Article

The Low-Threshold Calcium Channel Cav3.2

Determines Low-Threshold Mechanoreceptor
Function
Graphical Abstract

Authors
Amaury Francois, Niklas Schuetter, ...,
Olaf Pongs, Emmanuel Bourinet

Correspondence
emmanuel.bourinet@igf.cnrs.fr

In Brief
Francois et al. show that Cav3.2 calcium
channels are potent and specific
regulators of low-threshold
mechanoreceptors (LTMRs) setting the
efficient detection, conduction, and
transfer of tactile/pain information. In CLTMRs, this mechanism crucially
contributes to neuropathic pain,
highlighting the utility of a therapeutic
Cav3.2-inhibition strategy to improve
patient quality of life.

Highlights
d

Cav3.2 calcium channels are markers of both C- and Ad- lowthreshold mechanoreceptors

Cav3.2 in receptive fields and axons facilitates action

potential initiation and conduction

Cav3.2 in C-LTMR governs innocuous touch, noxious

mechanical cold, and chemical pain

Cav3.2 channels in C-LTMRs are molecular substrates of

allodynia in neuropathic pain

Francois et al., 2015, Cell Reports 10, 370382

January 20, 2015 2015 The Authors
http://dx.doi.org/10.1016/j.celrep.2014.12.042

Cell Reports

Article
The Low-Threshold Calcium Channel Cav3.2
Determines Low-Threshold Mechanoreceptor Function
Amaury Francois,1,2,3,4 Niklas Schuetter,5 Sophie Laffray,1,2,3,4 Juan Sanguesa,1,2,3,4 Anne Pizzoccaro,1,2,3,4
Stefan Dubel,1,2,3,4 Annabelle Mantilleri,6 Joel Nargeot,1,2,3,4 Jacques Noel,1,7 John N. Wood,8 Aziz Moqrich,6 Olaf Pongs,5
and Emmanuel Bourinet1,2,3,4,*
1Laboratories of Excellence, Ion Channel Science and Therapeutics, Institut de Ge
nomique Fonctionnelle, 141 rue de la Cardonille, 34094
Montpellier, France
2CNRS UMR5203, 34095 Montpellier, France
3INSERM, U661, 34095 Montpellier, France
4Universite
de Montpellier, 34095 Montpellier, France
5Department of Physiology, University of Saarland, School of Medicine, Kirrberger Strae 1, 66424 Homburg, Germany
6Aix-Marseille-Universite
, CNRS, Institut de Biologie du Developpement de Marseille, UMR 7288, 13288 Marseille, France
7Universite
de Nice Sophia Antipolis, Institut de Pharmacologie Moleculaire et Cellulaire, UMR7275 CNRS, 660 route des lucioles, 06560
Valbonne, France
8Wolfson Institute for Biomedical Research, University College London, Gower Street, London WC1E 6BT, UK
*Correspondence: emmanuel.bourinet@igf.cnrs.fr
http://dx.doi.org/10.1016/j.celrep.2014.12.042
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

SUMMARY

The T-type calcium channel Cav3.2 emerges as a key

regulator of sensory functions, but its expression
pattern within primary afferent neurons and its
contribution to modality-specific signaling remain
obscure. Here, we elucidate this issue using a unique
knockin/flox mouse strain wherein Cav3.2 is replaced by a functional Cav3.2-surface-ecliptic GFP
fusion. We demonstrate that Cav3.2 is a selective
marker of two major low-threshold mechanoreceptors (LTMRs), Ad- and C-LTMRs, innervating the
most abundant skin hair follicles. The presence of
Cav3.2 along LTMR-fiber trajectories is consistent
with critical roles at multiple sites, setting their
strong excitability. Strikingly, the C-LTMR-specific
knockout uncovers that Cav3.2 regulates light-touch
perception and noxious mechanical cold and chemical sensations and is essential to build up that debilitates allodynic symptoms of neuropathic pain, a
mechanism thought to be entirely A-LTMR specific.
Collectively, our findings support a fundamental
role for Cav3.2 in touch/pain pathophysiology, validating their critic pharmacological relevance to
relieve mechanical and cold allodynia.
INTRODUCTION
Somatosensory perception is dependent on a variety of primary
sensory neurons that are able to detect and transmit sensations,
which range from hedonic to painful (Basbaum et al., 2009; Lallemend and Ernfors, 2012; Liu and Ma, 2011). Deciphering this
complex neuronal diversity is a key challenge in sensory neuro370 Cell Reports 10, 370382, January 20, 2015 2015 The Authors

biology. One intriguing facet is the role of the cutaneous mechanosensors that discriminate innocuous and noxious mechanical
forces and respectively translate them into touch or pain
sensations. Moreover, in the development of pathological pain,
normal tactile signaling of low-threshold mechanoreceptors
(LTMRs) is converted into sensations of pain. As LTMRs encompass different neuron subclasses, decoding the mechanisms
that determine their sensitivity and their contribution to pain
pathologies is critical to the development of future analgesic
treatments.
Fast-conducting myelinated axons and large cell bodies characterize A-fiber-associated sensory neurons (Aa, b, or d), which
are involved mostly in touch and proprioception. Slowly conducting unmyelinated fibers with small cell bodies correspond
to C-fibers involved in not only nociception but also conveying
tactile information. LTMRs are heterogeneous; comprise Ab-,
Ad-, and C-fiber subtypes; and are further subdivided according
to their rate of adaptation to mechanical stimuli and by their distal
and central terminal anatomy (Li et al., 2011). However, molecular markers of LTMRs remain insufficient to fully decipher their
pathophysiological roles. Recently, a small population of unmyelinated dorsal root ganglia (DRG) neurons has been described
for their unique expression of tyrosine hydroxylase (TH), vesicular transporter of glutamate 3 (VGLUT3), and chemokine like protein TAFA4 (Brumovsky et al., 2006; Delfini et al., 2013; Seal
et al., 2009). These neurons represent the poorly understood
population of C-LTMRs involved in apparently opposing sensations of pleasant touch and hypersensitivity during chronic pain
(Bessou et al., 1971; Seal et al., 2009). Another specific population of LTMRs is the TrkB-expressing neurons innervating D-hair
follicles and described as Ad-LTMRs. Interestingly, these neurons express the Cav3.2 isoform of low-voltage-activated calcium channels (Shin et al., 2003).
This channel belongs to a gene family composed of three
members (Cav3.1, 3.2, and 3.3), with Cav3.2 being predominant
in DRGs. They activate at low voltages close to the membrane

resting potentials and therefore are important tuners of cell excitability (Perez-Reyes, 2003). A constitutive Cav3.2 knockout (KO)
in mice impairs Ad-LTMRs excitability (Wang and Lewin, 2011).
In addition, electrophysiological analyses shows that Cav3.2
has a proexcitatory impact in small-diameter nociceptors expressing mechanoactivated channels (Coste et al., 2007). In
Cav3.2 KO mice, intrathecal antisense-mediated Cav3.2 knockdown and the use of T-type channel antagonists induce a
marked analgesia in vivo (Bourinet et al., 2005; Choe et al.,
2011; Choi et al., 2007; Francois et al., 2014). Conversely,
enhanced Cav3.2 activity parallels chronic pain states (GarcaCaballero et al., 2014; Jagodic et al., 2008; Marger et al., 2011;
Messinger et al., 2009; Takahashi et al., 2010).
However, to unravel the complete role of Cav3.2 within all the
sensory neuron classes, several fundamental anatomical and
functional questions still remain to be addressed. The aim of
the present work is to determine the contribution of Cav3.2 to
pathophysiological somatosensory signaling within DRG neuron
subtypes by elucidating how its subcellular location affects information processing in sensory fibers.
To investigate this, we designed a knockin (KI) strategy by inserting a two loxP site-flanked ecliptic-GFP tag into exon 6 of the
Cav3.2 locus to allow genetic marking of Cav3.2 channels as
well as tissue-specific loss of function. Interestingly, our analysis
shows that Cav3.2 is a selective marker for mechanoreceptors
comprising the C- and the Ad-LTMRs expressing TH/VGLUT3/
TAFA4 and TrkB, respectively, in addition to a third population
of Ret-positive neurons that do not coexpress any of the classical markers for peptidergic and nonpeptidergic C-fibers. We
also show that Cav3.2 is localized within the putative zone of action potential (AP) generation, in the vicinity of hair follicles in
distal nerve endings. Furthermore, it is also expressed along
the axons of afferent C-LTMR fibers and in the nodes of Ranvier
of Ad-LTMRs and also both pre- and postsynaptically in the
dorsal horn lamina IIIII of the spinal cord. Functional analysis
of C-LTMR in cutaneous fibers shows that T-type channels
determine their low threshold of mechanical activation and
accelerate AP conduction along the axon.
Finally, the pathophysiological role of Cav3.2 within C-LTMRs
was assessed by generating a conditional Cav3.2 KO using
Nav1.8-Cre mice to eliminate Cav3.2 selectively from C-fibers.
We show here that Cav3.2 determines the basal high sensitivity
of C-LTMRs, regulating innocuous and noxious mechanical and
cold detection as well as chemically induced pain. Moreover, our
data demonstrate that C-LTMRs expressing Cav3.2 contribute
to both mechanical and cold allodynia in peripheral neuropathy.
In contrast to the canonical view that only myelinated tactile
afferences encode allodynia, our work refines this view by
providing molecular clues for the contribution of C-tactile fibers
and validates Cav3.2 as a pertinent target to dampen debilitating
aspects of pathological pain.
RESULTS
A Mouse Model to Investigate Cav3.2 Function in
Sensory Neurons
To understand Cav3.2 function within sensory neuron subtypes,
we engineered a knockin mouse with a fusion of a pH-sensitive

GFP (ecliptic-GFP; Miesenbock et al., 1998) within the large

extracellular loop preceding the P loop in the first transmembrane domain of the channel (Figures 1A1C). Homozygous
mice harboring this insertion (hereafter named KI) are viable
and fertile, born in a Mendelian ratio, and indistinguishable
from their control littermates. Importantly, the recombinant allele
allowed correct transcription and translation of the fusion protein
(Figures 1D, 1E, and S1AS1C) as well as fine localization of expressed channels (Figures S1D, S1E, and S2). Electrophysiological analysis of native currents in Ad-LTMRs cells, recognizable
by their rosette morphology in culture, showed no differences
in T-type calcium current amplitudes and kinetics between
wild-type (WT) and homozygous KI/KI mice (Figures 1F and
1G) or in voltage-dependent properties (Figures S1F and S1G).
Altogether, these results suggest no abnormalities in function
and membrane expression of Cav3.2-GFP channels in KI mice.
Cav3.2-Expressing Neurons Display Properties of Adand C-LTMRs
As the Cav3.2-GFP protein faithfully recapitulates Cav3.2 transcript expression, we deciphered the cellular distribution of
Cav3.2 protein in DRG neurons and correlated this pattern with
specific sensory modalities. Double immunostaining of GFP
with PGP9.5 (a general neuronal marker) identified GFP immunoreactivity (GFP-IR) in 33% of all L4 DRGs (Figure 2A). Cav3.2GFP+ neurons segregate into two classes: neurofilament
NF200+ representing 28% and peripherin+ representing 66%
of all Cav3.2-expressing neurons (Figure S2A). The subset of
large NF200+/Cav3.2-GFP+ neurons exclusively express TrkB,
the main marker of Ad-LTMR (Figure 2B), while those expressing
peripherin (an unmyelinated fiber marker) could be further subdivided into neurons containing C-LTMR markers such as TAFA4,
VGLUT3, and TH and an as yet to be identified population of Ret+
neurons that express none of the markers tested so far (Figures 2
and S2). Coexpression with TrkA, CGRP, and TRPV1 remained
minor in all cases (Figure 2G). Hence, Cav3.2 is expressed in major populations of mechanoreceptors.
We next used live GFP labeling of Cav3.2-expressing neurons
to decipher their physiological properties in culture. Mechanical
stimulation of the Cav3.2-GFP+ neuron cell soma evoked inward
mechanoactivated (MA) currents in all neurons tested. As previously described, MA currents have different adaptation rates
(Figure 3A) (Delmas et al., 2011). Interestingly, rapidly and intermediately adapting currents (RA and IA, respectively) were predominantly found in large GFP+ neurons, whereas slowly adapting (SA) and ultraslowly adapting (USA) currents predominate in
smaller GFP+ neurons (Figure 3B). Among the candidate substrates for MA channels, piezo2, which shows RA kinetics in recombinant systems, is highly expressed in Tafa4/Cav3.2-positive neurons (Figure S3), suggesting their possible contribution
to the SA/USA MA currents as well (but see Lou et al., 2013). Kinetics of MA currents uniquely shapes cell firing from phasic to
tonic spiking, properties identified in the distinct LTMR fiber subtypes (Woodbury et al., 2001). Given that Cav3.2-GFP marks CLTMRs and Ad-LTMRs, we tested the occurrence of these firing
properties in GFP+ neurons. Consistent with the slow adaptation
of MA currents, mechanical stimulation triggered tonic AP firing
in small GFP+ neurons. Furthermore, cooling also elicited firing in
Cell Reports 10, 370382, January 20, 2015 2015 The Authors 371

