Synaptic scaffolding protein Homer1a protects againstchronic inflammatory painAnke Tappe1, Matthias Klugmann2,5,6, Ceng Luo1,6, David Hirlinger1, Nitin Agarwal1, Justus Benrath3,Markus U Ehrengruber4,5, Matthew J During2,5 & Rohini Kuner1

Glutamatergic signaling and intracellular calcium mobilization

in the spinal cord are crucial for the development of nociceptive

plasticity, which is associated with chronic pathological pain1,2.

Long-form Homer proteins anchor glutamatergic receptors to

sources of calcium influx and release at synapses3–5, which is

antagonized by the short, activity-dependent splice variant

Homer1a. We show here that Homer1a operates in a negative

feedback loop to regulate the excitability of the pain pathway

in an activity-dependent manner. Homer1a is rapidly and

selectively upregulated in spinal cord neurons after peripheral

inflammation in an NMDA receptor–dependent manner.

Homer1a strongly attenuates calcium mobilization as well as

MAP kinase activation induced by glutamate receptors and

reduces synaptic contacts on spinal cord neurons that process

pain inputs. Preventing activity-induced upregulation of

Homer1a using shRNAs in mice in vivo exacerbates

inflammatory pain. Thus, activity-dependent uncoupling of

glutamate receptors from intracellular signaling mediators is a

novel, endogenous physiological mechanism for counteracting

sensitization at the first, crucial synapse in the pain pathway.

Furthermore, we observed that targeted gene transfer of

Homer1a to specific spinal segments in vivo reduces

inflammatory hyperalgesia. Thus, Homer1 function is

crucially involved in pain plasticity and constitutes a

promising therapeutic target for the treatment of chronic

inflammatory pain.

The long-form Homer1 proteins3, Homer1b and Homer1c (hereinreferred to collectively as Homer1b/c), assemble group I metabotropicglutamate receptors (mGluR1 and mGluR5, herein referred to collec-tively as mGluR1/5) in large macromolecular complexes with inositol1,4,5-triphosphate (IP3) receptors on the endoplasmatic reticulum5,TrpC channels6, calcium channels7,8 and components of the NMDAreceptor complex9 at the cell surface. Homer1a, the short, activity-dependent splice variant of Homer1b/c, lacks the ability to link

mGluR1/5 to synaptic proteins and competitively disassemblessynaptic glutamatergic signaling complexes3–9. Using immunohisto-chemistry, we found that Homer1b/c were strongly expressedthroughout the spinal cord (Fig. 1a and Supplementary Fig. 1online), whereas Homer1a was expressed at very low levels in naiverats. But expression of Homer1a was rapidly induced in the spinalcords of rats that developed peripheral hindpaw inflammation causedby unilateral, intraplantar injection of Complete Freund Adjuvant10

(CFA; Fig. 1a and Supplementary Table 1 online), which was furtherverified at the mRNA level by in situ hybridization (Fig. 1b) andreverse-transcription PCR (Fig. 1c) and at the protein level byimmunoprecipitation (Fig. 1d). Nociceptive neurons in the corre-sponding dorsal root ganglia (DRG) did not show inductionof Homer1a mRNA or protein (Supplementary Fig. 1 online),indicating that spinal neurons, but not presynaptic primary sensoryneurons, upregulate Homer1a after peripheral injury. Intrathecallumbar application of pharmacological inhibitors showed that block-ade of spinal NMDA receptors, extracellular signal–regulated proteinkinases (ERK1 and ERK2, herein referred to collectively as ERK1/2)or src kinases blocked spinal induction of Homer1a in the CFA model(P o 0.05 in all cases), whereas inhibition of mGluR1/5 had noeffect (Fig. 1e).

