TR ERVH TIVAP EASH TS

J.M.aDedadRobPhPaicDeBel

Abisoptrofroincwhtinmeaftaft(Colatdu(10inflwitmipriincTEvlPlowrattemthrjoiWe hypothesize that the effects of TENS are mediatedthrough the vlPAG that sends projections through the RVM tothe 2

Kean

Thmiinanboseerapasiove19

anneeleal.apanveanmianetbeisjthThtio

theenretproutePAG divisions, but are absent in the dorsolateral PAGdivision (Reichling and Basbaum, 1991). Inactivation of the

*CoE-mAbbuedneogre

Ne

030doispinal cord to produce an opioid-mediated analgesia.009 IBRO. Published by Elsevier Ltd. All rights reserved.

y words: pain, TENS, hyperalgesia, opioid, inflammation,algesia.

RVM disrupts antinociception mediated by stimulation ofthe PAG (Prieto et al., 1983; Sandkuhler and Gebhart,1984). Further, microinjection of morphine into the RVMproduces antinociception (Jensen and Yaksh, 1986; Mor-gan et al., 1998; Morgan and Whitney, 2000). Thus, opioid-ANSCUTANEOUS ELECTRICAL NIGH AND LOW FREQUENCIES ACERIAQUEDUCTAL GREY TO DECRYPERALGESIA IN ARTHRITIC RA

M. DESANTANA,a* L. F. S. DA SILVA,b

A. DE RESENDEc AND K. A. SLUKAb

partment of Physical Therapy, Federal University of Sergipe, Ci-e Universitria Professor Jos Alosio de Campos. Av. Marechalndon, s/n, Jardim Rosa Else, So Cristvo/Sergipe, Brazil

ysical Therapy and Rehabilitation Science Graduate Program,n Research Program, University of Iowa, Iowa City, IA, USA

partment of Physical Therapy, Federal University of Minas Gerais,o Horizonte, Minas Gerais, Brazil

stractTranscutaneous electric nerve stimulation (TENS)widely used for the treatment of pain. TENS produces anioid-mediated antinociception that utilizes the rostroven-medial medulla (RVM). Similarly, antinociception evokedm the periaqueductal grey (PAG) is opioid-mediated andludes a relay in the RVM. Therefore, we investigatedether the ventrolateral or dorsolateral PAG mediates an-ociception produced by TENS in rats. Paw and knee jointchanical withdrawal thresholds were assessed before ander knee joint inflammation (3% kaolin/carrageenan), ander TENS stimulation (active or sham). Cobalt chlorideCl2; 5 mM) or vehicle was microinjected into the ventro-eral periaqueductal grey (vlPAG) or dorsolateral periaque-ctal grey (dlPAG) prior to treatment with TENS. Either high0 Hz) or low (4 Hz) frequency TENS was then applied to theamed knee for 20 min. Active TENS significantly increasedhdrawal thresholds of the paw and knee joint in the groupcroinjected with vehicle when compared to thresholdsor to TENS (P

sp20-oac20freera(Coidetoppodu

Allteeaccuseof aDawelighFo

Ind

ImmscrthecarandtheinflmaditinanweWiedeas(Sl

Ca

Intr(vlPbefthe50sec(17Aftdril1 mral:belabommthesec(Urinflwaweiora

dlPthreuti.p.thisfround

Dr

Vedismicdosanddisina200volthumitneuof198198

iltomicovewatheto aappwh

Be

Thweoldtionwitweminplamafilaappfrowapawdrachapertes

witdesacc2 dpriowitjoinandforcon

J. M. DeSantana et al. / Neuroscience 163 (2009) 123312411234inal cord and RVM (Sluka et al., 1999; Kalra et al.,01). Specifically, low (4 Hz) frequency TENS activatespioid receptors and high (100 Hz) frequency TENS

tivates -opioid receptors (Sluka et al., 1999; Kalra et al.,01). Further, repeated application of TENS, low or highquency, produces analgesic tolerance and a cross-tol-nce to - and -opioid receptors spinally, respectivelyhandran and Sluka, 2003). As TENS produces an opi--mediated antinociception that utilizes the RVM (Kalraal., 2001) and antinociception evoked from the PAG isioid-mediated and includes a relay in the RVM, we hy-thesized that the PAG mediates the antinociception pro-ced by TENS.

EXPERIMENTAL PROCEDURES

experiments were approved by Animal Care and Use Commit-at the University of Iowa (Iowa City, IA, USA) and are inordance with the guidelines of National Institutes of Health onof laboratory animals. This study used the minimum numbernimals to obtain statistical significance. Adult male Spraguewley rats (n64; 225350 g, Harlan, Indianapolis, IN, USA)re used for this study. The animals were housed in a 12-ht/dark cycle, and the testing was done only in the light cycle.od and water were available to the animals ad libitum.

uction of inflammation

ediately after baseline behavioral measurements that are de-ibed below, rats were anesthetized with isoflurane (24%) andleft knee joint was injected intra-articularly with a mixture of 3%rageenan and 3% kaolin (0.1 ml in sterile saline, pH 7.4) (SlukaWestlund, 1993). The inflammation is considered acute forfirst 24 h, when there is neutrophil infiltration. By 1 week, theammation converts to chronic, as identified histologically bycrophage infiltration. This model is used to mimic arthritic con-ons and shows good predictability for drug effects (Radhakrish-et al., 2003). After induction of knee inflammation, the rats

re returned to their cages and allowed to recover for 24 h.thin 24 h, the animals exhibit signs of inflammation such asmatous and warm knee joints and also behavioral signs suchguarding and decreased weight bearing on the inflamed limbuka and Westlund, 1993).

