Downloaded from hyperalgesia through spinal …jpet.aspetjournals.org/content/jpet/early/2004/06/02/jpet.104... · Dialysis experiments were conducted in conscious, unrestrained rats

Koetzner et al JPET # 69484 1

Intrathecal protease–activated receptor stimulation produces thermal

hyperalgesia through spinal cyclooxygenase activity

Lee Koetzner1,2, Joshua A. Gregory1 and Tony L. Yaksh1,3

Laboratory of origin:

Anesthesiology Research Laboratory 0818

University of California–San Diego

9500 Gilman Drive

La Jolla, CA 92093–0818

JPET Fast Forward. Published on June 2, 2004 as DOI:10.1124/jpet.104.069484

Copyright 2004 by the American Society for Pharmacology and Experimental Therapeutics.

This article has not been copyedited and formatted. The final version may differ from this version.JPET Fast Forward. Published on June 2, 2004 as DOI: 10.1124/jpet.104.069484

at ASPE

T Journals on M

ay 4, 2018jpet.aspetjournals.org

Dow

nloaded from

http://jpet.aspetjournals.org/


Running title: Spinal PARs, hyperalgesia and PGE2 release

Corresponding author for revisions: Lee Koetzner, PhD

Department of Neuropharmacology

Purdue Pharma, LP

6 Cedar Brook Drive

Cranbury, NJ 08512

Voice (609) 409-4321 alternate: (609) 888-4713

Fax (609) 409-6922

Electronic mail [email protected] alternate: [email protected]

Text pages: 13

Tables: none

Figures: 8

References: 40

Words in abstract: 229

Words in introduction: 510

Words in discussion: 1448

Nonstandard abbreviations: AOC, area over curve; AUFS, absorbance units for full-scale deflection; C18,

octadecylsilane; COX, cyclooxygenase; cPLA2, cytoplasmic phospholipase A2; HPLC, high–performance

liquid chromatography; MAP, microtubule–associated protein; mRNA, messenger ribonucleic acid; MSP,

myelencephalon-specific protease; PAR, protease–activated receptor; PE, polyethylene; PGE1-3, pros-

taglandins E1-3; SC58,125, 5-(4-chlorophenyl)-1-(4-methoxyphenyl)-3-(trifluoromethyl)-1H-pyrazole;SC

58,560, 5-(4-fluorophenyl)-1-[4-(methylsulfonyl)phenyl]-3-(trifluoromethyl)-1H-pyrazole

Section: Neuropharmacology


at ASPE

T Journals on M


Dow

nloaded from



ABSTRACT

Activation of protease–activated receptors (PARs) in non–neural tissue results in prostaglandin produc-

tion. Since PARs are found in the spinal cord and increased prostaglandin release in the spinal cord causes

thermal hyperalgesia, we hypothesized that activation of these spinal PARs would stimulate prostaglandin

production and cause a cyclooxygenase–dependent thermal hyperalgesia. PARs were activated using ei-

ther thrombin or peptide agonists derived from the four PAR subtypes, delivered to the lumbar spinal

cord. Dialysis experiments were conducted in conscious, unrestrained rats using loop microdialysis

probes placed in the lumbar intrathecal space. Intrathecal thrombin stimulated release of prostaglandin E2

(PGE2) but not aspartate or glutamate. Intrathecal delivery of the PAR 1–derived peptide SFLLRN–NH2

and the PAR 2–derived peptide SLIGRL both stimulated PGE2 release; PAR 3–derived TFRGAP and

PAR 4–derived GYPGQV were inactive. Intrathecal thrombin had no effect upon formalin induced

flinching or tactile sensitivity but resulted in a thermal hyperalgesia. Intrathecal SFLLRN–NH2 and

SLIGRL both produced thermal hyperalgesia. Consistent with their effects on spinal PGE2, hyperalgesia

from these peptides was blocked by pretreatment with the cyclooxygenase inhibitor ibuprofen. SLIGRL–

induced hyperalgesia was also blocked by the selective inhibitors SC 58,560 (COX1) and SC 58,125

(COX 2). These data indicate that activation of spinal PAR 2 and possibly PAR 1 results in the stimula-

tion of the spinal cyclooxygenase cascade and a prostaglandin–dependent thermal hyperalgesia.


at ASPE

T Journals on M


Dow

nloaded from



Protease–activated receptors (PARs) are a family of guanine nucleotide binding protein coupled

receptors that are activated by enzymatic activity of serine proteases such as thrombin (cf. (Macfarlane et

al., 2001; O'Brien et al., 2001). The native form of each receptor includes consensus sites for proteolytic

cleavage in the extracellular portion of the amino terminus. After cleavage, the remaining amino terminus

activates the receptor. A comparable activation is achieved by synthetic peptides with corresponding se-

quences (e.g., (Chen et al., 1994). These receptor–derived ligand peptides show a general homology

across species (Macfarlane et al., 2001; O'Brien et al., 2001). Once activated, these receptors typically

increase intracellular calcium concentrations (Macfarlane et al., 2001). Such increases in calcium initiate

many signaling events, including activation of the phospholipase A2–cyclooxygenase cascade (Svensson

and Yaksh, 2002). In some cell types, PAR activation has indeed been shown to stimulate prostaglandin

production (e.g., (Neufeld and Majerus, 1983).

