injection in the rat hind paw. Neurosci. Lett. 210, 201^204. Li, D., Zhou, N., Johns, R.A., 1999. Soluble guanylate cyclase gene expression and localization in rat lung after exposure to hypoxia. Am. J. Physiol. 277, L841^L847. Malmberg, A.B., Chen, C., Tonegawa, S., Basbaum, A.I., 1997. Preserved acute pain and reduced neuropathic pain in mice lacking PKCQ. Nature 278, 279^283. Malmberg, A.B., Yaksh, T.L., 1993. Spinal nitric oxide synthesis inhibition blocks NMDA-induced thermal hyperalgesia and produces antinociception in the formalin test in rats. Pain 54, 291^300. Meller, S.T., Dykstra, C., Gebhart, G.F., 1992. Production of endogenous nitric oxide and activation of soluble guanylate cyclase are required for N-methyl-D-aspartate-produced facilitation of the nociceptive tail-£ick re£ex. Eur. J. Pharmacol. 214, 93^96. Meller, S.T., Dykstra, C., Gebhart, G.F., 1996. Acute thermal hyperalgesia in the rat is produced by activation of N-methyl-D-aspartate receptors and protein kinase C and production of nitric oxide. Neuroscience 50, 327^335. Moore, P.K., Babbedge, R.C., Wallace, P., Ga¡en, Z.A., Hart, S.L., 1993. 7-Nitroindazole, an inhibitor of nitric oxide synthase, exhibits antinociceptive activity in the mouse without increasing blood pressure. Br. J. Pharmacol. 108, 296^297. Okuda, K., Sakurada, C., Takahashi, M., Yamada, T., Sakurada, T., 2001. Characterization of nociceptive responses and spinal releases of nitric oxide metabolites and glutamate evoked by di¡erent concentrations of formalin in rats. Pain 92, 107^115. Pajewski, T.N., Difazio, C.A., Moscicki, J.C., Johns, R.A., 1996. Nitric oxide synthase inhibitor, 7-nitroindazole and nitroG -L-arginine methyl ester, dose dependently reduce the threshold for iso£urane anesthesia. Anesthesiology 85, 1111^1119. Pavel, J., Lukacova, N., Marsala, J., 2000. Regional changes of cyclic 3P,5P-guanosine monophosphate in the spinal cord of the rabbit following brief repeated ischemic insults. Neurochem. Res. 25, 1131^1137. Roche, A.K., Cook, M., Wilcox, G.L., Kajander, K.C., 1996. A nitric oxide synthesis inhibitor (L-NAME) reduces licking behavior and Foslabeling in the spinal cord of rats during formalin-induced in£ammation. Pain 66, 331^341. Saito, S., Kidd, G.J., Trapp, B.D., Dawson, T.M., Bredt, D.S., Wilson, D.A., Traystman, R.J., Snyder, S.H., Hanley, D.F., 1994. Rat spinal cord neurons contain nitric oxide synthase. Neuroscience 59, 447^456. Shibuta, S., Mashimo, T., Zhang, P., Ohara, A., Yoshiya, I., 1996. A new nitric oxide donor, NOC-18, exhibits a nociceptive e¡ect in the rat formalin model. J. Neurol. Sci. 141, 1^5. Tao, Y.-X., Hassan, A., Haddad, E., Johns, R.A., 2000. Expression and action of cyclic GMP-dependent protein kinase IK in in£ammatory hyperalgesia in rat spinal cord. Neuroscience 95, 525^533. Tao, Y.-X., Johns, R.A., 2000. Activation of cGMP-dependent protein kinase IK is required for N-methyl-D-aspartate- or nitric oxide-produced spinal thermal hyperalgesia. Eur. J. Pharmacol. 392, 141^145. Tao, Y.-X., Wei, F., Zhao, Z.-Q., 1997. A contribution of neurokinin-1 receptor to formalin-induced c-fos expression in the rat spinal dorsal horn. Neurosci. Lett. 221, 105^108. Tao, Y.-X., Zhao, Z.-Q., 1997. Ultrastructure of Fos-labeled neurons relating to nociceptive primary a¡erent and substance P terminals in rat spinal super¢cial laminae. Neuropeptides 31, 327^332. Terenghi, G., Riveros-Moreno, V., Hudson, L.D., Ibrahim, N.B., Polak, J.M., 1993. Immunohistochemistry of nitric oxide synthase demonstrates immunoreactive neurons in spinal cord and dorsal root ganglia of man and rat. J. Neurol. Sci. 118, 34^37. Tjolsen, A., Berge, O.G., Hunskaar, S., Rosland, J.H., Hole, K., 1992. The formalin test: an evaluation of the method. Pain 51, 5^17. Vles, J.S.H., de Louw, A.J., Steinbusch, H., Markerink-van Ittersum, M., Steinbush, H.W.M., Blanco, C.E., Axer, H., Troost, J., de Vente, J., 2000. Localization and age-related changes of nitric oxide- and ANP-mediated cyclic-GMP synthesis in rat cervical spinal cord: an immunocytochemical study. Brain Res. 857, 219^234. Wheeler-Aceto, H., Porreca, F., Cowan, A., 1990. The rat paw formalin test: comparison of noxious agents. Pain 40, 229^238. (Accepted 16 January 2002)

NSC 5527 24-5-02