We generated virions of chimeric recombinant adeno-associatedvirus (AAV1/2)11 expressing both a small interfering hairpin RNA(shRNA) targeted against the 3¢ untranslated region of Homer1amRNA, which is unique to the Homer1a splice variant12, and EGFPbicistronically (AAV-shH1a). AAV-shRNA reduced expression ofHomer1a in cultured spinal neurons and in spinal cord in vivo incomparison with virions expressing both control shRNA and EGFPbicistronically (AAV-conRNA; Supplementary Fig. 2 online). Injec-tion of CFA into the hindpaw produced an acute upregulation ofHomer1a in the spinal dorsal horn of AAV-conRNA–expressing micebut not in AAV-shH1a–expressing mice, although the expression ofHomer1b/c remained similar across all groups (Supplementary Fig. 2and Supplementary Table 1 online). We then addressed how loss of

Received 6 June 2005; accepted 10 April 2006; published online 21 May 2006; doi:10.1038/nm1406

1Pharmacology Institute, University of Heidelberg, Im Neuenheimer Feld 366, 69120 Heidelberg, Germany. 2Department of Molecular Medicine & Pathology,University of Auckland, 85 Park Road, Auckland, New Zealand. 3Department of Anesthesiology and General Intensive Care, Medical University of Vienna,Waehringer Guertel 18-20, A-1090 Vienna, Austria. 4Brain Research Institute, University of Zurich, Winterthurerstrasse 190, CH-8057 Zurich, Switzerland.5Present addresses: Institute for Neurobiology, IZN, University of Heidelberg, Im Neuenheimer Feld 364, 69120 Heidelberg, Germany (M.K.).Kantonsschule Hohe Promenade, Promenadengasse 11, CH-8001 Zurich, Switzerland (M.U.E.). Department of Neurological Surgery, Weill Medical Collegeof Cornell University, New York, New York 10021, USA (M.J.D.). 6These authors contributed equally to this work. Correspondence should be addressedto R.K. (rohini.kuner@pharma.uni-heidelberg.de).

NATURE MEDICINE VOLUME 12 [ NUMBER 6 [ JUNE 2006 677

L E T T ERS©

2006

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

emed

icin

e

spinal Homer1a expression affects behavioral correlates of sensitiza-tion in pain pathways. In the plantar formalin test for early postin-flammatory hypersensitivity13, AAV-shH1a–expressing mice showed asignificant increase in the duration of phase II nocifensive behaviors, ameasure of spinal sensitization, compared with mice expressing AAV-conRNA (P ¼ 0.008), whereas phase I responses to formalin, ameasure of acute nociceptor activation, were not affected (Fig. 1f).In the CFA model of chronic inflammatory pain, expression of AAV-shH1a led to a long-lasting exacerbation of thermal hyperalgesia incomparison with AAV-conRNA (P r 0.005 at 24, 40, 48, 58 and 74 hafter CFA; Fig. 1g), although paw edema was comparable across allgroups (data not shown). Basic physiological pain responses weresimilar in all groups (Supplementary Fig. 3 online). Thus, whenactivity-induced upregulation of Homer1a is prevented, tissue inflam-mation leads to exaggerated pain plasticity and hyperalgesia in vivo.

How does Homer1a modulate inflammatory pain? Calcium ima-ging of acute Fura-2–loaded spinal slices of rats acutely expressingEGFP via recombinant Semliki Forest Virus (SFV) virions14 showedphasic and dose-dependent calcium responses of lamina I and IIneurons to bath application of glutamate, to the selective mGluR1/5agonist (RS)-3,5-dihydroxyphenylglycine (DHPG) as well as toNMDA (Fig. 2a–c). In contrast, lamina I and II neurons acutely

a b

c d e f g

Control

Hom

er1b

/c-s

peci

ficH

omer

1a-s

peci

fic

Hom

er1a

GA

BA

B(1

)

Hom

er1a

IP: H

omer

1a-s

peci

ficIB

: Hom

er1a

-spe

cific

Lysa

teIB

: tub

ulin

-spe

cific

GA

BA

B(1

) Dur

atio

n of

noci

fens

ive

beha

vior

(s)

Per

cent

cha

nge

inpa

w w

ithdr

awal

late

ncy

InflamedIpsilateral toinflammation

Contralateral toinflammation Control

Contro

l

Infla

med

Ipsilateral toinflammation

Contralateral toinflammation

662 bp

0 h 6 h 12 h 80

60

40

*

*

**

*

**

**20

0

Saline

DMSO

4-CPG

D-APV

PD9805

9PP1

0 h

AAV-conRNA

AAV-shH1a

AAV-conRNA

AAV-shH1a

12 h

Per

cent

cha

nge

over

cont

rala

tera

l sid

e

631 bp

30 kDa

50 kDa

240200160

Phase 1 Phase 2

1208040

0

0Pre

Time after CFA injection (h)