Figure 1. Generation and

Cav3.2-GFP Knockin Mice

these neurons (n = 9), a hallmark of C-LTMR fibers (Li et al., 2011;

Takahashi et al., 2003) (Figure 3C). In contrast, none of the neurons tested responded to warming (not shown). Comparatively,
large GFP+ neurons systematically responded to mechanical
stimuli with a single spike phasic to the stimulus onset (Figure 3D). Neither cooling nor warming excited these neurons
(n = 10). In parallel, we were unable to detect any intracellular calcium level variations when we challenged neurons with thermo
TRP channels or Mrgp receptor agonists, with the exception of
the TRPA1 agonist AITC that induced small responses in some
small GFP+ neurons (not shown).
To confirm the distinct cellular firing of Cav3.2-GFP-expressing LTMRs, we used mechanical stimuli of various velocities.
Small Cav3.2-GFP+ neurons responded better to slow-motion
stimuli consistent with the physiological properties of C-LTMRs
(Figure 3E), as previously described for TAFA4+ neurons (Delfini
et al., 2013) and further demonstrated in Cav3.2-GFP+/TAFA4GFP+ neurons from double-heterozygote mice (Figure 3F). In
contrast, large GFP+ neurons only responded to fast mechanical
stimuli, as Ad-LTMR do (Abraira and Ginty, 2013). Figure 3F summarizes these observations by presenting the probe velocity
372 Cell Reports 10, 370382, January 20, 2015 2015 The Authors

Analysis

of

(A) Scheme of the Cav3.2-ecliptic-GFP fusion

structure.
(B) Segment of the murine Cav3.2 protein
sequence. The ecliptic-GFP is inserted in the
extracellular loop between the fifth transmembrane segment (IS5) and the pore region.
(C) Schematic diagram to illustrate the cacna1h
targeting strategy. Cacna1h locus: black vertical
bars depict the exons (Ex) structure from 2 to 9 of
the murine cacna1h gene. Targeting vector: the 50
and 30 arms each consist of 2.5 kb of genomic
sequence. Open triangles mark FRT sites surrounding neomycin resistance cassette (AFN).
eGFP displays ecliptic-GFP insertion. Filled triangles mark loxP sites inserted before exon 6 and
after exon 7. Black bars beneath the genomic
sequences illustrate Southern probes and PCR
fragments used for genotyping. Targeted locus:
scheme displays insertion of targeting vector into
the cacna1h locus. Targeted locus AFN removed:
sequence between FRT sites after AFN excision
with Flipase. Lines with arrowheads depict XbaI
restriction fragments.
(D) Southern analysis of wild-type (WT/WT), heterozygous (WT/KI), and homozygous mutant (KI/
KI) mouse DNA. Genomic DNA was digested with
XbaI. The 30 probe was used for the detection of
restriction fragments. Arrows mark positions of
wild-type (10 kb) and mutant (5 kb) bands.
(E) Genotyping results of a PCR assay performed
on genomic DNA. A 0.5 kb band demonstrates
excision of AFN cassette.
(F) Analysis of Cav3.2-mediated low-voltageactivated calcium currents in medium-sized DRG
neurons of KI/KI and WT/WT genotypes.
(G) Mean currentdensities in WT/WT-, WT/KI-, and
KI/KI-derived cells identified by their rosette
morphology.
Data are presented as mean SEM.

eliciting the first spike, as a function of cell capacitance. Two

populations can be distinguished from each other: the large neurons stimulated by fast stimuli, and the small neurons stimulated
preferentially by slow mechanical stimuli. Of note in this latter
group, the TAFA4-GFP+ and the TAFA4-GFP- neurons are similarly activated, and they all behave as C-LTMRs.
Cav3.2 Is Expressed in Fiber Compartments Critical for
AP Initiation and Propagation
Primary sensory neurons consist essentially of peripheral and
central terminal arbors as well as long axons within nerve bundles. Therefore, deciphering the Cav3.2-GFP subcellular localization is essential to understanding channel function(s). Analyses of GFP-IR in distal nerve terminals revealed the channel
presence in fibers innervating hairy skin. Using S100 as a marker
of terminal Schwann cells surrounding nerve endings, we could
detect GFP-IR in close proximity to lanceolate structures around
hair follicles (Figure 4A). Magnification shows an intense GFP-IR
nearby the branching points where daughter nerve endings join
each other. Interestingly, Cav3.2-GFP detection along teased fibers of sciatic nerve, revealed strong GFP-IR in the nodes of

GFP / PGP9.5

GFP

(305)

PGP9.5

E GFP

IB4

Merge

TH

Merge

(947)

=
33%

Merge

G
Trk A

T
H/VGLUT3/ TAFA4
A
TH/VGLUT3/

60

n=6 ; 8/174

Cav3.2 GFP KI adult DRG

n=8 ; 93/286

20

IB4

n=6 ; 14/158

40

Trk C

n=6 ; 0/160

Merge

80

n=8 ; 82/338

Ret

Trk
T
rk
kB

Merge

n=7 ; 0/153

GFP

TrkA

Ret
R
et / G
GFR

GFP

33% Cav3.2 +

n=9 ; 105/305

n=10 ; 39/392

100m

n=5 ; 0/149

TrkB

GFP

n=11 ; 230/391

GFP

(n=8)

n=11 ; 213/330

100m

% Colloc Cav3.2 / Markers

Cav3.2 KI

Figure 2. Sensory Neuron Classes Expressing Cav3.2

Adult L4 DRG sections from Cav3.2-GFP KI mice costained with anti-GFP antibodies (green) and anti-PGP9.5 (red) (A), anti-TrkB (B), anti-TrkA (C), anti-Ret (D),
isolectin B4 (E), and anti-tyrosine hydroxylase (TH) (F). In all panels, insets are magnified views of indicated areas within white boxes. Scale bar, 100 mm.
Quantitative data of colocalization between markers are shown in (G). In each histogram bar, n indicates the number of animals used, followed by the total
number of neurons colabeled with the specific marker over the total number of Cav3.2-GFP cells. Data are presented as mean SEM. The cartoon is a
representative view of the Cav3.2 population in adult DRG, in accordance with the different markers described.

Ranvier of some medium caliber axons flanked with paranodin

labeling and overlapping with KCNQ2 staining (paranode and
node markers, respectively) (Figure 4B). Concerning thin unmyelinated C-fibers, the GFP-IR signal was observed all along
axons (Figure 4B, empty arrows) overlapping with peripherin
(Figure S4A). We found a similar localization pattern in dorsal
roots purely composed of sensory afferences (Figure 4C). In
contrast, efferent axons in ventral roots did not show any GFPIR (Figure S4B). In the ganglia, GFP-IR was detected in the proximal unmyelinated segment linking the cell soma to the axon in all
Cav3.2-GFP+ neurons (Figure 4D). Thus, the channel is present
from the periphery, throughout the length of the axons with a
specific enrichment at locations key to cell excitability.
GFP-IR was examined in the dorsal horn of the spinal cord and
a strong signal was observed in the laminae IIIII with an overlap
with PKCg-containing interneurons (Figure 4E). Although the
GFP-IR may result from both afferent endings and spinal neurons, the labeling is consistent with Ad- and C-LTMR input zones
with little overlap with IB4+ or CGRP+ C-nociceptive inputs in the
outer lamina II or lamina I or with parvalbumin+ Ab inputs in laminas IIIIV (Figures S4C and S4D). More precise localization using

transmission electron microscopy (TEM) revealed GFP-IR in the

lamina IIIII synapses. Examination of more than 50 different
identified synapses shows that the most intense labeling was detected on the postsynaptic area compared to the presynaptic
zone (Figure 4F).
Cav3.2 Controls Mechanosensitivity Threshold and AP
Propagation in C-LTMRs
Since we aimed to investigate Cav3.2 function in C-fibers in vivo
using a conditional loss-of-function approach, we next examined the role of peripheral Cav3.2 channels in C-LTMRs using
the skin-nerve recording technique. Application of light mechanical stimuli on C-fibers successfully identified C-LTMRs with
firing thresholds below 10 mN (Figures 5A and 5B), slowly adapting firing, and pronounced after discharges as previously reported (Vallbo et al., 1999). Next, we used TTA-A2, a selective
T-type calcium channel antagonist (Francois et al., 2013). Local
blockade of Cav3.2 in fiber receptive fields induced a significant
elevation of the C-LTMR activation threshold and a reduction of
AP firing frequency (Figures 5A, 5B, and S5), therefore shifting
the firing characteristics toward those of high threshold
Cell Reports 10, 370382, January 20, 2015 2015 The Authors 373

Ultra-Slowly
Adapting

20

n=13

500pA

Slowly
Adapting

40

n=17

Intermediately
Adapting

60

n=16

Rapidly
Adapting

Figure 3. Mechanical Modalities in Cav3.2Expressing Neurons

80

Neuron capa. (pF)

6 m

20pA

6 m

10pA

30ms

4 m

10pA

6 m

n=17

0
RA IA SA USA

Small GFP+

0mV

Large GFP+

0 mV

20mV

0 mV

0 mV
20mV

4s

200ms

200ms
-60mV

-61mV

10s
-62mV

-64mV
22C
31C

7m

35C

6m

Large GFP+

Small GFP+

0 mV
20mV

0 mV

200ms

-59mV

-60mV
7m

7m

Velocity increase

Velocity increase

14C

Probe velocity (m.ms-1)

15C

100

10-1

10-2

10

Cav3.2-GFP
Cav3.2-GFP/Tafa4-GFP

-3

50

(A) Representative traces of mechanical-activated

(MA) currents in GFP+ DRG neuron soma in culture, elicited by a piezoelectric-driven probe
stimulation. The traveled distance by the probe is
indicated on the top of each trace (in mm). Time
constants of current decay discriminate four types
of MA currents classified as rapidly adapting (RA),
intermediately adapting (IA), slowly adapting (SA),
and ultraslowly adapting (USA).
(B) Cell size classification of recorded GFP+
neurons, according to their related capacitance
(in pF, 40 pF being the limit between small
and large neurons) as function of the MA
current type they express. Data are presented as
mean SEM.
(C and D) Action potentials (AP) generated by
mechanical stimulation (left) and cooling (right) in
small (C) and large (D) GFP+ neurons. Mechanical
and thermal stimulations are indicated below
each trace.
(E) AP generation in small (left) and large (right)
Cav3.2-GFP+ neurons with constant subthreshold mechanical stimulation applied at various velocities. Each color represents a different probe
velocity applied sequentially from slow to fast.
(F) Probe velocity threshold eliciting an AP according to the cell size of the recorded Cav3.2GFP+ neuron from Cav3.2-GFP KI (open black
circles) and Cav3.2-GFP KI 3 TAFA-4-GFP double-heterozygote mice (red filled circles).