While much work has focused on the peripheral distribution and function of PARs, other work

indicates the presence of these receptors in the central nervous system. Reversible binding of radiolabeled

human α–thrombin has been shown in human brain and spinal cord (McKinney et al., 1983). Thrombin

binding to cultured murine spinal ventral horn cells has been demonstrated using histological techniques

(Means and Anderson, 1986). In situ hybridization and Northern blot analysis revealed substantial expres-

sion of PAR 1 mRNA in the rat brain as well as spinal neurons, dorsal root ganglia and the sciatic nerve

(Niclou et al., 1998). An in situ hybridization study found PAR 1 mRNA in the spinal gray matter

(Weinstein et al., 1995). PAR 2 has also been found in neurons, with PAR 1 and PAR 2 mRNA found in

cultured hippocampal neurons (Smith-Swintosky et al., 1997). PAR 1 and PAR 2 immunoreactivity have

both been demonstrated in cells of the dorsal root ganglion (Steinhoff et al., 2000). PAR 3 mRNA was

detected by Northern blot analysis of human spinal cord tissue (Ishihara et al., 1997). Recently, RT–PCR

and in situ hybridization demonstrated all four PAR subtypes in cultured hippocampal slices and explants

(Striggow et al., 2001). Finally, RT–PCR and immunohistochemistry demonstrated all four PARs in pri-

mary rat astrocyte cultures (Wang et al., 2002a).


at ASPE

T Journals on M


Dow

nloaded from



The presence of PARs in the central nervous system and the ability of PARs to stimulate pros-

taglandin synthesis led us to consider whether PAR activation in the spinal cord would also lead to pros-

taglandin release and nociception. Spinal prostaglandin release is a common feature of many nociceptive

stimuli, and inhibition of cyclooxygenase activity blocks their action in animal models of nociception (see

(Svensson and Yaksh, 2002) for review). Administration of PGE2 produces both hyperalgesia and allo-

dynia in animals (e.g., (Minami et al., 1994). Accordingly, we hypothesized that activating spinal PARs

would produce prostaglandin E2 release. Moreover, those PARs that evoked prostaglandin release would

also produce cyclooxygenase–dependent thermal hyperalgesia. Both thrombin and peptides were used to

activate PARs. Spinal transmitter release was measured using microdialysis. Nociception was assessed

using thermal testing, von Frey filaments and formalin–induced paw flinching.

METHODS

All studies described herein were carried out under protocols approved by the Institutional Ani-

mal Care and Use Committee of the University of California, San Diego.

Preparation

Intrathecal catheters were prepared from PE–10 polyethylene tubing (0.6 mm outside diameter,

0.3 mm inside diameter; Becton–Dickinson, Sparks, MD). Tubing was tied in a half–hitch of approxi-

mately 3 mm diameter and the knot was coated with dental acrylic (Dentsply, Milford, DE). After the

acrylic dried, the intrathecal side of the tubing was stretched to reduce its diameter. The final product was

trimmed to a length of 85 mm.

The intrathecal portion of the dialysis probe consisted of a tubular 4 cm cellulose dialysis fiber

(Filtral AN69HF, Cobe Laboratories, Denver, CO) bent in half and connected by its ends to 7 cm of triple

lumen polyethylene tubing (Spectranetics, Colorado Springs, CO) using fused silica tubing (Polymicro

Technologies, Phoenix, AZ) as stents with methacrylate adhesive (LocTite North America, Rocky Hill,

CT). Fine gauge wire (A–M Systems, Carlsborg, WA) inside the dialysis fiber prevented crimping. The


at ASPE

T Journals on M


Dow

nloaded from



center lumen of the polyethylene tubing was used as an injection line. Each of the three lumens was heat–

fused to short lengths of single lumen polyethylene tubing.

Male Holtzman Sprague–Dawley rats (Harlan, Madison, WI) weighing about 300–350 g at the

beginning of the experiment were prepared with lumbar intrathecal injection catheters or dialysis probes.

Rats were anesthetized with isoflurane (Abbott Laboratories, North Chicago, IL) in a flow–through

chamber. After induction, isoflurane (2-3% in air) was delivered through a nose cone. After shaving and

cleaning the scalp, the rat was supported on a pad and secured via ear bars and the nose cone. Sharp dis-

section was used to separate muscles from the occipital bone and expose the atlantooccipital membrane.

A stab blade was used to make an incision in the dura mater. The dialysis probe or intrathecal catheter

was then inserted and advanced caudad. Connections were tunneled under the skin to exit over the frontal

bones; the free ends were plugged with short lengths of wire (V. Mueller, McGaw Park, IL). Catheters

and the injection line of dialysis probes were flushed with saline; the dialysis circuit was flushed with di-

lute penicillin–streptomycin solution (GIBCO/Invitrogen Life Technologies, Carlsbad, CA). After the

incision was closed, rats were placed in a warm chamber to recover from anesthesia. On demonstrating

intact neurological function, rats were returned to the vivarium.

Spinal microdialysis

Dialysis experiments were conducted in conscious rats after two or three days’ recovery from

surgery. Previous data suggested that this recovery period resulted in relatively consistent, detectable

dialysate concentrations of analytes (Malmberg and Yaksh, 1995). Tubing from the rat was connected to a

two channel fluid swivel (Eicom, Kyoto, Japan; Bioanalytical Systems, West Lafayette, IN; Instech Labo-

ratories, Plymouth Meeting, PA). The effluent was routed to a programmable, refrigerated fraction collec-

tor (Eicom). Syringe drivers (Harvard Apparatus, South Natick, MA) delivered artificial cerebrospinal

fluid at 10 µl/minute. The final composition of the fluid was 151 mM sodium, 2.6 mM potassium, 132

mM chloride, 1.3 mM calcium, 0.9 mM magnesium, 2.5 mM phosphate and 21 mM carbonate (all salts,

EM Science, Gibbstown, NJ). The solution was bubbled with 5 % CO2 in air until clarity and filtered for

sterility. Fractions (10 min.) were collected at 4°C and frozen immediately (–20°C for short–term storage,


at ASPE

T Journals on M


Dow

nloaded from



–70°C for long–term storage). Initial experiments and previous studies indicated that analyte levels reach

a steady state within thirty minutes of beginning perfusion (Koetzner, Marsala and Yaksh, unpublished

data; (Malmberg and Yaksh, 1995). Rats were only tested once in order to avoid any possibility of carry-

over effects.

Chemistry: Amino acid measurements

Samples were thawed and internal standard (400 picomoles of methionine sulfone; Sigma, St.

Louis, MO) was immediately added. The samples were dried under vacuum, then reconstituted in metha-

nol–aqueous sodium acetate–triethylamine (all solvents and HPLC reagents, Fisher Chemical, Fair Lawn,

NJ) and dried again. Amines were converted to chromogenic derivatives by conjugation with phenyli-

sothiocyanate (5 min. at room temperature in methanol–water–triethylamine). The reaction mixture was

dried under vacuum and reconstituted in acetonitrile–aqueous sodium phosphate. 30 µl samples of this

mixture were injected for HPLC using a Waters 712 autosampler (Waters Corporation, Milford, MA).