16 24 40 48 58 74 96

–20

–40

–60

–80

Inflamed

Figure 1 Nociceptive activity–induced changes in spinal expression of Homer1 variants in a model of unilateral hindpaw inflammation. (a) Immuno-

histochemical analysis of expression of Homer1b/c and Homer1a in the spinal cord of rats in control or inflamed states. (b) In situ hybridization for

expression of mRNA encoding Homer1a or GABAB(1) on spinal cord of rats in control or inflamed states. In a and b, boxed areas are magnified in the

panels to the right. (c) Reverse-transcription PCR for expression of cDNA encoding Homer1a or GABAB(1), performed on cDNAs derived from spinal cords

of rats in control or inflamed states. (d) Western blot analysis shows an increase in Homer1a immunoprecipitated from lumbar spinal cord of mice in naive

or inflamed states. b-tubulin served as loading control. (e) Effects of lumbar intrathecal administration of pharmacological inhibitors D-APV, PD98059,

PP1 and 4-CPG to rats on inflammation-induced upregulation of Homer1a in the ipsilateral over contralateral spinal cord in comparison with vehicleadministration (n ¼ 5 rats per group). *P o 0.05. Scale bars, 200 mm. (f) Enhanced cumulative responses in the formalin test upon spinal expression

of Homer1a-specific shRNA (shH1a; n ¼ 10 rats) as compared to control RNA (conRNA; n ¼ 12 rats; P o 0.05). (g) Enhanced thermal hyperalgesia

in mice expressing shH1a spinally as compared to conRNA after CFA-induced hindpaw inflammation (*P o 0.005; n ¼ 9 for shH1a group and 12 for

conRNA group).

b

c d

SFV-EGFP

SFV-EGFP-Homer1a

DHPG Glutamate NMDAKCl

30 s

0.1

∆F34

0/F

380

0.1

0.2

0.3

0.1

0.2

0.3

Pea

k ca

lciu

m r

espo

nse

∆F34

0/F

380

Pea

k ca

lciu

m r

espo

nse

∆F34

0/F

380

DHPG

Glutam

ate

NMDA

KCl

DHPG

Glutam

ate

NMDA

KCl

DHPG

Glutam

ate

NMDA

KCl

* * **

**

* * *5

10

15

20

25

Inte

grat

ed c

alci

umre

spon

se (

AU

C)

AAV-EGFPSFV-EGFP SFV-EGFP-Homer1aAAV-Homer1a

a

SF

V-E

GF

P

SF

V-E

GF

P-H

omer

1a

EGFPfluorescence

Peak afterglutamateBasal

EGFPfluorescence

Peak afterglutamateBasal

480 nm480 nm 340 nm340 nm

Figure 2 Calcium imaging using Fura-2 on acute spinal slices derived from

rats expressing EGFP or EGFP-tagged Homer1a via gene delivery mediated

by SFV or AAV virions to the spinal dorsal horn in vivo. (a) Visualization of

location of SFV-EGFP (control) or SFV-EGFP-Homer1a injection via EGFP

fluorescence and typical examples of changes in F340 induced by glutamate.

Boxed areas are magnified in the panels at the right. Dashed lines delineate

gray matter. Scale bars, 100 mm. (b) Typical examples of F340/380 traces

showing calcium mobilization upon application of glutamate, NMDA, DHPG

or KCl. Bars represent period of agonist application. (c) Average DF340/380

ratios showing peak calcium responses or average integrated calcium

responses calculated as the area under the curve (AUC) are significantlyreduced in spinal slices expressing SFV-Homer1a in comparison with SFV-

EGFP in response to DHPG (P o 0.001), glutamate (P o 0.001) or NMDA

(P o 0.001), whereas KCl-evoked responses remain unchanged

(P ¼ 0.3). (d) Similar results were obtained in spinal slices expressing AAV-

Homer1a and AAV-EGFP. Data are derived from at least five slices per group.

