A r t i c l e s

822 VOLUME 17 | NUMBER 7 | JULY 2011 nAture medicine

Individuals suffering from nerve injury or those undergoing treatment for cancer or AIDS can experience persistent pain that is resistant to commonly used analgesics. Some conditions of severe intractable clinical pain necessitate intrathecal delivery of Prialt (Azur Pharma, also known as ziconotide), a synthetic ω-conotoxin that blocks neu-ronal calcium channels. The clinical efficacy of Prialt has identified the N-type Ca2+ channel (CaV2.2) as a key target for the treatment of chronic pain1–4. However, the use of Prialt for pain management is limited by its method of delivery, narrow therapeutic window and adverse effects such as hypotension and memory loss1,4.

The discovery of new small-molecule inhibitors of CaV2.2 for use as analgesics may lead to improved therapeutic pharmacology. CRMP-2 is a modulator of CaV2.2 (refs. 5,6). It is a cytosolic phosphoprotein that was originally identified as a mediator of growth cone collapse7 that can also modify axon number, length8 and neuronal polarity9–11. CRMP-2 interacts with CaV2.2, and overexpression of CRMP-2 leads

to increased surface expression of CaV2.2, enhanced Ca2+ currents and an increase in stimulated release of calcitonin gene-related peptide (CGRP) from dorsal root ganglia (DRG)5,6. By contrast, knockdown of CRMP-2 markedly reduces Ca2+ currents and transmitter release5,6. Here we report that uncoupling the interaction between CRMP-2 and CaV2.2 led to a physiologically relevant decrease in Ca2+ current and neurotransmitter release (Supplementary Fig. 1a) and suppressed persistent inflammatory and neuropathic hypersensitivity.

RESULTSA peptide that uncouples CRMP-2 and Ca2+ channelsTo disrupt the interaction between CRMP-2 and the CaV2.2 com-plex in vivo, we synthesized peptides that covered the entire length of CRMP-2, including three CaV-binding domains (CBD1–CBD3) that are involved in the CRMP-2–CaV2.2 interaction5. We found that a CRMP-2 peptide (CBD3; residues 484–498) bound CaV2.2 (Fig. 1a).

1Program in Medical Neurosciences, Paul and Carole Stark Neurosciences Research Institute, Indiana University School of Medicine, Indianapolis, Indiana, USA. 2Department of Pharmacology and Toxicology, Indiana University School of Medicine, Indianapolis, Indiana, USA. 3Department of Biochemistry and Molecular Biology, Indiana University School of Medicine, Indianapolis, Indiana, USA. 4Department of Psychiatry, Indiana University School of Medicine, Indianapolis, Indiana, USA. 5Department of Neurological Surgery, Indiana University School of Medicine, Indianapolis, Indiana, USA. 6Department of Anatomy and Cell Biology, Indiana University School of Medicine, Indianapolis, Indiana, USA. 7Department of Anesthesia, Indiana University School of Medicine, Indianapolis, Indiana, USA. 8Sensory Plasticity Laboratory, Pain Research Center, Department of Anesthesiology, Perioperative and Pain Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts, USA. 9Department of Neurology, Indiana University School of Medicine, Indianapolis, Indiana, USA. 10Department of Indiana Clinical and Translational Sciences Institute, Indiana University School of Medicine, Indianapolis, Indiana, USA. 11These authors contributed equally to this work. Correspondence should be addressed to R.K. (khanna5@iupui.edu).

Received 14 September 2010; accepted 7 March 2011; published online 5 June 2011; doi:10.1038/nm.2345

Suppression of inflammatory and neuropathic pain by uncoupling CRMP-2 from the presynaptic Ca2+ channel complexJoel M Brittain1,11, Djane B Duarte1,11, Sarah M Wilson1, Weiguo Zhu1,2, Carrie Ballard1,3, Philip L Johnson4, Naikui Liu1,5, Wenhui Xiong1,6, Matthew S Ripsch1,7, Yuying Wang1,2, Jill C Fehrenbacher1,2,7, Stephanie D Fitz4, May Khanna3, Chul-Kyu Park8, Brian S Schmutzler1,2, Bo Myung Cheon1,7, Michael R Due1,7, Tatiana Brustovetsky2, Nicole M Ashpole1,3, Andy Hudmon1,2,3, Samy O Meroueh1,3, Cynthia M Hingtgen1,2,9, Nickolay Brustovetsky1,2, Ru-Rong Ji8, Joyce H Hurley1,3, Xiaoming Jin1,6, Anantha Shekhar4,10, Xiao-Ming Xu1,6,9, Gerry S Oxford1,2, Michael R Vasko1,2,7, Fletcher A White1,7 & Rajesh Khanna1,2

The use of N-type voltage-gated calcium channel (CaV2.2) blockers to treat pain is limited by many physiological side effects. Here we report that inflammatory and neuropathic hypersensitivity can be suppressed by inhibiting the binding of collapsin response mediator protein 2 (CRMP-2) to CaV2.2 and thereby reducing channel function. A peptide of CRMP-2 fused to the HIV transactivator of transcription (TAT) protein (TAT-CBD3) decreased neuropeptide release from sensory neurons and excitatory synaptic transmission in dorsal horn neurons, reduced meningeal blood flow, reduced nocifensive behavior induced by formalin injection or corneal capsaicin application and reversed neuropathic hypersensitivity produced by an antiretroviral drug. TAT-CBD3 was mildly anxiolytic without affecting memory retrieval, sensorimotor function or depression. At doses tenfold higher than that required to reduce hypersensitivity in vivo, TAT-CBD3 caused a transient episode of tail kinking and body contortion. By preventing CRMP-2–mediated enhancement of CaV2.2 function, TAT-CBD3 alleviated inflammatory and neuropathic hypersensitivity, an approach that may prove useful in managing chronic pain.

© 2

011

Nat

ure

Am

eric

a, In

c. A

ll ri

gh

ts r

eser

ved

.

A r t i c l e s

nAture medicine VOLUME 17 | NUMBER 7 | JULY 2011 823

Immunoprecipitations from rat spinal cord lysates showed that CBD3 inhibited the interaction between CRMP-2 and CaV2.2 but did not affect the interaction between tubulin and CRMP-2 (ref. 12; Fig. 1b). As CRMP-2 binds the first intracellular loop (L1) and the distal C terminus (Ct-dis) of CaV2.2 (ref. 5), we investigated whether CBD3 bound these regions. We used surface plasmon resonance to show that CBD3, but not a scrambled peptide, bound immobilized L1 and

Ct-dis (Fig. 1c). Moreover, the CBD3 peptide disrupted the inter-action between CRMP-2 and the L1 or Ct-dis regions (Fig. 1d and Supplementary Fig. 1b,c).

Because CRMP-2 facilitates trafficking of CaV2.2 to the surface5,6, we tested whether CBD3 could uncouple CRMP-2 from CaV2.2 to affect the trafficking, surface expression and activity of CaV2.2, as well as Ca2+ influx. Expression of both CaV2.2 and CBD3 in the

Figure 1 A CRMP-2 peptide suppresses the interaction between CaV2.2 and CRMP-2. (a) Normalized binding of CaV2.2 to 15-mer peptides (overlapping by 12 amino acids) encompassing full-length CRMP-2 overlaid with spinal cord lysates. The sequence of peptide 96, designated CBD3, is shown. (b) Immunoprecipitation (IP) of CaV2.2 (top), CRMP-2 (middle) and β-tubulin (bottom) with recombinant CRMP-2 or CaV2.2 antibody from spinal cord lysates in the presence of scramble or CBD3 peptides. (c) Sensorgram of CBD3 (1, 3, 5 µM) or scramble peptide (1, 3, 5 µM traces) binding to CaV2.2 cytosolic loop 1 (L1) and distal C terminus (Ct-dis). Dissociation was monitored for 4 min. RU, resonance units. (d) Binding of L1-GST and Ct-dis-GST fusion proteins to CRMP-2 in the presence of scramble or CBD3 peptides (10 µM). CRMP-2 binding to L1 and Ct-dis was probed with a CRMP-2–specific antibody. (e,f) Top, immunocytochemistry of expressed CaV2.2 in CAD cells without (e) or with (f) CBD3 overexpression. Scale bars, 10 µm. Bottom, normalized surface intensity (SI) between the arrows demarcating surfaces of cells in e and f. (g) Percentage of cells showing surface CaV2.2 expression (n > 100). (h) Immunoblots of biotinylated (surface) fractions of CAD cells expressing vector (scrambled), an N-terminal region of CRMP-2 (CBD1) or CBD3 probed with CaV2.2 antibody (n = 3). (i) Top, voltage protocol. Bottom, exemplar traces from hippocampal neurons overexpressing vector (EGFP), CRMP-2 or CRMP-2 + CBD3. (j) Peak current density (pA/pF) at +10 mV for vector- for CRMP-2– or CRMP-2 + CBD3–transfected neurons. *P < 0.05 versus CRMP-2, Student’s t test. Error bars represent means ± s.e.m.