Analytes were separated using a C18 column (MetaChem Polaris (MetaChem Technologies, Torrance,

CA), 5 µm x 25 cm, maintained at 46°C) and gradient elution with a gradient cleanup step (60 min.; Wa-

ters 510 pump, 1 ml/min.), controlled by computer. Peaks were detected by monitoring absorbance at 254

nm (0.05 AUFS, 2 samples/sec.; Agilent 1100 detector, Agilent Technologies, Palo Alto, CA). Ratios of

analyte peak area to internal standard peak area were calculated for all samples; analyte mass was esti-

mated by comparison to a 400 pmol standard solution. Estimates of aspartate, glutamate, serine and

taurine concentrations were obtained from the same chromatographic separation.

Chemistry: Prostaglandin analysis

PGE2 concentrations were assayed by competitive enzyme–linked immunosorbent assay using a

commercially available kit (Assay Designs 90001, Assay Designs, Ann Arbor, MI). This kit uses a mono-

clonal antibody to PGE2 (anchored to the plate via anti–IgG antibody) and PGE2–alkaline phosphatase

conjugate. The PGE2 antibody has 70 % cross–reactivity with PGE1 and 16 % cross–reactivity with PGE3;

no other cross–reactivities exceeding 2 % are known (manufacturer’s specifications). After incubation of

the reagents and sample, activity bound to the solid phase is measured with para–nitrophenylphosphate as


at ASPE

T Journals on M


Dow

nloaded from



absorbance at 450 nm. The amount of analyte in the sample is inferred from displacement of the PGE2–

alkaline phosphatase conjugate and sample values are determined by reference to a standard curve. The

linear detection range is generally 40–5000 picograms per ml sample. Some rats were excluded due to

elevated baseline PGE2 concentrations (>1000 fmol/100 µl fraction) or unmeasurably low PGE2 concen-

trations (<40 fmol/100 µl fraction).

Behavior: Thermal escape latency

Hindpaw withdrawal in response to thermal stimulation was measured using a modification of the

Hargreaves procedure, as previously reported (Dirig et al., 1997). Briefly, rats were allowed to habituate

to a temperature-controlled glass floor. Latency to respond to a heat stimulus, applied to the glass surface

under a hind paw, was measured with an electronic timer. Rats were tested one week after catheter im-

plant. Baseline latencies were measured, and stimulus intensity was adjusted to give latencies between ten

and twelve seconds. Treatments were injected and post–injection latencies were recorded for up to two

hours. Rats were also observed for untoward effects of treatments using a behavior checklist. Repeated

dosage of thrombin at four day intervals did not produce untoward effects. However, repeated administra-

tion of peptides was associated with a broad spectrum of toxicities, with ataxia predominant. Accordingly,

peptide treatments were only tested once in each rat.

Behavior: Tactile allodynia

Tactile allodynia was determined using von Frey filaments presented in a method of limits pat-

tern. The plantar surface of the paw was tested, with a brisk withdrawal of the paw defined as a behav-

ioral response. In addition, an area on the back lateral to the spine at the level of the third lumbar vertebra

was tested, with vocalization or rapid orientation of the head to the filament defined as a behavioral re-

sponse.

Behavior: Formalin–induced flinching

Flinching behavior following the intraplantar injection of 50 µl of 0.5 % formalin (v/v in saline)

was counted using the automated device described by Yaksh and colleagues (Yaksh et al., 2001b). A

metal band was affixed to the paw before injection using methacrylate adhesive. After habituation, the rat


at ASPE

T Journals on M


Dow

nloaded from



was placed on an electromagnetic transducer in a polycarbonate cylinder (approx. 15 cm diameter x 30

cm height). The signal from the transducer was digitized and signals characteristic of flinching behavior

were counted using a proprietary algorithm (Yaksh et al., 2001b).

Treatments

Bovine thrombin (E.C. 3.4.21.5, approx. 75 NIH units/mg; Sigma) was dissolved in saline. Typi-

cally, aliquots were frozen; due to the lability of the enzyme, aliquots were never refrozen or used after

more than two weeks’ storage at –20°C. Peptides were selected to mimic the cryptic ligands of rodent

PARs (PAR 1, SFLLRN–NH2; PAR 2, SLIGRL; PAR 3, TFRGAP and PAR 4, GYPGQV, all from

Bachem, Torrance, CA). For some studies, TFLLR (Bachem) was used as a selective activator of PAR 1

(Blackhart et al., 1996). Peptides and the sodium salt of (+)–ibuprofen (Ethyl Corporation, Orangeburg,

SC) were dissolved in saline. COX inhibitors SC 58,560 (selective for COX 1) and SC 58,125 (selective

for COX 2) were provided as a gift by Monsanto-Searle (St. Louis, Mo). COX inhibitors were dissolved

in 10% DMSO, 66% Cremophor EL (Sigma), 24% saline for systemic delivery. Controls were carried out

with corresponding vehicles. For intrathecal injection, treatments were prepared so that the total dose was

given in a volume of 10 µl with a 10 µl saline flush. All doses are presented in terms of salt forms.

Data analysis

Due to the presence of differences in baseline amino acid release, each rat’s amino acid microdi-

alysis data are expressed as the percentage of the rat’s baseline release. This baseline is the average of the

two fractions collected immediately before injection. Differences in baseline PGE2 release were small, so

PGE2 data were not normalized. Statistical significance was assessed using analysis of variance on inte-

grated release data (i.e., the sum of post-injection values); post hoc testing used Fisher’s protected least

significant difference test. Raw data and area over curve data were used for analysis of thermal escape

latencies; areas over the curve were calculated using the trapezoidal method. Analysis of variance was

generally used with Tukey's multiple comparison test; however, in analysis of raw data for peptide ef-

fects, error values were not evenly distributed and group sizes were small. In this subset of studies,

Kruskal–Wallis tests were used in order to eliminate the effects of heteroscedacity. Response thresholds


at ASPE

T Journals on M


Dow

nloaded from



for tactile allodynia testing were calculated as the force necessary to produce a probability of behavioral

response of 0.5. Either a statistically significant change in threshold or a decrease of threshold below four

grams was considered allodynia.