nnula implantation and microinjections

acerebral guide cannulae were placed in the ventrolateralAG) or dorsolateral (dlPAG) periaqueductal grey 3 to 5 daysore induction of knee joint inflammation. The rats were anes-tized with an i.p. injection of sodium pentobarbital (Nembutal,mg/kg, Ovation Pharmaceucticals, Deerfield, IL, USA) andured in a stereotaxic head holder to implant the guide cannula.5 mm in length, 26 gauge; Plastics One, Roanoke, VA, USA).er the midline incision, the skull was exposed, and a small holeled for placement of the guide cannula. The guide cannula wasm above the vlPAG, using the following coordinates: interau-1.7 mm; mediolateral: 0.6 mm; and dorsoventral: 5.0 mmow the skull surface. For dlPAG, the guide cannula was 1 mmve the dlPAG, using the following coordinates: interaural: 1.7; mediolateral: 0.6 mm; and dorsoventral: 4.8 mm belowskull surface (Paxinos and Watson, 2005). Cannulae wereured to the skull by stainless-steel screws and dental cementban and Smith, 1994). Cannula was implanted ipsilateral to theamed knee joint. A dummy cannula (33 gauge, Plastics One)s inserted into the guide cannula to maintain its patency. All rats

re allowed to recover 3 to 5 days after surgery before behav-l testing.

appoldTo examine placement of the cannula into the vlPAG orAG, an equivalent volume of methylene blue dye was injectedough the cannula at the end of the experiment. Rats were thenhanized with an overdose of sodium pentobarbital (150 mg/kg) and transcardially perfused with 4% paraformaldehyde. After, the brain was removed, stored in 30% sucrose solution,zen, cross-sectioned at 40 m on a cryostat and examineder a light microscope for placement of the cannula.

ug administration

hicle (0.5 l, 0.9% sterile saline) or 5 mM CoCl2 solution (0.5 l,solved in 0.9% sterile saline, Fisher Scientific, NJ, USA) wasroinjected into vlPAG or dlPAG through the guide cannula. Thee of CoCl2 was selected from a prior study (Cavun et al., 2004)through preliminary experiments. Microinjections of CoCl2 in

crete brain areas have been used for reversible functionalctivation (Kretz, 1984; Nuseir et al., 1999; Fisk and Wyss,0; Pajolla et al., 2005). Co2 obstructs the ionophore of thetage-gated Ca2 channel (Hagiwara and Byerly, 1981) ands induces blockade of Ca2-dependent release of neurotrans-ter from presynaptic terminals (Kretz, 1984). This blockade ofrotransmitter release therefore causes a reversible blockadeneuronal pathways that synapse in the targeted area (Kretz,4) and fibers of passage are not affected by CoCl2 (Kretz,4).A 33-gauge injection cannula was connected to a 10-l Ham-

n syringe through PE10 tubing backfilled with sterile saline. Theroinjection (0.5 l) of either CoCl2 or vehicle was performedr a 2-min period and the travel of the air bubble in the tubings carefully observed to ensure that the drug solution enteredinjection cannula. The needle was left in position for a minutellow diffusion of drug before the needle was withdrawn. TENSlication was performed 1 h after injection of CoCl2, a timeen preliminary studies show a maximal effect of CoCl2.

havioral assessment

e paw withdrawal threshold and the joint withdrawal thresholdre tested for all groups of rats. Paw and joint withdrawal thresh-s were assessed before and 24 h after induction of inflamma-, and 1 h after TENS application. Rats were tested for PWTh von Frey filaments applied to the paw. Initially, the animalsre maintained in their home cages in the behavior room for 30to acclimate to the environment. Then, the animals were

ced in transparent Lucite cubicles over a wire mesh and accli-ted for another 30 min before testing. A series of von Freyments with increasing bending forces (9.4495.8 mN) waslied to the plantar surface of the hind paw until the rat withdrew

m the stimulus (Gopalkrishnan and Sluka, 2000). Each filaments applied twice. The lowest force at which the rat withdrew itsfrom one of two applications was recorded as the paw with-

wal threshold for mechanical hyperalgesia. A reduction in me-nical withdrawal threshold was interpreted as cutaneous hy-algesia. This testing method has shown significant statisticaltretest reliability (Sluka et al., 1999).Rats were also tested for knee joint withdrawal thresholds

h a pair of forceps applied to the knee joints as previouslycribed (Vance et al., 2007; DeSantana et al., 2008). Rats werelimated in a restraining device three times a day 1 h apart forays, each acclimatizing session consisting of 5 min (two daysr to the induction of inflammation). The forceps were equippedh two strain gauges to measure force. To measure the kneet withdrawal threshold, animals were placed in the restrainer,the experimenter compressed the knee joint with the tip of the

ceps while the hind limb was extended. Compression wastinued until the animal withdrew the leg. The maximum force

lied at withdrawal was recorded as the joint withdrawal thresh-. Three trials 5 min apart at each time period were performed

andin whyp

TE12TEbyinfl(Sl

Ad

EMwauseanddiarecconens

admtionstraconinteconjusthowaplatresammethealwrec

Ex

BasurtionpawtheAntizewa

thr

forCo(n(nShHigLowanibehblinme

Sta

Sinwemicexadatthuforbet(paPawethrsid

Effwi

Jothrcami24thepamithrjecbesynrestheproeff

Figafte

J. M. DeSantana et al. / Neuroscience 163 (2009) 12331241 1235averaged to obtain one reading per time period. A decreaseithdrawal threshold of the knee joint was interpreted as jointeralgesia.We previously showed that (1) both low and high frequency

NS reduce hyperalgesia induced by kaolin and carrageenan for24 h after administration, (2) application of halothane withoutNS has no effect on the paw withdrawal latency to heat inducedjoint inflammation, and (3) application of TENS to a non-amed knee joint has no effect on the paw withdrawal latencyuka et al., 1998).

