Antidepressants suppress neuropathic pain by a peripheral β2-adrenoceptor mediated anti-TNFα...

1

2

3Q1

4

5

6789Q310

11

1213141516171819202122232425262728293031

50

51

52

53

54

55

56

57

58

59

60

Neurobiology of Disease xxx (2013) xxx–xxx

YNBDI-03029; No. of pages: 12; 4C: 5, 6, 9, 10

Contents lists available at ScienceDirect

Neurobiology of Disease

j ourna l homepage: www.e lsev ie r .com/ locate /ynbd i

Antidepressants suppress neuropathic pain by a peripheralβ2-adrenoceptor mediated anti-TNFα mechanism☆

RO

OFYohann Bohren a,b, Luc-Henri Tessier a, SalimMegat a,b, Hugues Petitjean a,b, Sylvain Hugel a, Dorothée Daniel a,

Mélanie Kremer a,b, Sylvie Fournel b,c, Lutz Hein d,e, Rémy Schlichter a,b, Marie-José Freund-Mercier a,b,Ipek Yalcin a,1, Michel Barrot a,⁎,1

a Institut des Neurosciences Cellulaires et Intégratives, Centre National de la Recherche Scientifique, 67084 Strasbourg, Franceb Université de Strasbourg, 67084 Strasbourg, Francec Laboratoire de Conception et Application de Molécules Bioactives, Centre National de la Recherche Scientifique, 67401 Illkirch, Franced Institute of Pharmacology, University of Freiburg, 79104 Freiburg, Germanye BIOSS Centre for Biological Signalling Studies, University of Freiburg, 79104 Freiburg, Germany

U

Abbreviations: Arbp, acidic ribosomal phosphoproteiCGRP, calcitonin gene related peptide; GS, glutamine synneurofilament 200; NP, neuropathic pain group; TCA,tumor necrosis factor α; TH, tyrosine hydroxylase.☆ Conflict of interest: The authors declare no competing⁎ Corresponding author at: Institut des Neuroscienc

UPR3212CNRS, 21 rueDescartes, 67084 Strasbourg cedex,E-mail address: [email protected] (M. BarroAvailable online on ScienceDirect (www.sciencedir

1 These authors contributed equally to this work.

0969-9961/$ – see front matter © 2013 Published by Elsehttp://dx.doi.org/10.1016/j.nbd.2013.08.012

Please cite this article as: Bohren, Y., et al.,mechanism, Neurobiol. Dis. (2013), http://d

Pa b s t r a c t
a r t i c l e i n f o
32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

Article history:Received 5 April 2013Revised 26 July 2013Accepted 14 August 2013Available online xxxx

Keywords:AllodyniaAntidepressantAdrenoceptorCytokineDorsal root gangliaSatelliteNeuropathyNoradrenergicPainTNFα

47
RECTED
Neuropathic pain is pain arising as a direct consequence of a lesion or disease affecting the somatosensorysystem. It is usually chronic and challenging to treat. Some antidepressants are first-line pharmacological treat-ments for neuropathic pain. The noradrenaline that is recruited by the action of the antidepressants on reuptaketransporters has been proposed to act through β2-adrenoceptors (β2-ARs) to lead to the observed therapeuticeffect. However, the complex downstream mechanism mediating this action remained to be identified. In thisstudy, we demonstrate in a mouse model of neuropathic pain that an antidepressant's effect on neuropathicallodynia involves the peripheral nervous system and the inhibition of cytokine tumor necrosis factor α(TNFα) production. The antiallodynic action of nortriptyline is indeed lost after peripheral sympathectomy,but not after lesion of central descending noradrenergic pathways. More particularly, we report thatantidepressant-recruited noradrenaline acts, within dorsal root ganglia, on β2-ARs expressed by non-neuronalsatellite cells. This stimulation of β2-ARs decreases the neuropathy-induced production of membrane-boundTNFα, resulting in relief of neuropathic allodynia. This indirect anti-TNFα action was observed with the tricyclicantidepressant nortriptyline, the selective serotonin and noradrenaline reuptake inhibitor venlafaxine and theβ2-AR agonist terbutaline. Our data revealed an original therapeutic mechanism that may open novel researchavenues for the management of painful peripheral neuropathies.

© 2013 Published by Elsevier Inc.

4849
R
61

62

63

64

65

66

67

68

69

70

NCO

Introduction

Neuropathic pain is defined as pain arising as a direct consequenceof a lesion or disease affecting the somatosensory system (Jensenet al., 2011). This complex syndrome implicates maladaptive changesin injured sensory neurons and along the entire nociceptive pathwaywithin the central nervous system (von Hehn et al., 2012). In the clinic,tricyclic antidepressants (TCAs) as well as selective serotonin andnoradrenaline reuptake inhibitors are recommended among first-linetreatments (Attal et al., 2006, 2010; Dworkin et al., 2007; Saarto and

71

72

73

74

75

76

77

78

79

80

n P0; β2-AR, β2-adrenoceptor;thetase; IL, interleukin; NF200,tricyclic antidepressant; TNFα,

financial interests.es Cellulaires et Intégratives,France. Fax:+333 8861 33 47.t).ect.com).

vier Inc.

Antidepressants suppress neux.doi.org/10.1016/j.nbd.2013.

Wiffen, 2007). In contrast, selective serotonin reuptake inhibitors arepoorly effective (Attal et al., 2006; Benbouzid et al., 2008a; Dworkinet al., 2007), suggesting that the noradrenergic component of antide-pressants has a key role in their action on neuropathic pain. Despitethe fact that the therapeutic effect of antidepressants is well document-ed, the precise mechanism by which neuropathic pain is alleviatedremains poorly understood (Mico et al., 2006).

Sustained antidepressant treatment is necessary to be effectiveagainst neuropathic pain, suggesting the recruitment of an indirectmechanism. Pharmacological and genetic approaches showed thatantidepressant-recruited noradrenaline selectively acts through β2-adrenoceptors (β2-ARs) to relieve neuropathic allodynia (Yalcin et al.,2009a, 2009b), and that repeated stimulation of these receptors bydirect agonists is sufficient to reach a therapeutic effect (Choucair-Jaafar et al., 2009, 2011; Yalcin et al., 2010). A clinical case reportappears to support the action of β2-mimetics against neuropathic pain(Cok et al., 2010).

To progress into the pain reliefmechanism, it is important to identifyits neuroanatomical substrate. A critical step is to determine the sourceof noradrenaline recruited by antidepressants. Indeed, noradrenaline

ropathic pain by a peripheral β2-adrenoceptor mediated anti-TNFα08.012

http://dx.doi.org/10.1016/j.nbd.2013.08.012

mailto:[email protected]


http://www.sciencedirect.com/science/journal/09699961


T

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

2 Y. Bohren et al. / Neurobiology of Disease xxx (2013) xxx–xxx

UNCO

RREC

can be released within supraspinal structures (El Mansari et al., 2010),at the spinal level by descending noradrenergic pathways (Millan,2002; Yoshimura and Furue, 2006), and at the peripheral level in dorsalroot ganglia following neuropathy-induced noradrenergic sprouting ofsympathetic nerve fibers (McLachlan et al., 1993; Ramer and Bisby,1998). Previous data with local administration of a β2-AR antagonistsuggested that the initial substrate for antidepressants' antiallodynicaction might be localized at the spinal cord and/or at the dorsal rootganglia level rather than at the supraspinal level (Yalcin et al., 2009b).

Although β2-ARs are critical for the antiallodynic action of antide-pressants, the downstream mechanism has not yet been identified.Experimental evidence supports a role of glial and/or immune cells inthe pathophysiology of neuropathic pain, particularly through therecruitment of cytokines (Austin and Moalem-Taylor, 2011; Thackeret al., 2007; Vallejo et al., 2010). Clinical studies also support cytokineimplication in neuropathic pain (Empl et al., 2001; Lindenlaub andSommer, 2003). It is however not knownwhether antidepressant treat-ment targets these neuroimmune actors.

In this study, we demonstrate that the peripheral nervous system isessential for the antiallodynic effect of nortriptyline. We show that thenoradrenaline recruited by this antidepressant acts within dorsal rootganglia, on β2-ARs expressed by non-neuronal satellite cells. We alsoshow that both nortriptyline and venlafaxine inhibit local TNFα produc-tion. These findings reveal a novel cellular and molecular substrate forthe antiallodynic action of antidepressants.

Material and methods

Animals

Experiments were done using male C57BL/6J mice (Charles River,L'Arbresle, France), or with mice lacking β2-ARs and their littermatecontrols. In the latter case, the mice were created in the laboratory ofBrian Kobilka (Stanford University, CA) and have been described previ-ously (Chruscinski et al., 1999). Heterozygotes (Adrb2+/−) were bred inour animal facilities, mice were genotyped upon weaning, and theexperiments were conducted on male Adrb2+/+ and Adrb2−/− litter-mate mice. All experiments started with 8- to 9-week-old mice. Theywere group-housed four to five per cage and kept under a 12-hourlight/dark cycle (lights on at 6:00 AM) with food and water ad libitum.All animals received proper care in agreement with the guidelines foranimal experimentation of the International Association for the Studyof Pain and the European Communities Council Directive 86/6609/EEC.The animal facilities are registered for animal experimentation withapproved procedures by Alsace regional veterinary office (agreementC67-482-1) and the scientists in charge of the experiments possessthe French legal certificate authorizing experimentation on livinganimals.