RESULTS

Baseline dialysis data

Baseline PGE2 concentrations in spinal dialysates were consistent and detectable. For thrombin

experiments and controls (n = 7–10), baseline dialysate PGE2 ranged from 102 ± 12 (mean ± standard

error) to 124 ± 21 fmol per 10 min. fraction. For peptide treatment groups (n = 6–9), dialysate PGE2

ranged from 101 ± 9 to 169 ± 35 fmol per 10 min. fraction. Baseline concentrations of amino acids in spi-

nal dialysates were within the reliable linear range of the assay in all cases (means ± standard errors, n =

8: aspartate, 37 ± 13 pmol/10 min. fraction; glutamate, 91 ± 14; serine, 381 ± 63; taurine, 199 ± 24).

Thrombin and microdialysis

Rats treated with 30 ng thrombin showed a substantial and long–lasting increase in dialysate

PGE2, which peaked between 30 and 90 minutes after injection (Fig. 1). Analysis of variance revealed a

significant treatment effect (F (2,23) = 4.873, p < 0.05), with thrombin–treated rats having higher dialys-

ate PGE2 than vehicle–injected rats (p < 0.05) or uninjected control rats (p < 0.01). Rats in the two control

groups did not differ significantly (p = 0.27). Spinal release of amino acids was also measured by mi-

crodialysis before and after injection of 100 ng thrombin (n = 8). None of the amino acids showed a sub-

stantial increase following thrombin administration (Fig. 2).

PAR–derived peptides and microdialysis

The effects of the PAR 1–derived peptide SFLLRN–NH2 were tested at intrathecal doses of 3, 30

and 200 µg. SFLLRN–NH2 produced a substantial, dose–dependent increase in PGE2 release, tending to

come in a rapid peak in the first twenty minutes following injection and a second peak at approximately

one hour after injection (Fig. 3). When the post–injection PGE2 release was summed, the effect of

SFLLRN–NH2 was statistically significant (F (3,29) = 8.542, p < 0.001) (Fig. 6). Compared to the vehi-


at ASPE

T Journals on M


Dow

nloaded from



cle–injected rats, the effects of 30 µg (p < 0.001) and 200 µg (p < 0.01) were significant, while those of 3

µg were not (p = 0.72).

The PAR 2–derived peptide SLIGRL was tested at intrathecal doses of 3, 30 and 200 µg.

SLIGRL increased PGE2 release with a biphasic time course (Fig. 4). Integrated post–injection PGE2 re-

lease showed a biphasic and statistically significant effect of SLIGRL (F (3,26) = 33.172, p < 0.001) (Fig.

6). Both 3 and 30 µg of SLIGRL stimulated more release than vehicle (p < 0.01 and p < 0.001, respec-

tively), but the effect of 200 µg was not statistically significant (p = 0.25).

The PAR 3–derived peptide TFRGAP and the PAR 4–derived GYPGQV were each tested at 200

µg (Fig. 5). Compared with vehicle, no statistically significant treatment effects were observed (F (2,20) =

0.521, p = 0.60).

Thrombin and behavior

Thrombin–induced thermal hyperalgesia was tested using withdrawal latencies in a modified

Hargreaves device. Pilot studies indicated that intrathecal doses of up to 10 ng of thrombin were without

effect. At doses from 300 ng to 1 µg, rats developed lethargy combined with chromodacyorrhea, ptosis

and piloerection. These high dose effects are reminiscent of the behavior seen with high doses of PGE2

((Minami et al., 1994); Svensson and Yaksh, unpublished data). For full studies, thermal escape latencies

were tested using thrombin doses from 30 to 300 ng and compared with 300 ng bovine serum albumin

(n=7 observations per treatment). Treatment with 30 ng thrombin decreased paw withdrawal latencies

relative to baseline and relative to the control treatment (Fig. 7a). The magnitude of the hyperalgesic ef-

fect peaked sixty minutes after injection and remained substantial at ninety minutes; at two hours after

injection, latencies were not different from those of control rats (data not shown). This effect was dose–

dependent (Fig. 7b; F (3,24) = 3.304, p < 0.05), as rats treated with 300 ng thrombin showed a general

suppression of behavior, including an increase in paw withdrawal latency.

Tactile sensitivity was measured using von Frey filaments before and twenty, forty or sixty min-

utes following the intrathecal injection of 100 ng thrombin (n = 6). The plantar surface of the paw and a

lumbar dorsal paramedian site were both tested. Allodynia did not develop at either site; paw withdrawal


at ASPE

T Journals on M


Dow

nloaded from



thresholds never decreased below 11.6 g and dorsum response thresholds never decreased below 14.2 g

after injection.

The spontaneous flinching response to intraplantar formalin injection was tested in rats receiving

either 100 ng thrombin or vehicle (n = 7 per group). Pretreatments were injected via intrathecal catheters

fifteen minutes before formalin injection (0.5 % v/v). Neither phase 1 flinching behavior (the first nine

minutes after formalin injection) nor phase 2 flinching behavior (from ten to sixty minutes after formalin

injection) was affected (data not shown).

PAR–derived peptides and behavior

Intrathecal injection of saline or 30 µg of SFLLRN–NH2 or SLIGRL (n = 6 per treatment) re-

sulted in reductions in hindpaw withdrawal latency (i.e., hyperalgesia) at the limit of statistical signifi-

cance (Kruskal–Wallis H = 5.66, 2 df; p = 0.059). Intrathecal injection of the PAR 1–selective peptide

TFLLR did not produce thermal hyperalgesia (data not shown). In rats pretreated with (+)–ibuprofen (30

µg intrathecal, twenty minutes before peptide injection), no significant effect of SFLLRN–NH2 or

SLIGRL administration was observed (Fig. 8b; H = 1.21, 2 df; p = 0.55). The effects of SLIGRL were

investigated further with isozyme–specific inhibitors. Both the COX 1 inhibitor (SC 58,560) and the COX

2 inhibitor (SC 58,125) blocked the thermal hyperalgesia produced by SLIGRL (Fig. 9). Areas over the

curve were significantly different (F(2,15) = 8.09; p < 0.005), with significant effects of both SC 58,560

(Tukey's q = 4.97 vs. saline, p < 0.01) and SC 51,245 (q = 4.88 vs. saline, p < 0.01).