ministration of TENS

PI Select TENS units with an asymmetrical biphasic squareve (EMPI Inc., MN, USA) and half-inch circular electrodes wered. Under isoflurane anesthesia, the left knee joint was shavedround pre-gelled surface electrodes were applied to the me-

l and lateral aspects of the inflamed knee joint in the groupseiving active TENS or sham TENS. Animals were observedtinuously during TENS to ensure adequate anesthesia and toure that the electrodes remained in contact with the skin.Either high (100 Hz) or low (4 Hz) frequency TENS wasinistered keeping other parameters constant, i.e. pulse dura-(100 s), sensory intensity and 20 min for stimulation. Thistegy allowed a comparison of frequency differences withoutfounding differences in pulse duration or amplitude. Sensorynsity was determined by increasing the intensity until a muscletraction was visibly observed and then reducing the intensity tot below this level. The parameters were selected to modelse used clinically (Sluka et al., 1998). The sham TENS groups anesthetized with 12% isoflurane and electrodes wereced on their shaved knee joint, but did not receive TENSatment. Importantly, three rats always were anesthetized withe vaporizer; at least one rat receiving the sham TENS treat-

nt and one rat receiving the active TENS treatment were anes-tized at the same time. This procedure ensured that thereays were animals in the sham TENS treatment groups thateived the same dose of anesthesia as the active TENS groups.

perimental design

seline paw and knee joint withdrawal thresholds were mea-ed bilaterally prior to the induction of the knee joint inflamma-. Twenty-four hours after induction of knee joint inflammation,and knee joint withdrawal thresholds were reassessed and

n, the animals were microinjected with either CoCl2 or saline.hour following the microinjection, the rats were lightly anesthe-d with 12% isoflurane for placement of the electrodes. TENSs then applied for 20 min.

Fig. 1. Time line for the experiment.Following baseline and post-inflammation (24 h) withdrawalesholds, rats (n64) were randomly divided into 12 groups, six

threinjevlPAG: (1) Sham TENSvehicle (n6); (2) Sham TENSCl2 (n6); (3) High TENSvehicle (n6); (4) High TENSCoCl26); (5) Low TENSvehicle (n6); (6) Low TENSCoCl26); and six for dlPAG: (7) Sham TENSvehicle (n5); (8)

am TENSCoCl2 (n4); (9) High TENSvehicle (n6); (10)h TENSCoCl2 (n5); (11) Low TENSvehicle (n4); (12)TENSCoCl2 (n4). One hour after application of TENS,

mals were retested for withdrawal thresholds. Importantly, allavioral tests were done by the same experimenter who wasded to the drug injection and to the TENS group. All experi-ntal design is shown in Fig. 1.

tistical analysis

ce the data for mechanical withdrawal thresholds of the pawre not evenly distributed, and were on a discontinuous logarith-scale, non-parametric analysis with the KruskalWallis testmined differences between groups. Joint withdrawal thresholda were evenly distributed and on a continuous scale and weres examined for differences with a repeated measures ANOVAdifferences across time and between groups. Post hoc testingween individual groups was performed with a Tukeys testrametric) or signed rank test (nonparametric) as appropriate.rametric t-test and non-parametric Wilcoxon matched pairs testre used to analyze changes in the paw and joint withdrawalesholds within groups, respectively. P value 0.05 was con-ered significant.

RESULTS

ects of CoCl2 microinjection into the vlPAG onthdrawal thresholds

int inflammation significantly decreased the withdrawalesholds of the paw 24 h after injection of kaolin andrrageenan (Fig. 2). In preliminary experiments (n6),croinjection of 5 mM CoCl2 into the vlPAG ipsilaterallyh after induction of inflammation significantly increasedwithdrawal threshold to mechanical stimulation of the

w (Fig. 2). The effect of CoCl2 peaked 30 min aftercroinjection and lasted through 90 min. Withdrawalesholds returned to pre-CoCl2 values 2 h after microin-tion of CoCl2 (Fig. 2). Thus, we applied TENS for 20 minginning 1 h after CoCl2 to have an adequate block ofaptic transmission during TENS, and to test behavioralponses 1 h after the end of the TENS treatment so thateffects of CoCl2 were no longer present. The analgesiaduced by TENS would still be present at this time as theects of TENS last a minimum of 12 h (Sluka et al., 1998).

. 2. Microinjection of cobalt chloride (CoCl2) into the vlPAG 24 hr induction of inflammation increased the mechanical withdrawal

shold of the paw. The effect of CoCl2 peaked 30 min after micro-ction and lasted through 90 min.

DisdlP

Hisdisinjesiteincansitethe

EffTE

Thmeeithfousigthr

hig

priwhtre(Pthequthrthrinwhclegro5AtoeffTEaftwitbamu15pa434B

IninvpadeanintagTEce

Op

ThanbeetcroquGeAdcicvlPmiha19knstrtiorepanrec

FigPaxjecdisin odatbolvehlow

J. M. DeSantana et al. / Neuroscience 163 (2009) 123312411236tribution of microinjection sites in the vlPAG andAG

tological analysis showed that microinjection sites weretributed in the vlPAG or the dlPAG. Fig. 3A displays thection sites in all groups in the vlPAG, and Fig. 3B showss in the dlPAG. Sites outside the vlPAG or dlPAGluding the lateral PAG (n2), superior colliculus (n6)d aqueduct (n1) were removed from analysis. Thes plotted show the area of maximum concentration ofdye.

ects of the blockade of the vlPAG and dlPAG onNS antinociception

ere were no significant differences between groups forchanical withdrawal threshold of the paw or knee jointer before or 24 h after induction of inflammation. Twenty-r hours after the induction of inflammation there was anificant decrease in both paw and knee joint withdrawalesholds (P0.01; Figs. 4A and 5A).