Neuropathy and nociceptive testing

The neuropathy was induced by cuffing the main branch of theright sciatic nerve (Benbouzid et al., 2008b; Mosconi and Kruger,1996). Surgeries were done under ketamine (17 mg/mL)/xylazine(2.5 mg/mL) anesthesia (i.p., 4 mL/kg) (Centravet, Taden, France). Thecommon branch of the right sciatic nerve was exposed and a 2-mmsection of split PE-20 polyethylene tubing (Harvard Apparatus, Les Ulis,France) was placed around it (neuropathic pain group: NP). The shavedskin was closed using suture. Sham-operated mice underwent the samesurgical procedure described without implantation of the cuff (Shamgroup). The mechanical allodynia was tested using von Frey hairs andresults were expressed in grams. Mice were placed in clear Plexiglasboxes (7 cm × 9 cm × 7 cm) on an elevated mesh screen. Calibratedvon Frey microfilaments (Bioseb, Vitrolles, France) were applied to theplantar surface of each hindpaw until they just bowed, in a series ofascending forces up to the mechanical threshold. Each filament was

Please cite this article as: Bohren, Y., et al., Antidepressants suppress neumechanism, Neurobiol. Dis. (2013), http://dx.doi.org/10.1016/j.nbd.2013.

ED P

RO

OF

tested five times per paw and the threshold was defined as the lower oftwo consecutive filaments for which three or more withdrawals out ofthe five trials were observed (Barrot, 2012; Bohren et al., 2010). Tostudy the antiallodynic effect of chronic drug treatments, the mice weretested in the morning, before drug injections when applicable, thusreflecting the therapeutic action of previous treatment days, as previouslydescribed (Bohren et al., 2010; Yalcin et al., 2009a, 2009b).

Treatment procedures

Treatments with the TCA nortriptyline, the selective serotonin andnoradrenaline reuptake inhibitor venlafaxine, the β2-mimetic terbuta-line, or the anti-TNFα etanercept or infliximab, began fifteen or sixteendays after the neuropathy was induced, and they were maintained foreither 2 or 3 weeks. During the treatments, themice received two injec-tions per day (morning and evening) of nortriptyline (5 mg/kg),venlafaxine (5 mg/kg) or terbutaline (0.5 mg/kg) (Sigma-Aldrich,Saint-Quentin Fallavier, France), one injection per day (morning,5 mg/kg) of etanercept (ENBREL®, Wyeth Pharmaceuticals, Blois,France), one injection every other day (2 mg/kg) of etanercept, or oneunique injection of infliximab (10 mg/kg, REMICADE®, Schering-Plough, Levallois-Perret, France). Drugs were dissolved in 0.9% NaClsolution that was also used for control injections, and theywere admin-istered intraperitoneally in a volume of 5 mL/kg. For cellular andmolec-ular experiments, the nervous tissues were collected five weeks post-surgery, i.e. after three weeks of treatment.

Noradrenergic lesions

Chemical noradrenergic lesion was done using guanethidinemonosulfate (ISMELIN®, Laboratoires Genopharm, Saint-Thibault-des-Vignes, France) (Neil et al., 1991). Three different paradigms were used,each targeting different anatomical sites. (1) A group of mice underwentfive daily intraperitoneal injections of guanethidine (30 mg/kg) in avolumeof 5 mL/kg; (2) another received three daily intrathecal injectionsat the lumbar level (30 μg in 10 μL); and (3) a last group received threedaily intrathecal injections at the thoracic level (30 μg in 10 μL). Intrathe-cal injections were performed under gaseous anesthesia (halothane 3%)as previously described. Briefly, for lumbar intrathecal administration, a27-gauge needle connected to a 50 μL Hamilton syringe was insertedbetween the L5 and L6 vertebrae, into the sub-arachnoidal space. Place-ment of the needlewas verified by the elicitation of a tail flickmovement.For thoracic administration, the needlewas inserted at the level of the lastthoracic vertebrae, just above the lower rib used asmark. All guanethidineadministrations were done two weeks before nerve injury. A controlgroup of non-lesionedmice underwent the same nerve cuffing procedureas the lesioned groups.

Immunostaining

Five weeks post-surgery, neuropathic and sham mice were deeplyanesthetized with pentobarbital and perfused intracardially. Lumbardorsal root ganglia (L4, L5 and L6) and spinal cord were dissected,postfixed, cryoprotected, embedded in OCT compound (Sakura Finetek,Villeneuve d'Ascq, France), frozen and cut into 14 μm thick sections thatwere then mounted on Superfrost®Plus slides (O. Kindler GmbH,Freiburg, Germany). To evaluate tyrosine hydroxylase (TH) expression,we used standardized procedures (Kaufling et al., 2010), incubating thesectionswith Sheep anti-mouse TH antibody (1:500, Millipore, AB1542,France Millipore, Molsheim, France). The sections were then incubatedwith secondary Cy3-conjugated antibodies done in donkey (1:400, Jack-son ImmunoResearch, Newmarket, UK). Thereafter, the sections werewashed, mounted with VECTASHIELD, and viewed under a Nikon E80imicroscope with ×10 and ×20 objectives, and images acquired withMBFBioscience camera CX9000using Picture Frame acquisition software(MBF Bioscience, Magdeburg, Germany). For TNFα detection, mice



202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250Q4

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270Q5

271

272

273

274

275

276

277

278

279

280

281

282

3Y. Bohren et al. / Neurobiology of Disease xxx (2013) xxx–xxx

received an acute infliximab injection (10 mg/kg, i.p.) at least twoweeksafter sciatic nerve surgery, and were perfused 6 h post-injection.Cy3-conjugated anti-human antibodies (Jackson ImmunoResearch,709-165-149) were used to detect infliximab. Costaining was doneeither by using Rabbit anti-glutamine synthetase (GS) (1:500, SigmaAldrich, G2781) for satellite cells, NeuroTrace™ fluorescent Nissl forneurons (1:200, Molecular Probes, N-21479), or by using Rabbit anti-calcitonin gene related peptide (CGRP) (1:20,000, Amersham,RPN1842), Rabbit anti-neurofilament 200 (NF200) (1:2000, SigmaAldrich, N4142) or Rabbit anti-substance P (1:15,000, gift from Dr.Tramu, Bordeaux) antibodies, or isolectin IB4 Alexa Fluor® 488 conjugate(2 μg/mL, Molecular Probes, I21411) for staining of different neuronalcell populations. When an additional fluorescent secondary antibodywas required, we used Alexa Fluor® 488-conjugated antibodies(Molecular Probes). Microphotographs for detection of infliximab onlywere done with a Nikon E80i microscope with ×4, ×10 and ×40 objec-tives, and images acquired with MBF Bioscience camera CX9000 usingPicture Frame acquisition software. Microphotographs for fluorescencedouble labeling were taken using a Leica confocal microscope TCS SP5II, under ×20 objective, and analyzed using the Leica application suiteadvanced fluorescence software. For all pictures, Adobe Photoshop CS3(Adobe, San José, CA, USA) was used to adjust contrast and brightness.

T283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

REC

Dorsal root ganglia cell culture

Dorsal root ganglia cells were prepared from adult mice. Dorsal rootganglia from lumbar levels were dissected out, collected in phosphate-buffered solution, and enzymatically dissociated for 25 min at 37 °Cwith trypsin–EDTA (0.5 g/L, Seromed) and 2 mg/mL of collagenase IA/dispase (Invitrogen). Enzymatic dissociation was stopped by the addi-tion of the culturemediumwhich consisted ofMEMα (minimumessen-tial medium alpha, Gibco, France) supplemented with 10% v/v heat-inactivated horse serum (Gibco) and 50 IU/mL penicillin–streptomycin(Gibco). The cells were then mechanically dissociated by triturationwith fire-polished Pasteur pipettes of decreasing tip diameter. Dorsalroot ganglia cells were plated on 15 mm glass coverslips previouslycoated with poly-D-lysine (0.02 mg/mL). Cultures were maintained at37 °C in a water-saturated atmosphere (95% air and 5% CO2). Cellcultures were suitable for single cell PCR or calcium imaging experi-ments and were respectively used between 2 and 5 h or between 12and 15 h after the seeding.

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

UNCO

R

Single cell PCR

Single-cell PCR experiments were conducted on dissociated dorsalroot ganglia cell culture of naïve mice. Cells were collected individuallyfrom the coverslip using a patch clamp pipette filled with 8 μL of buffersolution (mM): 140 KCl, 1MgCl2, 0.5 EGTA, pH 7.4. The glass pipette tipwas then broken off into a thin-walled PCR reaction tube containing40 μL of the reaction mix SuperScriptIII One-Step RT-PCR system(Invitrogen) and the first set of primers. This first set included couplesof primers for NF200 (accession number: NM_010904), β2-AR (accessionnumber: NM_007420) and GS (accession number: AY044241). 2 μL ofSuperScriptIII RT/Platinum Taq mix was added to each sample andcDNAs were synthesized by incubation at 50 °C for 30 min and the firstPCR consisted of an initial denaturation at 94 °C for 2 min followed by37 cycles (94 °C for 30 s, 52 °C for 45 s, 68 °C for 45 s) and a final exten-sion at 68 °C for 5 min. The resulting product was diluted 1:100 and re-amplified by Taq PCR (Invitrogen) using a single couple of nested primersfor each specific gene studied. The second amplification consisted of aninitial denaturation step at 94 °C for 2 min followed by 35 cycles (94 °Cfor 30 s, 55 °C for 45 s, 72 °C for 45 s) and a final extension at 72 °C for5 min. One-fifth of the PCRproductwas run on an agarose gel and stainedwith ethidium bromide.