DISCUSSION

In peripheral tissues, activation of PARs stimulates intracellular messengers associated with acti-

vation of cyclooxygenases (Macfarlane et al., 2001; O'Brien et al., 2001; Svensson and Yaksh, 2002) and

evokes prostaglandin production (Neufeld and Majerus, 1983). The CNS and spinal cord express all of

the proteins necessary for PAR signaling, including all currently identified PARs (Ishihara et al., 1997;

Niclou et al., 1998; Steinhoff et al., 2000; Wang et al., 2002a), proteases (Dihanich et al., 1991; Chen et


at ASPE

T Journals on M


Dow

nloaded from



al., 1995; Scarisbrick et al., 1997; Scarisbrick et al., 2001) and regulators of protease activity (e.g., Niclou

et al., 1998). Accordingly, we sought to determine which spinal PARs might be coupled to the formation

of PGE2.

Activating spinal PARs evokes PGE2 release

Using the intrathecal loop dialysis system, we examined spinal PAR effects on extracellular con-

centrations of PGE2. Intrathecal thrombin resulted in PGE2 release with no effect on aspartate, glutamate,

serine or taurine (Figs. 1 and 2). PGE2 release was observed after intrathecal delivery of agonist peptides

derived from PAR 1 and PAR 2 (Figs. 3, 4 and 6). In contrast, peptides derived from PARs 3 and 4 had

no effect, even at high doses (Figs. 5 and 6). The initial time course of prostaglandin release suggests that

de novo cyclooxygenase synthesis is not required—i.e., that spinal cyclooxygenases supporting PAR sig-

naling are constitutively expressed. The failure of PAR 3 and PAR 4 peptides to stimulate PGE2 produc-

tion demonstrates the specificity of the effects of the PAR 1 and PAR 2 peptides.

The data from these studies strongly suggest that, at the spinal level, PAR 2 is involved. The data

are less clear regarding the role of PAR 1. Ligand specificity is currently a limitation in the interpretation

of PAR studies in general. Few PAR ligands have been validated across the entire family of PARs

(Macfarlane et al., 2001; O'Brien et al., 2001). Both SLIGRL and SFLLRN will activate a homogenous

population of PAR 2, although PAR 1 is only activated by SFLLRN (e.g., (Lerner et al., 1996)). Throm-

bin shows a clear preference for PAR 1 (Nystedt et al., 1994; Blackhart et al., 1996). While initial reports

suggested PAR 3 was inactive, several studies have reported efficacy of TFRGAP at PAR 3 (Wang et al.,

2002a; Wang et al., 2002b). The current biochemical literature strongly suggests that PAR 3– and PAR 4–

derived peptides have several fold lower affinity for their receptors than do PAR 1– and PAR 2–derived

peptides do for their receptors (e.g., Xu et al., 1998). Thus, negative findings in these experiments with

TFRGAP and GYPGQV strongly suggest that PAR 3 and PAR 4 are not involved.

Activating spinal PARs evokes hyperalgesia

Consistent with the increase in PGE2 release, intrathecal thrombin and PAR 2–derived peptides

caused thermal hyperalgesia. This hyperalgesia was blocked by ibuprofen and both COX 1– and COX 2–


at ASPE

T Journals on M


Dow

nloaded from



selective inhibitors. Although thrombin and the PAR 1–derived peptide SFLLRN-NH2 produced thermal

hyperalgesia, the more PAR 1–selective peptide TFLLR did not. Consistent with these behavioral obser-

vations, the dialysis data suggest that PAR 2 and, with lesser certainty, PAR 1 are involved. More specific

ligands will be necessary to conclusively define the role of spinal PAR 1.

Previous work has emphasized that the effect of PGE2 may reflect some combination of enhanced

release of transmitters from primary afferents, depolarization of second order afferents or reduced activa-

tion of inhibitory interneurons (see Svensson and Yaksh, 2002) for a review). The locus of PAR activity

is not clear, however. Thrombin resulted in a significant increase in PGE2 but not amino acids, implicat-

ing mechanisms other than activation of primary afferents. This selectivity is unusual in our experience

with nociceptive spinal activators such as intraplantar formalin (Malmberg and Yaksh, 1995) and in-

trathecal dynorphin (Koetzner et al., 2004). A lack of primary afferent activity contrasts with the activity

of peripheral nerve PARs. For example, in superfused rat dorsal spinal cord tissue, trypsin, tryptase and

SLIGRL stimulate release of substance P (Steinhoff et al., 2000). Intrathecal administration of neurokinin

receptor 1 antagonists block hyperalgesia following intraplantar or intracolonic administration of SLIGRL

(Vergnolle et al., 2001; Coelho et al., 2002).

Role of COX 1 and COX 2

The present observations show that both COX 1 and COX 2 inhibtors will reduce PAR–evoked

hyperalgesia and establish a crucial downstream linkage with the cPLA2–COX cascade. In many systems,

COX–mediated hyperalgesia appears to be produced exclusively by COX 2, e.g. following intraplantar

carrageenan and intrathecal substance P (Yaksh et al., 2001a). The limited role of COX 1 in spinal PGE2

release has been surprising. The absence of an effect has been suggested to reflect different distributions

or differences in substrate requirements for the two isozymes. More recently, there has been a growing

appreciation of the role for COX 1, as after nerve injury (Ma et al., 2002) or after spinal delivery of

dynorphin (Koetzner et al., 2004). Spinal PAR 2 activation may also involve the downstream events that

are a critical for these other COX 1–activating stimuli.


at ASPE

T Journals on M


Dow

nloaded from



Physiological significance of PAR activation

The physiological significance of PAR activation in the periphery can be inferred from experi-

ments that have blocked the protease–PAR interaction. For example, soybean trypsin inhibitor pretreat-

ment nearly eliminated carrageenan–induced hyperalgesia and inflammation (Van Arman et al., 1966).