0

0.25

0.50

0.75

1.00

20 40 60 80 100CRMP-2 peptide number

CaV

2.2

bind

ing

to C

RM

P-2

(nor

mal

ized

to m

ax.)

1009080

0

10

20

30

CRMP-2

Cur

rent

den

sity

(pA

/pF

) *

CRMP-2 + CBD3

–80 mV+10 mV, 200 ms

40 ms100 pA

(18) (9)

0

25

50

75

CaV

2.2

surf

ace

exp

ress

ion

(% o

f cel

ls)

*

g

CaV2.2260

135

95

IP: streptavidin

h

CBD3

a

i

b

75

75

52

IP: Ct-dis52

IP: L1

CR

MP

-2C

RM

P-2

e

SI

SI1

0

1

0

ARSRLAELRGVPRGL

d

L1

Ct-dis 100 s

10 R

U 1

3

5 µM

+CBD3

+Scramble

10 R

U

+CBD3

1

3

5

+Scramble

CaV2.2

β-tubulin

IP: CRMP-2

IP: CRMP-2

52

260

MW(kDa)

c

(9)

*

Vector

CR

MP

-2

IP: CaV2.2

52

f

j

Scram

ble

Scram

ble

Scram

ble

CBD3

CBD3

CBD3

+ CBD3

+CBD3

Vecto

r

CaV2.

2

CRMP-2

CBD1

Figure 2 TAT-CBD3 reduces Ca2+ currents in DRGs and excitatory synaptic transmission in lamina II neurons from spinal cord slices. (a) Representative differential interference contrast/fluorescence images showing robust penetration of FITC-TAT-CBD3 into DRGs (arrowheads) but not other cells (arrows). Nuclei are stained with Hoechst dye in the bottom image. Scale bars, 10 µm. (b) Representative current traces from a DRG incubated for 15 min with TAT-Scramble (10 µM; green) or TAT-CBD3 (10 µM; purple) in response to the voltage steps illustrated at the top. (c) Current-voltage relationships for the currents shown in b fitted to a b-spline line. Peak currents were normalized to the cell capacitance. (d) Peak current density measured at −10 mV for DRGs incubated with TAT-Scramble, TAT-CBD3 or TAT-CBD3 + 1 µM ω-conotoxin (CTX). The numbers in parentheses represent numbers of cells tested. *P < 0.05 versus TAT-Scramble. (e) Top, representative traces of spontaneous EPSCs (sEPSCs) in lamina II neurons in spinal cord slices before treatment (left) or after application of 10 µM TAT-Scramble (middle) or 10 µM TAT-CBD3 (right). Bottom, enlarged traces. Voltage-clamp recordings (holding voltage, −70 mV) were used to record synaptic responses. (f) Ratio of sEPSC frequency and amplitude. *P < 0.05 compared with baseline. Error bars represent means ± s.e.m.

b+40 mV, 200 ms

–50 mV↑∆10 mV

a

50 ms 250 pA

–80 mV

TAT-Scramble

TAT-CBD3

dc

Cur

rent

den

sity

(pA

/pF

)

TAT-Scramble TAT-CBD3

*

NS

(8) (9) (4)

10

30

40

0

20

Pea

k cu

rren

t den

sity

(pA

/pF

)

*

00.20.40.60.81.01.2

Rat

io o

f sE

PS

C

FrequencyAmplitudef

5 s20 pA

e TAT-Scramble TAT-CBD3Control

TAT-Scr.(n = 7)

TAT-CBD3(n = 6)

*

0

–10

–20

–30

–40

–60 –40 –20 0 20 40Membrane potential (mV)

TAT-Scr

amble

TAT-CBD3

+

ω-CTX

TAT-CBD3

© 2

011

Nat

ure

Am

eric

a, In

c. A

ll ri

gh

ts r

eser

ved

.

A r t i c l e s

824 VOLUME 17 | NUMBER 7 | JULY 2011 nAture medicine

CAD neuronal cell line13 resulted in almost complete retention of the channel in cytoplasmic aggregates (Fig. 1e–g). Surface biotinylation in CAD cells showed that the expression of CBD3, but not of scrambled peptide or CBD1, prevented the surface expression of co-expressed CaV2.2 (Fig. 1h). Furthermore, expression of CRMP-2 with CBD3 in rat hippocampal neurons eliminated the CRMP-2–mediated increase in CaV2.2 current density5 (Fig. 1i,j) and the expression of CBD3, but not of a scrambled control peptide, reduced depolarization-induced calcium influx in hippocampal neurons (Supplementary Fig. 2a–c). Thus, in vitro, CBD3 disrupts the interaction between CRMP-2 and CaV2.2, decreases the plasma membrane trafficking of CaV2.2 and reduces Ca2+ current density.

We fused CBD3 with the transduction domain of the HIV-1 TAT protein14 to generate the cell-permeant TAT-CBD3, which readily entered DRG neurons (Fig. 2a). Application of TAT-CBD3 to DRGs for 15 min reduced Ca2+ currents by ~60% (Fig. 2b), and currents were not further blocked by addition of the CaV2.2 blocker ω-conotoxin GVIA (1 µM; Fig. 2c,d); this suggests that TAT-CBD3 is selective for N-type channels. We obtained qualitatively similar results from calcium imaging in rat DRGs: TAT-CBD3 reduced K+-evoked Ca2+ influx selectively through N-type channels (Supplementary Fig. 2d). TAT-CDB3 did not affect sodium current density or gating in DRGs (Supplementary Fig. 3).

To determine whether uncoupling CRMP-2 from CaV2.2 with TAT-CBD3 modulates synaptic transmission, we used patch-clamp recording in spinal cord slices to measure synaptic responses in mouse lamina II neurons (Fig. 2e,f) that received input from C-fiber primary afferents expressing CaV2.2 (refs. 15,16). The spon-taneous excitatory postsynaptic currents (sEPSCs) in these neurons are caused by glutamate release and reflect presynaptic and post-synaptic mechanisms (frequency change and amplitude change, respectively)17. Perfusion of mouse spinal cord slices with TAT-CBD3 reduced the frequency of sEPSCs by 57% without changing their amplitudes, supporting the idea that it has a presynaptic action (Fig. 2f). By contrast, the control TAT-Scramble had no effect on eEPSC frequency (Fig. 2e,f). Recordings from layer V pyramidal neurons in cortical brain slices also showed that TAT-CBD3 reduced the probability of glutamate release from stimulated presynaptic terminals (Supplementary Fig. 4).

TAT-CBD3 reduces evoked CGRP releaseCalcium entry through presynaptic CaV2.2 on small-diameter sensory neurons18,19 is directly coupled to transmitter release20,21. Consequently, we investigated whether TAT-CBD3 could modulate the release of immunoreactive CGRP (iCGRP) from dissociated mouse DRG neurons treated with 10 µM of peptide. Pretreatment with TAT-CBD3, but not with TAT-Scramble, for 20 min or 12 h reduced the CGRP release evoked by 50 mM potassium chlo-ride without affecting resting release (Supplementary Fig. 5a–d). Furthermore, total CGRP content was unaffected by the peptides. Cell viability, measured after 12 h incubation, was not affected by any treatments (Supplementary Fig. 5).