ED P

RO

OF

Calcium imaging

Calcium imaging experiments were conducted on dissociated dorsalroot ganglia cell culture of sham and neuropathic mice. During calciummeasurements, the cells were continuously superfused with a salinesolution containing (mM): 135 NaCl, 5 KCl, 2.5 CaCl2, 1 MgCl2, 10glucose, 5 HEPES, pH 7.4. High KCl solution had the same composition,except: KCl, 50 mM and NaCl, 90 mM. The whole dish was perfusedwith saline solution, and the imaged field was perfused by a single-tipmultichannel gravity-fed system, allowing a switch between variousperfused solutions. Cells were loadedwith Fura-2 by 60-minute incuba-tion at room temperature in the recording solution with 2 μM Fura-2acetoxymethyl ester (Fura-2/AM; Molecular Probes) and 0.001% (w/v)pluronic acid (Molecular Probes). Cells were washed three times withsaline solution before and after loading. Fluorescence measurementson individual cellswere performed on an invertedmicroscope (Axiovert35; Zeiss) with an oil-immersion ×40 objective (Fluor 40, NA 1.30;Nikon) using a quantitative real-time imaging system comprising acooled CCD camera (CoolSnap HQ, Photometrics) and an image analysissoftware package (Imaging Workbench 4.0 Software, Axon Instru-ments). Fluorescence was alternately excited at wavelengths of 350and 380 nm with a lambda-10 filter wheel (Sutter Instruments), andemitted light was collected above 520 nm. Pairs of images wereacquired every 2 s. Intracellular calcium is expressed throughout asthefluorescence ratio F350/F380, calculated after background subtraction.Experiments were performed at room temperature. All drugs wereprepared as 1000× concentrated stock solutions. ATP (50 μM), ADP(50 μM), UTP (50 μM) and terbutaline (100 μM) were diluted intosaline solution immediately before use.

Real-time quantitative PCR

Quantification of gene expression in dorsal root ganglia in eachgroup was assessed with a real-time quantitative PCR method. TotalRNA was extracted using RNeasy (QIAGEN), and cDNA was generatedwith High capacity RNA-to-cDNA master mix from Applied Biosystems.Quantitative real-time PCR was performed on a StepOne Real-TimePCR system (Applied Biosystems) using SYBR Green PCR Master Mixassay (Applied Biosystems). The amplification reaction was performedfor 40 cycles with denaturation at 95 °C for 10 min, followed byannealing at 95 °C for 15 s, and extension and detection at 60 °C for1 min. All experiments were done with triplicate sample deposits onthe amplification plates. The relative RNA abundance of each targetgene transcript was normalized against endogenous gene control acidicribosomal phosphoprotein P0 (Arbp). Data were analyzed according torelative gene expression by 2−ΔΔCt method, where Ct represents thethreshold cycle for each transcript (Schmittgen and Zakrajsek, 2000).Primers (designed from Mus musculus gene data bank using AppliedBiosystems Primer Express Software version 3.0) are as follows: Arbp(or Rp1p0, acidic ribosomal phosphoprotein): 5′-GCCAGCTCAGAACACTGGTCTA-3′ and 5′-ATGCCCAAAGCCTGGAAGA-3′; Tnf (TNFα, tumornecrosis factor): 5′-CAGCCGATGGGTTGTACCTT-3′ and 5′-GGCAGCCTTGTCCCTTGA-3′; Il10 (IL-10, interleukin10): 5′-GATGCCCCAGGCAGAGAA-3′ and 5′CACCCAGGGAATTCAAATGC-3′; Tnfrsf1a (TNF-R1, tumornecrosis factor receptor 1): 5′-TCCGCTTGCAAATGTCACA-3′ and 5′-GGCAACAGCACCGCAGTAC-3′; Adbr2 (β2-AR): 5′-TGGGAACGGCTACTCTAGCAA-3′ and 5′-TGTTTGGCTCCCCTGTGTAGT-3′; Ptgs2 (Cox-2, pros-taglandin G/H synthase 2): 5′-TTCGGGAGCACAACAGAGTGT-3′ and 5′-GCTCATCACCCCACTCAGGAT-3′; Il1a (IL-1α, interleukin 1 alpha): 5′-GGAGAGCCGGGTGACAGTATC-3′ and 5′-TCAGCCGTCTCTTCTTCAGAATC-3′; Il1B (IL-1β, interleukin 1 beta): 5′-AGTTGACGGACCCCAAAAGA-3′ and 5′-GGACAGCCCAGGTCAAAGG-3′; and Il6 (IL-6, interleukin 6):5′-CCACGGCCTTCCCTACTTC-3′ and 5′-TTGGGAGTGGTATCCTCTGTGA-3′.

Two procedures were used depending on the set of experiments.One was done with a total of n = 9 mice per group (Fig. 4B). To limit



TED

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349Q6

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

Sham, Sal NP, Sal

Sham, Treatment NP, Treatment

Pre

ssur

e (g

)

Time (days)0 6 12 18 24 30 36

0

1

23

45

67

Sal or Nor

0

1

23

45

67

Time (days)

0 6 12 18 24 30 36

Sal or Ter

Pre

ssur

e (g

)

Time (days)0 6 12 18 24 30 36

0

1

23

45

67

Sal or Nor

* * * * * * *

Sal or Ter

Time (days)

0 6 12 18 24 30 360

1

23

45

67

A

B

Contra

Contra

Ipsi

Ipsi

* * * * * * *

++ +

+ + +

Fig. 1. Chronic nortriptyline or terbutaline treatments relieve neuropathic allodynia. Twoweeks after nerve injury (NP, neuropathic pain group), treatment with the TCA nortripty-line (Nor, 5 mg/kg, twice a day), the β2-AR agonist terbutaline (Ter, 0.5 mg/kg, twice aday) or their saline control (0.9% NaCl) started and was maintained for 3 weeks. Thehindpaw mechanical threshold was tested using von Frey filaments. (A) Long-termnortriptyline treatment did not affect the threshold of the contralateral paw, but itsuppressed the ipsilateral neuropathy-induced allodynia. (B) Similar results wereobserved with terbutaline. (Data from one experiment with n = 6 mice per group,*p b 0.001 vs. Sham, +p b 0.05 NP-treated vs. NP-Saline.) Data are expressed asmean ± SEM.


UNCO

RREC

potential interindividual differences due to eithermice or the dissectionprocedure, we pooled samples from the 3 lumbar dorsal root ganglia of3 mice from the same experimental group (i.e. 9 dorsal root ganglia).These pooled samples were processed for real-time PCR, each samplebeing present as triplicate on the plate. We repeated the experimenttwo other times, with pooled sets of 3 other animals each time. Theother sets of experiments were conducted on samples from individualmice. All individual samples were compared to a reference composedof a mix of the controls' cDNAs and results were standardized bythe 2−ΔΔCt method as described above.

Immunoblotting analysis

Due to differences in optimized extraction procedures, real-timequantitative PCR and Western blot were always done on independentsets of mice. For Western blot, total protein was extracted in 150 μLlysis buffer (20 mM Tris pH 7.5; 150 mM NaCl; 10% glycerol; 1%NP40; Protease Inhibitors Cocktail, Roche). The proteinswere quantitat-ed with DC protein assay kit (Bio-Rad) and 15 μg of total protein fromindividual animalswas resolved by12% SDS-polyacrylamide gel electro-phoresis under reducing conditions and then transferred to PVDFmembrane. The blots were incubated for 1 h in blocking agent (ECLkit, Amersham Biosciences) and subsequently, overnight with thespecific antibodies for TNFα (1:1000, R&D Systems, AF-410-NA) andβ-tubulin (1:5000, Santa Cruz Biotechnology Inc., sc-9935) followedby anti-goat HRP-conjugated secondary antibodies (1:5000, Chemicon,AP307P and AP106P). Blotswere evaluated by chemiluminescence (ECLAdvance Western Blotting Detection System Kit, Amersham Biosciences,RPN 2135) using Hyperfilm substrates (Amersham Biosciences, RPN1674K). Relative protein expression was determined using the densi-tometry tool of Adobe Photoshop CS3 software.

ELISA assays

Commercially available ELISA reagents were used to evaluateconcentrations of TNFα (BD OptEIA™, BD Biosciences, San Diego, CA).All procedures were performed following the manufacturer's instruc-tions. Results were expressed as cytokine concentration in pg/mL. Thedetection limit was 15 pg/mL.

Statistics

Data are expressed asmean ± standard error of themean (SEM). Sta-tistical analyses were performed with STATISTICA 8 software (StatSoft,Tulsa, OK, USA), using multifactor analysis of variance (ANOVA). Thesurgery procedure (sham or cuff) and the treatments (saline vs. druginjections) were taken as between-group factors. When needed, thetime of measurement (time course data) was taken as a within-subjectfactor. The Duncan test was used for post hoc comparisons. In calciumimaging experiments, the fraction of cells answering or not to terbutalineapplication in sham and neuropathic conditions were compared by acontingency table with an Exact Fischer's test. Finally, for quantitativePCR and immunoblotting experiments, statistical analyses wereperformedwith anon-parametric test of Kruskal–Wallis and comparisonsbetween groups were done by U Mann–Whitney test. The significancelevel was set at p b 0.05.