100

Neuron capa. (pF)

mechanoreceptor C-fibers (C-HTMR). In contrast, application of

TTA-A2 had no effect on genuine C-HTMR (not shown). The
presence of Cav3.2 along C-LTMR axons suggests a potential
role in AP conduction velocity (CV). Indeed, application of
TTA-A2 on the entire skin-nerve preparation to target axonal
Cav3.2, slowed significantly the CV with minimal effect on the
AP shape, supporting the presence of functional axonal Cav3.2
channels (Figures 5C and 5D).
Generation of a Conditional Cav3.2 Knockout Mouse
loxP sites within the KI were designed to trigger tissue-specific
Cav3.2 gene inactivation. Partnering the KI animals with mice expressing Cre recombinase under the Nav1.8 promoter (Stirling
et al., 2005) allowed deletion of Cav3.2 specifically in smalldiameter DRG neurons (mice hereafter named SNS KO). As for
the KI, SNS KO are born in a Mendelian ratio, survive to
adulthood, and are indistinguishable from their KI littermates.
Figure 6A shows a perfect overlap between Nav1.8 and smalldiameter Cav3.2-GFP+ neurons in KI DRGs while this double labeling is completely lost in SNS KO. Consistently, the number of
Cav3.2-GFP+ neurons dropped from 33% to 13.9% 1.7% (Figure 6B). GFP costaining with specific markers indicated that
recombination mainly occurred in small-diameter Ret+ neurons
(Figure 6E) and, most importantly, in all TH+ C-LTMRs (Figure 6F).
This indicates effective Cav3.2 loss of function in this population,
without affecting TrkB+ neurons (Figures 6C, 6D, 6G, 6H, and
374 Cell Reports 10, 370382, January 20, 2015 2015 The Authors

S6AS6E). Accordingly, T-type currents are unaffected in SNS

KO Ad-LTMRs (Figures S6A and S6B), and live staining of
surface-expressed Cav3.2-GFP revealed a selective channel
disappearance in small neurons sparing the Ad-LTMRs (Figures
S6CS6E). Taken together, our data demonstrate that the SNS
KO mouse model represents a powerful tool to decipher the
role of peripheral Cav3.2 in the pathophysiology of C-LTMRs.
Cav3.2 in C-LTMR Controls Innocuous and Noxious
Mechanical/Cold Perception and Pathological Pain
To investigate whether SNS KO had modified mechanical
perception, we subjected these mice to a wide range of innocuous and noxious mechanical stimuli. We first used three static
von Frey filaments corresponding to innocuous, intermediate,
and noxious stimulations (0.07 g, 0.6 g, and 1.4 g respectively)
and generated a normalized withdrawal score. SNS KO mice
had a reduced score for the intermediate filament, indicative of
a dampened light-touch sensitivity (Figure 7A). Similarly, determination of the von Frey paw withdrawal threshold (PWT; up
and down method) showed a tactile deficit in SNS KO (Figure 7H,
PWT before spared nerve injury model [SNI] surgery). Differences between SNS KO mice and their KI littermates are not
due to motor or coordination deficits (Figure S7D). The animals
ability to respond to painful mechanical stimuli of stronger intensities was tested with two additional paradigms: the dynamic
von Frey (esthesiometer) and the Randall-Selitto test. In both

c
DRG
Hair
Follicule

e
Spinal
Cord

b
A

Hairy skin

GFP

S100

Merge

Magnification

50m

Sciatic nerve

GFP

Paranodin

Merge

KCNQ2

Merge

Paranodin

Merge

KCNQ2

Merge

20m

GFP
20m

Dorsal roots

GFP
20m

GFP

Figure 4. Subcellular Localization of Cav3.2

in Primary Sensory Neurons
Representation of the different sensory neuron
subcellular localizations explored for Cav3.2-GFP
expression in adult KI mice.
(A) Hairy skin whole-mount immunostaining of
Cav3.2-GFP (green) and Schwann cell marker
S100 (red) in KI mice. Empty arrows indicate
nerves innervating hair follicles forming lanceolate
endings. The right panel is a magnification of the
box on the merge image.
(B and C) Teased fibers costained with Cav3.2GFP (green), paranodin (red, top), or KCNQ2 (red,
bottom) in the sciatic nerve (B) or dorsal roots (C).
Nodes of Ranvier that express Cav3.2 in their
center are indicated by filled arrows, while Cav3.2
negative nodes of Ranvier are indicated by asterisks. Empty arrows represent Cav3.2-GFP immunoreactivity in small unmyelinated fibers.
(D) Cav3.2-GFP+ DRG soma positive for TrkB (left)
or Ret (right) at high magnification. Empty arrows
indicate the presence of Cav3.2-GFP in the axon
proximal segment.
(E) Lumbar spinal cord dorsal horn section from
Cav3.2-GFP KI mice costained for GFP (green),
IB4 (in red, marking lamina II inner), and PKCg
(blue, marking lamina II/inner-III).
(F) Transmission electron microscopy (TEM) images from pre-embedded GFP silver staining of
spinal cord dorsal horn coronal sections. Red dots
represent GFP antibody complexes around individual silver dot (25 mm). Areas in yellow are
presynaptic regions containing dense synaptic
vesicles. Green area is a postsynaptic region
recognized by the postsynaptic density. Right
panel is a magnification of the box in the left image.

20m

Proximal
segment

TrkB / GFP

GFP

25m

25m

Dorsal horn

Dorsal horn

pre
GFP

100m

IB4
post

PKC

Merge

100 nm

tests, SNS KO displayed dampened responses to paw and tail

pressure, respectively (Figures 7B and 7C). Altogether, our
data demonstrate that Cav3.2 is required to respond to innocuous and noxious mechanical stimuli.
We next assessed the thermosensation phenotype in SNS KO
mice. We subjected mice to a range of temperature sensation
paradigms including paw immersion in hot and cold water baths,

Ret / GFP
the dynamic cold- and hot-plate tests,
and a thermotaxis assay. In agreement
with the expression of Cav3.2-GFP in
cold-sensitive small DRG soma (Figure 3C), SNS KO exhibited a decreased
cold sensitivity in the paw immersion
test (Figure 7D), a diminished escape
behavior to noxious cold temperatures in
the dynamic cold-plate test (Figure 7E),
pre
and a strong preference for cooler temperatures in the gradient test paradigm
(Figures 7F and S7A), whereas acute
innocuous cool stimulation with the
acetone evaporation test (3 C4 C
Anti-GFP
Area (25nm)
drop of skin temperature) did not reveal
behavioral changes in control animals
(Figure 7I, day 2 before SNI surgery).
To test whether Cav3.2 in C-LTMR fibers also plays a role in
pathological pain, we first subjected mice to the formalin test.
SNS KO exhibited significantly reduced first and second phase
responses following formalin injection (Figure 7G), indicating
the importance of Cav3.2-expressing C-LTMRs in this process
related to both acute and inflammatory pain. Next, we assayed
neuropathic pain with the SNI model (Decosterd and Woolf,
Cell Reports 10, 370382, January 20, 2015 2015 The Authors 375

C-LTMR

Threshold
15

Threshold (mN)

Ctrl

TTA-A2 500nM

0.05 mV
Mech. Stim.

40

**

10

*
CV (m/s)

Ctrl

0.05 mV

20
10

Conduction Velocity
0.6

TTA-A2 500nM

30

Ctrl
TTA-A2 500nM
D

CV: 0.6 m/s

***

2 sec / 2 ms

Elec. Stim.

Frequency
Spike nb / Stimulation

0.4

0.2

CV: 0.31 m/s

0.0

30 ms

2000). We measured the increase in tactile and cold sensitivity in

which innocuous stimuli illicit a pain response (allodynia), a phenomenon characteristic of neuropathic pain. Development of
mechanical allodynia, demonstrated by the robust lowering of
tactile thresholds in the KI mice (92% decrease) was significantly
reduced in SNS KO by about a quarter (Figure 7H, day 14, and
Figure S7B, 73% decrease only). However, since sensory
perception is not linear but follows a logarithmic rule (Webers
law), the observed dampened allodynia in the lower end of the
von Frey detection range has a great biological relevance (Mills
et al., 2012). Correction for this law (as in Mills et al.s statistics
on the log PWT) accordingly increases the significance of SNS
KO dampened mechanical allodynia (from p < 0.05 (Figure 7H,
D14) to p < 0.001). Similarly, assessing sensitivity to innocuous
skin cooling by the acetone evaporation test revealed a strong
phenotype, since cold allodynia was abolished in SNS KO
mice (Figure 7I). Although sex differences in neuropathic pain
exist, both genders revealed the same phenotypes (Figures
S7B and S7C), and thus male and female data were pooled (Figures 7H and 7I). Taken together, these results show that Cav3.2
channel activity in C-LTMR fibers is required for fine perception
of multiple sensory modalities under both basal and pathological
conditions.
DISCUSSION
Low-voltage-activated T-type calcium channels encoded by the
Cav3.2 subunit have emerged as key elements of sensory
neuron excitability in shaping both pain and tactile perception
(Bourinet et al., 2014). However, deciphering how Cav3.2
376 Cell Reports 10, 370382, January 20, 2015 2015 The Authors

Figure 5. Cav3.2 Tunes Mechanical Responses of Wild-Type Mouse C-LTMR

(A) Representative nerve-skin recording of a
C-LTMR fiber in response to a mechanical stimulation of its receptive field in the hairy skin.
Response to a suprathreshold stimulation in
control condition (top trace) or after local application of 500 nM TTA-A2 to the receptive field
(bottom trace). Inset on the right is superposed
magnifications of isolated AP from traces on the
left in control and TTA-A2 situation.
(B) Summary of C-LTMR mechanical thresholds
(left) and APs frequency (right) before and after
application of 500 nM TTA-A2. Data are presented
as mean SEM (**p < 0.01 [n = 16] for thresholds;
***p < 0.001 [n = 16] for frequency; Wilcoxon
signed rank test).
(C and D) Action of TTA-A2 on conduction velocity. The tonic action of TTA-A2 on the conduction
velocity of a C-LTMR fiber (resting CV 0.6 m/s)
is shown in (C), where 500 nM of TTA-A2 is applied
to the entire recording chamber to have access
to the nerve truck and to the receptive field.
Note that conduction velocity is decreased
(CV 0.3 m/s in TTA-A2). (D) Histogram representation of conduction velocity before and after
application of 500 nM TTA-A2. Data presented
mean SEM. *p < 0.05 (n = 4); Wilcoxon signed
rank test.