We next examined the effect of TAT-CBD3 on capsaicin-evoked release of CGRP from rat spinal cord slices. This release occurs pri-marily from the central terminals of neurons that express the tran-sient receptor potential vanilloid type 1 (TRPV1) channel22, which is important for pain transduction. Perfusion with peptides did not change basal iCGRP release (Fig. 3a,b). However, perfusion with 20 µM (but not 10 µM; Supplementary Fig. 5e,f) of TAT-CBD3 led to a decrease in capsaicin-evoked iCGRP release compared to that seen in cells perfused with TAT-Scramble (Fig. 3c). The total iCGRP content did not differ (Fig. 3d). The fact that TAT-CBD3 had no effect on TRPV1 current recordings (peak amplitudes and activation rates)

3 6 9 12 15 18 21 24 27 30 33 36 39 420

0.2

0.4

0.6

0.8

iCG

RP

(%

con

tent

per

min

)

TAT-Scramble (20 µM)

a

* **

3 6 9 12 15 18 21 24 27 30 33 36 39 42

TAT-CBD3 (20 µM)

0

0.2

0.4

0.6

0.8

iCG

RP

(%

con

tent

per

min

)

c

b HEPESCap (500 nM)

*

*

*

NS

n = 7

n = 7

0

2

4

6

iCG

RP

(%

con

tent

per

9 m

in)

TAT-Scramble (20 µM)

TAT-CBD3 (20 µM)

100

200

300

400

Tot

al iC

GR

P (

fmol

mg–1

)

0

d

(7) (7)

20 µM 500

n = 7

NS

Time(min)

Time(min)

Basal

Basal

Stimula

ted

Stimula

ted

TAT-

Scram

ble

TAT-CBD3

Evoke

d

Evoke

d

HEPESCap (500 nM)

**

Figure 3 TAT-CBD3 reduces capsaicin-stimulated release of iCGRP from spinal cord slices. (a–d) iCGRP release was measured in three 3-min exposures to HEPES buffer alone (white bars), to HEPES buffer with peptides (white), to HEPES buffer containing 500 nM capsaicin (Cap) with peptides (yellow), then to HEPES alone to re-establish baseline (white). Each column represents the mean ± s.e.m. of iCGRP levels in each 3-min perfusate sample, expressed as percentage of total peptide content in the tissues per minute (n = 7 rats). TAT-Scramble (a) or TAT-CBD3 (b), was included as indicated by the horizontal bars. *P < 0.05 versus basal iCGRP release in the absence of capsaicin (analysis of variance (ANOVA), Dunnett’s post hoc test). Neither peptide altered basal iCGRP release (not significant, NS). (c) Basal release is the amount of iCGRP released in the three fractions exposed to HEPES plus peptides. Stimulated release is the amount of iCGRP released in the three fractions exposed to 500 nM capsaicin + peptides. The evoked release was obtained by subtracting iCGRP release during three basal fractions (12–18 min) from that during the three capsaicin-stimulated fractions (21–27 min) and expressed as percentage of total iCGRP content in each group. *P < 0.05 versus the respective TAT-Scramble using a Student’s t test. (d) Total content of iCGRP (in fmol mg−1) is the sum of CGRP released during perfusion and from spinal cord tissue measured at the end of the release experiments.

© 2

011

Nat

ure

Am

eric

a, In

c. A

ll ri

gh

ts r

eser

ved

.

A r t i c l e s

nAture medicine VOLUME 17 | NUMBER 7 | JULY 2011 825

from DRG neurons (Supplementary Fig. 6) shows that TAT-CBD3 does not act through direct inhibition of TRPV1 channels.

TAT-CBD3 affects vasodilatation in rat dura mater in vivo The dura mater is innervated by trigeminal, capsaicin-sensitive sen-sory neurons that mediate meningeal vascular responses related to headache pain23. As the release of CGRP from sensory nerve endings causes vasodilation, we tested the potential involvement of CRMP-2 using in vivo laser Doppler blood flowmetry in rats to measure capsaicin-induced changes in blood flow24 (Fig. 4a). Capsaicin induced CGRP-dependent dilation of meningeal blood vessels (Fig. 4b,c) that faded to baseline values within minutes. Dural application of TAT-CBD3 before nasal administration of capsaicin inhibited cap-saicin-induced changes in blood flow in a dose-dependent manner (Fig. 4c,d). TAT-CBD3 alone did not alter basal blood flow: changes in blood flow without capsaicin were as follows (in tissue perfusion units): vehicle, −6 ± 1 (n = 5); TAT-Scramble, −9 ± 3 (n = 5); and TAT-CBD3, −4 ± 3 (n = 4).

The capsaicin-induced blood flow changes were CGRP dependent, as they could be blocked by prior dural administration of the CGRP antagonist CGRP8–37 (Fig. 4c). The concentration-response curve of percentage inhibition (versus averaged TAT-Scramble) of blood flow by TAT-CBD3 yielded a half-maximal inhibitory concentration of 3.1 ± 1.1 µM (Fig. 4d).

TAT-CBD3 reduces evoked nocifensive behaviorsAs the inhibition of CaV2.2 is antinociceptive25, we examined whether TAT-CBD3 could attenuate nociceptive responses in rat pain models. We first investigated the effects of peptides on formalin-induced nocifensive behavior. In rats that received a subcutaneous injection (to the dorsal surface of the hindpaw) of vehicle (20 µl 0.5% vol/vol DMSO) 30 min before injection of formalin (2.5% vol/vol in 50 µl), we observed the expected biphasic formalin response26 (Fig. 5a). Immediately after formalin injection, the rats showed a high degree of flinching (phase 1) for about 10 min followed by a second period of flinching (phase 2) which subsided by 60 min. Pretreatment with 30 or 100 µM TAT-Scramble did not change either phase of the for-malin test. By contrast, rats pretreated with 30 or 100 µM TAT-CBD3 showed blunted nociceptive behaviors in both phases (Fig. 5a,b), which suggests that TAT-CBD3 inhibits nociception mediated by

direct activation of sensory neurons (phase 1) and, to some extent, nociception associated with inflammation27 and spinal involvement (phase 2)28,29. Pretreatment with 3 µM TAT-CBD3 did not affect the formalin-induced behavior (Fig. 5a,b). Injection of peptides alone, before the formalin injection, did not induce any nocifensive behavior. Formalin (2.5%) produced a fourfold change in paw thickness (ipsi-lateral minus contralateral) compared to saline (Fig. 5c), consistent with the edema that typically follows inflammation. TAT-CBD3 did not inhibit formalin-induced edema.

To determine whether TAT-CBD3 inhibits capsaicin-induced noci-ception, we used the capsaicin eye-wipe test. The cornea is inner-vated by trigeminal afferent nerves, of which ~25% express TRPV1 (refs. 30,31). Application of TAT-CBD3 alone to the rat cornea did not induce nocifensive behavior. Pretreatment with 30 or 100 µM TAT-CBD3 for 30 min attenuated capsaicin-induced nocifensive behavior (Fig. 5d), suggesting that TAT-CBD3 is antinociceptive at a peripheral site. Pretreatment with 3 µM TAT-CBD3 did not affect nocifensive behavior; however, 100 µM TAT-Scramble showed a nonspecific effect, increasing the nocifensive response time (Fig. 5d).