Results

Antidepressant action on neuropathic pain requires peripheral noradrenaline

To mimic human neuropathy resulting from a trauma of peripheralnerves, we used chronic sciatic nerve cuffing in mice (Benbouzid et al.,2008b; Mosconi and Kruger, 1996). Mechanical allodynia is one of thesymptoms distressing the patients. In the neuropathic mice, the ipsilat-eral allodynia appears on the first day post-surgery and persists over


PRO

OF

three months (Benbouzid et al., 2008b). This allodynia is relieved bychronic treatment with the TCA nortriptyline or the β2-AR agonistterbutaline (Benbouzid et al., 2008a; Choucair-Jaafar et al., 2009)(Fig. 1) (Nortriptyline, right paw: F7,140 = 3.77, p b 0.001; post-hoc:NP-Nortriptyline N NP-Saline at p b 0.0001 on post-surgery days 28 to35. Terbutaline, right paw: F7,140 = 3.09, p b 0.005; post-hoc:NP-Terbutaline N NP-Saline at p b 0.0001 on post-surgery days 28 to35). As observed in other neuropathic pain models (McLachlan et al.,1993; Ramer and Bisby, 1998), the cuff insertion also induced asprouting of noradrenergic fibers in the lumbar dorsal root ganglia(Fig. 2A).

Since preclinical (Yalcin et al., 2009a) and clinical (Krell et al., 2005)data showed that thenoradrenergic system is essential for the therapeuticaction of antidepressant drugs, we investigated the possible source ofendogenous noradrenaline which is targeted by antidepressant drugs toalleviate neuropathic allodynia. For this purpose, noradrenergic lesionswere done twoweeks before induction of the neuropathy, with guaneth-idinewhich does not cross the blood–brain barrier (Neil et al., 1991). Thisdrug enters noradrenergic terminals through the noradrenaline trans-porter. While guanethidine at low dose is clinically used for peripheraldepletion of noradrenaline storage sites, larger doses of guanethidinelead to the selective destruction of sympathetic fibers. This depletion ofnoradrenergic fibers did not affect per se the basal mechanical sensitivityor the neuropathic allodynia (Fig. 2B). However, both the intraperitonealand the lumbar intrathecal injections of guanethidine prevented theantiallodynic action of nortriptyline (Fig. 2C) (Intraperitoneal: F7,175 =2.72, p b 0.015; post-hoc: (NP-Nortriptyline = NP-Saline) b Sham-Saline at p b 0.0001 on post-surgery days 3 to 35. Lumbar: F7,175 =2.47, p b 0.02; post-hoc: (NP-Nortriptyline = NP-Saline) b Sham-



ORRECTED P

RO

OF

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

Sham, Sal NP, Sal

Sham, Treatment NP, Treatment

Peripheral (Guanethidine i.p.)

Lumbar (Guanethidine i.t.)

Sal or Nor

Peripheral (Guanethidine i.p.)

Thoracic (Guanethidine i.t.)

36Time (days)

0 6 12 18 24 30

Sal or Nor

Pre

ssur

e (g

) Sal or Nor

01234567 Sal or Ter

* * * * * * * * * * * * * ** * * * * * * * * * * * * *01234567

01234567

01234567

36Time (days)

0 6 12 18 24 3036Time (days)

0 6 12 18 24 3036Time (days)

0 6 12 18 24 30controlsguanethidine, i.p.guanethidine, lumbarguanethidine, thoracic

Baseline

NP(2 weeks)

Pre

ssur

e (g

)

01234567

Pre

ssur

e (g

)

01234

Lumbar DRG

Lumbarspinal cord

NPlumbar lesion

NPnon lesioned

NPperipheral lesion

Shamnon lesioned

NPthoracic lesion

+ + + + +

A

B C

THTHTHTHTH

THTHTHTHTH

DRG DRG DRG DRG DRG

SC SC SC SC SC

Fig. 2. Peripheral noradrenergic fibersmediate allodynia relief by antidepressant drug. Guanethidine injections allowed lesioning the noradrenergic fibers. (A) Tyrosine hydroxylase (TH)immunostaining in the lumbar dorsal root ganglia (DRG; scale bars: 100 μm) and the spinal cord (SC; scale bars: 50 μm). The peripheral neuropathy (NP) induced a sympathetic sproutingin the DRG (white arrows). Intraperitoneal guanethidine (peripheral lesion) suppressed these fibers without affecting the spinal cord ones. Lumbar intrathecal guanethidine suppressedboth DRG and dorsal horn noradrenergic fibers. Thoracic intrathecal guanethidine suppressed descending noradrenergic fibers, sparing DRG sprouting. (B) Twoweeks after guanethidineinjections, no difference was observed between control mice and mice that received intraperitoneal, lumbar intrathecal or thoracic intrathecal guanethidine (baseline, top graph).Similarly, these injections of guanethidinedid not affect theneuropathic allodynia (NP at 2 weeks, bottom graph). (C) Theperipheral and the lumbar intrathecal injections of guanethidine,but not the thoracic intrathecal one, prevented the antiallodynic action of the TCA nortriptyline (Nor). The antiallodynic action of the β2-AR agonist terbutaline (Ter) was still present inmice which received peripheral guanethidine. (Data pooled from 2 independent experiments, n = 5–9 per group, *p b 0.001 vs. Sham, +p b 0.05 NP-treated vs. NP-Saline).


UNC

Saline at p b 0.0001 on post-surgery days 3 to 35). While intraperito-neal guanethidine induced a sympathectomy, suppressing peripher-al noradrenergic fibers only, the lumbar intrathecal administrationaffected both the spinal cord and noradrenergic fibers in the lumbardorsal root ganglia nearby the guanethidine injection site (Fig. 2A).This suggests a critical role of dorsal root ganglia noradrenergicsprouting in the TCA action, but it does not exclude a contributionof spinal descending pathways. To address this question, spinalnoradrenergic fibers were suppressed without affecting the lumbardorsal root ganglia sprouting, by doing thoracic intrathecal adminis-tration of guanethidine. This led to suppression of descending norad-renergic fibers, including at the lumbar level (Fig. 2A). In this case theantiallodynic action of nortriptylinewasmaintained (Fig. 2C) (Thoracic:F7,168 = 6.55, p b 0.0001; post-hoc: NP-Nortriptyline N NP-Saline atp b 0.005 on post-surgery days 28 to 35). Together, these results indi-cate that noradrenergic sprouting in the dorsal root ganglia is a criticalsubstrate for the antiallodynic action of a TCA.


In order to test the location of β2-ARs in the therapeuticmechanism,we tested terbutaline in sympathectomized mice. This β-mimeticremained effective in the absence of sympathetic fibers in the dorsalroot ganglia (Fig. 2C) (F7,147 = 4.14, p b 0.0005; post-hoc: NP-Terbutaline N NP-Saline at p b 0.0001 on post-surgery days 32 and35). It showed that the implicated β2-ARswere neither on noradrenergicterminals nor upstream of the noradrenergic system; thus confirming adownstream, postsynaptic role of β2-ARs.

β2-ARs are located on non-neuronal dorsal root ganglia cells

Antidepressant-recruited noradrenaline acts on β2-ARs to relieveneuropathic allodynia (Yalcin et al., 2009a, 2009b). To identify thetype of cells expressing mRNA encoding β2-ARs in dorsal root ganglia,we performed single cell PCR on dissociated dorsal root ganglia cellculture (Fig. 3A). NF200 neurofilament was used as a neuronal marker(Ho and O'Leary, 2010) and GS as a marker of satellite glial cells in the



443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479


dorsal root ganglia (Hanani, 2005). About a third of the tested cells werepositive for β2-AR mRNA, but they were also positive for both NF200and GS (Fig. 3A), suggesting that individual dorsal root ganglia neuronscould not be dissociated from the satellite cells tightly covering them(Hanani, 2005). We overcame this problem by using a functionalapproach with calcium imaging on dorsal root ganglia cell culture.High KCl was used as neuronal activator, a response to ADP and/orUDP in the absence of response to KCl allowed identifying non-neuronal cells, and terbutaline was used as selective β2-AR agonist. Insham mice, around 3% of non-neuronal cells expressed functional β2-ARs, whereas 25% of non-neuronal cells from neuropathic animalsdisplayed a calcium response to terbutaline (Figs. 3B, C) (Fisher's exacttest: p b 0.002). The pathological condition thus increased functionalβ2-ARs in dorsal root ganglia, without altering mRNA expression ofthese receptors (Fig. 3D). Contrasting with these non-neuronal cells,neurons from either sham or neuropathic mice never responded toterbutaline application (Fig. 3C), suggesting that functional β2-ARsincreasing intracellular calcium concentrations are selectively locatedin the plasma membrane of non-neuronal cells.

UNCO

RRECT

A

Cell proportion

NF200+

GS+

NF200+

GS+

Adrb2 +

Adrb2 -

31%(22/70)

69%(48/70)

M

NF200Adrb2

GS

M1

1.4

0 200 400 600 800 1000

0.8

1

1.2

1.6

Time (s)

R=

F 350

/F38

0

C

2 3 C

KCl(4 s)

ADP(10 s)

Ter(20 s)

Res

pond

to T

er (

% o

f cel

l)

0

10

20

30

40

50Sh

N

0/88

non neuronaneuron (N)

Fig. 3.β2-Adrenoceptors are on dorsal root ganglia non-neuronal cells. (A) Single cell PCR analy3 cells; C, negative control). (B) Calcium imaging on dissociated dorsal root ganglia cell cultureKCl application, and after 20 second terbutaline (Ter) (N, neuron; nnc, non-neuronal cell; scalechange influorescence emission during bath application of high KCl (50 mM), ADP (50 μM) or tneuronal cells, and in a higher proportion of cells in dorsal root ganglia culture from cuff mice (*quantitative real-time PCR (n = 3 × 3 per group). (Single cell PCR data are pooled from 8 indeReal-time PCR data represent 3 independent experiments, each of them pooling samples from


OF

Antidepressants and a β2-AR agonist display anti-TNFα action

Cytokines produced by non-neuronal cells, such as glial and immunecells, participate in the development and maintenance of neuropathicpain (Austin and Moalem-Taylor, 2011; Thacker et al., 2007; Vallejoet al., 2010).