achieves these functions requires assessment of their cellular

and subcellular localization as well as the analysis of the impact
of their functional loss in primary afferent neurons. To this end,
we have used mouse genetics to knock in a surface GFP-tagged
Cav3.2 channel together with an insertion of loxP sites into the
cacna1h gene to generate tissue-specific conditional knockout
animals. Our data demonstrate that this modification preserves
the channel function, as the knockin mice express a fully functional ecliptic-GFP surface-tagged Cav3.2. Although several
mouse models have been generated using cytosolic GFPtagged membrane receptors or ligand gated channels, such an
all-in-one extracellular ecliptic-GFP knockin-floxed mouse
has not been reported for an ion channel before. Using this versatile model, we provide a comprehensive examination of the
cellular distribution and functional relevance of the Cav3.2 channel in sensory processing, focusing on primary afferent neurons
and the mechanistic rationale of its specific expression in
defined subsets of LTMR neurons.
In the Cav3.2-GFP KI mouse, the tagged-channel expression
is subjected to the same genetic regulation as the WT
sequence since the 30 and the 50 regulatory regions are kept
intact. Expression overlap between GFP signal and Cav3.2
mRNA, as well as similarity with previous in situ hybridization
(Talley et al., 1999), unaffected current density, pharmacological properties, and normal animal behavior, indicated that the
tagged channel is processed in the same manner as the wildtype Cav3.2.
GFP immunolabeling showed that Cav3.2-expressing neurons
can be split into two distinct subsets of LTMRs: the Ad- and the
C-LTMRs, the first encompassing TrkB-positive and the latter

E
H

Ret-positive/IB4-negative neurons. These data are in sharp

contrast with the recently described molecular characterization
of Cav3.2-expressing neurons using a commercial anti-Cav3.2
antibody in DRG cultures (Rose et al., 2013). Limits of antibody
specificity could explain this discrepancy, since the same antiCav3.2 antibody identified bands unrelated to Cav3.2 in our experiments (Figure S1B).
Functional analysis using electrophysiological recordings
confirmed the expression pattern of Cav3.2-GFP neurons.
They displayed MA currents with distinct adaptation rates. Medium-sized neurons had MA currents with rapid adaptation,
typical of Ad-LTMRs, while the smaller neurons displayed SA
MA currents, similar to that described for C-LTMRs. It is worth
noting that the molecular characterization of these SA neurons
can be split into TAFA4+/TH+ and a yet to be identified Ret+/
IB4 population, suggesting a previously postulated heterogeneity within C-LTMRs (Delfini et al., 2013; Lou et al., 2013; Vrontou et al., 2013). Another feature of C-fibers expressing Cav3.2 is
their enhanced excitability to cold/cool stimulation independently of the presence of TRPM8, as described for Nav1.8-containing neurons (Abrahamsen et al., 2008; Zimmermann et al.,
2007). Altogether, we propose that C-LTMRs is a neuronal population of distinct identities, the majority of which express
Cav3.2.

Figure 6. Sensory Neuron Classes Expressing Cav3.2 in Cav3.2 SNS KO

(AF) Colocalization of Cav3.2-GFP with markers of
sensory neurons in adult DRG sections from
Cav3.2-GFP KI (A) or Cav3.2 SNS KO mice (AF).
Cav3.2-GFP-positive neurons costained by antiGFP (green) and by anti-Nav1.8 (red) antibodies in
(A). Top panels are DRG sections from Cav3.2-GFP
KI mice, and bottom panels are DRG sections from
Cav3.2 SNS KO mice. (B) Percentage of Cav3.2GFP-positive neurons in sections from Cav3.2-GFP
KI mice or Cav3.2-SNS KO mice. Data represent
mean SEM. Cav3.2-GFP positive neurons (green)
in Cav3.2 SNS KO costained by anti-TrkA, antiTrkB, anti-Ret, or anti-TH in red, respectively, in (C),
(D), (E), and (F).
(G) Size distribution (in mm2) of Cav3.2-GFP-positive neuron somas of Cav3.2-GFP KI mice (black
bars) or Cav3.2 SNS KO mice (white bars).
(H) Representative view of Cav3.2-GFP repartition
in adult sensory neuron subclasses in the Cav3.2
SNS KO lumbar DRGs in accordance with the
different markers described (CF).

Ad- and C-LTMRs mainly innervate the

hairy skin in the periphery with forkshaped lanceolate endings interdigitated
around the same two hair types, namely
the zigzag and the awl/auchene hairs
(98% of mouse fur hairs; Li et al., 2011).
Moreover, the multiple sensory afferent
terminal arbor morphologies revealed
that C-tactile fibers are extremely
branched, connect to a large number of
distant hairs, and cover a large skin territory (Wu et al., 2012). This organization is well matched for an
optimal stimulation by slowly moving stimuli. Generator potentials likely originate in axonal compartments, within or adjacent
to lanceolate structures, when mechanosensors open following
hair deflection. These potentials then spread from the daughter
branches down into the parental nerve arbor where the AP
firing threshold is reached (Hall and Treinin, 2011). While the
site of AP initiation is not well defined in sensory endings,
branching points are functionally important (Peng et al.,
1999). In this context, we provide evidence for a preferential
Cav3.2-GFP localization in distal nerve regions around fiber
branching sites. Considering that T-type calcium channels activate before sodium channels, they are well positioned for
lowering the excitability threshold and helping AP generation.
This is indeed what isolated skin nerve experiments demonstrated as pharmacological T-type channel inhibition in CLTMR receptive fields elevates their mechanical threshold.
Similarly TTA-A2 increases the threshold of electrically evoked
AP firing in Cav3.2 expressing DRG in culture (Francois et al.,
2013). Although we did not functionally investigate Ad-LTMRs
fibers here, previous reports using Mibefradil, a nonselective
T-type channel blocker, or Cav3.2-KO mice demonstrated
that Cav3.2 affects Ad-LTMR excitability (Wang and Lewin,
2011).
Cell Reports 10, 370382, January 20, 2015 2015 The Authors 377

A Von-Frey filaments

B Aesthesiometer

60
40
20

0.6g

1.4g

Paw immersion

5
4
3
2
1

50
0

***

n=7/7

15

cut off

80

0-90 min
10

10

****
5

***
*

60

40
5
20

0
5C

10C

40C

0
10

SNI mech. allodynia

90

***

300

60

200

30

100

0
10

20 30 40
Time (min)

50

**

VonFrey PWT (g)

0.1

1st
2nd
Phases

n=15/14

-4

12

Response to acetone (s/2min)

20
30
40
50
Floor temperature (C)

SNI cold allodynia

n=34/36
400

10

10 15 20 30 35 40 45
Floor temperature (C)

n=8/8

20

0
0

46C

G Formalin
120

F Temperature gradient

n=13/19

Cumulative Jumps

Paw withdrawal latency (s)

Cav3.2 GFP KI
Cav3.2 SNS KO

100

Dynamic cold / hot plate

Nocifensive behavior (s)

150

n=11/11
15

200

SNI surgery

Occupancy (%)

0.07g

n=7/9
250

Withdrawal threshold (g)

80

RS tail

n=11/15
7

Paw withdrawal threshold (g)

Paw withdrawal (%)

n=20/16
100

**

NS
a

Ctrl SNI CtrlSNI

Days

Figure 7. Acute Light-Touch, Mechanical Cold, and Chemical Pain Deficits and Profound Reduction of Allodynia Produced by Neuropathic
Nerve Injury in Cav3.2 SNS KO Mice
(AC) Cav3.2-GFP KI and Cav3.2 SNS KO paw withdrawal response to punctate mechanical stimulation of the hindpaw with three distinct von Frey filaments
(innocuous 0.07 g, intermediate 0.6 g, and noxious 1.4 g) (A) or with the electronic von Frey apparatus (B) and tail withdrawal to noxious compression with the
Randall-Selitto (RS) test (C).
(D) Withdrawal latency to paw immersion in noxious cold (5 C, 10 C) and hot (40 C, 46 C) water.
(E) Responses to dynamic cold and hotplate tests are shown as the cumulative number of jumps for Cav3.2-SNS KO and KI mice according to plate temperature.
(F) Thermotaxis with the temperature gradient test is shown as the time spent by mice on each of these zones for the 90 min duration of the experiment.
(G) Presence of nocifensive behavior (in s) following injection of 2% formalin (10 ml) into the hindpaw is recorded in both characteristic phases of this test (first
phase from 0 to 10 min, and second phase from 10 to 50 min). Kinetics of the experiment by 5 min intervals (left) and cumulative behavior for both phases (right).
(H and I) In the spared nerve injury neuropathic pain model (SNI), mice were tested before (day 2) and after (days 414) the surgery for their paw withdrawal
response to von Frey filaments (up-and-down method, H) or for their sensitivity to cool stimulation with acetone ejected on the plantar face of hindpaws
(cumulative time spent to flick or lick the paw in s [I]).
All data presented are mean SEM. *p < 0.05 **p < 0.01 ***p < 0.001 ****p < 0.0001. a indicates nonsignificant (day 2 SNS KO versus GFP KI), Mann-Whitney U
test or two-way ANOVA with Bonferroni post hoc test.

378 Cell Reports 10, 370382, January 20, 2015 2015 The Authors

DRG neurons are extremely compartmentalized with a very

long axonal element representing more than 99% of membrane
surface (Devor, 1999). Exploration of Cav3.2-GFP axonal localization revealed two patterns: one confined in the node of Ranvier
of Ad-LTMRs, the other with a diffuse distribution along some
thin unmyelinated C-fibers, corresponding presumably to CLTMRs neurons. In both cases, Cav3.2 is expected to regulate
AP conduction parameters. A restricted nodal localization of
ion channels is finely controlled by intracellular scaffolds, and
the same is also true for the axon initial segment (AIS) in the
CNS. Interestingly, T-type channels have been functionally identified in the AIS, thereby tuning neuronal intrinsic plasticity
(Bender et al., 2010; Grubb and Burrone, 2010). The function of
AIS located T-type channels in the control of neuronal output is
certainly conserved for axonal Cav3.2 within unmyelinated regions of DRGs, namely in the nodes of Ranvier for Ad-LTMRs
and along the fibers for C-LTMRs, where they could boost the
sodium spike conduction. Investigating the impact of T-type
channel blockade on C-LTMRs axonal conduction velocity, we
showed that Cav3.2 activity contributes to a faithful conduction
of AP. In vivo studies using microneurography showed that CLTMRs diverge from other C-fibers, with a stable CV upon
increased stimulation frequency and some degree of spontaneous activity that leads to efficient sensory information
signaling (George et al., 2007). These features might result
from the tuning of their excitability by Cav3.2 directly or indirectly
via intracellular calcium regulation of other membrane conductances. Elucidating these mechanisms would be particularly
interesting in pathological conditions that commonly alter axonal
excitability. As a follow up, a dynamic view of Cav3.2 localization
will be informative using the properties of the ecliptic-GFP designed to explore such phenomenon.
To complete the picture of Cav3.2 subcellular expression,
exploration of central terminals within the dorsal horn corroborated the robustness of the model. Indeed, C-LTMRs and AdLTMRs connect to the ventral lamina IIi and laminae IIi/III,
respectively (Li et al., 2011), where Cav3.2-GFP is detected overlapping with the PKC-g labeled layer known to receive LTMR inputs. Moreover, Cav3.2-GFP is also robustly expressed in spinal
cord neurons in these regions. Electron microscope analysis
within the lamina II confirms the presence of Cav3.2 at both
pre- and postsynapses in glomeruli structures. The presynaptic
presence substantiates the emerging role of low-voltage-gated
calcium influx in neurotransmitters release, including in dorsal
horn afferent inputs (Garca-Caballero et al., 2014; Jacus et al.,
2012). Understanding the postsynaptic contribution will need
further exploration. Since spinal neuron wiring and function are
still largely unknown, visualization of Cav3.2-GFP will be a
powerful tool to address such questions.
Finally, the Cav3.2 SNS KO provides insights into its functional
significance in C-fibers. First, removing Cav3.2 from C-fibers is
dispensable for their development and for their peripheral and
central innervations. Furthermore, it does not affect Cav3.2
expression in Ad-LTMRs or the CNS. We show here that the
sensitivity to perceive innocuous and noxious mechanical stimuli
is dampened in the SNS KO mice. This phenotype is combined
with an altered perception of cold in line with C-LTMR characteristics (Li et al., 2011). Moreover, the early and the late phases of