TAT-CBD3 attenuates ddC-induced neuropathic pain behaviorWe next examined the effects of the peptide on chronic nociceptive behavior in an animal model of painful neuropathy induced by AIDS therapy32,33. Nucleoside reverse transcriptase inhibitors such as 2′,3′ dideoxycytidine (ddC), which are commonly used to treat AIDS, pro-duce side effects that include painful neuropathies. We evaluated the ability of peptides to reverse tactile hypersensitivity in rats 7 d after a single injection of ddC. TAT-CBD3 alone had no effect on paw withdrawal threshold (PWT). TAT-CBD3, but not TAT-Scramble, caused a dose-dependent increase in PWT when administered intra-peritoneally after ddC treatment (Fig. 5e). Tactile hypersensitivity was completely reversed 1 h after intraperitoneal injection of 1 mg kg−1 TAT-CBD3. Four hours after injection, the TAT-CBD3–induced reversal of hypersensitivity had diminished by 50%, which may be accounted for by degradation and biodistribution of the peptide34. To investigate the distribution of peptides after intraperitoneal injection, we collected tissue samples from animals injected with FITC–TAT-CBD3. After the injection we detected the peptide in the DRG at 15 min (Fig. 5f–i), spinal cord at 15 min (Fig. 6) and brain at 1 h (Supplementary Fig. 7a–c). We observed transient contortions in

Figure 4 TAT-CBD3 reduces changes in meningeal blood flow induced by capsaicin. (a) Experimental paradigm for the laser Doppler flowmetry measurements. (b) Representative normalized traces of middle meningeal blood flow changes in response to nasally administered capsaicin (100 nM) in the presence of TAT-Scramble (30 µM, green trace) or TAT-CBD3 pretreatment (30 µM, purple trace, applied durally 15 min before capsaicin administration). Laser Doppler flowmetry measurements were collected at 1 Hz and binned by averaging every 10 samples for graphical representation. The data from each rat were normalized to the first 3 min of basal data and the horizontal dashed line indicates the calculated baseline. The ordinate represents red blood cell flux measurements in arbitrary units (AU). (c) Summary of blood flow changes after nasal administration of capsaicin in the absence or presence of previous administration of TAT-CBD3 (3, 10 or 30 µM) or TAT-Scramble to the dura. *P < 0.05 versus vehicle (unpaired Student’s t test). The number of rats tested for each condition is indicated in parentheses. (d) Concentration-response curve of percentage inhibition (versus averaged TAT-Scramble) of blood flow by TAT-CBD3. Error bars represent means ± s.e.m.

a SIF (dural) Peptides (dural)

Stimulated

15 min15 min 15 min

Cap (nasal)

PretreatmentBasal blood flow

d100

75

50

25

01 10

Inhi

bitio

n of

blo

od fl

ow (

%)

com

pare

d to

TA

T s

cram

ble

TAT-CBD3(µM)

n = 4–5

50

b

100 nM Cap (nasal)

TAT-CBD3 (30 µM) 3 minTAT-Scramble

5 AU

Peptides (dural)

c

0

10

20

30

40

Cha

nge

in b

lood

flow

(%

)

Cap (100 nM): + + +

TAT-CBD3

+ + +

TAT-Scramble

CGRP 8–37

Vehicl

e

(5) (5) (4)(5)(3)

**

(5)

*

1 µM 30 µM 30 µM 3 µM 10 µM

© 2

011

Nat

ure

Am

eric

a, In

c. A

ll ri

gh

ts r

eser

ved

.

A r t i c l e s

826 VOLUME 17 | NUMBER 7 | JULY 2011 nAture medicine

rats injected intraperitoneally with 20 mg per kg body weight of TAT-CBD3, but not at lower doses that inhibited hypersensitivity (Fig. 5e and Supplementary Fig. 7d–h).

Thus, our results indicate that TAT-CBD3, which interferes with the interaction between CaV2.2 and CRMP-2, reduces acute inflam-matory and neuropathic pain behaviors.

TAT-CBD3 does not cause neurobehavioral deficitsWe investigated whether TAT-CBD3 had any effect on mouse motor coor-dination, locomotor function, sedation (rotarod test35) and hippocampus-dependent memory (Morris water-maze test36,37). Impaired locomotor function did not account for reduced flinching and paw withdrawal as TAT-CBD3 (10 and 50 mg per kg body weight; intraperitoneal injec-tion) had no effect in the accelerating rotarod test (Fig. 6a). There was also no effect of TAT-CBD3 (10 mg per kg body weight; intraperitoneal

injection) on motor coordination (rotarod test) and spatial memory retrieval, as measured by the Morris water maze, between 1 h and 7 d after administration (Fig. 6b,c). We observed a single episode of kinking at the base of the tail and whole body contortion in rats imme-diately after single injections of doses higher than 10 mg per kg body weight of TAT-CBD3 (Supplementary Fig. 7d–g).

As pharmacological blockade of N-type channels has been clinically linked to anxiety and depression38, we tested whether TAT-

CDB3 could alter these behaviors in mice using the elevated plus maze (EPM) test and light-dark box test (LDBT). These paradigms assess the conflict between hiding in enclosed dark areas (dark box or closed arm) and exploring new environments (white box or open arm)39,40. In the EPM test, neither time spent in the open or closed arms nor the frequency of entries into the open or closed arms was altered by any dose of TAT-CBD3, compared to TAT-Scramble (Fig. 6h and Supplementary Table 1). In the LDBT, although time spent in the white and dark boxes did not differ between any of the condi-tions, the number of transitions between the light and dark boxes was higher in mice injected with 1 mg per kg body weight TAT-CBD3 compared to TAT-Scramble (Fig. 6i). These results suggest that TAT-CBD3 does not affect anxiety-associated behaviors apart from increasing transitions in LDBT, supporting the idea that it may have slightly anxiolytic properties.

ba

d e

Hoechst

NeuNFITC-TAT-CBD3

f g

h i

Merged

c

Saline

0

20

40

60

80

100

0

100

200

300

400

Flin

ches

Flinches

Phase1(0–10 min)

Phase2(15–60 min)120 600

500

Vehicl

e

No

form

alin

30 µM

100

µM3

µM30

µM

100

µMSali

ne

Vehicl

e

30 µM

100

µM3

µM30

µM

100

µM

No

form

alin

TAT-Scr. TAT-CBD3 TAT-Scr. TAT-CBD3

* *

n = 4–10

* *

Vehicl

e

TAT-Scr

.

TAT-CBD3

n = 4–10

Paw

thic

knes

s(ip

si –

con

tra,

mm

)

0

1

2

Saline

* * *

Formalin (2.5%)

n = 4–10

0

20

40

60

80

100

10 20 30 40 50 60Time (min)

Flin

ches

Saline Formalin

TAT-CBD3

3 µM30 µM100 µM

TAT-Scramble

30 µM 100 µM

Vehicle TAT-Scr. TAT-CBD3

(30 µM) (100 µM)DMSO (3 µM) (30 µM) (100 µM)Cap (3 µM): + ++ + –––––– +

Noc

ifens

ive

resp

onse

(s)

* *

n = 3–7

0

50

100

150

200

250

+

Treatment:

#

Paw

with

draw

al th

resh

old

(mN

)

80

60

40

20

0

1 h after i.p. injection 4 h afteri.p. injection

*

* *

* *

n = 7

Pre-o

p m

ean

TAT-CBD3

ddC P

ID7

TAT-Scr

.

TAT-CBD3

TAT-CBD3

TAT-CBD3

TAT-CBD3

TAT-CBD3

PID7 0.1 mg kg–1no ddC; 1 mg kg–1No ddC

PID7 10 mg kg–1PID7 1 mg kg–1

Figure 5 TAT-CBD3 reduces acute, inflammatory and neuropathic pain. (a) Time course of the number of flinches after subcutaneous (dorsal surface of paw) injection of formalin (2.5% in 50 µl saline) in rats pretreated with peptides (3–100 µM; 20 µl into dorsal surface of paw) 30 min before formalin (n = 4–10). (b) The effect of peptides on number of flinches on formalin-induced phase 1 (0–10 min) and phase 2 (15–60 min). *P < 0.05 versus formalin-injected rats. (c) Paw thickness, as measured 1 h after injection of saline, formalin or formalin + peptides (100 µM). *P < 0.05 versus saline-injected rats. (d) Pretreatment with TAT-CBD3 peptide attenuates capsaicin-evoked nocifensive behavior. Vehicle (0.3% DMSO) or peptides (concentrations as indicated) in saline (40 µl) was instilled corneally and nocifensive behavior noted. Five minutes later, capsaicin (3 µM in 40 µl saline) was applied corneally and nocifensive behavior noted. *P < 0.05 versus 30 or 100 µM TAT-scramble or 3 µM TAT-CBD3; #P < 0.05 versus all conditions except 3 µM TAT-CBD3 (ANOVA with Dunnett’s post hoc test). (e) PWT (mN) of rats injected once with ddC and treated with TAT peptides on day 7 after injection (PID7). Response of ddC alone at PID7 is shown in the brown bar. *P < 0.05 versus ddC or TAT-Scramble (ANOVA with Student-Newman-Keuls post hoc test). (f–i) Fluorescent imaging of DRGs, isolated 15 min after injection of FITC-TAT-CBD3 (f, FITC; green) and immunolableled with a NeuN-specific antibody (g, NeuN; red). Cells were also stained with Hoechst (h, blue), which labels cell nuclei. (i) Merged image. Scale bars, 100 µm (f–i). Error bars represent means ± s.e.m.