L E T TERS

678 VOLUME 12 [ NUMBER 6 [ JUNE 2006 NATURE MEDICINE

©20

06 N

atur

e P

ublis

hing

Gro

up

http

://w

ww

.nat

ure.

com

/nat

urem

edic

ine

expressing EGFP-tagged Homer1a via SFV virions showed a largereduction in peak as well as integrated calcium responses to glutamate,DHPG and NMDA (P o 0.05 compared to EGFP group in all cases;Fig. 2a–c). Furthermore, long-term expression of EGFP or Homer1ain the spinal dorsal horn via AAV virions15 led to very similar effects(Fig. 2d), providing a link to behavioral experiments in vivo. Indissociated cultured spinal neurons, we obtained similar results withrespect to calcium mobilization induced by DHPG or glutamate(Supplementary Fig. 4 online), although NMDA-induced calciumresponses were not affected by expression of Homer1a, suggesting arequirement for precise neuronal connectivity and receptor assemblyfor modulation of NMDA receptor–mediated calcium mobilization byHomer1a. The preserved integrity of slices or cultured neuronsexpressing Homer1a was shown by their normal responses to bathapplication of KCl (Fig. 2b,c) or agonists of other Gq/11-coupledreceptors, which are unrelated to Homer1 protein interactions, suchsubstance P, carbachol or bradykinin (Supplementary Fig. 4).Although expression of Homer1a has been reported to be associatedwith constitutive activation of heterologously expressed mGluR5 incultured cerebellar neurons16, we found that basal levels ofcalcium were not substantially altered upon expression ofHomer1a in spinal neurons in dissociated cultures or in spinal slices(Supplementary Fig. 4).

To address whether Homer1a could potentially influence spinalnociceptive processing by modulating synaptic spines on spinalneurons, we imaged spines on lamina V neurons in acutely preparedspinal slices from rats expressing enhanced green fluorescent protein(EGFP) or EGFP-Homer1a in vivo as described above (Fig. 3a). Spinalneurons expressing Homer1a showed a significantly reduced densityof spines on proximal dendrites as compared to neurons expressingEGFP alone (P o 0.001; Fig. 3a,b). In dissociated cultures, the

dendrites of Homer1a-expressing spinal neurons showed reduceddensity of puncta immunoreactive for the presynaptic marker proteinsynaptophysin or for the postsynaptic marker proteins PSD-95 or theNR1 subunit of NMDA receptors as compared with EGFP-expressingspinal neurons (P o 0.05 in each case; Fig. 3b), suggesting reducedsynaptic localization of NMDA receptors in Homer1a-expressingneurons. Recent studies show that molecular pathways linkingmGluR1/5 and NMDA receptors located in synaptic membranes tointracellular MAP kinases such as ERK1/2 require long-form Homerproteins such as Homer1b/c17. ERK1/2 are important messengerslinking synaptic activity to nuclear transcriptional control of plasti-city-related genes, including those involved in the development ofcentral sensitization and chronic pain18. We observed that bothDHPG- and NMDA-induced phosphorylation of ERK1/2 as well asnuclear translocation of phosphorylated, activated ERK1/2 (pERK1/2)were significantly lower in Homer1a-expressing neurons than inEGFP-expressing neurons (P o 0.05; Fig. 3c–e). Do these in vitrofindings hold true in the spinal cord in vivo upon physiologicalactivation of glutamate receptors by nociceptive activity? We observedthat the increase in the number of pERK+ cells in the ipsilateralsuperficial spinal dorsal horn induced by intraplantar injection offormalin was significantly higher in mice expressing EGFP than inmice expressing Homer1a in the spinal dorsal horn (P o 0.01;Fig. 3f,g). In contrast, mice expressing AAV-shH1a showed a signifi-cantly greater induction of pERK1/2 with formalin than mice injectedintraspinally with AAV-conRNA (P o 0.01; Fig. 3f,g). These resultsshow that both endogenously and exogenously expressed Homer1acounteracts activation of ERK1/2 induced by nociceptive activity inthe spinal dorsal horn in vivo.