. 3. Schematic coronal sections of the rat brain adapted frominos and Watson (2005) illustrating approximate sites of microin-tions into the vlPAG (A) and dlPAG (B). Numbers indicate thattance from the interaural in millimeters. Only rats with injection sitesr immediately adjacent to the vlPAG or dlPAG were included ina analysis. Symbols represent the microinjection sites; filled sym-s indicate animals microinjected with CoCl2 and open symbols,icle. Animals were stimulated with high frequency TENS (squares),frequency TENS (circles) or sham TENS (triangles).In the group of rats microinjected with saline, eitherh or low frequency TENS significantly reversed the

furTEmary (knee joint) and secondary (paw) hyperalgesiaen compared to the withdrawal thresholds prior to TENSatment (P0.001) or with sham TENS treatment0.001; Fig. 4A). However, microinjection of CoCl2 intovlPAG prior to application of either high or low fre-

ency TENS prevented the increases in withdrawalesholds normally observed by TENS. Withdrawaleshold of the paw and knee joint was significantly lowerthe groups treated with TENS and CoCl2 into the vlPAGen compared to the group treated with TENS and vehi-; and was not significantly different from sham TENSups or from the pre-TENS withdrawal thresholds (Fig.). However, microinjection of CoCl2 into the dlPAG priorapplication of either high or low frequency TENS had noect on the antihyperalgesia produced by TENS. All theNS groups after treatment were different from shamer treatment with TENS for both the muscle and pawhdrawal threshold. However no difference between co-lt and vehicle for each frequency was observed for thescle (14518% HFCoCl2 vs. 14318%, HFVehicle;220% LFCoCl2 vs. 1328%, LFVehicle), or thew (35566% HFCoCl2 vs. 606166%, HFVehicle;9112% LFCoCl2 vs. 682143% LFVehicle) (Figs.and 5B).

DISCUSSION

the current study, we injected CoCl2 into the PAG toestigate if the PAG was involved in the antinociceptivethway activated by stimulation with TENS. Our datamonstrated a complete blockade of the effects of highd low frequency TENS following microinjection of CoCl2o the vlPAG, but not the dlPAG. These data are inreement with prior data from our laboratory showing thatNS produces analgesia through activation of opioid re-ptors in the RVM (Kalra et al., 2001).

ioid analgesia in the PAG

ere is a growing literature indicating distinct dorsolaterald ventral antinociceptive systems within the PAG (Beh-hani, 1995; Cannon et al., 1982; Morgan, 1991; Morganal., 1989). Numerous studies show that morphine mi-injection into the ventral PAG produces analgesia (Jac-et and Lajtha, 1976; Yaksh et al., 1976; Lewis andbhart, 1977; Jensen and Yaksh, 1986; Siuciak andvokat, 1987; Behbehani, 1995). The behavioral antino-eption produced by microinjection of morphine into theAG, but not the dlPAG, decreases with repeated ad-nistration (Jacquet and Lajtha, 1976; Lewis and Geb-rt, 1977; Siuciak and Advokat, 1987; Tortorici et al.,99, 2001; Morgan et al., 2005a,b), a phenomenonown as opioid tolerance. This opioid tolerance is re-icted to the vlPAG, and does not occur with administra-n in the dlPAG (Tortorici et al., 1999). As with morphine,eated application of TENS results in analgesic toler-ce by the fourth day with a cross-tolerance at opioideptors in the spinal cord (Chandran and Sluka, 2003)

ther supporting a role for the opioid-analgesic system inNS analgesia.

agplaovthetricforterPA-

cothr(N1919an

Figandmicdiff

J. M. DeSantana et al. / Neuroscience 163 (2009) 12331241 1237Similar to morphine, microinjection of -opioid receptoronists produces antinociception as measured by the hotte and tail-flick test (Jensen and Yaksh, 1986). More-er, microinjection of a -opioid receptor antagonist intovlPAG prevented the antinociception produced by elec-al stimulation of the amygdala further supporting a role-opioid receptors in the PAG (Tershner and Helmstet-, 2000). These data suggest that activation of the ventralG produces an opioid-mediated analgesia that utilizesand -opioid receptors.The PAG does not have a major projection to the spinal

rd (Basbaum and Fields, 1984), but produces its effectsough a relay in the RVM, i.e. the nucleus raphe magnusRM) and adjacent structures (Kuypers and Maisky,75; Castiglioni et al., 1978; Mantyh and Peschanski,

. 4. Bar graph representing mechanical withdrawal threshold of the pa(B) dlPAG. Mechanical withdrawal thresholds are illustrated prior toroinjection of the vlPAG or dlPAG. Data are represented as meanerent from baseline time; significantly different from vehicle control82; Urban and Smith, 1994). The morphine-inducedalgesia from the PAG produces antinociception by acti-

inhVetion of both - and -opioid receptors in the RVM (Kiefelal., 1993; Urban and Smith, 1994). The RVM in turnjects to the spinal cord and reduces activity of nocicep-dorsal horn neurons, to result in an analgesic effectuo and Gebhart, 1997; Venegas and Schaible, 2004).e antinociception produced by microinjection of mor-ine in the vlPAG is prevented by blockade of NMDA, andth - and -opioid receptors in the RVM (Kiefel et al.,93; Spinella et al., 1996). These effects of activation ofvlPAG inhibitory pathway by morphine modulate neu-al activity in the RVM such that there was an increaseoff-cell activity and a decrease in on-cell activity (Chengal., 1986). Thus, the PAG projects through the RVM toduce inhibition. The RVM, then projects to the spinalrd using serotoninergic and non-serotoninergic cells to

nimals microinjected with either CoCl2 or vehicle into the (A) vlPAGof inflammation (Baseline), before application of TENS, and aftervalue 0.05 was considered statistically significant. * Significantlyvaetprotive(ZhThphbo19theroninetproco

w from ainductionSEM. Pgroups.ibit dorsal horn neurons (Zhuo and Gebhart, 1991;negas and Schable, 2004).

anRVacintsigcoadmu(R

anthrfretiocofreac

Figormicdiff

J. M. DeSantana et al. / Neuroscience 163 (2009) 123312411238The PAG and RVM work synergistically to producealgesia. Coadministration of -opioid agonists into theM and the PAG results in a profound synergistic inter-tion (Rossi et al., 1994). Coadministration of DAMGOo one region with deltorphin in the other also results in anificant synergy, whereas, if DAMGO and deltorphin areadministered together in the same brain area there is anditive effect. These findings suggest the existence of/mu and mu/delta synergy between the PAG and RVMossi et al., 1994).We therefore hypothesize that TENS utilizes the opioid

algesia system originating in the vlPAG which projectsough the RVM to the spinal cord. Both high and lowquency TENS produce their analgesia effects by activa-n of - and -opioid receptors in the RVM and the spinalrd (Sluka et al., 1999; Kalra et al., 2001). Further, low