© 2

011

Nat

ure

Am

eric

a, In

c. A

ll ri

gh

ts r

eser

ved

.

A r t i c l e s

nAture medicine VOLUME 17 | NUMBER 7 | JULY 2011 827

As rodents show immobile postures when placed in inescapable, stressful situations41, immobility behavior in the tail suspension test (TST) or forced swim test (FST) is used as a measure of ‘depression’ or ‘despair-associated’ behavior, which is attenuated by antidepres-sant treatments42. In the TST in mice, neither time spent immobile nor the frequency of immobile episodes was altered by any dose of TAT-CBD3, compared to TAT-Scramble (Fig. 6j and Supplementary Table 1). Overall, no dose of TAT-CBD3 altered depression or despair- associated behavior.

DISCUSSIONOur findings show that administration of TAT-conjugated CBD3 peptide interfered with CaV2.2 trafficking to the presynaptic membrane and that treatment with TAT-CBD3 peptide inhibited calcium currents, stimulus-evoked neuropeptide release from sensory neurons and excitatory synaptic transmission in dor-sal horn neurons in rats and mice. TAT-CBD3, which disrupts the enhancement of CaV2.2 function by CRMP-2, achieves a therapeutic window suitable for a number of pain states, both

inflammatory and neuropathic, with no impairment of motor function or higher order processes. TAT-CBD3 seems to affect pain signaling by regulating CaV2.2.

To block interactions between CRMP-2 and the N-type calcium channel, we designed a CaV2.2-binding peptide, CBD3, which is fully conserved between rodents and humans and has little or no sequence homology with other proteins. We conjugated the peptide to the HIV-1 TAT domain to overcome the obstacle of poor plasma membrane penetrance of peptides43,44.

Pharmacological block of CaV2.2 not only reduces presynaptic neurotransmitter release but also might decrease the excitability of the postsynaptic neurons in lamina I of the spinal cord45. The decrease in sEPSC frequency in the postsynaptic, spinal cord lamina II neurons after treatment with TAT-CBD3 may be due to both inhibition of glutamate release from sensory neurons and diminished vesicular recycling25.

It has been suggested that CaV2.2 instigates the increased excitability and neurotransmitter release that are associated with chronic and neuro-pathic pain conditions19,25,45,46. Genetic and pharmacological block

a b c

d e f

FITC-TAT-CBD3 NeuN FITC-TAT-CBD3 NeuN Hoechst

g

FIT

C-T

AT

-CB

D3

Neu

NH

oech

st

3 h 1 d 2 d 3 d0

5

10

15

20

25

Late

ncy

(s)

VehicleTAT-CBD3TAT-Scramble

Time after injection

n = 7

h i j

0

50

100

150

200

Tim

e (s

)

n = 12–14

Open arm time Closed arm time0

20

40

60

80

100

Tim

e (s

)

Immobilityduration

Immobilityfrequency

n = 12–14

*

0

10

20

30

40

Tim

e in

targ

et q

uadr

ant (

%)

3 d 3 d

n = 7

Time after injection3 d 3 d

Pat

h le

ngth

(m

)

0

1

2

3

n = 7

TAT-CBD3TAT-Scramble

0

20

40

60

80

100

Tim

e on

rot

arod

(s)

1 h 6 h 1 d 3 d 7 d0

20

40

60

80

100

Tim

e on

rot

arod

(s)

Time after injection1 h 6 h 1 d 3 d 7 d

Time after injection

n = 7n = 7

10 mg kg–1TAT-Scramble TAT-CBD3 10 mg kg–1Vehicle 50 mg kg–1

TAT-Scramble TAT-CBD310 mg kg–1

10 mg kg–1

0.1 mg kg–1

1 mg kg–1

0

50

100

150

200

Tim

e (s

)

n = 12–14

Duration in light box Duration in dark box Latency to enterdark box

Number oftransitions

0

4

12

8

Tim

e (s

)

n = 12–14

Figure 6 TAT-CBD3 has no effect on sensorimotor and cognitive functions but has a mild anxiolytic effect. (a) Latency to fall off a slow (left) or fast (right) rotating rod. There were no significant differences in rotarod performances between groups (ANOVA with Dunnett’s post hoc test). (b) Latency for mice to find a hidden platform in the Morris water maze was not different between groups. (c) Time spent in target quadrant. There were no significant differences in percentage time spent in target quadrant or path length between groups (Student’s t test). (d–g) Uptake of FITC-TAT-CBD3 into neurons in ventral horn 15 min after intraperitoneal (i.p.) injection (20 mg kg−1). TAT-CBD3 (d, green; FITC) accumulated in motor neurons (arrowheads), which co-labeled with NeuN (e, red). Nuclei are stained with Hoechst (blue). Merged images show co-labeling of FITC-TAT-CBD3–containing neurons with NeuN and Hoescht at low (f) and high magnifications (g). Scale bars, 100 µm (d–f); 40 µm (g). (h) Elevated plus maze test to evaluate anxiety-associated behaviors. (i) Light dark box test for anxiety-associated behaviors. (*P < 0.05 versus TAT-Scramble, one-way ANOVA with Dunnett’s post hoc test.) (j) Tail suspension test of depression- or despair-associated behaviors. Error bars represent means ± s.e.m.

© 2

011

Nat

ure

Am

eric

a, In

c. A

ll ri

gh

ts r

eser

ved

.

A r t i c l e s

828 VOLUME 17 | NUMBER 7 | JULY 2011 nAture medicine

of CaV2.2 after injury in rodents attenuates nociceptive behavior47,48. Moreover, expression of CaV2.2 is upregulated in several animal mod-els of neuropathic pain46. Inhibition of CaV2.2 is also one mechanism of morphine-induced analgesia49. A pro-nociceptive role for CaV2.2 is further highlighted by the identification of alternative splice vari-ants of CaV2.2 expressed on small-diameter nociceptive neurons50 that contribute to thermal and mechanical hyperalgesia51. For these reasons, CaV2.2 has become a prime therapeutic target in the treat-ment of chronic pain.

To determine the effects of TAT-CBD3 on nociception, we used a variety of rat pain models that included acute and inflammatory or chronic nociception states. Behavioral outcomes from the capsaicin eye-wipe test52 suggested that the peptide inhibited acute nociception. In addition, the administration of TAT-CBD3 into the dorsal surface of the paw significantly reduced the number of flinches in phase 1 and phase 2 of the formalin test. TAT-CBD3 had a greater effect in phase 1 of the formalin test than in phase 2, which at first may suggest that CaV2.2 has a more pronounced effect in primary nociception than in the perception of inflammatory pain. Phase 1 of the formalin test results from direct stimulation of nociceptors, whereas phase 2 involves a period of central sensitization during which inflammatory phenomena occur53. Therefore, because the peptide was injected in the periphery, the effect observed in phase 1 suggests that TAT-CBD3 either affects transmission of the nociceptive signal or inhibits the release of CGRP or other neuropeptides at the peripheral end of the nociceptors. The observation that the effects of TAT-CBD3 were anti-nociceptive in acute studies is consistent with the activity-dependent regulation that has been demonstrated for the interaction between CRMP-2 and CaV2.2 (ref. 5), consistent with a decrease in presynaptic neuronal excitability.