A crucial question that arises is whether these basic findings can beapplied to develop new therapeutic approaches for the treatment of

a

c d

f g

e

bSFV-EGFP SFV-EGFP

SFV-EGFP-Homer1a

SFV-EGFP

Synapto-physin

PSD-95

NR1Den

sity

of s

pine

s/10

µm

Fol

d in

crea

se o

ver

basa

l lev

el

Nuc

lear

tran

sloc

atio

n of

pE

RK

(per

cent

tota

l cel

ls)

Num

ber

of p

ER

K+ c

ells

Imm

unor

eact

ive

punc

ta/1

0 µm

4

3

2

1

210

10

20

30

20

30

40

4

6

8

10

12

1*

*

*

**

*

*

*

**

*

* *234567

in vivo in vitro

SFV-EGFP-Homer1a

pERK1Bas

al

DHPG

NMDA

Basal

DHPG

NMDA

DHPG

NMDA

DHPG

NMDA

DHPG

Untre

ated

NMDA

pERK2

pERK1 pERK2

ERK1

AAV-EGFP AAV-Homer1a

AAV-EGFP AAV-Homer1a

AAV-EGFP

AAV-EGFP

Basal Induced Induced

Basal Induced

Basal Induced

Induced

AAV-Homer1a

AAV-Homer1a

AAV-conRNA

AAV-conRNA

AAV-shH1a

AAV-shRNA

AAV-EGFPAAV-Homer1a

Figure 3 Effects of Homer1a expression on

spine density and activation of ERK1/2 in

the spinal cord. (a) Low-, medium- and

high-magnification (insets) views of lamina

V neurons (arrowhead) with dendritic spines

in spinal slices from rats expressing either

SFV-EGFP or SFV-EGFP-Homer1a in vivo.

Scale bars: left, 100 mm; middle, 40 mm; inset,

10 mm. (b) Reduced dendritic spine density in

neurons expressing EGFP-tagged Homer1a as

compared to EGFP judged either as EGFP+

spines in spinal slices labeled in vivo (left) or

spines immunoreactive for synaptophysin,

PSD-95 or NR1 in spinal neurons cultured

in vitro (right). *P o 0.05 in all cases.(c,d) Typical examples (c) and quantitative

summary (d) of western blot analysis of total

ERK1 or phosphorylated ERK1 and ERK2

(pERK1 and pERK2, respectively) on cultured

spinal neurons expressing AAV-EGFP or AAV-

Homer1a upon treatment with DHPG or NMDA

or a mock treatment (basal). (e) Quantification

of cultured spinal neurons showing a nuclear

localization of pERK1/2-specific immunoreactivity

from experiments described in a. (f,g) Typical

examples (f) and quantification (g) of intraplantar

formalin-induced pERK1/2 immunoreactivity in

spinal dorsal horns of mice expressing AAV-EGFP,

AAV-Homer1a, AAV-conRNA or AAV-shH1a.

*P o 0.01, at least three mice per group

per treatment. Scale bar in f, 200 mm.

L E T T ERS

NATURE MEDICINE VOLUME 12 [ NUMBER 6 [ JUNE 2006 679

©20

06 N

atur

e P

ublis

hing

Gro

up

http

://w

ww

.nat

ure.

com

/nat

urem

edic

ine

inflammatory pain. We addressed this question preclinically usinggene transfer with AAV1/2 virions to the spinal dorsal horns ofsegments receiving nociceptive inputs from the hindlimbs in mice(Fig. 4a). Titers of injected AAV-Homer1a virions were carefullyadjusted to achieve a level of Homer1a expression comparable tothe levels induced by peripheral inflammation (i.e., about twofold;Supplementary Table 1). Although mice expressing Homer1a,Homer1c or EGFP in the spinal dorsal horn were indistinguishablewith respect to nociceptive behavior in several tests such as the tail-flick reflex or hindpaw withdrawal in response to heat, cold orpressure10, responses to intraplantar capsaicin or with respect tomotor performance (Supplementary Fig. 3), mice expressingHomer1a showed a small but statistically significant reduction inphase I responses to intraplantar formalin (P o 0.05). Notably,Homer1a-overexpressing mice showed a significant reduction in themagnitude of the phase II response to formalin over mice over-expressing Homer1c (P ¼ 0.01 and P ¼ 0.014, respectively) orexpressing EGFP (P ¼ 0.003 and P ¼ 0.011, respectively; Fig. 4b),suggesting a reduction of acute short-term plasticity in the spinaldorsal horn by Homer1a. Moreover, in the CFA test, the magnitude aswell as the duration of thermal hyperalgesia was lower in Homer1a-overexpressing mice as compared to Homer1c-overexpressing orEGFP-expressing mice (P o 0.004; Fig. 4c).