By Western blot (p b 0.001) and mRNA analysis (p b 0.05), weobserved increased levels of TNFα in the lumbar dorsal root ganglia ofC57BL/6J neuropathic mice at 5–6 weeks post-injury (Fig. 4A). Tofurther examine the role of neuroinflammatorymediators in the neuro-pathic pain treatment by antidepressants or by a β2-AR agonist, geneexpression of TNFα, IL-10, Cox-2, IL-1 and IL-6was analyzed in the lum-bar dorsal root ganglia. The implication of some of these mediators iswell established in the early phase following nerve injury (Lee et al.,2004; Uceyler et al., 2007). However, their role in the dorsal root gangliais less clear atmuch later time-points, when chronic neuropathic pain iswell installed. Five weeks post-injury, we observed a preferential rolefor local TNFα. Indeed, an increased mRNA expression of TNFα wasobserved in neuropathic mice (Fig. 4B) (H3,12 = 9.72, p b 0.05),

ED P

RO

B

2

mR

NA

exp

ress

ion

0

1

Adrb2

NP, Nor

Sham, Sal

NP, Sal

R

0,8 1,2 1,5

nnc N nnc N

nnc NBF 350 nm

KCl Ter

nnc N

ns

Dam NP

nncnnc N

0/762/68

16/64*

l cell (nnc)

qPCR

sis for NF200,β2-AR (Adrb2) and GSmRNAs (M,molecularweight ladder; 1–3, examples of, example of a typical field in bright field (BF), at 350 nm fluorescence, after 4 second highbar: 20 μm). (C) Example of typical traces of calcium transients, represented as ratiometricheβ2-AR agonist terbutaline (100 μM). Response to terbutalinewas only observed in non-p b 0.002 vs. Sham-nnc). (D) β2-ARmRNA expression in dorsal root ganglia, evaluated bypendent experiments. Calcium imaging data are pooled from 6 independent experiments.3 animals per group. ns, not significant.)



CO

RRECTED P

RO

OF

480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

508

509

Control(saline)

LPS(20 mg/kg)

NP(5 weeks)

non detectable

non detectable

1375 +/- 61 pg/mL

NP, Nor

Sham, Nor

Sham, Sal

NP, Sal

NP, Ter

Sham, Sal

NP, Sal

Pre

ssur

e (g

)

Adrb2 +/+

0

1

2

3

4

5

6

7

Adrb2 -/-

0

1

2

Sham

, Sal

NP, Sal

NP, Nor

NP, Ter

Adrb2 +/+

TN

Fα

leve

l

0

1

2

NP, Nor

0

1

2

TN

Fα

leve

l

Sham

, Sal

NP, Sal

NP, Nor

Adrb2 -/-

Adrb2 +/+

Adrb2 -/-

NP, Ter

Sham, Sal

NP, Sal

NP, Nor

After treatmentBefore treatment

Pre

ssur

e (g

)

0

1

2

3

4

5

6

7

ShamSal

NPSal

NPVen

0

1

2

TN

Fα

leve

l

Sham

, Sal

NP, Sal

NP, Ven

Sham

, Sal

NP, Sal

Sham, Sal

NP, Sal

NP, Ven

0

1

2

TN

Fα

leve

lSham, Sal

NP, Sal 0

1

2

mR

NA

exp

ress

ion

0

1

2

mR

NA

exp

ress

ion

HSerum TNFα

B

D E

**

mR

NA

exp

ress

ion

TNFα IL-10 TNFR1

*

β-tub.(53 kDa)

TNFα(26 kDa)

*

Cox-2 IL-1α IL-1β IL-6

*

β-tub.(53 kDa)

TNFα(26 kDa)

**

**

F*

β-tub.(53 kDa)

TNFα(26 kDa)

G

β-tub.(53 kDa)

TNFα(26 kDa)

A

*

*

C* *

Sham, SalNP, Sal NP, Nor

Fig. 4. Antidepressant drugs display anti-TNFα action. (A) Western blot and mRNA expression analysis on dorsal root ganglia, showing increased TNFα levels in C57BL/6J mice at5–6 weeks after induction of the neuropathy (Western blot: n = 12 per group; mRNA: n = 6 per group; *p b 0.05). (B) mRNA expression of different neuroinflammatory-relatedgenes in dorsal root ganglia, evaluated by real-time PCR (n = 3 × 3 per group, *p b 0.05). (C) mRNA expression analysis confirming that the TCA nortriptyline suppressed theneuropathy-induced overexpression of TNFαmRNA (n = 7–9 per group, *p b 0.05). (D) Behavioral analysis of neuropathic allodynia treatment in Adrb2+/+ and Adrb2−/−mice. Twentydays of nortriptyline or terbutaline treatment suppressed the allodynia inwild-typemice, but nortriptyline was ineffective in Adrb2−/−mice (n = 5–6 per group, *p b 0.01 vs. Sham). (E)Western blot analysis of TNFα levels in the dorsal root ganglia of Adrb2+/+ and Adrb2−/− mice. Nortriptyline or terbutaline treatment suppressed the overexpression of mTNFα in wild-type mice, but nortriptyline was ineffective in Adrb2−/− mice (n = 4 per group, *p b 0.05 vs. Sham-Sal). (F) Effect of a 2-week treatment with the serotonin and noradrenaline reuptakeinhibitor venlafaxine on neuropathic allodynia (n = 10 per group, *p b 0.01 vs. pre-treatment). (G) Western blot analysis showing the anti-TNFα action of the long-term venlafaxinetreatment (n = 8 per group, *p b 0.05 vs. Sham-Sal). (H) ELISA analysis of serum TNFα, LPS injection was used as positive control (n = 6 per group).


UNwhereasmRNAexpression of TNFα receptor 1 (TNFR1)was not affected

by the neuropathy (Fig. 4B). No significant changewas observed for theother inflammatory related genes Cox-2, IL-1α, IL-1β or IL-6, and atrend to a decrease was observed for the anti-inflammatory cytokineIL-10 (Fig. 4B).

The treatment with nortriptyline suppressed the mRNA over-expression of TNFα associated with the long-term neuropathy (Fig. 4B).This effect was confirmed at the mRNA level in an independent experi-ment in C57BL/6J mice (Fig. 4C) (H2,25 = 9.11, p b 0.01), and it was alsoobserved at the protein level (Fig. 4E), using dorsal root ganglia fromwild type (Adrb2+/+) and β2-AR deficient (Adrb2−/−) mice (H5,24 =15.46, p b 0.01). Indeed, in wild type animals, the membrane-boundform of TNFα (mTNFα) increased in neuropathic mice (p b 0.05 againstSham-Saline), and this increase was reversed by either nortriptyline orterbutaline treatments (Fig. 4E). In Adrb2−/− mice, the antidepressant


treatmentwas ineffective to suppress the neuropathy-induced increasein mTNFα (Fig. 4E) (p b 0.05 against Sham-Saline), which is in agree-ment with its lack of antiallodynic action in these mice (Yalcin et al.,2009a; Fig. 4D) (Adrb2−/− mice: F2,15 = 77.5, p b 0.0001; post-hoc:(NP-Nortriptyline = NP-Saline) b Sham-Saline at p b 0.001). β2-ARsare thus indirectly mediating both the molecular and the behavioralactions of nortriptyline.

To test whether this action on TNFα may be generalized to otherantidepressants, we tested the selective serotonin and noradrenalinereuptake inhibitor venlafaxine. Two-week venlafaxine treatment sup-pressed both the neuropathic allodynia (Fig. 4F) (F2,26 = 5.63, p b 0.01;post-hoc: (NP-Ven = Sham-Saline) N NP-Saline at p b 0.001) and theneuropathic overexpression of mTNFα in the lumbar dorsal root ganglia(Fig. 4G) (H2,24 = 10.64, p b 0.005; (Sham-Saline = NP-Ven) b NP-Saline at p b 0.05).



T

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

552

553

554

555

556

557

558

559

560

561Q7

562

563

564

565

566

567

568

569

570

571

572

573

574

575

576

577

578

579

580

581

582

583

584

585

586

587

588

589

590

591

592

593

594

595

596

597

598

599

Q2


CO

RREC

The Western blot approach allows differentiating mTNFα (26 kDa)from soluble TNFα (17 kDa). While mTNFα changed with neuropathicpain and its treatments, the soluble form of TNFα remainedundetectable in the dorsal root ganglia. We also assessed soluble TNFαin serum by ELISA, and no soluble TNFα could be detected (Fig. 4H).These results support a preferential role of mTNFα in neuropathicpain (Zhou et al., 2010), and reveal mTNFα as a downstream targetfor antidepressant drug action.