chemical pain induced by formalin is reduced. This biphasic

response is believed to reflect an acute pain in the periphery
and then a central sensitization. Yet, recent reevaluation of this
scheme emphasized on a peripheral substrate of all the underlying mechanisms (Fischer et al., 2014). This further supports that
Cav3.2-expressing peripheral C-LTMRs are important for this
behavior, in contrast to many other C-fibers (Shields et al., 2010).
In a pathological context, LTMRs afferents are the prime candidates to support injury-mediated allodynia due to their high
cutaneous sensitivity. The weight of clinical data emphasizes
the implication of myelinated A-b LTMR afferents as the major
cellular substrate of allodynia (Campbell et al., 1988). Preclinical
findings corroborate this view with cellular and molecular evidences, such as recent data using GFP-delta-opioid-receptor
KI mouse (Bardoni et al., 2014). Nevertheless, we show that
Cav3.2 channels are not expressed in Ab-fibers. Additionally,
Ad-LTMRs could also be good candidates. However, these fibers
do not seem to be conserved between rodents and humans
(Lechner and Lewin, 2013), which also limits somewhat the translational value of murine Ad-LTMR data. Moreover, in Ad-fibers,
Cav3.2 expression is not reduced in the SNS KO. Finally, the
contribution of C-LTMRs to allodynia has emerged from recent
studies on C-tactile-fibers both in human patients and in animal
models (McGlone et al., 2014; Nagi et al., 2011). Furthermore, unlike Ab-fibers, C-LTMRs respond to both touch and cooling,
consistent with an allodynic phenomenon associated with these
two stimuli. Therefore, our data not only demonstrate the contribution of C-tactile-fibers to the pathophysiology of neuropathic
pain but also reinforce the status of peripherally expressed
Cav3.2 channels as a target of choice for future treatments of mechanical and cold allodynia,. These symptoms are commonly
associated with multiple neuropathies. Assessing the impact of
C-LTMR-expressed Cav3.2 in relevant preclinical models is
therefore an important perspective. It is noteworthy that cold allodynia is abolished by intrathecal Cav3.2 antisense treatment in
a rat model of diabetic neuropathy (Obradovic et al., 2014). Other
compelling data that reinforce our research come from the clinical traits of autism combining impaired social interactions with
hypo- or hypertouch and/or cold perception that possibly implicate altered affective function of C-tactile-fibers (Cascio et al.,
2008; McGlone et al., 2014). Strikingly, Cav3.2 reduction and
gain of function are linked to a range of autism spectrum disorder
mutations of the CACNA1H gene (Splawski et al., 2006). Whether
these mutations affect patients C-LTMR functions is yet to be
elucidated. Thus, allodynia is certainly more complex than
involving solely Ab-fibers, with a likely scenario where other afferents, such as Cav3.2-expressing C-LTMRs, are synergistically
orchestrating this maladaptive phenomenon.
Consistent with our study, the VGLUT3 KO that reduces glutamate release from C-LTMRs shows impaired injury-induced
mechanical hypersensitivity (Seal et al., 2009), although the
phenotype may result from CNS-expressed transporters (Lou
et al., 2013). Conversely, in the same fibers, when the release
of the analgesic TAFA4 protein is abolished, pain is exacerbated
(Delfini et al., 2013). An attractive hypothesis would be that
Cav3.2-mediated calcium influx is important for the release of
the pronociceptive glutamate but dispensable for TAFA4 secretion. Moreover, the proexcitatory effect of Cav3.2 contributes to
Cell Reports 10, 370382, January 20, 2015 2015 The Authors 379

the efficient firing of C-LTMRs and the subsequent release of

glutamate, further amplified presynaptically by low-voltagegated calcium entry. In a neuropathic states, a change in the
analgesic/painful balance of C-LTMR signaling could largely
support allodynia, with a potential reduction of TAFA4 secretion
resulting in central disinhibition, combined with an increased
Cav3.2 membrane expression exacerbating pain signaling
from the periphery. However, the conditional Runx1 KO driven
by the VGLUT3-Cre shows that the altered distal skin innervation
and mechanosensitivity of VGLUT3-positive C-LTMRs does not
affect acute and pathological pain (even though cold nociception
was not assessed) (Lou et al., 2013). One possible reconciliation
with our results is that the Cav3.2-expressing C-LTMR population is broader than the persistent VGLUT3-expressing DRGs
and that Cav3.2 removal from VGLUT3-negative C-LTMRs contributes to the phenotype described here. The use of a more cellspecific Cav3.2 ablation would be necessary to resolve this
issue.
The Cav3.2 SNS KO phenotype agrees with what is expected
from the channel loss in C-LTMR fibers. However, it is distinct
from the pain phenotype of the complete channel deletion in
the null mice (Choi et al., 2007) or in all DRG and some spinal neurons in rats with intrathecal antisense-mediated Cav3.2 knockdown (Bourinet et al., 2005; Messinger et al., 2009) regarding
noxious heat perception. Interestingly, we demonstrate that
heat-evoked pain is retained when Cav3.2 is deleted only in
Nav1.8-positive DRGs, a phenotype reminiscent of conditional
ablation of different genes in the same neuronal population.
Thus, additional roles of Cav3.2 may be mediated through neurons that do not express Nav1.8 (in Ad-LTMRs in a few Nav1.8negative C-fibers, or likely within spinal and supraspinal pain
pathways). Deciphering all the sensory modalities associated
with different sets of Cav3.2-expressing cells will require additional specific conditional ablations.
In conclusion, we provide evidence that LTMR subtypes innervating the majority of hair subtypes in the skin rely on Cav3.2
channel functioning for their high sensitivity to detect, faithfully
conduct, and transfer tactile/pain information. This peculiarity
may result from a common global specification program of Adand C-LTMRs, ranging from the expression of specific molecular
markers, including Cav3.2 in their portfolio of ion channels, to
their anatomy and functional organization (Li et al., 2011; Lou
et al., 2013). Since LTMRs neurons play a crucial role in the pathophysiology of multiple chronic pain syndromes linked to allodynia, hypersensitivity, and spontaneous pain that are largely resistant to present therapies, these data further validate the utility of
pharmacologically inhibiting Cav3.2 to improve quality of life.

formed using standard methods (see the Supplemental Experimental Procedures for details).
Electrophysiology
Mouse DRGs were prepared as described previously. Cav3.2 GFP+ positive
neurons were identified from their extracellular GFP with live staining using
mouse anti-GFP immunoglobulin G (IgG) (1/100, 5 min, Roche) and Alexa
488-coupled anti-mouse IgG (1/300, 5 min, Invitrogen). Voltage-gated calcium
currents, MA currents, or membrane potentials were recorded as in Delfini
et al. (2013) and Francois et al. (2013). The isolated paw hairy skin-saphenous
nerve preparation and single-fiber recording were used as previously
described (see the Supplemental Experimental Procedures for details).
Behavioral Assays
Experiments were conducted at room temperature with 8- to 12-week-old
littermate males and with both genders for neuropathic pain. Animals were
acclimated to their testing environment prior to all experiments. Experimenters
were blind to the genotype during testing. Pain scores were determined with
strict adherence to ethical guidelines (Zimmermann, 1983) and in respect to
local ethical committee (see the Supplemental Experimental Procedures for
details).
SUPPLEMENTAL INFORMATION
Supplemental Information includes Supplemental Experimental Procedures
and seven figures and can be found with this article online at http://dx.doi.
org/10.1016/j.celrep.2014.12.042.
AUTHOR CONTRIBUTIONS
A.F., N.S., S.L., O.P., and E.B. designed the experiments. A.F., N.S., S.L., A.P.,
J.S., S.D., A.M., J.N., and E.B. performed the experiments. J.W. provided reagents. A.F., N.S., S.L., A.P., S.D., A.D., J.N., O.P., and E.B. analyzed the data.
A.F., A.M., O.P., and E.B. wrote the manuscript.
ACKNOWLEDGMENTS
We are grateful to Montpellier RIO imaging; C. Cazevieille and C. Sanchez for
TEM microscopy (CRIC, Montpellier); J. Rothman for ecliptic GFP cDNA; V.
Uebele (Merck) for the TTA-A2; R. Seal for VGLUT3-GFP mice DRGs; L. Goutebroze, J. Devaux, and P. Carroll for antibodies; and C. Barre`re, J. Guillemin,
C. Jopling, and N. Lamb for assistance. This work was supported by grants
from the ANR (ANR-09-MNPS-037, ANR-08-NMPS-025), the AFM (AFMPainT) to E.B., and by a grant of the Deutsche Forschungsgemeinschaft
Po137/42-1 to O.P. and N.S. A.F., S.L., and J.S. are supported by fellowships
from the French ministry of research and education, the AFM, the Labex ICST,
and the FRM.
Received: October 27, 2014
Revised: November 14, 2014
Accepted: December 18, 2014
Published: January 15, 2015
REFERENCES

EXPERIMENTAL PROCEDURES
Mice Lines
To generate Cav3.2-GFP KI mice, we used the bacterial artificial chromosomebased homologous recombination in embryonic stem cells. The Nav1.8-Cre
and the Venus-TAFA4 mice lines had been described previously (see the Supplemental Experimental Procedures for details).
Immunohistofluorescence
Immunohistochemistry on tissue sections, in situ hybridization, hairy skin
whole-mount immunohistochemistry, and teased fiber labeling were per-

380 Cell Reports 10, 370382, January 20, 2015 2015 The Authors

Abrahamsen, B., Zhao, J., Asante, C.O., Cendan, C.M., Marsh, S., MartinezBarbera, J.P., Nassar, M.A., Dickenson, A.H., and Wood, J.N. (2008). The
cell and molecular basis of mechanical, cold, and inflammatory pain. Science
321, 702705.
Abraira, V.E., and Ginty, D.D. (2013). The sensory neurons of touch. Neuron 79,
618639.
Bardoni, R., Tawfik, V.L., Wang, D., Francois, A., Solorzano, C., Shuster, S.A.,
Choudhury, P., Betelli, C., Cassidy, C., Smith, K., et al. (2014). Delta opioid receptors presynaptically regulate cutaneous mechanosensory neuron input to
the spinal cord dorsal horn. Neuron 81, 13121327.

Basbaum, A.I., Bautista, D.M., Scherrer, G., and Julius, D. (2009). Cellular and
molecular mechanisms of pain. Cell 139, 267284.

Grubb, M.S., and Burrone, J. (2010). Activity-dependent relocation of the axon

initial segment fine-tunes neuronal excitability. Nature 465, 10701074.