Thus far, endogenous protective mechanisms, which seek to balanceand counter pronociceptive synaptic plasticity, have focused on the

release of inhibitory neuromodulators such as endorphins and endo-cannabinoids. Here, we propose a novel feedback mechanism, basedon activity-dependent remodeling of glutamatergic synapses throughmodulation of protein-protein interactions. Several factors, includingthe ability of Homer1a to attenuate inflammatory hypersensitivitywithout significantly affecting physiological pain or other neurologicalfunctions (for example, motor coordination or locomotion) favor theidea of therapeutically exploiting the functions of Homer1 proteins inchronic pain. Although intraparenchymal injections with recombinantAAV virions were used here mainly as proof-of-principle tools forillustrating the efficacy of Homer1a as a therapeutic target, they alsolay a basis for developing less invasive tools for future clinicalapplications. In conclusion, our observation that synaptic molecularscaffolds can be directly regulated by nociceptive activity to modulatepain processing underscores their potential relevance as novel classof therapeutic targets against chronic pain. In particular, we haveidentified and validated Homer1a as a therapeutic target in thetreatment of inflammatory pain.

METHODSFor details, see Supplementary Methods online.

Mouse models of postinflammatory pain. All animal procedures were

performed in accordance with ethical guidelines laid down by the local

governing body (Regierungsprasidium Karlsruhe, Germany). We carried out

all behavioral measurements in awake, unrestrained, age-matched adult (43

month old) male C57BL6 mice or Wistar rats. We administered unilateral

injections of CFA (Sigma Aldrich) into the intraplantar surface of the hindpaw

as previously described10 to rats (100 ml) or mice (20 ml), and injected control

rats and mice with 0.9% saline. We analyzed latency of paw withdrawal in

response to noxious heat, paw pressure and tail-flick reflex as previously

described19 (Ugo Basile, Inc.). We determined paw response to cold by

measuring the frequency of paw withdrawal in response to plantar application

of acetone, which we applied five times at intervals of 3 min. We performed the

intraplantar formalin test as previously described13,19.

Calcium imaging on spinal slices. We injected rats with SFV virions at a

titer of 106 in spinal dorsal horns at postnatal days (P)13–18, and sectioned

acute spinal cord slices (400 mm) at 16 h after injection as described below.

We incubated slices at 32 1C for 30 min with incubation buffer gassed with

carbogen (5% CO2 in 95% O2) and consisting of the following (in mM):

95 NaCl, 1.8 KCl, 1.2 KH2PO4, 0.5 CaCl2, 7 MgCl2, 26 NaHCO3, 15 glucose

and 50 sucrose. We loaded slices with Fura-2 AM (10 mM) in the presence

of pluronic acid (0.1%) and Hepes (20 mM), de-esterified them and trans-

ferred them to a perfusion chamber superfused (2.5 ml/min) with carbogen-

aerated perfusion buffer consisting of the following (in mM): 127 NaCl, 1.8

KCl, 1.2 KH2PO4, 2.4 CaCl2, 1.3 MgCl2, 26 NaHCO3 and 15 glucose. We

applied agonists by bath application for 10 s and used a 10-min washout

interval between subsequent applications. We calculated F340/380 ratios as

described previously20. Calcium imaging experiments on spinal slices of

mice injected intraspinally with AAV-EGFP or AAV-HA-Homer1a were

followed by immunohistochemistry to detect expression of hemagglutinin-

tagged Homer1a.

Injection of AAV virions. The construction of the AAV viruses expressing

Homer1a, Homer1c or EGFP used in this manuscript has been previously

described15. We anesthetized adult mice with Fentanyl:Domitor:Dormicum

(4:6:16; 0.7 ml/g; intraperitoneal) and we infused 300 nl of a 2:1 mixture of AAV

viral stocks with 20% mannitol (to enhance vector spread and improve

transduction13) containing 1.2 � 107 virions four times bilaterally (two times

on each side) directly into the spinal parenchyma of spinal segments L3–L5.