Similarly, mice with a targeted disruption of the PAR 2 gene show a diminished response to intraplantar

formalin or the mast cell degranulator compound 48/80 (Vergnolle et al., 2001). For PARs in the brain

and spinal cord, there are two possible sources of proteases. Blood–derived enzymes (e.g., thrombin) are

likely to play a role in cases of blood–brain barrier disruption; this could involve penetrating injury, but

diminished blood–brain barrier function may also be present in conditions ranging from stress to status

epilepticus (Gingrich and Traynelis, 2000). Small but pervasive breaches of the blood–brain barrier may

also be a feature of dural puncture headache (Jankowski, 2002). However, a more intriguing possibility

involves the production of serine proteases by CNS tissue.

Several laboratories have demonstrated a variety of serine proteases in central nervous system

tissues and suggested their physiological relevance. Prothrombin mRNA has been demonstrated in rat

brain; additionally, prothrombin mRNA has been found in dorsal root ganglia (Dihanich et al., 1991). Al-

though no reports to date have identified trypsin in the brain or spinal cord, a number of trypsin homologs

have been identified. Neuropsin mRNA is found in the hippocampus and cortex, and mRNA increased

substantially following electrical kindling (Chen et al., 1995). MSP is a trypsin homolog cloned from rat

spinal cord (Scarisbrick et al., 1997) with broad substrate specificity; the probable cleavage sequence fits

the cryptic PAR 2 peptide (Blaber et al., 2002). Kainate administration increased MSP expression in all

areas of the spinal cord, suggesting a link between excitatory amino acid receptors and MSP (Scarisbrick

et al., 1997). A more recent study has found MSP mRNA and immunoreactivity in human dorsal root

ganglia and spinal cord (Scarisbrick et al., 2001). Finally, a serine protease with trypsin–like substrate

preference and inhibitor sensitivity has been cloned from rat brain; it activates PAR 2 on glioblastoma

cells (Sawada et al., 2000). These data, combined with reports of protease regulators in the central nerv-


at ASPE

T Journals on M


Dow

nloaded from



ous system (e.g., Niclou et al., 1998), suggest that the CNS is capable of regulated proteolytic signaling

through PARs.

PARs are a new class of receptors with a relatively undeveloped pharmacology. PARs have a

complex regulatory cycle; for example, PAR 2 calcium signaling is polyphasic, with an initial peak and a

later plateau (Böhm et al., 1996). PAR activity couples to cyclooxygenases, which are subject to their

own complex regulation (e.g., Wu et al., 1999). The vagaries of prostanoid effects on behavior are well

known (Svensson and Yaksh, 2002). These uncertainties force us to make certain speculations in the in-

terpretation of our data. We attribute the increased response latency following high doses of thrombin

(Fig. 7) to a general suppression of behavior rather than analgesia on the basis of diminished locomotor

behavior and markers of stress observed following high dose thrombin (results, data not shown) which are

typical of high–dose PGE2 effects (Svensson and Yaksh, 2002). The complexities of the PGE2 response

(i.e., biphasic dose–response and time course; Figs. 3–6) could be a reflection of the regulation of PARs

and cyclooxygenases in the spinal cord. However, the state of our understanding of both PAR and

cyclooxygenase function means that this attribution is purely speculative.

In summary, we have demonstrated that stimulation of spinal PARs 1 and 2 produces cyclooxy-

genase––dependent thermal hyperalgesia. Further work will be necessary to map out the signal transduc-

tion pathways which link PARs with cyclooxygenase activity. PAR–stimulated calcium flux may drive

kinase activity; indeed, it has long been known that cellular tyrosine phosphorylation increases following

thrombin treatment (e.g., Ferrell and Martin, 1988). This suggests a role for MAP kinases (Wang et al.,

2002b), a question that will be resolved in future experiments.


at ASPE

T Journals on M


Dow

nloaded from



ACKNOWLEDGEMENTS

The authors would like to thank Alan Moore and Steve Rossi, PhD (UCSD Department of Anesthesiol-

ogy Analytical Laboratory) for assay support and Shelle Malkmus (UCSD Department of Anesthesiology

Research Laboratory) for her help in coordinating these experiments.


at ASPE

T Journals on M


Dow

nloaded from



REFERENCES

Blaber SI, Scarisbrick IA, Bernett MJ, Dhanarajan P, Seavy MA, Jin Y, Schwartz MA, Rodriguez M and

Blaber M (2002) Enzymatic properties of rat myelencephalon-specific protease. Biochemistry

41:1165-1173.

Blackhart BD, Emilsson K, Nguyen D, Teng W, Martelli AJ, Nystedt S, Sundelin J and Scarborough RM

(1996) Ligand cross-reactivity within the protease-activated receptor family. J Biol Chem

271:16466-16471.

Böhm SK, Khitin LM, Grady EF, Aponte G, Payan DG and Bunnett NW (1996) Mechanisms of desensi-

tization and resensitization of proteinase–activated receptor–2. J Biol Chem 271:22003-22016.

Chen J, Ishii M, Wang L, Ishii K and Coughlin SR (1994) Thrombin receptor activation. Confirmation of

the intramolecular tethered liganding hypothesis and discovery of an alternative intermolecular

liganding mode. J Biol Chem 269:16041-16045.

Chen Z-L, Yoshida S, Kato K, Momota Y, Suzuki J, Tanaka T, Ito J, Nishino H, Aimoto S, Kiyama H

and Shiosaka S (1995) Expression and activity-dependent changes of a novel limbic-serine prote-

ase gene in the hippocampus. J Neurosci 15:5088-5097.

Coelho A-M, Vergnolle N, Guiard B, Fioramonti J and Bueno L (2002) Proteinases and proteinase-

activated receptor 2: a possible role to promote visceral hyperalgesia in rats. Gastroenterology

122:1035-1047.

Dihanich M, Kaser M, Reinhard E, Cunningham D and Monard D (1991) Prothrombin mRNA is ex-

pressed by cells of the nervous system. Neuron 6:575-581.