Bender, K.J., Ford, C.P., and Trussell, L.O. (2010). Dopaminergic modulation
of axon initial segment calcium channels regulates action potential initiation.
Neuron 68, 500511.

Hall, D.H., and Treinin, M. (2011). How does morphology relate to function in
sensory arbors? Trends Neurosci. 34, 443451.

Bessou, P., Burgess, P.R., Perl, E.R., and Taylor, C.B. (1971). Dynamic properties of mechanoreceptors with unmyelinated (C) fibers. J. Neurophysiol. 34,
116131.
Bourinet, E., Alloui, A., Monteil, A., Barre`re, C., Couette, B., Poirot, O., Pages,
A., McRory, J., Snutch, T.P., Eschalier, A., and Nargeot, J. (2005). Silencing of
the Cav3.2 T-type calcium channel gene in sensory neurons demonstrates its
major role in nociception. EMBO J. 24, 315324.
Bourinet, E., Altier, C., Hildebrand, M.E., Trang, T., Salter, M.W., and Zamponi,
G.W. (2014). Calcium-permeable ion channels in pain signaling. Physiol. Rev.
94, 81140.
Brumovsky, P., Villar, M.J., and Hokfelt, T. (2006). Tyrosine hydroxylase is expressed in a subpopulation of small dorsal root ganglion neurons in the adult
mouse. Exp. Neurol. 200, 153165.

Jacus, M.O., Uebele, V.N., Renger, J.J., and Todorovic, S.M. (2012). Presynaptic Cav3.2 channels regulate excitatory neurotransmission in nociceptive
dorsal horn neurons. J. Neurosci. 32, 93749382.
Jagodic, M.M., Pathirathna, S., Joksovic, P.M., Lee, W., Nelson, M.T., Naik,
A.K., Su, P., Jevtovic-Todorovic, V., and Todorovic, S.M. (2008). Upregulation
of the T-type calcium current in small rat sensory neurons after chronic
constrictive injury of the sciatic nerve. J. Neurophysiol. 99, 31513156.
Lallemend, F., and Ernfors, P. (2012). Molecular interactions underlying the
specification of sensory neurons. Trends Neurosci. 35, 373381.
Lechner, S.G., and Lewin, G.R. (2013). Hairy sensation. Physiology (Bethesda)
28, 142150.
Li, L., Rutlin, M., Abraira, V.E., Cassidy, C., Kus, L., Gong, S., Jankowski, M.P.,
Luo, W., Heintz, N., Koerber, H.R., et al. (2011). The functional organization of
cutaneous low-threshold mechanosensory neurons. Cell 147, 16151627.

Campbell, J.N., Raja, S.N., Meyer, R.A., and Mackinnon, S.E. (1988). Myelinated afferents signal the hyperalgesia associated with nerve injury. Pain 32,
8994.

Liu, Y., and Ma, Q. (2011). Generation of somatic sensory neuron diversity and
implications on sensory coding. Curr. Opin. Neurobiol. 21, 5260.

Cascio, C., McGlone, F., Folger, S., Tannan, V., Baranek, G., Pelphrey, K.A.,
and Essick, G. (2008). Tactile perception in adults with autism: a multidimensional psychophysical study. J. Autism Dev. Disord. 38, 127137.

Lou, S., Duan, B., Vong, L., Lowell, B.B., and Ma, Q. (2013). Runx1 controls terminal morphology and mechanosensitivity of VGLUT3-expressing C-mechanoreceptors. J. Neurosci. 33, 870882.

Choe, W., Messinger, R.B., Leach, E., Eckle, V.S., Obradovic, A., Salajegheh,
R., Jevtovic-Todorovic, V., and Todorovic, S.M. (2011). TTA-P2 is a potent and
selective blocker of T-type calcium channels in rat sensory neurons and a
novel antinociceptive agent. Mol. Pharmacol. 80, 900910.

Marger, F., Gelot, A., Alloui, A., Matricon, J., Ferrer, J.F., Barre`re, C., Pizzoccaro, A., Muller, E., Nargeot, J., Snutch, T.P., et al. (2011). T-type calcium
channels contribute to colonic hypersensitivity in a rat model of irritable bowel
syndrome. Proc. Natl. Acad. Sci. USA 108, 1126811273.

Choi, S., Na, H.S., Kim, J., Lee, J., Lee, S., Kim, D., Park, J., Chen, C.C., Campbell, K.P., and Shin, H.S. (2007). Attenuated pain responses in mice lacking
Ca(V)3.2 T-type channels. Genes Brain Behav. 6, 425431.

McGlone, F., Wessberg, J., and Olausson, H. (2014). Discriminative and affective touch: sensing and feeling. Neuron 82, 737755.

Coste, B., Crest, M., and Delmas, P. (2007). Pharmacological dissection and
distribution of NaN/Nav1.9, T-type Ca2+ currents, and mechanically activated
cation currents in different populations of DRG neurons. J. Gen. Physiol. 129,
5777.
Decosterd, I., and Woolf, C.J. (2000). Spared nerve injury: an animal model of
persistent peripheral neuropathic pain. Pain 87, 149158.
Delfini, M.C., Mantilleri, A., Gaillard, S., Hao, J., Reynders, A., Malapert, P.,
Alonso, S., Francois, A., Barrere, C., Seal, R., et al. (2013). TAFA4, a chemokine-like protein, modulates injury-induced mechanical and chemical pain hypersensitivity in mice. Cell Rep. 5, 378388.

Messinger, R.B., Naik, A.K., Jagodic, M.M., Nelson, M.T., Lee, W.Y., Choe,
W.J., Orestes, P., Latham, J.R., Todorovic, S.M., and Jevtovic-Todorovic, V.
(2009). In vivo silencing of the Ca(V)3.2 T-type calcium channels in sensory
neurons alleviates hyperalgesia in rats with streptozocin-induced diabetic
neuropathy. Pain 145, 184195.
Miesenbock, G., De Angelis, D.A., and Rothman, J.E. (1998). Visualizing secretion and synaptic transmission with pH-sensitive green fluorescent proteins.
Nature 394, 192195.
Mills, C., Leblond, D., Joshi, S., Zhu, C., Hsieh, G., Jacobson, P., Meyer, M.,
and Decker, M. (2012). Estimating efficacy and drug ED50s using von Frey
thresholds: impact of webers law and log transformation. J. Pain 13, 519523.

Delmas, P., Hao, J., and Rodat-Despoix, L. (2011). Molecular mechanisms of

mechanotransduction in mammalian sensory neurons. Nat. Rev. Neurosci. 12,
139153.

Nagi, S.S., Rubin, T.K., Chelvanayagam, D.K., Macefield, V.G., and Mahns,
D.A. (2011). Allodynia mediated by C-tactile afferents in human hairy skin.
J. Physiol. 589, 40654075.

Devor, M. (1999). Unexplained peculiarities of the dorsal root ganglion. Pain 6

(6), S27S35.

Obradovic, A.Lj., Hwang, S.M., Scarpa, J., Hong, S.J., Todorovic, S.M., and
Jevtovic-Todorovic, V. (2014). CaV3.2 T-type calcium channels in peripheral
sensory neurons are important for mibefradil-induced reversal of hyperalgesia
and allodynia in rats with painful diabetic neuropathy. PLoS ONE 9, e91467.

Fischer, M., Carli, G., Raboisson, P., and Reeh, P. (2014). The interphase of the
formalin test. Pain 155, 511521.
Francois, A., Kerckhove, N., Meleine, M., Alloui, A., Barrere, C., Gelot, A., Uebele, V.N., Renger, J.J., Eschalier, A., Ardid, D., and Bourinet, E. (2013). Statedependent properties of a new T-type calcium channel blocker enhance Ca(V)
3.2 selectivity and support analgesic effects. Pain 154, 283293.
Francois, A., Laffray, S., Pizzoccaro, A., Eschalier, A., and Bourinet, E. (2014).
T-type calcium channels in chronic pain: mouse models and specific blockers.
Pflugers Arch. 466, 707717.
Garca-Caballero, A., Gadotti, V.M., Stemkowski, P., Weiss, N., Souza, I.A.,
Hodgkinson, V., Bladen, C., Chen, L., Hamid, J., Pizzoccaro, A., et al.
(2014). The deubiquitinating enzyme USP5 modulates neuropathic and inflammatory pain by enhancing Cav3.2 channel activity. Neuron 83, 11441158.
George, A., Serra, J., Navarro, X., and Bostock, H. (2007). Velocity recovery cycles of single C fibres innervating rat skin. J. Physiol. 578, 213232.

Peng, Y.B., Ringkamp, M., Campbell, J.N., and Meyer, R.A. (1999). Electrophysiological assessment of the cutaneous arborization of Adelta-fiber nociceptors. J. Neurophysiol. 82, 11641177.
Perez-Reyes, E. (2003). Molecular physiology of low-voltage-activated t-type
calcium channels. Physiol. Rev. 83, 117161.
Rose, K.E., Lunardi, N., Boscolo, A., Dong, X., Erisir, A., Jevtovic-Todorovic,
V., and Todorovic, S.M. (2013). Immunohistological demonstration of CaV3.2
T-type voltage-gated calcium channel expression in soma of dorsal root
ganglion neurons and peripheral axons of rat and mouse. Neuroscience 250,
263274.
Seal, R.P., Wang, X., Guan, Y., Raja, S.N., Woodbury, C.J., Basbaum, A.I., and
Edwards, R.H. (2009). Injury-induced mechanical hypersensitivity requires
C-low threshold mechanoreceptors. Nature 462, 651655.

Cell Reports 10, 370382, January 20, 2015 2015 The Authors 381

Shields, S.D., Cavanaugh, D.J., Lee, H., Anderson, D.J., and Basbaum, A.I.
(2010). Pain behavior in the formalin test persists after ablation of the great majority of C-fiber nociceptors. Pain 151, 422429.

Vallbo, A.B., Olausson, H., and Wessberg, J. (1999). Unmyelinated afferents

constitute a second system coding tactile stimuli of the human hairy skin.
J. Neurophysiol. 81, 27532763.

Shin, J.B., Martinez-Salgado, C., Heppenstall, P.A., and Lewin, G.R. (2003). A
T-type calcium channel required for normal function of a mammalian mechanoreceptor. Nat. Neurosci. 6, 724730.

Vrontou, S., Wong, A.M., Rau, K.K., Koerber, H.R., and Anderson, D.J. (2013).
Genetic identification of C fibres that detect massage-like stroking of hairy skin
in vivo. Nature 493, 669673.

Splawski, I., Yoo, D.S., Stotz, S.C., Cherry, A., Clapham, D.E., and Keating,
M.T. (2006). CACNA1H mutations in autism spectrum disorders. J. Biol.
Chem. 281, 2208522091.
Stirling, L.C., Forlani, G., Baker, M.D., Wood, J.N., Matthews, E.A., Dickenson,
A.H., and Nassar, M.A. (2005). Nociceptor-specific gene deletion using heterozygous NaV1.8-Cre recombinase mice. Pain 113, 2736.
Takahashi, K., Sato, J., and Mizumura, K. (2003). Responses of C-fiber low
threshold mechanoreceptors and nociceptors to cold were facilitated in rats
persistently inflamed and hypersensitive to cold. Neurosci. Res. 47, 409419.
Takahashi, T., Aoki, Y., Okubo, K., Maeda, Y., Sekiguchi, F., Mitani, K., Nishikawa, H., and Kawabata, A. (2010). Upregulation of Ca(v)3.2 T-type calcium
channels targeted by endogenous hydrogen sulfide contributes to maintenance of neuropathic pain. Pain 150, 183191.
Talley, E.M., Cribbs, L.L., Lee, J.H., Daud, A., Perez-Reyes, E., and Bayliss,
D.A. (1999). Differential distribution of three members of a gene family encoding low voltage-activated (T-type) calcium channels. J. Neurosci. 19, 1895
1911.