. 5. Bar graph representing mechanical withdrawal threshold of the kn(B) dlPAG. Mechanical withdrawal thresholds are illustrated prior to iroinjection of the vlPAG or dlPAG. Data are represented as meanerent from baseline time; significantly different from vehicle controlquency TENS releases 5-HT in the spinal cord andtivates serotoninergic receptors, 5-HT2 and 5-HT3

-2etadhakrishnan et al., 2003; Sluka et al., 2006). Highquency TENS does not utilize 5-HT but releases GABAt activates GABAA receptors in the spinal cordadhakrishnan et al., 2003; Maeda et al., 2007). Bothh and low frequency TENS reduce dorsal horn neuronnsitization after inflammation (Ma and Sluka, 2001), andconsequent hyperalgesia (Sluka et al., 1999). Thus,

NS, both high and low frequency requires activation ofurons in the vlPAG, the RVM, and spinal cord, andlizes opioid mechanisms to produce analgesia.Although the PAG and the RVM are clearly involved inanalgesia produced by both low and high frequency

NS, other mechanisms including peripheral, segmentalinal analgesia, or systemic effects are also possible.eed prior studies show that effects of low and/or highquency TENS can be prevented by blockade of opioid or

nimals microinjected with either CoCl2 or vehicle into the (A) vlPAGof inflammation (Baseline), before application of TENS, and afteralue 0.05 was considered statistically significant. * Significantly(Rfretha(RhigsetheTEneuti

theTEspIndfre

ee from anductionSEM. P vgroups.noradrenergic receptors at the site of stimulation (Kingal., 2005; Resende et al., 2004). Spinally, there is acti-

vacecodeRamuco

Fa

Ththeinflnothe20PAal.E2acricthemavetiatioinfltoc20mathecicRV

Cli

Thphsysceobshprohigdofluflu19aloforPATEwawitpotenlogopquavfre

viocohaeffcloaanfutcoTEaim(H

Acfrocomtheupoare

Bag

Bas

Beh

Bur

Ca

Ca

Ca

Ch

Ch

De

Fie

Fis

Ge

Go

J. M. DeSantana et al. / Neuroscience 163 (2009) 12331241 1239tion and release of GABA, activation of muscarinic re-ptors, and activation of opioid receptors, all of whichuld be related to either spinal segmental inhibition orscending inhibitory pathways (Maeda et al., 2007;dharkshinan and Sluka, 2003; Sluka et al., 1999). Thus,ltiple mechanisms acting in concert or independentlyuld be responsible for the analgesia produced by TENS.

cilitation from the PAG

e current study showed that microinjection of CoCl2 intovlPAG reversed the hyperalgesia induced by knee jointammation, suggesting a role for the PAG in facilitatingciception after injury. Recent studies support a role forPAG in facilitation of nociception (Heinricher et al.,

04; Guo et al., 2006). This facilitatory effect from theG involves prostaglandin E2 and BDNF (Heinricher et, 2004; Guo et al., 2006). Microinjection of prostaglandindecreases the paw withdrawal latency to heat, and

tivates facilitatory cells in the RVM, i.e. on-cells (Hein-her et al., 2004). Further there are increases in BDNF inPAG, the BDNF receptor TrkB in the RVM after inflam-tion, and blockade of TrkB receptors in the RVM re-rses hyperalgesia (Guo et al., 2006). There is substan-l evidence that the RVM mediates facilitation of nocicep-n after injury in numerous animal models including jointammation (Urban et al., 1996), pancreatitis (Vera-Por-arrero et al., 2006), neuropathic pain (Burgess et al.,02), visceral pain (Zhuo and Gebhart, 2002), and inflam-tory pain (Guan et al., 2004; Sugiyo et al., 2005). Thus,re is emerging evidence that the vlPAG facilitates no-eption, and this facilitation is mediated through theM.

nical significance

e use of TENS can therefore be thought of as a non-armacological tool to engage our endogenous analgesictem. It utilizes endogenous opioids acting on their re-ptors to produce analgesia without side effects normallyserved with exogenous opioids. Early clinical studiesow that low frequency TENS utilizes opioid receptors toduce analgesia (Sjolund and Eriksonn, 1979). Furtherh frequency TENS increases the concentration of -en-rphins increase in the bloodstream and cerebrospinalid, and methionineenkephalin in the cerebrospinalid, in human subjects (Han et al., 1991; Salar et al.,81). Together, these clinical studies in human subjects,ng with animals studies on mechanisms support a roleactivation of opioid-mediated analgesia utilizing theGRVM pathway for both high and low frequencyNS. Since TENS utilizes known pharmacological path-ys, particularly opioids, TENS should be administeredh similar principles. Clinicians should be aware of thetential for the development of tolerance, as well as po-tial interactions of TENS with the patients pharmaco-ical therapy. For example, if subjects have been takingioids long enough to develop tolerance then low fre-ency TENS, which utilizes -opioid receptors, should be

oided. Preclinical studies in rats support this since lowquency TENS is ineffective in rats that were made pre-usly tolerant to opioids (Sluka et al., 2000). Further,mbining pharmacological agents with TENS could en-nce the effectiveness of the treatment, and reduce sideects of the drug. For example combining morphine ornidine with TENS enhances the analgesic effect so thatlower dose of the drug produces a similar degree ofalgesia (Sluka, 2000; Sluka and Chandran, 2002). Thus,ure studies should examine the effects of combiningmmon pharmaceutical agents for treatment of pain withNS in both animal and human subjects, and should beed at developing mechanisms to prevent tolerance

ingne and Sluka, 2008; DeSantana et al., 2008).

knowledgmentThis study is supported by a competitive grantm EMPI, Inc. and National Institutes of Health AR052316. Nomercial party having a direct financial interest in the results ofresearch supporting this article has or will confer a benefitn the authors or upon any organization with which the authorsassociated.