Dirig DM, Salami A, Rathbun ML, Ozaki GT and Yaksh TL (1997) Characterization of variables defin-

ing hindpaw withdrawal latency evoked by radiant thermal stimuli. J Neurosci Methods 76:183-

191.

Ferrell JE, Jr. and Martin GS (1988) Platelet tyrosine-specific protein phosphorylation is regulated by

thrombin. Mol Cell Biol 8:3603-3610.


at ASPE

T Journals on M


Dow

nloaded from



Gingrich MB and Traynelis SF (2000) Serine proteases and brain damage - is there a link? Trends Neuro-

sci 23:399-407.

Ishihara H, Connolly AJ, Zeng D, Kahn ML, Zheng Y-W, Timmons C, Tram T and Coughlin SR (1997)

Protease-activated receptor 3 is a second thrombin receptor in humans. Nature 386:502-506.

Jankowski CJ (2002) Neuraxial anesthetic techniques, in Textbook of regional anesthesia (Raj PP ed) pp

829-852, Churchill Livingstone, New York.

Koetzner L, Hua X-Y, Lai J, Porreca F and Yaksh TL (2004) Nonopioid actions of intrathecal dynorphin

evoke spinal excitatory amino acid and prostaglandin E2 release mediated by cyclooxygenase-1

and -2. J Neurosci 24:1451-1458.

Lerner DJ, Chen M, Tram T and Coughlin SR (1996) Agonist recognition by proteinase-activated recep-

tor 2 and thrombin receptor. Importance of extracellular loop interactions for receptor function. J

Biol Chem 271:13943-13947.

Ma W, Du W and Eisenach JC (2002) Role for both spinal cord COX-1 and COX-2 in maintenance of

mechanical hypersensitivity following peripheral nerve injury. Brain Res 937:94-99.

Macfarlane SR, Seatter MJ, Kanke T, Hunter GD and Plevin R (2001) Proteinase-activated receptors.

Pharmacol Rev 53:245-282.

Malmberg AB and Yaksh TL (1995) Cyclooxygenase inhibition and the spinal release of prostaglandin

E2 and amino acids evoked by paw formalin injection: a microdialysis study in unanesthetized

rats. J Neurosci 15:2768-2776.

McKinney M, Snider RM and Richelson E (1983) Thrombin binding to human brain and spinal cord.

Mayo Clin Proc 58:829-831.

Means ED and Anderson DK (1986) Thrombin interactions with central nervous system tissue and impli-

cations of these interactions. Ann N Y Acad Sci 485:314-322.

Minami T, Uda R, Horiguchi S, Ito S, Hyodo M and Hayaishi O (1994) Allodynia evoked by intrathecal

administration of prostaglandin E2 to conscious mice. Pain 57:217-223.


at ASPE

T Journals on M


Dow

nloaded from



Neufeld EJ and Majerus PW (1983) Arachidonate release and phosphatidic acid turnover in stimulated

human platelets. J Biol Chem 258:2461-2467.

Niclou SP, Suidan HS, Pavlik A, Vejsada R and Monard D (1998) Changes in the expression of protease-

activated receptor 1 and protease nexin-1 mRNA during rat nervous system development and af-

ter nerve lesion. Eur J Neurosci 10:1590-1607.

Nystedt S, Emilsson K, Wahlestedt C and Sundelin J (1994) Molecular cloning of a potential proteinase

activated receptor. Proc Natl Acad Sci U S A 91:9208-9212.

O'Brien PJ, Molino M, Kahn M and Brass LF (2001) Protease activated receptors: theme and variations.

Oncogene 20:1570-1581.

Sawada K, Nishibori M, Nakaya N, Wang Z and Saeki K (2000) Purification and characterization of a

trypsin-like serine proteinase from rat brain slices that degrades laminin and type IV collagen and

stimulates protease-activated receptor-2. J Neurochem 74:1731-1738.

Scarisbrick IA, Isackson PJ, Ciric B, Windebank AJ and Rodriguez M (2001) MSP, a trypsin-like serine

protease, is abundantly expressed in the human nervous system. J Comp Neurol 431:347-361.

Scarisbrick IA, Towner MD and Isackson PJ (1997) Nervous system-specific expression of a novel serine

protease: regulation in the adult rat spinal cord by excitotoxic injury. J Neurosci 17:8156-8168.

Smith-Swintosky VL, Cheo-Isaacs CT, D'Andrea MR, Santulli RJ, Darrow AL and Andrade-Gordon P

(1997) Protease-activated receptor-2 (PAR-2) is present in the rat hippocampus and is associated

with neurodegeneration. J Neurochem 69:1890-1896.

Steinhoff M, Vergnolle N, Young SH, Tognetto M, Amadesi S, Ennes HS, Trevisani M, Hollenberg MD,

Wallace JL, Caughey GH, Mitchell SE, Williams LM, Geppetti P, Mayer EA and Bunnett NW

(2000) Agonists of proteinase-activated receptor 2 induce inflammation by a neurogenic mecha-

nism. Nat Med 6:151-158.

Striggow F, Riek-Burchardt M, Kiesel A, Schmidt W, Henrich-Noack P, Breder J, Krug M, Reymann KG

and Reiser G (2001) Four different types of protease-activated receptors are widely expressed in

the brain and are up-regulated in hippocampus by severe ischemia. Eur J Neurosci 14:595-608.


at ASPE

T Journals on M


Dow

nloaded from



Svensson CI and Yaksh TL (2002) The spinal phospholipase-cyclooxygenase-prostanoid cascade in no-

ciceptive processing. Annu Rev Pharmacol Toxicol 42:553-583.

Van Arman CG, Nuss GW, Winter CA and Flataker L (1966) Proteolytic enzymes as mediators of pain,

in Pharmacology of Pain (EG P ed) pp 25-32, Pergamon Press, Oxford.

Vergnolle N, Bunnett NW, Sharkey KA, Brussee V, Compton SJ, Grady EF, Cirino G, Gerard N, Bas-

baum AI, Andrade-Gordon P, Hollenberg MD and Wallace JL (2001) Proteinase-activated recep-

tor-2 and hyperalgesia: A novel pain pathway. Nat Med 7:821-826.