382 Cell Reports 10, 370382, January 20, 2015 2015 The Authors

Wang, R., and Lewin, G.R. (2011). The Cav3.2 T-type calcium channel regulates temporal coding in mouse mechanoreceptors. J. Physiol. 589, 2229
2243.
Woodbury, C.J., Ritter, A.M., and Koerber, H.R. (2001). Central anatomy of individual rapidly adapting low-threshold mechanoreceptors innervating the
hairy skin of newborn mice: early maturation of hair follicle afferents.
J. Comp. Neurol. 436, 304323.
Wu, H., Williams, J., and Nathans, J. (2012). Morphologic diversity of cutaneous sensory afferents revealed by genetically directed sparse labeling. eLife
1, e00181.
Zimmermann, M. (1983). Ethical guidelines for investigations of experimental
pain in conscious animals. Pain 16, 109110.
Zimmermann, K., Leffler, A., Babes, A., Cendan, C.M., Carr, R.W., Kobayashi,
J., Nau, C., Wood, J.N., and Reeh, P.W. (2007). Sensory neuron sodium channel Nav1.8 is essential for pain at low temperatures. Nature 447, 855858.

Cell Reports
Supplemental Information

The Low-Threshold Calcium Channel Cav3.2

Determines Low-Threshold Mechanoreceptor Function
Amaury Franois, Niklas Schetter, Sophie Laffray, Juan Sanguesa, Anne Pizzoccaro,
Stefan Dubel, Annabelle Mantilleri, Joel Nargeot, Jacques Nol, John N. Wood, Aziz
Moqrich, Olaf Pongs, and Emmanuel Bourinet

Supplemental information

Supplemental experimental procedures

Mice lines
To generate the Cav3.2-GFP KI, the genomic BAC clone RP24-263I14 containing the entire sequence of the
murine cacna1h gene was obtained from CHORI. A ~1.1 kb MfeI fragment containing exon 6 was used as
basis for the targeting construct. The ecliptic GFP ORF was inserted in frame into exon 6 by overlap PCR. 5
and 3 homologous arms (each 2.5kb in total) were amplified by PCR and ligated into BamH1 and SacII
restriction sites. LoxP sites were inserted into the restriction sites BoxI and Van91I with the following annealed
oligomers:
5-gggtcATAACTTCGTATAGCATACATTATACGAAGTTAT-3,
5-ATAACTTCGTATAATGTATGCTATACGAAGTTATgaccc-3 (Box I) and
5-ATAACTTCGTATAGCATACATTATACGAAGTTATctg-3,
5-ATAACTTCGTATAATGTATGCTATACGAAGTTATcag-3 (Van91I).
This construct was cloned into pKO scrambler vector and finally a neomycin resistance cassette containing
FRT sites (AFN) was cloned into the BoxI site. The targeting vector was linearized with NotI and
electroporated into the SV129 R1 ES cell line. Positive clones were detected with southern blot analysis with
probes 5 and 3 of the targeting insertion (Figure S1C) (targeting frequency 0.7%). ES clones harboring the
modified cacna1h allele were injected into blastocysts to obtain chimeric mice which were crossed to
C57Bl/6J females to obtain germline transmission. The AFN cassette was constructed to be selfexcised upon
male germline transmission (unpublished Data; (Bunting et al., 1999)). Since this was not observed, the
resistance cassette AFN was removed by crossing F1 progeny with a Flipase-expressing deleter line, and
was confirmed by Southern-Blot (DrdI, not shown) and PCR with following oligomers:
AGATGCTTCAGGTATGGCTTGTCAG-3,
5-GAGCATTTCTGCATGCCGTTGTCG-3( Figure S1D).
The mice were further backcrossed with C57/Bl6j and resulting progeny (> 6 generations) were used for
experimental analysis.
The Nav1.8 Cre mouse line and the Venus-TAFA4 mice had been described previously and were crossed
with the Cav3.2 KI line. Genotypes were identified by PCR analysis of genomic DNA extracted from mouse
tail.

Immunohistofluorescence
Mice (P-60 to P160) were anesthetized with pentobarbital injection and transcardially perfused with HBSS
(pH 7.4, 4 C). Brain, spinal cords, DRGs, nerves and hairy skin were dissected from the perfused mice.
DRGs and spinal cord were fixed in TBS containing 4% PFA (pH7.4) at room temperature for 30 and 60
minutes, respectively. Brain and hairy skin was fixed with TBS containing 4% PFA at 4C overnight. Nerves
were fixed with TBS containing 4% PFA at 4C for 30 minutes. Tissues were then washed three times 10
minutes in TBS at 4C. DRGs, brain and spinal cord were embedded in 4% agarose and sectioned at 20-30
m using a vibratome. Then, sections were proceeded in free floating section and blocked with TBS
containing plus 0.05% tween 20 (TBST), 0.2% Triton X 100 and 10% normal serum (goat or donkey) at room
temperature for 1 hr. Tissue sections were incubated with primary antibodies diluted in blocking solution at
4C overnight or 4 hours at 37C. Then, sections were washed four times with TBST, and incubated with
secondary antibodies diluted in blocking solution at room temperature for 1 hr or 2 hrs at 37C, washed again
four times with TBST and put on superfrost slices to be mounted with Vectashield (Vectors laboratories). IB4
was diluted at 1/200 and incubated together with secondary antibodies. The percentage of GFP neurons
were determined compared to the number of PGP9.5 positive neurons.

Nerve teasing
Nerves were teased on Super-Frost Plus slides, air-dried for 23 hrs, and then kept frozen at 20C until
used. Slides were post-fixed using ice-cold methanol for 5 min, washed with TBS. The blocking procedure,
incubation with primary and secondary antibodies and mounting is the same as described above for tissue
slices.

Whole-mount immunohistochemistry.
Whole-mount immunohistochemistry of hairy skin was performed as described previously (Delfini et al., 2013;
Li et al., 2011). After a transcardial HBSS perfusion, back hairy skin from adult mice was treated with
commercial hair remover, wiped clean with tissue paper and tape stripped until glistening. Skin was dissected,
cut into small pieces and fixed in 4% PFA in TBS at 4C for 2 hrs. Tissue was rinsed in TBS and washed with
TBST containing 0.4% Triton X-100 every 30 min for 58 hrs. Then, the skin was incubated with primary
antibodies in 0.4% TBST containing 5% goat/donkey serum and 20% DMSO at room temperature for 35
days. Tissues were washed with 0.4% TBST every 30 min for 58 hrs and transferred to secondary antibodies

in 0.4% TBST containing 5% goat/donkey serum and 20% DMSO and incubated at room temperature for 2
4 days. Tissues were washed with 0.4% TBST every 30 min for 58 hrs. Hairy skin was dehydrated in 50%
methanol for 5 min and 100% methanol for 20 min, three times, and lastly cleared in BABB (benzyl alcohol,
Sigma; benzyl benzoate, Sigma; 1:2) at room temperature for 20 min.

Antibodies
The primary antibodies used for this study were: chicken anti-GFP (Invitrogen, 1/700), rabbit anti-GFP
(Invitrogen, 1/500), rabbit anti-PGP 9.5 (Millipore 1/500); Alexa 594-conjugated IB4 (Invitrogen, 1/200), rabbit
anti-CGRP (Peninsula, 1/250), mouse anti-NF200 (Sigma, 1:200), rabbit anti-peripherin (Millipore, 1/250);
rabbit anti-parvalbumin (Swant, 1/500), rabbit anti-PKC (Santa cruz, 1/1000), goat anti-Ret (R&D system,
1:50), sheep anti-Tyrosine Hydroxylase (Millipore; 1/400), rabbit anti-TrkA (Milipore, 1/500), goat anti-TrkB
(R&D system, 1/500), goat anti-TrkC (R&D system, 1/500), rabbit anti-TrpV1 (gift from Dr. Patrik Caroll
,1/1000), mouse anti Nav1.8 (Neuromab, 1/250), rabbit anti-KCNQ2 (1/200, gift from Dr. Jerome Devaux ),
rabbit anti-paranodin (1/500, gift from Dr. Laurence Goutebroze).
The secondary antibodies used were: FITC conjugated donkey anti-chicken antibody (Jackson
ImmunoResearch), CF488A conjugated donkey anti-chicken antibody (Sigma), Cy3 or Cy5 conjugated goat
anti-mouse antibodies (Jackson ImmunoReserch), Cy3 or Cy5 conjugated goat anti-rabbit antibodies
(Jackson laboratories), Alexa 488, 568 or 647 conjugated goat anti-rabbit antibodies (Invitrogen), Alexa 568
conjugated donkey anti-goat antibody (Invitrogen). Cy3 conjugated donkey anti-sheep antibody (Jackson),
Alexa 647 conjugated donkey anti-sheep antibody (Invitrogen). They were all used at 1/500.

Tissue processing for transmission electron microscopy visualization

Mice were anesthetized with pentobarbital injection and transcardially perfused with TBS containing 4% PFA
(pH7.4) and 0.5 glutaraldehyde (pH 7.4, 4C). Spinal cords were dissected and fixed in TBS containing 4%
PFA (pH7.4) and 0.5 glutaraldehyde at 4C overnight. Tissues were washed three times 10 minutes in TBS
at 4C. Spinal cords were embedded in 4% agarose and sectioned at 500 m using a vibratome. Then,
sections were processed in free floating sections and blocked against endogenous peroxidase with 3% H2O2
in TBS for 15 minutes and then in TBS containing plus 0.05% tween 20 (TBST), 0.1% saponin and 10%
normal serum (goat or donkey) at room temperature for 2 hrs. Tissue sections were incubated with primary
antibodies diluted in blocking solution at 4C overnight. Sections were washed four times with TBST and

incubated with biotin-coupled secondary antibodies diluted in blocking solution at room temperature for 1 hr,
washed again four times with TBST, incubated with streptavidin-coupled horseradish peroxidase diluted in
blocking solution at room temperature for 10 minutes and washed again four times with TBST. Silver particle
precipitation reactions were performed according to the manufacturers instrunctions EnzMet (nanoprobe)
kit. Tissues were immersed in a solution of 2.5% glutaraldehyde in Sorensens buffer (0.1M, pH 7.4) overnight
at 4C and after rinsing in Sorensens buffer, post-fixed in a 0.1% osmic acid for 1 hr in the dark at room
temperature. After two rinses in Sorensens buffer, neurons were dehydrated in a graded series of ethanol
solutions (30-100%) before embedding in EmBed 812 using an Automated Microwave Tissue Processor for
Electronic Microscopy, Leica EM AMW. Thin sections (70 nm; Leica-Reichert Ultracut E) were collected at
different levels of each block. These sections were counterstained with uranyl acetate and observed using a
Hitachi 7100 transmission electron microscope in the Centre de Ressources en Imagerie Cellulaire de
Montpellier (France).