REFERENCES

ley EE, Chieng BC, Christie MJ, Connor M (2005) Opioid tolerancein periaqueductal gray neurons isolated from mice chronicallytreated with morphine. Br J Pharmacol 146(1):6876.baum AI, Fields HL (1984) Endogenous pain control systems:brainstem spinal pathways and endorphin circuitry. Annu Rev Neu-rosci 7:309338.behani MM (1995) Functional characteristics of the midbrain peri-aqueductal gray. Prog Neurobiol 46:575605.gess SE, Gardell LR, Ossipov MH, Malan TP Jr, Vanderah TW, LaiJ, Porreca F (2002) Time-dependent descending facilitation fromthe rostral ventromedial medulla maintains, but does not initiate,neuropathic pain. J Neurosci 22(12):51295136.nnon JT, Prieto GJ, Lee A, Liebeskind JC (1982) Evidence foropioid and non-opioid forms of stimulation-produced analgesia inthe rat. Brain Res 243:315321.stiglioni AJ, Gallaway MC, Coulter JD (1978) Spinal projectionsfrom the midbrain in monkey. J Comp Neurol 178:329346.vun S, Goktalay G, Millington WR (2004) The hypotension evokedby visceral nociception is mediated by delta opioid receptors in theperiaqueductal gray. Brain Res 1019(12):237245.andran P, Sluka KA (2003) Development of opioid tolerance withrepeated transcutaneous electrical nerve stimulation administra-tion. Pain 102(12):195201.eng ZF, Fields HL, Heinricher MM (1986) Morphine microinjectedinto the periaqueductal gray has differential effects on 3 classes ofmedullary neurons. Brain Res 375(1):5765.Santana JM, Santana-Filho VJ, Sluka KA (2008) Modulation be-tween high- and low-frequency transcutaneous electric nerve stim-ulation delays the development of analgesic tolerance in arthriticrats. Arch Phys Med Rehabil 89(4):754760.lds HL, Heinricher MM, Mason P (1991) Neurotransmitters in no-ciceptive modulatory circuits. Annu Rev Neurosci 14:219245.k GD, Wyss JM (2000) Descending projections of infralimbic cortexthat mediate stimulation-evoked changes in arterial pressure.Brain Res 859(1):8395.bhart GF, Jones SL, Besson JM (1988) Effects of morphine given inthe brain stem on the activity of dorsal horn nociceptive neurons.In: Progress in brain research: pain modulation (Fields HL, ed), pp229243. Amsterdam: Elsevier.palkrishnan P, Sluka KA (2000) Effect of varying frequency, inten-sity, and pulse duration of transcutaneous electrical nerve stimu-

lation on primary hyperalgesia in inflamed rats. Arch Phys MedRehabil 81(7):984990.

Gu

Gu

Ha

Ha

He

Hin

Ho

Jac

Jen

Jen

Kal

Kie

Kin

Kre

Krz

Kuy

Lew

Ma

Ma

Ma

Mo

Mo

Mo

Mo

Mo

Mo

Mo

Nu

Oliv

Os

Paj

Pax

Prie

Ra

Ra

Re

Re

Ro

Sal

San

J. M. DeSantana et al. / Neuroscience 163 (2009) 123312411240an Y, Guo W, Robbins MT, Dubner R, Ren K (2004) Changes inAMPA receptor phosphorylation in the rostral ventromedial me-dulla after inflammatory hyperalgesia in rats. Neurosci Lett 366(2):201205.o W, Robbins MT, Wei F, Zou S, Dubner R, Ren K (2006) Supraspi-nal brain-derived neurotrophic factor signaling: a novel mechanismfor descending pain facilitation. J Neurosci 26(1):126137.giwara S, Byerly L (1981) Calcium channel. Annu Rev Neurosci4:69125.n JS, Chen XH, Sun SL, Xu XJ, Yuan Y, Yan SC, Hao JX, TereniusL (1991) Effect of low and high frequency TENS on met-enkepha-lin-arg-phe and dynorphin a immunoreactivity in human lumbarCSF. Pain 47:295298.inricher MM, Martenson ME, Neubert MJ (2004) Prostaglandin E2in the midbrain periaqueductal gray produces hyperalgesia andactivates pain-modulating circuitry in the rostral ventromedial me-dulla. Pain 110(12):419426.gne PM, Sluka KA (2008) Blockade of NMDA receptors preventsanalgesic tolerance to repeated transcutaneous electrical nervestimulation (TENS) in rats. J Pain 9(3):217225.sobuchi Y, Adams JE, Linchitz R (1977) Pain relief by electricalstimulation of the central gray matter in humans and its reversal bynaloxone. Science 197:183186.quet YF, Lajtha A (1976) The periaqueductal gray: site of morphineanalgesia and tolerance as shown by 2-way cross tolerance be-tween systemic and intracerebral injections. Brain Res 103:501513.sen TS, Yaksh TL (1989) Comparison of the antinociceptive effectof morphine and glutamate at coincidental sites in the periaque-ductal gray and medial medulla in rats. Brain Res 476(1):19.sen TS, Yaksh TL (1986) Comparison of the antinociceptive actionof mu and delta opioid receptor ligands in the periaqueductal graymatter, medial and paramedial ventral medulla in the rats as stud-ied by the microinjection technique. Brain Res 372:301312.ra A, Urban MO, Sluka KA (2001) Blockade of opioid receptors inrostral ventral medulla prevents antihyperalgesia produced bytranscutaneous electrical nerve stimulation (TENS). J PharmacolExp Ther 298(1):257263.fel JM, Rossi GC, Bodnar RJ (1993) Medullary mu and delta opioidreceptors modulate mesencephalic morphine analgesia in rats.Brain Res 624(12):151161.g EW, Audette K, Athman GA, Nguyen HO, Sluka KA, FairbanksCA (2005) Transcutaneous electrical nerve stimulation activatesperipherally located alpha-2A adrenergic receptors. Pain 115(3):364373.tz R (1984) Local cobalt injection: a method to discriminate pre-synaptic axonal from postsynaptic neuronal activity. J NeurosciMethods 11(2):129135.anowska EK, Bodnar RJ (1999) Morphine antinociception elicitedfrom the ventrolateral periaqueductal gray is sensitive to sex andgonadectomy differences in rats. Brain Res 821:224230.pers HGJM, Maisky VA (1975) Retrograde axonal transport ofhorseradish peroxidase from spinal cord to brainstem cell groups incats. Neurosci Lett 1:914.is VA, Gebhart GF (1977) Morphine-induced and stimulation-pro-duced analgesias at coincident periaqueductal central gray loci:evaluation of analgesic congruence, tolerance, and cross-toler-ance. Exp Neurol 57:934955.YT, Sluka KA (2001) Reduction in inflammation-induced sensiti-zation of dorsal horn neurons by transcutaneous electrical nervestimulation in anesthetized rats. Exp Brain Res 137(1):94102.eda Y, Lisi TL, Vance CG, Sluka KA (2007) Release of GABA andactivation of GABA(A) in the spinal cord mediates the effects ofTENS in rats. Brain Res 1136(1):4350.ntyh PW, Peschanski M (1982) Spinal projections from the periaq-