As our results in other models showed that CGRP release was inhibited, we hypothesized that this could explain the effect of TAT-CBD3 in phase 1 of the formalin test. If TAT-CBD3 inhibits the peripheral release of CGRP, then one would expect a reduction in edema. However, TAT-CBD3 did not inhibit formalin-induced edema. Although the peripheral inflammatory process and its relation-ship to nociception are not completely understood, edema occurs in response to several inflammatory mediators that could be released by cells other than neurons. Therefore, inhibiting the release of only CGRP is probably insufficient to reduce edema. The differential effects of TAT-CBD3 on nociceptive behavior and edema suggest that these inflammatory components do not share a common mechanism. In support of this assertion, a serotonin receptor antagonist inhibits nociceptive behavior induced by 2.5% formalin but does not inhibit edema54. Moreover, morphine administered peripherally inhibits carrageenan-induced hyperalgesia without inhibiting edema, whereas systemically injected morphine reduced edema, plasma extravasation and inflammatory hyperalgesia55. Thus, although TAT-CBD3 did not inhibit formalin-induced edema, an effect on neurotransmitter release cannot be ruled out as a mechanism of its peripheral antinociceptive effect in our pain models.

In addition, we found that TAT-CBD3 suppressed tactile hypersensitivity in an animal model of HIV treatment–induced peripheral neuropathy, a chronic model of neuropathic pain. This model uses the antiretroviral treatment ddC to induce the dying-back neuropathy of small fibers that is seen in patients with AIDS after treatment, which has been attributed to reduced cal-cium buffering32,33,56–58. Systemic administration of TAT-CBD3 reversed ddC-induced nociceptive behaviors, suggesting that the interaction between CRMP-2 and CaV2.2 has a continued role in

neurotransmitter release. Consistent with this hypothesis, CaV2.2 mediates enhanced release of neurotransmitters in the spinal cord that is important for the maintenance of inflammatory pain59.

Despite the potential of pharmacological inhibitors of N-type channels in the treatment of intractable or chronic pain, they have a narrow therapeutic window25. Intrathecal delivery of Prialt in animal and clinical studies results in a multitude of deleterious side-effects including impairments in learning and memory and motor coordina-tion, and increased anxiety or depression4,19,25,48. At doses more than 50-fold higher than that required to reduce hypersensitivity in vivo, TAT-CBD3 produced mild motor impairment (transient tail kink-ing and body contortion) but had no effect on motor coordination, memory retreival or anxiety and depression-associated behaviors in these animals. TAT-CBD3 had a mild anxiolytic effect consistent with that observed in mice lacking CaV2.2 (ref. 47). The relative lack of toxicity observed with systemic delivery of TAT-CBD3 provides evidence that it has therapeutic promise.

From these findings, we propose that TAT-CBD3 suppresses pain hypersensitivity without directly blocking CaV2.2, but rather by inhibiting the binding of a regulator of CaV2.2 function, CRMP-2. Thus, our findings point toward an approach that might be useful for managing clinical pain.

METHODSMethods and any associated references are available in the online version of the paper at http://www.nature.com/nm/.

Note: Supplementary information is available on the Nature Medicine website.

ACKNOWLeDGMeNTSThis work is supported by grants from the US National Institutes of Health: Dental and Craniofacial Research (DE14318-06 to J.C.F. and DE017794 to R.-R.J.), Drug Abuse (DA026040 to F.A.W.), Neurological Disorders and Stroke (NS051668 to C.M.H. and NS050131 to N.B.) and Environmental Health Sciences (ES017430 to G.S.O. and J.H.H.); the Indiana State Department of Health−Spinal Cord and Brain Injury Fund (A70-0-079212 to N.B. and A70-9-079138 to R.K.) and the Indiana University Biomedical Committee–Research Support Funds (2286501 to R.K.); a National Scientist Development Grant from the American Heart Association (SDG5280023 to R.K.); and the Elwert Award in Medicine to R.K. J.M.B. is the recipient of a Larry Kays Medical Neuroscience fellowship. S.M.W. is a Stark Scholar. We thank A. Molosh and members of the Pain and Sensory Group for discussions, S.K. Ahuja for assistance with behavioral experiments and C. Kohn for comments on the manuscript.

AUTHOR CONTRIBUTIONSJ.M.B. performed molecular biology, biochemistry and calcium imaging experiments and analyzed the data. D.B.D. carried out the spinal cord slice release and formalin behavior experiments and helped to write the manuscript. S.M.W. performed immunocytochemistry and wrote the manuscript. C.B. carried out the laser Doppler blood flowmetry. J.H.H. analyzed the blood flow data. P.L.J. and S.D.F. performed anxiety and despair behavior experiments. W.Z. and Y.W. performed DRG and hippocampal patching. C.-K.P. conducted electrophysiology in spinal cord slices. W.X. and X.J. performed electrophysiology on brain slices. B.S.S. carried out the DRG release assays. T.B., N.B. and J.M.B. performed and analyzed the calcium imaging experiments. B.M.C., M.R.D. and M.S.R. performed DRG immunocytochemistry and ddC behavior experiments. M.K. and S.O.M. performed the surface plasmon resonance experiments and analyzed the data. N.L. performed the rotarod and water maze experiments. J.C.F. performed the nocifensive behavior experiments and editing of the manuscript. N.M.A. and A.H. synthesized the peptide blot. X.-M.X., C.M.H., M.R.V., G.S.O. and A.S. contributed to editing of the manuscript. R.-R.J contributed to electrophysiology of spinal cord slices and editing of the manuscript. F.A.W. analyzed the ddC behavior data and contributed to writing and editing the manuscript. R.K. identified the peptide, conceived the study, designed and supervised the overall project and wrote the manuscript.

COMPeTING FINANCIAL INTeReSTSThe authors declare no competing financial interests.

© 2

011

Nat

ure

Am

eric

a, In

c. A

ll ri

gh

ts r

eser

ved

.

A r t i c l e s

nAture medicine VOLUME 17 | NUMBER 7 | JULY 2011 829

Published online at http://www.nature.com/naturemedicine/. Reprints and permissions information is available online at http://www.nature.com/reprints/index.html.

1. Staats, P.S. et al. Intrathecal ziconotide in the treatment of refractory pain in patients with cancer or AIDS: a randomized controlled trial. J. Am. Med. Assoc. 291, 63–70 (2004).

2. Rauck, R.L. et al. A randomized, double-blind, placebo-controlled study of intrathecal ziconotide in adults with severe chronic pain. J. Pain Symptom Manage. 31, 393–406 (2006).

3. Wallace, M.S. Intrathecal ziconotide for severe chronic pain: safety and tolerability results of an open-label, long-term trial. Anesth. Analg. 106, 628–637 (2008).

4. Schmidtko, A., Lotsch, J., Freynhagen, R. & Geisslinger, G. Ziconotide for treatment of severe chronic pain. Lancet 375, 1569–1577 (2010).

5. Brittain, J.M. et al. An atypical role for collapsin response mediator protein 2 (CRMP-2) in neurotransmitter release via interaction with presynaptic voltage-gated Ca2+ channels. J. Biol. Chem. 284, 31375–31390 (2009).

6. Chi, X.X. et al. Regulation of N-type voltage-gated calcium (CaV2.2) channels and transmitter release by collapsin response mediator protein-2 (CRMP-2) in sensory neurons. J. Cell Sci. 122, 4351–4362 (2009).

7. Goshima, Y., Nakamura, F., Strittmatter, P. & Strittmatter, S.M. Collapsin-induced growth cone collapse mediated by an intracellular protein related to UNC-33. Nature 376, 509–514 (1995).