Wang H, Ubl JJ and Reiser G (2002a) Four subtypes of protease-activated receptors, co-expressed in rat

astrocytes, evoke different physiological signaling. Glia 37:53-63.

Wang H, Ubl JJ, Stricker R and Reiser G (2002b) Thrombin (PAR-1)-induced proliferation in astrocytes

via MAPK involves multiple signaling pathways. Am J Physiol Cell Physiol 283:C1351-1364.

Weinstein JR, Gold SJ, Cunningham DD and Gall CM (1995) Cellular localization of thrombin receptor

mRNA in rat brain: expression by mesencephalic dopaminergic neurons and codistribution with

prothrombin mRNA. J Neurosci 15:2906-2919.

Wu G, Wei C, Kulmacz RJ, Osawa Y and Tsai A-L (1999) A mechanistic study of self-inactivation of the

peroxidase activity in prostaglandin H synthase-1. J Biol Chem 274:9231-9237.

Xu W-F, Andersen H, Whitmore TE, Presnell SR, Yee DP, Ching A, Gilbert T, Davie EW and Foster DC

(1998) Cloning and characterization of human protease–activated receptor 4. Proc Natl Acad Sci

U S A 95:6642-6646.

Yaksh TL, Dirig DM, Conway CM, Svensson C, Luo ZD and Isakson PC (2001a) The acute antihyperal-

gesic action of nonsteroidal, anti-inflammatory drugs and release of spinal prostaglandin E2 is

mediated by the inhibition of constitutive spinal cyclooxygenase-2 (COX-2) but not COX-1. J

Neurosci 21:5847-5853.

Yaksh TL, Ozaki G, McCumber D, Rathbun M, Svensson C, Malkmus S and Yaksh MC (2001b) An

automated flinch detecting system for use in the formalin nociceptive bioassay. J Appl Physiol

90:2386-2402.


at ASPE

T Journals on M


Dow

nloaded from




at ASPE

T Journals on M


Dow

nloaded from



FOOTNOTES

This work was supported by NIH grant NS16541 to TLY; LK was supported by training grant NS07407

to TLY.

Reprint requests to:

Tony L. Yaksh, PhD

Anesthesiology Research Laboratory 0818

University of California—San Diego

9500 Gilman Drive

La Jolla, CA 92093-0818

Voice (619) 543-3597

Fax (619) 543-6070

Electronic mail [email protected]

1: Department of Anesthesiology, University of California—San Diego

2: Present address: Department of Neuropharmacology, Purdue Pharma, LP

3: Department of Pharmacology, University of California—San Diego


at ASPE

T Journals on M


Dow

nloaded from



FIGURE LEGENDS

Figure 1. PGE2 release after thrombin

Intrathecal injection of 30 ng of thrombin stimulated prostaglandin E2 release. Integrated postinjection

data showed a significant difference between thrombin and vehicle (p < 0.05) or no–injection control (p <

0.01). There was no significant difference between controls.

Figure 2. Amino acid release after thrombin

Intrathecal injection of 100 ng of thrombin did not result in substantial amino acid release.

Figure 3. PGE2 release after SFLLRN–NH2

The time course of dialysate PGE2 concentrations was measured following intrathecal administration of 3,

30 or 200 µg of the PAR 1–derived peptide SFLLRN–NH2.

Figure 4. PGE2 release after SFLLRN–NH2

The time course of dialysate PGE2 concentrations was measured following intrathecal administration of 3,

30 or 200 µg of the PAR 2–derived peptide SLIGRL.

Figure 5.PGE2 release after TFRGAP and GYPGQV

The time course of dialysate PGE2 concentrations was measured following administration of 200 µg of

the PAR 3–derived peptide TFRGAP or 200 µg of the PAR 4–derived peptide GYPGQV.

Figure 6. PGE2 release: summary

Intrathecal injection of the peptides SFLLRN–NH2 and SLIGRL stimulated significant prostaglandin E2

release; injection of TFRGAP and GYPGQV did not. Compared to vehicle controls (dotted line), inte-

grated postinjection release was significantly increased for 30 and 200 µg of SFLLRN–NH2 (p < 0.001

and 0.01, respectively) but not 3 µg. Prostaglandin E2 release was significantly increased after 3 and 30


at ASPE

T Journals on M


Dow

nloaded from



µg of SLIGRL (p < 0.01 and 0.01, respectively) but not 200 µg. No significant effect of TFRGAP or

GYPGQV was observed.

Figure 7. Thermal hyperalgesia after thrombin

A. The time course of paw withdrawal latencies following intrathecal administration of either 30 ng

thrombin or 300 ng bovine serum albumin suggested that hyperalgesia reaches a maximum within 60

minutes of injection.

B. Analysis of data at 60 minutes after injection indicated a dose–dependent effect of thrombin, with a

decrease in latency at 30 ng and an increase at 300 ng.

Figure 8.Thermal hyperalgesia after peptides

A. Paw withdrawal latencies were followed after intrathecal administration of either saline or 30 µg of

SFLLRN–NH2 or SLIGRL; both peptides produced a trend toward hyperalgesia (p = 0.059).

B. Pretreatment with ibuprofen (30 µg, intrathecal) eliminated any hyperalgesic effect of the peptides (p =

0.55).

C. Comparison of hyperalgesia produced by SLIGRL following saline, the COX 1–specific inhibitor SC

58,560 or the COX 2–specific inhibitor SC 58,125 showed significant effects of both COX 1 and COX 2

(p < 0.01, both comparisons).


at ASPE

T Journals on M


Dow

nloaded from



at ASPE

T Journals on M


Dow

nloaded from



at ASPE

T Journals on M


Dow

nloaded from



at ASPE

T Journals on M


Dow

nloaded from



at ASPE

T Journals on M


Dow

nloaded from



at ASPE

T Journals on M


Dow

nloaded from



at ASPE

T Journals on M


Dow

nloaded from



at ASPE

T Journals on M


Dow

nloaded from



at ASPE

T Journals on M


Dow

nloaded from


Documents

Downloaded from hyperalgesia through spinal …jpet.aspetjournals.org/content/jpet/early/2004/06/02/jpet.104... · Dialysis experiments were conducted in conscious, unrestrained rats