Anti-TNFα relief of neuropathic allodynia

The concomitant loss of anti-TNFα and antiallodynic actions ofnortriptyline in Adrb2−/− mice, suggested that mTNFα could becritical for both the pathophysiology and the treatment of neuro-pathic pain. Thus, we tested the influence of etanercept andinfliximab – clinically used anti-TNFα drugs that do not cross theblood–brain barrier – on neuropathic allodynia. Both drugssuppressed this symptom, with infliximab displaying the fastestonset of therapeutic action and maintaining its antiallodynic actionover 3 weeks following a single administration (Fig. 5A) (Infliximab:F24,192 = 7.68, p b 0.0001; post-hoc: NP-Infliximab N NP-Saline atp b 0.001 on post-surgery days 16 to 35. Etanercept: F24,192 = 7.34,p b 0.0001; post-hoc: NP-Etanercept N NP-Saline at p b 0.001 on post-surgery days 18 to 35, and b Sham-Saline until post-surgery day 24).This difference between infliximab and etanercept may be related notonly to differences in their half-lives (8 to 9.5 days versus 70 h in humans;notice instructions), but also to the fact that infliximab binds to mTNFαwith higher avidity and stability than etanercept (Scallon et al., 2002).Together, the results support a critical role of peripheral mTNFα inthe maintenance of neuropathic allodynia. Moreover, the therapeu-tic efficacy of etanercept was still observed in Adrb2−/− mice(Fig. 5B), supporting the idea that mTNFα acts downstream ofthese adrenoceptors (Etanercept, Adrb2+/+ mice: F3,20 = 14.80,p b 0.0001; post-hoc: NP-Etanercept N NP-Saline at p b 0.001. Eta-nercept, Adrb2−/− mice: F3,20 = 30.22, p b 0.0001; post-hoc: NP-Etanercept N NP-Saline at p b 0.001).

Satellite cells are the source of TNFα in dorsal root ganglia

Infliximab is an artificial mouse/humanmonoclonal antibody direct-ed against TNFα, and we observed that it was effective in vivo tosuppress neuropathic allodynia. We thus perfused neuropathic miceafter in vivo administration of infliximab, and detected this antibodyto visualize its binding to mTNFα. The presence of infliximab in dorsalroot ganglia after intraperitoneal injection (Fig. 6A) and its absencewithin the spinal cord and the brain (Figs. 6B, C) confirmed that thisdrug does not cross the blood–brain barrier and that its antiallodynicaction is peripherally mediated. Co-staining with markers of satellite

UN

Sham, Sal NP, Sal NP, Trea

Pre

ssur

e (g

)

Time (days)0 6 12 18 24 30 36

01234567

Time (days0 6 12 18 24

01234567

**

* * * *** **

++ ++

A

* ** ** * *** **

EtaneInfliximab

++++++++

Fig. 5. Direct anti-TNFα treatments relieve neuropathic allodynia. (A) The direct anti-TNFα treother day from day 15 onwards) relieved neuropathic allodynia (n = 6–7 per group, *p b 0.05 vmice. Etanercept (Eta, 5 mg/kg, once a day for 3 weeks, from day 15 onwards) was still effecti


cells (GS, Fig. 7A) or neuronal cells (NT, Fig. 7A) (CGRP, NF200, SP, IB4,Fig. 7B) revealed that mTNFα was more selectively expressed by thesatellite cells in the dorsal root ganglia.

ED P

RO

OF

Discussion

The need to improve neuropathic pain therapies is important withrespect to the side effects and to the relative efficacy of existing treat-ments. In this context, decrypting themechanismof existing treatmentsmay help in improving them or in discovering new therapeutic targets.Our data show that the antiallodynic action of chronic antidepressanttreatment involves the peripheral nervous system. We provide evi-dence that β2-ARs present on dorsal root ganglia satellite cells areresponsible for this action, and that local mTNFα expression can beimportant for the maintenance and treatment of neuropathic pain.

Noradrenaline is a major actor in the action of antidepressant treat-ment (Barrot et al., 2009, 2010). Previous studies in murine models ofperipheral neuropathy showed that this therapeutic effect can beblocked by lumbar intrathecal manipulations (Suzuki et al., 2008;Yalcin et al., 2009b), which supports a common hypothesis implyingthe recruitment of noradrenergic descending pathways in the analgesicaction of antidepressants (Millan, 2002; Yoshimura and Furue, 2006).However, intrathecalmanipulations also affect the closest nearby dorsalroot ganglia, preventing to distinguish between central and peripheralaction. By comparing noradrenergic lesions at different levels of thenervous system, we show that the peripheral noradrenergic system isnecessary for the antiallodynic property of nortriptyline. In this casethe local source of noradrenaline would be provided by sympatheticfibers sprouting in dorsal root ganglia that accompanies peripheralnerve injury (McLachlan et al., 1993; Ramer and Bisby, 1998). Converse-ly, the antiallodynic action of the antidepressant is still present afterdestruction of noradrenergic descending pathways. While thesedescending pathways are well known to exert an inhibitory controlover nociceptive information (Millan, 2002), our data show that theydo not play a critical role in antidepressant action. Nevertheless, theperipheral mechanism for the relief of peripheral neuropathic paindoes exclude that the central component of antidepressants also partic-ipates to their therapeutic action, particularly concerning anxio-depressive and cognitive co-morbidities that often accompanies neuro-pathic pain (Attal et al., 2011). Indeed, neuropathic pain does altersupraspinal noradrenergic systems (Alba-Delgado et al., 2012).

Dorsal root ganglia are composed of pseudo-unipolar type neuronswhose afferent axons relay sensory information into the central nervoussystem, and of non-neuronal cells such as resident immune cells andsatellite glial cells around the soma of neurons (Hanani, 2005; vonHehn et al., 2012). β2-AR mRNA was previously detected in dorsalroot ganglia (Maruo et al., 2006). In our study, we found that thesereceptors are more precisely expressed in non-neuronal cells, which

01234567

Pre

ssur

e (g

)

01234567Adrb2 +/+ Adrb2 -/-

tment

)30 36

* * *

+++

B

NPSham

Sal Eta Sal Eta

*

rcept

*

atments of infliximab (10 mg/kg, injected once on day 15) or etanercept (2 mg/kg, everys. Sham, +p b 0.05 NP-treated vs.NP-Sal). (B) Etanercept effect in Adrb2+/+ and Adrb2−/−

ve in Adrb2−/− mice (n = 6 per group, *p b 0.001).



T

PRO

OF

600

601

602

603

604

605

606

607

608

609

610

611

612

613

DRG

1 2 3 4

LV

Cg

1 2 3 4

4321 dh

vh

B

C

ATNFα TNFαHoechst Hoechst

DorsalRoot

Ganglia

SpinalCord

Brain

Fig. 6.Detection of TNFα by peripheral injection of infliximab. Cuffmice received acute infliximab (10 mg/kg, i.p., n = 3), an artificialmouse/humanmonoclonal antibody directed againstTNFα, and were perfused 6 h post-injection. Pictures numbered 1 and 3 show TNFα detection (in red), and columns 2 and 4 show the Hoechst 33342 staining of nuclei of the same sec-tions. Columns 3 and 4 show a larger view of the boxed areas in column 1 pictures. (A) A reticular TNFα staining was detected in the dorsal root ganglia after peripheral administration ofinfliximab.No stainingwas detected in the spinal cord (B) or brain (C) of the same animals, confirming that infliximabdoesnot cross the blood–brain barrier. Scale bars: 500 μm inB1,2 andC1,2, 200 μm in A1,2, B3,4 and C3,4, and 50 μm in A3,4. Abbreviations: Cg, cingulate cortex; dh, dorsal horn; DRG, dorsal root ganglion; LV, lateral ventricle; vh, ventral horn.


are likely satellite cells expressing mTNFα. These satellite cells are theperipheral analogs of the central nervous system's astrocytes whichare also known to express β2-AR on their membrane in various speciesincluding human (Mantyh et al., 1995; Trimmer et al., 1984). Moreover,we observed that the efficacy of β2-AR coupling to a transduction path-way increasing intracellular calcium is enhanced in the lumbar dorsalroot ganglia under neuropathic condition. Together, our data indicate

UNCO

RREC

TNFα+/GS+: 88.4% of GS+ cells (489/553 cells)

GSTNFα

TNFα NT

TNFα+/NT+: 7.0% of NT+ cells (37/527 cells)

A B

Fig. 7. Detection of TNFα in the dorsal root ganglia. Neuropathic mice received acute infliximabTNFα, and were perfused 6 h post-injection. (A) Confocal microscopy allowed to visualize TNFTNFαwas detected inGS-positive satellite cells (top graph,white arrows) indorsal root ganglia.the confocal images show examples of non-neuronal TNFα staining. For both GS and NT staininScale bars: 50 μm. (B) Neuronal staining and detection of TNFα in the dorsal root ganglia. Doubl(n = 3). Thefirst line shows theneuronal staining, the second line shows the TNFα detection, an200 (NF200), substance P (SP) and the isolectin IB4 (IB4) were used as markers of different neudorsal root ganglia formed a reticular network surrounding the neurons. Scale bar: 20 μm.


EDthat the peripheral nerve injury does create the substrate – i.e. norad-

renergic sprouting and enhanced β2-AR functionality – through whichthe antidepressant treatment acts to relieve neuropathic allodynia.