8. Inagaki, N. et al. CRMP-2 induces axons in cultured hippocampal neurons. Nat. Neurosci. 4, 781–782 (2001).

9. Yoshimura, T. et al. GSK-3beta regulates phosphorylation of CRMP-2 and neuronal polarity. Cell 120, 137–149 (2005).

10. Morita, T. & Sobue, K. Specification of neuronal polarity regulated by local translation of CRMP2 and Tau via the mTOR-p70S6K pathway. J. Biol. Chem. 284, 27734–27745 (2009).

11. Arimura, N. et al. CRMP-2 directly binds to cytoplasmic dynein and interferes with its activity. J. Neurochem. 111, 380–390 (2009).

12. Fukata, Y. et al. CRMP-2 binds to tubulin heterodimers to promote microtubule assembly. Nat. Cell Biol. 4, 583–591 (2002).

13. Wang, Y. et al. In silico docking and electrophysiological characterization of lacosamide binding sites on collapsin response mediator protein 2 (CRMP-2) identifies a pocket important in modulating sodium channel slow inactivation. J. Biol. Chem. 285, 25296–25307 (2010).

14. Schwarze, S.R., Ho, A., Vocero-Akbani, A. & Dowdy, S.F. In vivo protein transduction: delivery of a biologically active protein into the mouse. Science 285, 1569–1572 (1999).

15. Andrade, A., Denome, S., Jiang, Y.Q., Marangoudakis, S. & Lipscombe, D. Opioid inhibition of N-type Ca2+ channels and spinal analgesia couple to alternative splicing. Nat. Neurosci. 13, 1249–1256 (2010).

16. Todd, A.J. Neuronal circuitry for pain processing in the dorsal horn. Nat. Rev. Neurosci. 11, 823–836 (2010).

17. Liu, T., Xu, Z.Z., Park, C.K., Berta, T. & Ji, R.R. Toll-like receptor 7 mediates pruritus. Nat. Neurosci. 13, 1460–1462 (2010).

18. Yamamoto, T. & Takahara, A. Recent updates of N-type calcium channel blockers with therapeutic potential for neuropathic pain and stroke. Curr. Top. Med. Chem. 9, 377–395 (2009).

19. Zamponi, G.W., Lewis, R.J., Todorovic, S.M., Arneric, S.P. & Snutch, T.P. Role of voltage-gated calcium channels in ascending pain pathways. Brain Res. Rev. 60, 84–89 (2009).

20. Catterall, W.A. & Few, A.P. Calcium channel regulation and presynaptic plasticity. Neuron 59, 882–901 (2008).

21. Dodge, F.A. Jr. & Rahamimoff, R. Co-operative action a calcium ions in transmitter release at the neuromuscular junction. J. Physiol. (Lond.) 193, 419–432 (1967).

22. Patapoutian, A., Tate, S. & Woolf, C.J. Transient receptor potential channels: targeting pain at the source. Nat. Rev. Drug Discov. 8, 55–68 (2009).

23. Goadsby, P.J. Calcitonin gene-related peptide (CGRP) antagonists and migraine: is this a new era? Neurology 70, 1300–1301 (2008).

24. Kunkler, P.E., Ballard, C.J., Oxford, G.S. & Hurley, J.H. TRPA1 receptors mediate environmental irritant-induced meningeal vasodilatation. Pain 152, 38–44 (2010).

25. Snutch, T.P. Targeting chronic and neuropathic pain: the N-type calcium channel comes of age. NeuroRx 2, 662–670 (2005).

26. Dubuisson, D. & Dennis, S.G. The formalin test: a quantitative study of the analgesic effects of morphine, meperidine, and brain stem stimulation in rats and cats. Pain 4, 161–174 (1977).

27. Hunskaar, S. & Hole, K. The formalin test in mice: dissociation between inflammatory and non-inflammatory pain. Pain 30, 103–114 (1987).

28. Dickenson, A.H. & Sullivan, A.F. Subcutaneous formalin-induced activity of dorsal horn neurones in the rat: differential response to an intrathecal opiate administered pre or post formalin. Pain 30, 349–360 (1987).

29. Coderre, T.J., Vaccarino, A.L. & Melzack, R. Central nervous system plasticity in the tonic pain response to subcutaneous formalin injection. Brain Res. 535, 155–158 (1990).

30. Guo, A., Vulchanova, L., Wang, J., Li, X. & Elde, R. Immunocytochemical localization of the vanilloid receptor 1 (VR1): relationship to neuropeptides, the P2X3 purinoceptor and IB4 binding sites. Eur. J. Neurosci. 11, 946–958 (1999).

31. Nakamura, A. et al. Morphological and immunohistochemical characterization of the trigeminal ganglion neurons innervating the cornea and upper eyelid of the rat. J. Chem. Neuroanat. 34, 95–101 (2007).

32. Joseph, E.K., Chen, X., Khasar, S.G. & Levine, J.D. Novel mechanism of enhanced nociception in a model of AIDS therapy-induced painful peripheral neuropathy in the rat. Pain 107, 147–158 (2004).

33. Bhangoo, S.K., Ripsch, M.S., Buchanan, D.J., Miller, R.J. & White, F.A. Increased chemokine signaling in a model of HIV1-associated peripheral neuropathy. Mol. Pain. 5, 48 (2009).

34. Cai, S.R. et al. The kinetics and tissue distribution of protein transduction in mice. Eur. J. Pharm. Sci. 27, 311–319 (2006).

35. Hamm, R.J., Pike, B.R., O’Dell, D.M., Lyeth, B.G. & Jenkins, L.W. The rotarod test: an evaluation of its effectiveness in assessing motor deficits following traumatic brain injury. J. Neurotrauma 11, 187–196 (1994).

36. D’Hooge, R. & De Deyn, P.P. Applications of the Morris water maze in the study of learning and memory. Brain Res. Brain Res. Rev. 36, 60–90 (2001).

37. Gage, F.H., Chen, K.S., Buzsaki, G. & Armstrong, D. Experimental approaches to age-related cognitive impairments. Neurobiol. Aging 9, 645–655 (1988).

38. Rauck, R.L., Wallace, M.S., Burton, A.W., Kapural, L. & North, J.M. Intrathecal ziconotide for neuropathic pain: a review. Pain Pract. 9, 327–337 (2009).

39. Bourin, M. & Hascoet, M. The mouse light/dark box test. Eur. J. Pharmacol. 463, 55–65 (2003).

40. Lalonde, R. & Strazielle, C. Relations between open-field, elevated plus-maze, and emergence tests as displayed by C57/BL6J and BALB/c mice. J. Neurosci. Methods 171, 48–52 (2008).

41. Varty, G.B., Cohen-Williams, M.E. & Hunter, J.C. The antidepressant-like effects of neurokinin NK1 receptor antagonists in a gerbil tail suspension test. Behav. Pharmacol. 14, 87–95 (2003).

42. Cryan, J.F., Mombereau, C. & Vassout, A. The tail suspension test as a model for assessing antidepressant activity: review of pharmacological and genetic studies in mice. Neurosci. Biobehav. Rev. 29, 571–625 (2005).

43. Schwarze, S.R. & Dowdy, S.F. In vivo protein transduction: intracellular delivery of biologically active proteins, compounds and DNA. Trends Pharmacol. Sci. 21, 45–48 (2000).

44. Yoon, J.S. et al. Characteristics of HIV-Tat protein transduction domain. J. Microbiol. 42, 328–335 (2004).

45. Winquist, R.J., Pan, J.Q. & Gribkoff, V.K. Use-dependent blockade of Cav2.2 voltage-gated calcium channels for neuropathic pain. Biochem. Pharmacol. 70, 489–499 (2005).

46. Cizkova, D. et al. Localization of N-type Ca2+ channels in the rat spinal cord following chronic constrictive nerve injury. Exp. Brain Res. 147, 456–463 (2002).

47. Saegusa, H. et al. Suppression of inflammatory and neuropathic pain symptoms in mice lacking the N-type Ca2+ channel. EMBO J. 20, 2349–2356 (2001).

48. Malmberg, A.B. & Yaksh, T.L. Voltage-sensitive calcium channels in spinal nociceptive processing: blockade of N- and P-type channels inhibits formalin-induced nociception. J. Neurosci. 14, 4882–4890 (1994).