Infusion rate was maintained at about 66 nl/min using a microprocessor-

controlled minipump (World Precision Instruments), which limits tissue

damage. By regulating titer and volume of the injected vectors and the angle

of micropipette placement, we achieved selective expression of proteins at the

desired levels in spinal dorsal horn over 2–3 mm along the rostrocaudal axis

D

L

V

M

a

b

c

Injection:

Antibody:

AAV-HA-Homer1a

AAV-HA-Homer1c AAV-EGFP

HA-specific EGFPfluorescence

0

0

40

80

120

160

200

240

–20

–40

–60

–80

Dur

atio

n of

noci

fens

ive

beha

vior

(s)

Per

cent

cha

nge

in p

aw w

ithdr

awal

late

ncy

Phase 1 Phase 2

Pre 16 24 40 48

*

*

* **

*

Time after CFA injection (h)

AAV-EGFP

AAV-Homer1c

AAV-Homer1a

AAV-EGFP AAV-Homer1c AAV-Homer1a

Figure 4 Effects of spinal expression of AAV-Homer1a, AAV-Homer1c

(both tagged with hemagglutinin epitope) or EGFP (control) on nocifensive

behavior associated with inflammatory pain in mice in vivo.

(a) Representative examples of EGFP fluorescence or immunohistochemistry

with an HA-specific antibody in transduced spinal dorsal horn. Scale bar,

200 mm. (b) Reduced cumulative responses in the formalin test upon spinal

expression of AAV-Homer1a as compared to AAV-Homer1c or AAV-EGFP

*P o 0.05 (n ¼ 10, 11 and 10 for the EGFP, Homer1c and Homer1a

groups, respectively). (c) Spinal expression of Homer1a, but not of

Homer1c, reduces the magnitude of thermal hyperalgesia produced by

intraplantar injection of CFA. *P o 0.002 for all time points (n ¼ 10,

11 and 11 for the EGFP, Homer1c and Homer1a groups, respectively).

L E T TERS

680 VOLUME 12 [ NUMBER 6 [ JUNE 2006 NATURE MEDICINE

©20

06 N

atur

e P

ublis

hing

Gro

up

http

://w

ww

.nat

ure.

com

/nat

urem

edic

ine

without significant spread into the ventral horn. We allowed mice to recover for

3 weeks before commencing behavioral analysis. For calcium imaging

experiments, we injected rats at P11 and analyzed them at P19.

shRNA experiments. For generation of mouse Homer1a-specific shRNAs, we

searched the 3¢ untranslated region of the Homer1a mRNA12 using a shRNA

Design Algorithm (Cenix Biosciences). We generated oligonucleotides corre-

sponding to Homer1a-specific shRNA (shH1a; sense strand, 5¢-GGAGCAUUG

AGCUAAUUAUTT-3¢; antisense strand, 5¢-AUAAUUAGCUCAAUGCUC

CTT-3¢; Sigma Genosyses), annealed them and cloned them into an AAV

plasmid that simultaneously drives the expression of the shRNAs and EGFP as

previously described15. This sequence is targeted against Exon 5¢ of the Homer1

gene, which is unique to the Homer1a transcript and functions as an intron in

the other Homer slice variants12. A BLAST search with the sequences did not

uncover any homology with other genes. Similarly, a control shRNA (sense

strand, 5¢-GUACUGCUUACGAUACGGTT-3¢; antisense strand, 5¢-CCGUAU

CGUAAGCAGUACUTT-3¢) was cloned and packaged (AAV-conRNA) to serve

as a negative control for potential unspecific effects associated with delivery of

shRNA. We added approximately 108 viral genomes to the culture medium

of cultured spinal neurons (106) at 14 d in vitro in 10-cm dishes and

injected approximately 108 viral particles in the spinal parenchyma in mice,

as described above.

Data analysis and statistics. All data are presented as mean ± s.e.m. Analysis of

variance (ANOVA) for random measures was performed followed by post hoc

Fisher test to determine statistically significant differences. P o 0.05 was

considered significant.

Accession codes. GenBank: Homer1a splice variant, AF425674.

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

ACKNOWLEDGMENTSThe authors are grateful to P.F. Worley for the gift of rat Homer1a and Homer1ccDNAs; P.H. Seeburg for the gift of a plasmid containing the Homer1a 3¢ UTR;M. Schaefer, B. Heinke and O. Friedrich for help with calcium imagingexperiments; and to T. Kuner and S. Offermanns for critically reading an earlierversion of this manuscript. This work was supported by an Emmy Noether-program fellowship from the Deutsches Forschunggemeinschaft to R.K.