Recent data indicate that neuropathic painmay partly be consideredas a neuro-immune disorder, with a critical involvement of glial and/orimmune cells following nerve injury (Austin andMoalem-Taylor, 2011;Thacker et al., 2007; Vallejo et al., 2010). The activation of these cells at

CGRP

TNFα

TNFαCGRP

NF200

TNFα

TNFαNF200

SP

TNFα

TNFαSP

IB4

TNFα

TNFαIB4

(10 mg/kg, i.p., n = 3), an artificial mouse/humanmonoclonal antibody directed againstα (in red), cell nuclei (Hoechst 33342 in blue) and the satellite cell marker GS (in green).TNFαwas rarely detected in neurons (NT in blue, bottomgraph,NeuroTrace™). Arrows ong, a quantification of TNFα-positive cells was done on 9 sections (3 sections × 3 animals).e-staining for infliximab detection and neuronalmarkers was done on DRG from cuff miced the third line is amerged picture. Calcitonin gene related peptide (CGRP), neurofilamentronal cell populations. TNFαwas not expressed by neurons, but the TNFα staining in the



CTED P

RO

OF

614

615

616

617

618

619

620

621

622

623

624

625

626

627

628

629

630

631

632

633

634

635

636

637

638

639

640

641

642

643

644

645

646

647

648

649

650

651

652

653

654

655

656

657

658

659

660

661

662

663

664

665

666

667

668

669

Peripheral nerve injury

Sciatic nerve cuffing

Mechanical allodynia

Dorsal root ganglia neuron

L4

L5

L6

β2-AR

NETNET

NET NET

NE

T

NE

T

NA NA

NA

Sympatheticsprouting

TNFR mTNFα Satellite cellsTransduction

pathway

β2-mimetic

Antidepressantdrug

anti-TNFα

Fig. 8. Schematic of the proposed peripheral action of antidepressant drugs in the context of a peripheral neuropathy. Following sciatic nerve injury, a noradrenergic sprouting is observedin dorsal root ganglia. Antidepressant drugs inhibit noradrenaline (NA) reuptake in sympathetic varicosities (NET: transporter), thus locally increasing noradrenaline levels. Noradrenalinestimulates β2-AR expressed by non-neuronal satellite cells, which lead to decreased mTNFα production.


UNCO

RREthe lesion site or in the dorsal root ganglia for the peripheral nervous

system, or in the spinal cord or supraspinal structures for the centralnervous system, results in the production of cytokines (Austin andMoalem-Taylor, 2011; White et al., 2009). Pro-inflammatory cytokinesproduced in the peripheral nervous system after injury participate inthe pathophysiology of neuropathic pain (Leung and Cahill, 2010;Sommer et al., 2001a; Wagner and Myers, 1996), particularly in itsinitiation. In installed neuropathic pain, some cytokines such as TNFαstill display enhanced expression, which has pathophysiologicalrelevance. Indeed, blocking TNFα can relieve neuropathic painsymptoms both in animals ((Sommer et al., 2001a, 2001b); presentresults) and in humans (Korhonen et al., 2005; Tobinick andDavoodifar, 2004). Our results suggest that antidepressant drugswould act as indirect anti-TNFα drugs to relieve neuropathic allodynia.This anti-TNFα effect of antidepressants is partial as it suppressesneuropathy-induced TNFα overexpression without affecting basalexpression of the cytokine. The mechanism is β2-AR mediated, asindicated by the loss of both anti-TNFα and behavioral actions ofnortriptyline in Adrb2−/− mice.

TNFα is primarily membrane-bound (mTNFα, 26 kDa), and thismTNFα is cleaved by the TNFα converting enzyme TACE to release thesoluble peptide (sTNFα, 17 kDa). The diffusible sTNFα is the moststudied form of TNFα. However, mTNFα is also biologically active asself-assembling non-covalent bound trimers (MacEwan, 2002). Afterchronic neuropathy, we did not detect sTNFα in the serum or in thedorsal root ganglia, butwe found a significant increase ofmTNFαwithinthe dorsal root ganglia. While we cannot exclude the presence ofconcentrations of sTNFα that would be below detection threshold, our


data however support a preferential role for mTNFα. This preferentialrecruitment of mTNFα in neuropathic pain has also been observed byother groups in the spinal cord (Hao et al., 2007; Peng et al., 2006),where it is produced by glial cells (Zhou et al., 2010). In the dorsalroot ganglia, we observed that the mTNFα was produced by satellitecells, which might participate in neuropathic allodynia by influencingthe nearby nociceptors through a local cell–cell signaling communica-tion (Cabal-Hierro and Lazo, 2012; McCoy and Tansey, 2008; Santelloand Volterra, 2012). Moreover, this local mTNFα recruitment by theneuropathic condition should not be considered as a classical inflamma-tory cascade. Indeed, neither IL-1 nor IL-6 expression was affected bythe neuropathy at late time-points (present data), and the inhibitionof the inflammatory cascade through Cox targeting by ketoprofen hasno antiallodynic action (Benbouzid et al., 2008a).

The noradrenaline, at the start point of the antidepressant's thera-peutic cascade, originates from sympathetic fibers sprouting in thedorsal root ganglia. Interestingly, the sympathetic nervous system hasalso been shown to be a potent regulator of TNFα production after lipo-polysaccharide (LPS) exposure (Elenkov et al., 2000; Szelenyi et al.,2000). In LPS-induced inflammation, noradrenaline released fromsympathetic nerve fibers or β-AR agonists inhibits the production ofTNFα by immune and glial cells (Severn et al., 1992; van der Poll et al.,1994) and this modulatory effect may be mediated via β2-AR (Hetieret al., 1991; Nakamura et al., 1998). Our results support the idea thatantidepressant drugs would act on neuropathic allodynia through asimilar cascade of events (Fig. 8).

While a therapeutic action of anti-TNFα treatments can be observedin neuropathic pain, direct anti-TNFα therapies such as etanercept or



T

670

671

672

673

674

675

676

677

678

679

680

681

682

683

684

685

686

687

688

689

690

691

692

693

694

695

696

697

698

699

700

701

702

703

704

705

706

707

708

709

710

711712Q8713714715Q9716717718719720721722723724725726727728729730731732733734735736737

738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823


UNCO

RREC

infliximab may not be appropriate for a large clinical use in the neuro-pathic pain context. Indeed, the balance of benefits/risks requiresimprovements as these compounds display major side effects, affectingthe immune system and its ability to fight infections or cancer. Some ofthese risks are life-threatening with a notable incidence. Direct anti-TNFα therapy may thus be acceptable for the major inflammatory orauto-immune diseases for which these compounds are presentlyprescribed as second choice treatments (i.e. rheumatoid polyarthritisor psoriasis which is resistant to other known treatments), but theywould be less acceptable for neuropathic pain which is a more wide-spread condition. These direct anti-TNFα treatments are fusion proteinscapturing TNFα. They may be considered as too effective, blocking highpathological TNFα levels as well as physiologically relevant TNFα.Present data on the neuropathic pain models suggest that the indirectblunting of TNFα overexpression (as observed with antidepressants orβ2-mimetics) could be of interest, as it may preserve physiologicallevels of TNFα and thus prevent some of the adverse effects of directanti-TNFα targeting. While this hypothesis is still speculative at thispoint, the exploration of other partial inhibitors of TNFα productionwarrants further research.

In conclusion, this study highlights a novel mechanistic substrate forantidepressant drug action against neuropathic pain. By recruitingnoradrenaline from sympathetic fibers sprouting in the dorsal rootganglia, antidepressant drugs stimulate local β2-ARs on non-neuronalcells. This action decreases mTNFα production and leads to theantiallodynic effect. This finding is an important step to understandthemechanism bywhich antidepressant drugs relieve this neurologicalcondition. Our data obtained in a mouse model suggest that directly orindirectly targeting TNFαmay potentially offer alternative therapeuticsto antidepressant drugs for neuropathic pain management. It is impor-tant to note that this study was conducted in an animal model withintraperitoneal treatments. It will now require confirmation with oraladministration of antidepressant drugs and a clinical validation. Suchfuture validation will also be critical to confirm the value of this ap-proach for translational research on neuropathic pain and its treatment.

Acknowledgments

This work was supported by the Centre National de la RechercheScientifique (contract UPR3212), the University of Strasbourg andNeurex. Y.B. was supported by a Neurex fellowship.We thank StéphaneDoridot for animal care and genotyping.

References

Alba-Delgado, C., et al., 2012. Chronic pain leads to concomitant noradrenergic impair-ment and mood disorders. Biol. Psychiatry.

Attal, N., et al., 2006. EFNS guidelines on pharmacological treatment of neuropathic pain.Eur. J. Neurol. 13, 1153–1169.

Attal, N., et al., 2010. EFNS guidelines on the pharmacological treatment of neuropathicpain: 2010 revision. Eur. J. Neurol. 17, e1113–e1188.

Attal, N., et al., 2011. The specific disease burden of neuropathic pain: results of a Frenchnationwide survey. Pain 152, 2836–2843.

Austin, P.J., Moalem-Taylor, G., 2011. The neuro-immune balance in neuropathic pain:involvement of inflammatory immune cells, immune-like glial cells and cytokines.J. Neuroimmunol. 229, 26–50.

Barrot, M., 2012. Tests and models of nociception and pain in rodents. Neuroscience 211,39–50.

Barrot, M., et al., 2009. From antidepressant drugs to beta-mimetics: preclinical insightson potential new treatments for neuropathic pain. Recent Pat. CNS Drug Discov. 4,182–189.

Barrot, M., et al., 2010. Antidepressant treatment of neuropathic pain: looking for themechanism. Future Neurol. 5, 247–257.

Benbouzid, M., et al., 2008a. Chronic, but not acute, tricyclic antidepressant treatmentalleviates neuropathic allodynia after sciatic nerve cuffing in mice. Eur. J. Pain 12,1008–1017.

Benbouzid, M., et al., 2008b. Sciatic nerve cuffing in mice: a model of sustained neuro-pathic pain. Eur. J. Pain 12, 591–599.

Bohren, Y., et al., 2010. Mu-opioid receptors are not necessary for nortriptyline treatmentof neuropathic allodynia. Eur. J. Pain 14, 700–704.

Cabal-Hierro, L., Lazo, P.S., 2012. Signal transduction by tumor necrosis factor receptors.Cell. Signal. 24, 1297–1305.


ED P

RO

OF

Choucair-Jaafar, N., et al., 2009. Beta2-adrenoceptor agonists alleviate neuropathicallodynia in mice after chronic treatment. Br. J. Pharmacol. 158, 1683–1694.