49. Seward, E., Hammond, C. & Henderson, G. Mu-opioid-receptor-mediated inhibition of the N-type calcium-channel current. Proc. Biol. Soc. 244, 129–135 (1991).

50. Bell, T.J., Thaler, C., Castiglioni, A.J., Helton, T.D. & Lipscombe, D. Cell-specific alternative splicing increases calcium channel current density in the pain pathway. Neuron 41, 127–138 (2004).

51. Altier, C. et al. Differential role of N-type calcium channel splice isoforms in pain. J. Neurosci. 27, 6363–6373 (2007).

52. Farazifard, R., Safarpour, F., Sheibani, V. & Javan, M. Eye-wiping test: a sensitive animal model for acute trigeminal pain studies. Brain Res. Brain Res. Protoc. 16, 44–49 (2005).

53. Le Bars, D., Gozariu, M. & Cadden, S.W. Animal models of nociception. Pharmacol. Rev. 53, 597–652 (2001).

54. Doak, G.J. & Sawynok, J. Formalin-induced nociceptive behavior and edema: involvement of multiple peripheral 5-hydroxytryptamine receptor subtypes. Neuroscience 80, 939–949 (1997).

55. Whiteside, G.T., Boulet, J.M. & Walker, K. The role of central and peripheral mu opioid receptors in inflammatory pain and edema: a study using morphine and DiPOA ([8-(3,3-diphenyl-propyl)-4-oxo-1-phenyl-1,3,8-triaza-spiro[4.5]dec-3-yl]-acetic acid). J. Pharmacol. Exp. Ther. 314, 1234–1240 (2005).

56. Yarchoan, R. et al. Phase 1 studies of 2′,3′-dideoxycytidine in severe human immunodeficiency virus infection as a single agent and alternating with zidovudine (AZT). Lancet 1, 76–81 (1988).

57. Dubinsky, R.M., Yarchoan, R., Dalakas, M. & Broder, S. Reversible axonal neuropathy from the treatment of AIDS and related disorders with 2′,3′-dideoxycytidine (ddC). Muscle Nerve 12, 856–860 (1989).

58. Bhangoo, S.K. et al. CXCR4 chemokine receptor signaling mediates pain hypersensitivity in association with antiretroviral toxic neuropathy. Brain Behav. Immun. 21, 581–591 (2007).

59. Nakanishi, O., Ishikawa, T. & Imamura, Y. Modulation of formalin-evoked hyperalgesia by intrathecal N-type Ca channel and protein kinase C inhibitor in the rat. Cell. Mol. Neurobiol. 19, 191–197 (1999).

© 2

011

Nat

ure

Am

eric

a, In

c. A

ll ri

gh

ts r

eser

ved

.

nAture medicine doi:10.1038/nm.2345

ONLINE METHODSRats and mice. Procedures involving rats and mice and their care were in accord-ance with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health publication 85-23) and approved by the Institutional Animal Care and Use Committee of Indiana University School of Medicine. Sprague-Dawley rats and C57BL/6 and CD1 mice were purchased from Harlan.

Cell culture. Primary hippocampal cultures from postnatal day 1 were prepared as described5. Rat DRG neuron cultures were prepared as previously described6. Neuronal CAD cells were cultured as previously described13.

Peptides. TAT-Scramble (YGRKKRRQRRRWEAKEMLYFEALVIE; a random sequence with no homology to any known sequence (TAT sequence shown in underlined text)) and TAT-CBD3 (YGRKKRRQRRRARSRLAELRGVPRGL) were synthesized by Antagene and verified by mass spectroscopy (Department of Chemistry, Indiana University School of Medicine).

We produced peptide spot arrays and far western peptide arrays (10–15-mer) spanning the entire length of rat CRMP-2 in 20–30 arrays onto polyvinylidene fluoride membranes using the SPOTS-synthesis method. We overlaid the mem-branes with purified synaptosome fraction enriched in Ca2+ channels for 1 h and then probed in a far western with a CaV2.2 antibody (Calbiochem).

Biochemistry and immunocytochemistry experiments. We performed immunoprecipitation, in vitro binding assays and surface biotinylation as pre-viously described5,6. For immunoprecipitation experiments, we isolated spinal cords from adult male rats by pressurized hydraulic ejection with ice-cold PBS. We purchased CRMP-2–specific antibody from Chemicon, and βIII tubu-lin–specific antibody from Promega. We determined the binding between TAT peptides and CaV2.2 L1- or Ct-dis-GST fusion proteins by surface plasmon resonance using a BIAcore3000 instrument (Biacore AB). We describe details of this assay as well as immunocytochemistry in CAD cells, DRG and spinal cord in the Supplementary Methods.

Electrophysiological recordings. We performed whole-cell recordings at 25 ± 1 °C from hippocampal5 and sensory neurons6. We recorded Na+ and Ca2+ cur-rents from DRGs using an EPC 10 Amplifier (HEKA Electronics) as previously described6. Recordings from slices prepared from lumbar spinal cord (L4-L5) were as previously described17. We made whole-cell patch-clamp recordings from lamina II neurons in voltage clamp mode17.

Neurotransmitter release radioimmunoassay. We measured stimulus-evoked release and content of iCGRP from isolated sensory neurons as described6. We examined the release of iCGRP from spinal cord slices as described in Supplementary Methods.

Dural blood flow analyses. We measured dural blood flow with a laser Doppler flowmeter (TSI) as described24. We placed a needle-type probe over a large branch of the middle meningeal artery, distant from visible cortical blood vessels. We kept the cranial window moist with synthetic interstitial solution consisting of (in mM): 135 NaCl, 5 KCl, 5 CaCl2, 1 MgCl2, 10 HEPES and 10 d-glucose (pH 7.3). We recorded blood flow on-line at 1 Hz using Axoscope (Axon Instruments). Detailed methodology and references are described in the Supplementary Methods.

Behavioral analyses. We conducted all behavioral experiments with observ-ers blinded to treatments. We subcutaneously injected formalin (2.5%, 50 µl) into the dorsal surface of the hindpaw. For the formalin test, we observed two animals in adjacent chambers at one time, for flinching behaviors27. We present data as a time course or the cumulative number of flinches during phase 1 (0–10 min) or phase 2 (15–60 min). For the eye-wipe test, we pipetted vehicle, TAT-Scramble or TAT-CDB3 peptides into the right eye, and measured the resulting nocifensive behavior. We defined nocifensive behavior as time the rats spent holding the eye shut, actively grooming or wiping the treated eye. We recorded this nocifensive behavior for 5 min and followed by capsaicin application into the pretreated eye. We again observed nocifensive behavior for 5 min after capsaicin application.

We established hyperalgesia and allodynia by a single intraperitoneal injection (25 mg kg−1) of the antiretroviral drug ddC (Sigma). We performed the von Frey test on the area of the hindpaws as described; details of this behavioral test and associated references are described in the Supplementary Methods.

We performed the rotarod test for motor coordination as described in the Supplementary Methods. After intraperitoneal injection of vehicle or peptides, we tested mice with three trials each of fast and slow acceleration. We used the Morris water maze test to examine reference or spatial memory36. We trained mice before the intraperitoneal injection of TAT-Scramble or TAT-CBD3 pep-tides for four consecutive days (four trials per day). We evaluated performance 3 d after injection.

We used the light-dark box test and the elevated plus maze as measures of anxiety-associated behaviors. We performed the tail suspension test to evaluate despair- and depression-associated behavior. Detailed methods for these tests and associated references are available in the Supplementary Methods.

Statistical analyses. All sample means are reported ± s.e.m. We used Student’s t test or ANOVA for comparison of multiple groups with Dunnett’s post hoc analysis to determine statistically significant differences between sample groups at the significance level indicated.

Additional methods. Detailed methodology is described in the Supplementary Methods.

© 2

011

Nat

ure

Am

eric

a, In

c. A

ll ri

gh

ts r

eser

ved

.