ueductal grey and dorsal raphe in the rat, cat and monkey. Neu-roscience 7(11):27692776.rgan MM, Clayton CC, Boyer-Quick JS (2005a) Differential suscep-tibility of the PAG and RVM to tolerance to the antinociceptiveeffect of morphine in the rat. Pain 113:9198.rgan MM, Depaulis A, Liebeskind JC (1987) Diazepam dissociatesthe analgesic and aversive effects of periaqueductal gray stimula-tion in the rats. Brain Res 423(12):395398.rgan MM, Tierney BW, Ingram SL (2005b) Intermittent dosingprolongs tolerance to the antinociceptive effect of morphine micro-injection into the periaqueductal gray. Brain Res 1059:173178.rgan MM (1991) Differences in antinociception evoked from dorsaland ventral regions of the caudal periaqueductal gray matter. In:The midbrain periaqueductal gray mattered (DePaulis A, BandlerR, eds), pp 139150. New York: Plenum.rgan MM, Whitney PK (2000) Immobility accompanies the antino-ciception mediated by the rostral ventromedial medulla of the rat.Brain Res 872(12):276281.rgan MM, Whitney PK, Gold MS (1998) Immobility and flight asso-ciated with antinociception produced by activation of the ventraland lateral/dorsal regions of the rat periaqueductal gray. Brain Res804:159166.rgan MM, Sohn JH, Liebeskind JC (1989) Stimulation of the peri-aqueductal gray matter inhibits nociception at the supraspinal aswell as spinal level. Brain Res 502:6166.seir K, Heidenreich BA, Proudfit HK (1999) The antinociceptionproduced by microinjection of a cholinergic agonist in the ventro-medial medulla is mediated by noradrenergic neurons in the A7catecholamine cell group. Brain Res 822:17.eras JL, Besson JM, Guilbaud G, Liebeskind JC (1974) Behavioraland electrophysiological evidence of pain inhibition from midbrainstimulation in the cat. Exp Brain Res 20:3244.borne PB, Vaughan CW, Wilson HI, Christie MJ (1996) Opioidinhibition of rat periaqueductal grey neurones with identified pro-jections to rostral ventromedial medulla in vitro. J Physiol 490(2):383389.olla GP, Tavares RF, Pelosi GG, Corra FM (2005) Involvement ofthe periaqueductal gray in the hypotensive response evoked byL-glutamate microinjection in the lateral hypothalamus of unanes-thetized rats. Auton Neurosci 122(12):8493.inos G, Watson SJ (2005) The rat brain, in stereotaxic coordinates.San Diego: Elsevier Academic Press.to GJ, Cannon JT, Liebeskind JC (1983) N. raphe magnus lesionsdisrupt stimulation-produced analgesia from ventral but not dorsalmidbrain areas in the rat. Brain Res 261:5357.dhakrishnan R, Moore SA, Sluka KA (2003) Unilateral carrageenaninjection into muscle or joint induces chronic bilateral hyperalgesiain rats. Pain 104(3):567577.dhakrishinan R, Sluka KA (2003) Spinal muscarinic receptors areactivated during low and high frequency TENS-induced antihyper-algesia in rats. Neuropharmacology 45:11111119.ichling DB, Basbaum AI (1991) Collateralization of periaqueductalgray neurons to forebrain or diencephalon and to the medullarynucleus raphe magnus in the rat. Neuroscience 42(1):183200.sende MA, Sabino GG, Cndido CR, Pereira LS, Francischi JN(2004) Local transcutaneoous electrical stimulation (TENS) effectsin experimental inflammatory edema and pain. Eur J Pharmacol504(3):217222.ssi GC, Pasternak GW, Bodnar RJ (1994) Mu and delta opioidsynergy between the periaqueductal gray and the rostro-ventralmedulla. Brain Res 665:8593.ar G, Job I, Mingrino S, Bosio A, Trabucchi M (1981) Effect oftranscutaneous electrotherapy on CSF -endorphin content in pa-tients without pain problems. Pain 10:169172.dkuhler J, Gebhart GF (1984) Relative contributions of the nucleusraphe magnus and adjacent medullary reticular formation to theinhibition by stimulation in the periaqueductal gray of a spinal

nociceptive reflex in the pentobarbital-anesthetized rat. Brain Res305:7787.