COMPETING INTERESTS STATEMENTThe authors declare that they have no competing financial interests.

Published online at http://www.nature.com/naturemedicine/

Reprints and permissions information is available online at http://npg.nature.com/

reprintsandpermissions/

1. Sandkuhler, J. Learning and memory in pain pathways. Pain 88, 113–118 (2000).2. Woolf, C.J. & Salter, M.W. Neuronal plasticity: increasing the gain in pain. Science 288,

1765–1768 (2000).3. Xiao, B., Cheng, T.J. & Worley, P.F. Homer: a link between neural activity and glutamate

receptor function. Curr. Opin. Neurobiol. 10, 370–374 (2000).4. Ehrengruber, M.U., Kato, A., Inokuchi, K. & Hennou, S. Homer/Vesl proteins and their

roles in CNS neurons. Mol. Neurobiol. 29, 213–228 (2004).5. Tu, J.C. et al. Homer binds a novel proline-rich motif and links group 1 metabotropic

glutamate receptors with IP3 receptors. Neuron 21, 717–726 (1998).6. Yuan, J.P. et al. Homer binds TRPC family channels and is required for gating of TRPC1

by IP3 receptors. Cell 114, 777–789 (2003).7. Fagni, L., Chavis, P., Ango, F. & Bockaert, J. Complex interactions between mGluRs,

intracellular Ca2+ stores and ion channels in neurons. Trends Neurosci. 23, 80–88(2000).

8. Feng, W. et al. Homer regulates gain of ryanodine receptor type 1 channel complex.J. Biol. Chem. 277, 44722–44730 (2002).

9. Tu, J.C. et al. Coupling of mGluR/Homer and PSD-95 complexes by the Shank family ofpostsynaptic density proteins. Neuron 23, 583–592 (1999).

10. Hargreaves, K., Dubner, F., Brown, F., Flores, C. & Joris, J. A new and sensitive methodfor measuring thermal nociception in cutaneous hyperalgesia. Pain 32, 77–88 (1988).

11. During, M.J., Young, D., Baer, K., Lawlor, P. & Klugmann, M. Development andoptimization of adeno-associated virus vector transfer into the central nervous system.Methods Mol. Med. 76, 221–236 (2003).

12. Bottai, D. et al. Synaptic activity-induced conversion of intronic to exonic sequence inHomer 1 immediate early gene expression. J. Neurosci. 22, 167–175 (2002).

13. Tjolsen, A., Berge, O.G., Hunskaar, S., Rosland, J.H. & Hole, K. The formalin test: anevaluation of the method. Pain 51, 5–17 (1992).

14. Hennou, S. et al. Homer-1a/Vesl-1S enhances hippocampal synaptic transmission.Eur. J. Neurosci. 18, 811–819 (2003).

15. Klugmann, M. et al. AAV-mediated hippocampal expression of short and long Homer 1proteins differentially affect cognition and seizure activity in adult rats. Mol. Cell.Neurosci. 28, 347–360 (2005).

16. Ango, F. et al. Agonist-independent activation of metabotropic glutamate receptors bythe intracellular protein Homer. Nature 411, 962–965 (2001).

17. Yang, L. et al. A novel Ca2+-independent signaling pathway to extracellular signal-regulated protein kinase by coactivation of NMDA receptors and metabotropic gluta-mate receptor 5 in neurons. J. Neurosci. 24, 10846–10857 (2004).

18. Ji, R.R., Baba, H., Brenner, G.J. & Woolf, C.J. Nociceptive-specific activation of ERK inspinal neurons contributes to pain hypersensitivity. Nat. Neurosci. 2, 1114–1119(1999).

19. Hartmann, B. et al. The AMPA receptor subunits GluR-a and GluR-b reciprocallymodulate spinal synaptic plasticity and inflammatory pain. Neuron 44, 637–650(2004).

20. Grynkiewicz, G., Poenie, M. & Tsien, R.Y. A new generation of Ca2+ indicators withgreatly improved fluorescence properties. J. Biol. Chem. 260, 3440–3450 (1985).

L E T T ERS

NATURE MEDICINE VOLUME 12 [ NUMBER 6 [ JUNE 2006 681

©20

06 N

atur

e P

ublis

hing

Gro

up

http

://w

ww

.nat

ure.

com

/nat

urem

edic

ine