Choucair-Jaafar, N., et al., 2011. Cardiovascular effects of chronic treatment with a β2-adrenoceptor agonist relievingneuropathic pain inmice.Neuropharmacology61, 51–60.

Chruscinski, A.J., et al., 1999. Targeted disruption of the beta2 adrenergic receptor gene.J. Biol. Chem. 274, 16694–16700.

Cok, O.Y., et al., 2010. Is there a place for beta-mimetics in clinical management of neuro-pathic pain? Salbutamol therapy in six cases. Anesthesiology 112, 1276–1279.

Dworkin, R.H., et al., 2007. Pharmacologic management of neuropathic pain: evidence-based recommendations. Pain 132, 237–251.

El Mansari, M., et al., 2010. Relevance of norepinephrine–dopamine interactions in thetreatment of major depressive disorder. CNS Neurosci. Ther. 16, e1–e17.

Elenkov, I.J., et al., 2000. The sympathetic nerve—an integrative interface between twosupersystems: the brain and the immune system. Pharmacol. Rev. 52, 595–638.

Empl, M., et al., 2001. TNF-alpha expression in painful and nonpainful neuropathies.Neurology 56, 1371–1377.

Hanani, M., 2005. Satellite glial cells in sensory ganglia: from form to function. Brain Res.Brain Res. Rev. 48, 457–476.

Hao, S., et al., 2007. Gene transfer to interfere with TNFalpha signaling in neuropathicpain. Gene Ther. 14, 1010–1016.

Hetier, E., et al., 1991. Modulation of interleukin-1 and tumor necrosis factor expressionby beta-adrenergic agonists in mouse ameboid microglial cells. Exp. Brain Res. 86,407–413.

Ho, C., O'Leary, M.E., 2010. Single-cell analysis of sodium channel expression in dorsal rootganglion neurons. Mol. Cell. Neurosci. 46, 159–166.

Jensen, T.S., et al., 2011. A new definition of neuropathic pain. Pain 152, 2204–2205.Kaufling, J., et al., 2010. Pharmacological recruitment of the GABAergic tail of the ventral

tegmental area by acute drug exposure. Br. J. Pharmacol. 161, 1677–1691.Korhonen, T., et al., 2005. The treatment of disc herniation-induced sciatica with

infliximab: results of a randomized, controlled, 3-month follow-up study. Spine(Phila Pa 1976) 30, 2724–2728.

Krell, H.V., et al., 2005. Evaluation of reboxetine, a noradrenergic antidepressant, for thetreatment of fibromyalgia and chronic low back pain. Psychosomatics 46, 379–384.

Lee, H.L., et al., 2004. Temporal expression of cytokines and their receptors mRNAs in aneuropathic pain model. Neuroreport 15, 2807–2811.

Leung, L., Cahill, C.M., 2010. TNF-alpha andneuropathic pain—a review. J. Neuroinflammation7, 27.

Lindenlaub, T., Sommer, C., 2003. Cytokines in sural nerve biopsies from inflammatoryand non-inflammatory neuropathies. Acta Neuropathol. 105, 593–602.

MacEwan, D.J., 2002. TNF receptor subtype signalling: differences and cellular conse-quences. Cell. Signal. 14, 477–492.

Mantyh, P.W., et al., 1995. Beta 2-adrenergic receptors are expressed by glia in vivo in thenormal and injured central nervous system in the rat, rabbit, and human. J. Neurosci.15, 152–164.

Maruo, K., et al., 2006. Modulation of P2X receptors via adrenergic pathways in rat dorsalroot ganglion neurons after sciatic nerve injury. Pain 120, 106–112.

McCoy, M.K., Tansey, M.G., 2008. TNF signaling inhibition in the CNS: implications fornormal brain function and neurodegenerative disease. J. Neuroinflammation 5, 45.

McLachlan, E.M., et al., 1993. Peripheral nerve injury triggers noradrenergic sproutingwithin dorsal root ganglia. Nature 363, 543–546.

Mico, J.A., et al., 2006. Antidepressants and pain. Trends Pharmacol. Sci. 27, 348–354.Millan, M.J., 2002. Descending control of pain. Prog. Neurobiol. 66, 355–474.Mosconi, T., Kruger, L., 1996. Fixed-diameter polyethylene cuffs applied to the rat sciatic

nerve induce a painful neuropathy: ultrastructural morphometric analysis of axonalalterations. Pain 64, 37–57.

Nakamura, A., et al., 1998. Regulation of tumour necrosis factor and interleukin-6 genetranscription by beta2-adrenoceptor in the rat astrocytes. J. Neuroimmunol. 88,144–153.

Neil, A., et al., 1991. Effects of guanethidine on sensitization to natural stimuli and self-mutilating behaviour in rats with a peripheral neuropathy. Brain Res. 565, 237–246.

Peng, X.M., et al., 2006. Tumor necrosis factor-alpha contributes to below-level neuro-pathic pain after spinal cord injury. Ann. Neurol. 59, 843–851.

Ramer, M.S., Bisby, M.A., 1998. Normal and injury-induced sympathetic innervation of ratdorsal root ganglia increases with age. J. Comp. Neurol. 394, 38–47.

Saarto, T., Wiffen, P.J., 2007. Antidepressants for neuropathic pain. Cochrane DatabaseSyst. Rev. CD005454.

Santello, M., Volterra, A., 2012. TNFalpha in synaptic function: switching gears. TrendsNeurosci. 35, 638–647.

Scallon, B., et al., 2002. Binding and functional comparisons of two types of tumor necrosisfactor antagonists. J. Pharmacol. Exp. Ther. 301, 418–426.

Schmittgen, T.D., Zakrajsek, B.A., 2000. Effect of experimental treatment on housekeepinggene expression: validation by real-time, quantitative RT-PCR. J. Biochem. Biophys.Methods 46, 69–81.

Severn, A., et al., 1992. Regulation of tumor necrosis factor production by adrenaline andbeta-adrenergic agonists. J. Immunol. 148, 3441–3445.

Sommer, C., et al., 2001a. Anti-TNF-neutralizing antibodies reduce pain-related behaviorin two different mouse models of painful mononeuropathy. Brain Res. 913, 86–89.

Sommer, C., et al., 2001b. Etanercept reduces hyperalgesia in experimental painfulneuropathy. J. Peripher. Nerv. Syst. 6, 67–72.

Suzuki, T., et al., 2008. Antiallodynic and antihyperalgesic effect of milnacipran in micewith spinal nerve ligation. Anesth. Analg. 106, 1309–1315 (table of contents).

Szelenyi, J., et al., 2000. Differential involvement of sympathetic nervous system andimmune system in the modulation of TNF-alpha production by alpha2- and beta-adrenoceptors in mice. J. Neuroimmunol. 103, 34–40.

Thacker, M.A., et al., 2007. Pathophysiology of peripheral neuropathic pain: immune cellsand molecules. Anesth. Analg. 105, 838–847.



824825826827828829830831832833834835836837838

839840841842843844845846847848849850851852853

855


Tobinick, E., Davoodifar, S., 2004. Efficacy of etanercept delivered by perispinal adminis-tration for chronic back and/or neck disc-related pain: a study of clinical observationsin 143 patients. Curr. Med. Res. Opin. 20, 1075–1085.

Trimmer, P.A., et al., 1984. Combination of immunocytochemistry and radioligand receptorassay to identify beta-adrenergic receptor subtypes on astroglia in vitro. J. Neurosci. 4,1598–1606.

Uceyler, N., et al., 2007. Early cytokine expression in mouse sciatic nerve after chronicconstriction nerve injury depends on calpain. Brain Behav. Immun. 21, 553–560.

Vallejo, R., et al., 2010. The role of glia and the immune system in the development andmaintenance of neuropathic pain. Pain Pract. 10, 167–184.

van der Poll, T., et al., 1994. Noradrenaline inhibits lipopolysaccharide-induced tumor ne-crosis factor and interleukin 6 production in human whole blood. Infect. Immun. 62,2046–2050.

von Hehn, C.A., et al., 2012. Deconstructing the neuropathic pain phenotype to revealneural mechanisms. Neuron 73, 638–652.

UNCO

RRECT

854


Wagner, R., Myers, R.R., 1996. Endoneurial injection of TNF-alpha produces neuropathicpain behaviors. Neuroreport 7, 2897–2901.

White, F.A., et al., 2009. Chemokine signaling and the management of neuropathic pain.Mol. Interv. 9, 188–195.

Yalcin, I., et al., 2009a. Beta(2)-adrenoceptors are critical for antidepressant treatment ofneuropathic pain. Ann. Neurol. 65, 218–225.

Yalcin, I., et al., 2009b. Beta2-adrenoceptors are essential for desipramine, venlafaxine orreboxetine action in neuropathic pain. Neurobiol. Dis. 33, 386–394.

Yalcin, I., et al., 2010. Chronic treatment with agonists of beta(2)-adrenergic receptors inneuropathic pain. Exp. Neurol. 221, 115–121.

Yoshimura, M., Furue, H., 2006. Mechanisms for the anti-nociceptive actions of the de-scending noradrenergic and serotonergic systems in the spinal cord. J. Pharmacol.Sci. 101, 107–117.

Zhou, Z., et al., 2010. A novel cell–cell signaling by microglial transmembrane TNFalphawith implications for neuropathic pain. Pain 151, 296–306.

ED P

RO

OF



Antidepressants suppress neuropathic pain by a peripheral β2-adrenoceptor mediated anti-TNFα...

Documents

Transcript of Antidepressants suppress neuropathic pain by a peripheral β2-adrenoceptor mediated anti-TNFα...