Siu

Sjo

Slu

Slu

Slu

Slu

Slu

183190.Sluka KA (2000) Systemic morphine in combination with TENS pro-

Slu

Spi

Sug

Ter

vitro. J Physiol 498(2):463472.Vera-Portocarrero LP, Xie JY, Kowal J, Ossipov MH, King T, Porreca

Yak

Zhu

Zhu

Zhu

7 June 2ine 2 Jul

J. M. DeSantana et al. / Neuroscience 163 (2009) 12331241 1241duces an increased antihyperalgesia in rats with acute inflamma-tion. J Pain 1(3):204211.ka KA, Westlund KN (1993) Behavioral and immunohistochemicalchanges in an experimental arthritis model in rats. Pain 55(3):367377.nella M, Cooper ML, Bodnar RJ (1996) Excitatory amino acidantagonists in the rostral ventromedial medulla inhibit mesence-phalic morphine analgesia in rats. Pain 64(3):545552.iyo S, Takemura M, Dubner R, Ren K (2005) Trigeminal transitionzone/rostral ventromedial medulla connections and facilitation oforofacial hyperalgesia after masseter inflammation in rats. J CompNeurol 493(4):510523.shner SA, Helmstetter FJ (2000) Antinociception produced by muopioid receptor activation in the amygdala is partly dependent onactivation of mu opioid and neurotensin receptors in the ventralperiaqueductal gray. Brain Res 865(1):1726.

(Accepted 2(Available onlF (2006) Descending facilitation from the rostral ventromedial me-dulla maintains visceral pain in rats with experimental pancreatitis.Gastroenterology 130(7):21552164.sh TL, Yeung JC, Rudy TA (1976) Systematic examination in therat of brain sites sensitive to the direct application of morphine:observation of differential effects within the periaqueductal gray.Brain Res 114:83103.o M, Gebhart GF (2002) Facilitation and attenuation of a visceralnociceptive reflex from the rostroventral medulla in the rat. Gas-troenterology 122(4):10071019.o M, Gebhart GF (1991) Spinal serotonin receptors mediate de-scending facilitation of a nociceptive reflex from the nuclei reticu-laris gigantocellularis and gigantocellularis pars alpha in the rat.Brain Res 550(1):3548.o M, Gebhart GF (1997) Biphasic modulation of spinal nociceptivetransmission from the medullary raphe nuclei in the rat. J Neuro-physiol 78(2):746758.

009)y 2009)ciak JA, Advokat C (1987) Tolerance to morphine microinjections inthe periaqueductal gray (PAG) induces tolerance to systemic, butnot intrathecal morphine. Brain Res 424:311319.lund BH, Eriksson MB (1979) The influence of naloxone on anal-gesia produced by peripheral conditioning stimulation. Brain Res173(2):295301.ka KA, Deacon M, Stibal A, Strissel S, Terpstra A (1999) Spinalblockade of opioid receptors prevents the analgesia produced byTENS in arthritic rats. J Pharmacol Exp Ther 289(2):840846.ka KA, Lisi TL, Westlund KN (2006) Increased release of serotoninin the spinal cord during low, but not high, frequency transcutane-ous electric nerve stimulation in rats with joint inflammation. ArchPhys Med Rehabil 87(8):11371140.ka KA, Bailey K, Bogush J, Olson R, Ricketts A (1998) Treatmentwith either high or low frequency TENS reduces the secondaryhyperalgesia observed after injection of kaolin and carrageenaninto the knee joint. Pain 77(1):97102.ka KA, Judge MA, McColley MM, Reveiz PM, Taylor BM (2000)Low frequency TENS is less effective than high frequency TENS atreducing inflammation-induced hyperalgesia in morphine-tolerantrats. Eur J Pain 4(2):185193.ka KA, Chandran P (2002) Enhanced reduction in hyperalgesia bycombined administration of clonidine and TENS. Pain 100(12):

Tershner SA, Mitchell JM, Fields HL (2000) Brainstem pain modulatingcircuitry is sexually dimorphic with respect to mu and kappa opioidreceptor function. Pain 85:153159.

Tortorici V, Morgan MM, Vanegas H (2001) Tolerance to repeatedmicroinjection of morphine into the periaqueductal gray is associ-ated with changes in the behavior of off- and on-cells in the rostralventromedial medulla of rats. Pain 89:237244.

Tortorici VT, Robbins CS, Morgan MM (1999) Tolerance to the antino-ciceptive effect of morphine microinjections into the ventral but notlateral-dorsal periaqueductal gray of the rat. Behav Neurosci113:833839.

Urban MO, Smith DJ (1994) Nuclei within the rostral ventromedialmedulla mediating morphine antinociception from the periaque-ductal gray. Brain Res 652(1):916.

Urban MO, Jiang MC, Gebhart GF (1996) Participation of centraldescending nociceptive facilitatory systems in secondary hyperal-gesia produced by mustard oil. Brain Res 737(12):8391.

Vance CG, Radhakrishnan R, Skyba DA, Sluka KA (2007) Transcu-taneous electrical nerve stimulation at both high and low frequen-cies reduces primary hyperalgesia in rats with joint inflammation ina time-dependent manner. Phys Ther 87(1):4451.

Vaughan CW, Christie MJ (1997) Presynaptic inhibitory action ofopioids on synaptic transmission in the rat periaqueductal grey in

TRANSCUTANEOUS ELECTRICAL NERVE STIMULATION AT BOTH HIGH AND LOW FREQUENCIES ACTIVATES VENTROLATERAL PERIAQUEDUCTAL GREY TO DECREASE MECHANICAL HYPERALGESIA IN ARTHRITIC RATSEXPERIMENTAL PROCEDURESInduction of inflammationCannula implantation and microinjectionsDrug administrationBehavioral assessmentAdministration of TENSExperimental designStatistical analysis

RESULTSEffects of CoCl2 microinjection into the vlPAG on withdrawal thresholdsDistribution of microinjection sites in the vlPAG and dlPAGEffects of the blockade of the vlPAG and dlPAG on TENS antinociception

DISCUSSIONOpioid analgesia in the PAGFacilitation from the PAGClinical significance

AcknowledgmentREFERENCES