In situ hybridization
In situ hybridization was performed following standard protocols (Delfini et al., 2013). To obtain adult tissues,
mice were deeply anesthetized with a mix of ketamine/xylazine and then transcardially perfused with an icecold solution of paraformaldehyde 4% in PBS (PAF). After dissection, they were post-fixed for at least 24 hrs
in the same fixative at 4C. P0 were collected in ice-cold PBS 1X, gently washed, and fixed for 24 hrs in 4%
PAF. RNA probes were synthesized using gene-specific PCR primers and cDNA templates from embryonic
or adult mouse DRG. Double fluorescent in situ hybridization was performed using a combination of
digoxigenin and fluorescein/biotin labeled probes. Probes were hybridized overnight at 55C and the slides
incubated with the horseradish peroxidase anti digoxigenin/fluorescein/biotin antibody (Roche). Final
detection was achieved using fluorescein/cy3/cy5 TSA plus kit (Perkin Elmer).

Western blotting
Experiments were performed on homozygote wild type (WT/WT), homozygote Cav3.2-GFP (KI/KI) and
heterozygotes (WT/KI) animals, as well as on total Knock-out for the Cav3.2 (KO)(Chen et al., 2003).
Tissues (brain, spinal cord, DRGs) were dissected in cold PBS on ice and ground into lysis buffer (150mM
NaCl, 20mM Tris HCl pH7.5, Triton X100 1%, Protease inhibitor cocktails (Roche). After 30 minutes left aside,
samples were passed through a 29 gauge needle 10 times and centrifuged at 1000g, 4C, 10 minutes. After

quantifying the supernatant protein concentration, 40 to 80 g total protein was loaded on a 6%

polyacrylamide gel. After transfer, membranes were incubated at 4C overnight with an anti-GFP rabbit
polyclonal antibody (1/20000; Torrey-pine) in blocking solution (5% milk/0.2% Tween 20/PBS). A goat
horseradish peroxidase-linked anti-rabbit antibody (1:2000; Jackson ImmunoResearch) was used as a
secondary antibody, and detection was performed with ECL SuperSignal West Femto maximum
chemiluminescent substrate (Thermo Scientific).

Mechanical and thermal stimulation on whole cell recordings

Mechanical stimulations were performed with a glass probe attached to a piezo-electric device (PZ-150 M;
Burleigh) and rapid changes of the bath temperature were achieved using the WAS-2 perfusion system
(Dittert et al., 2006).

Skin nerve recordings

The isolated dorsal paw hairy skin-saphenous nerve preparation and single-fiber recording technique were
used as described (Noel et al., 2009; Zimmermann et al., 2009). The skin was placed with the corium side
up in the organ bath superfused with synthetic interstitial fluid (SIF buffer) consisting of (in mM): NaCl, 120;
KCl, 3.5; MgSO4, 0.7; NaH2PO4, 1.7; Na2HCO3, 5; CaCl2, 2.0; Na-gluconate, 9.5; glucose, 5.5; sucrose,
7.5; and HEPES, 10 at a pH of 7.4. Fibers in the skin were stimulated with a mechanical search stimulus
applied with a glass rod and classified by conduction velocity (C fibers CV<1.2 m/s, A CV<10 m/s, and A
CV>10 m/s), von Frey hair thresholds and adaptation properties to supra-threshold stimuli. Receptive fields
were isolated with a stainless steel ring 0.8 cm diameter, which isolated receptive fields from the surrounding
bath chamber. Drugs were perfused inside the ring. TTA-A2 (Merck) was prepared from a 10mM stock
solution and applied to the preparation at 500 nM. For repetitive mechanical stimulation of identified C-LTMR
fibers, calibrated ramp and hold stimuli of 5 s duration were applied on the receptive fields with a cylindrical
probe, 2 mm diameter, mounted on a micromanipulator MP-285 Sutter instrument. The probe vertical
displacement was 100 m at the speed of 2,000 m/s. Extracellular potentials were recorded with a DAM 80
AC differential amplifier DAM 80 (WPI) and Spike2 software (Cambridge Electronic Design). Spikes were
discriminated off-line with the principal component analysis extension of the Spike2 software.

Behavioral assays
For experiments on naive mice, acute noxious and non-noxious mechanical perception was assessed using
the von Frey hair filaments of three different bending forces (0.07, 0.6 and 1.4 g) (Descoeur et al., 2011).
Mechanical threshold was evaluated using the aesthesiometer Bioseb apparatus with a ramping force of 10
s up to 10 g. Animal hind paws were assayed three times with at least 5 min of latency between trials to
calculate an averaged withdrawal. Noxious mechanical threshold was measured using Randall & Selitto
instruments on the animal tail. Tails were pinched three times near the base with, at least 5 min of latency
between trials, to calculate an averaged withdrawal. Threshold reflex responses to noxious and innocuous
temperatures were assessed using paw immersion in a water bath set at 5, 10, 40 or 46C (Allchorne et al.,
2005). Animals were habituated to manipulation and paw immersion at 25C two days before each
experiment. The cut off was set to 15 seconds to avoid tissue damage. Noxious cold and hot tolerance was
assessed using a dynamic cold/hot plate (Bioseb, France) as we previously described (Descoeur et al.,
2011). Animals were placed on the test arena with the floor temperature progressively cooled or heated from
30 to 1C or 30 to 50C at a rate of 1C or +1C/min respectively. Nocifensive behaviors (jumps) were
quoted. The thermal gradient test has been described previously (Moqrich et al., 2005) to determine mouse
thermotaxis. Briefly, mice were individually tracked for 90 min in an isolated corridor (2 corridors available in
the thermal gradient apparatus from Bioseb) ranging from 10C to 53.5C. Each session was video tracked
and the time spent at each position within the gradient, the time of immobility, and the distance walked by the
mice were recorded.
For the formalin test, mice were housed individually in Plexiglas chambers surrounded by mirrors, after
habituation to the testing environment for 30 minutes, mice were injected subcutaneously with formalin (10
l 2% in PBS) and the bouts of licking the injected paw counted at 5 minute intervals for 60 minutes. Time
spent exhibiting these pain behaviors was recorded for the first (0-10 min) and second phases (10-60 min).
For experiments on neuropathic pain, 8-12 week old male and female Cav3.2 KI and SNS KO mice were
used. Male and females were separately used on alternate days. For surgical procedures, animals were
deeply anesthetized with Ketamine-Xylasine (respectively 100 mg/kg and 10 mg/kg) and unilateral spared
nerve injury (SNI) surgery was performed (Decosterd and Woolf, 2000). Briefly, the left hindlimb was
immobilized in a lateral position and slightly elevated. Skin incision was made at mid-thigh level using the
femur as a landmark and muscle layers were separated to bring out the sciatic nerve and its three branches
(sural, common peroneal, and tibial nerves). Both tibial and common peroneal nerves were tightly ligated

with a #6.0 silk thread (Ethicon, Johnson, and Johnson Intl, Brussels, Belgium), transected together and a
12 mm section of the two nerves was removed. The sural branch was carefully preserved by avoiding any
nerve stretch or contact with surgical tools. Muscle layer was closed by careful apposition and skin sutured
with vicryl #4.0 threads. A sc. injection of NaCl 0.9% (10 ml/1 kg) was finally performed to prevent for
dehydratation. Mechanical sensitivity was assessed by placing animals on an elevated wire mesh grid and
stimulating the hind paw with von Frey hairs by using the up-down paradigm (Chaplan et al., 1994). The left
hindpaw was tested twice, separated by 30 min. The schedule of the experiment is as follows. After testing
for baseline mechanical sensitivity for two or three days, mice were subjected to surgery (SNI model). Note
that, as this model spared the sural nerve territory, von Frey stimuli were targeted to the region comprising
of the lateral aspect of the hindpaw during the baseline and post-operative measurements. Mice were then
tested for mechanical sensitivity on postoperative days 4, 7, and 14. Potential locomotor deficits were
evaluated with a rotarod accelerating procedure (0-16 rpm for 3 min) before and after SNI surgery at day 14.
Sensitivity to innocuous skin cooling was assessed by the acetone drop evaporation assay (Hulse et al.,
2012). Prior to surgery during the baseline and 14 days post SNI, 10 minutes following von Frey testing a
drop of acetone was applied to the sural nerve plantar territory of the left hind paw using a 1 ml syringe. The
duration of flinching/pain like behavior (seconds) was recorded immediately following acetone application for
a total period of 2 minutes. This was performed three times from which the mean was calculated.
Statistical analysis was performed by one- or two-way analysis of variance (ANOVA) and posthoc analysis
performed using Tukeys, Newman-Keuls or Mann-Whitney U test using GraphPad prism.

Supplemental Figure legends

Figure S1 (related to figure 1).

Expression analysis of mutant knock-in mRNA (A) and protein (B) in mouse brain preparations. A. Bars show
relative amounts of Cav3.2 (grey bars) and GFP (green bars) mRNAs. Internal standard for RNA
quantification assay was GAPDH mRNA. Values were normalized to KI/KI mRNA level. B, Cav3.2-GFP
protein expression in total brain membrane preparations. Proteins from mouse brain lysates were separated
by PAGE and after transfer, probed with anti-GFP antibody (Torey-Pine). -Tubulin served as a loading
control. Parallel analysis of protein extracted from Cav3.2 WT and KO mice confirmed the migration size of
the Cav3.2 protein with an anti-Cav3.2 antibody (Sigma). C. Cav3.2-GFP protein expression in total DRG
and spinal cord membranes analysed as in B with an anti-GFP antibody (Torey-Pine). D. Fluorescence insitu hybridization of Cav3.2 mRNA (in red) with anti-GFP immunofluorescence staining (in green) in an adult
DRG section from Cav3.2-GFP KI mice. E. Adult DRG sections from Cav3.2-GFP KI mice (KI; left panel) or
wild type mice (WT; right panel) analyzed by immunofluorescence with anti-GFP antibodies (Invitrogen). F.
Current voltage relationships for low voltage activated (LVA) T-type calcium currents in wild type (WT/WT,
open circles) and homozygous Cav3.2GFPe KI (KI/KI, closed circles) in rosette DRG neurons (almost
entirely supported by the Cav3.2 subunit in these neurons). Currents were elicited by depolarizations from 100mV to various test potentials (from -80 to -30 mV by 5mV increments during 100ms). Maximum inward
current was plotted as function of the test depolarization. Data are presented as mean SEM. G. Normalized
current availability curves of LVA Ca2+current in rosette DRG neurons of wild type mice (open circles) and
homozygous Cav3.2-GFP KI mice (filled circles). Membrane potential was held at -100 mV and a 10 s
conditional pre-pulse from -110 to -45 mV was followed by a 100 ms test pulse to -40 mV. Sweep interval
was 10 s. Maximum inward current during the test pulse was normalized and plotted as function of the prepulse depolarization, and fitted with the Boltzmann sigmoidal equation.

Figure S2 (related to figure 2)

Adult L4 DRG sections from Cav3.2-GFP KI mice co-stained with anti-GFP antibodies (green) and with antiNF 200 (red) and with peripherin (blue) in (A), anti-TrkC (red) in (B), anti-parvalbumin (red) in (C), anti-CGRP
(red) in (D), anti-TRPV1 (red) in (E). F-G. In situ hybridization of Cav3.2 mRNA in adult DRG sections from
VGLUT3-GFP mice (F) or from TAFA4-GFP mice (G), co stained with anti-GFP antibodies (green).

Figure S3 (related to figure 3)

In situ hybridization of Piezo2 mRNA (red) in adult L4 DRG section from Tafa4-GFP heterozygote mice
stained with anti-GFP antibodies (green).

Figure S4 (related to figure 4)

A. Co-staining of Cav3.2-GFP (green) and peripherin (red) in unmyelinanted teased fibers of sciatic nerve.
B. Absence of co-staining of Cav3.2-GFP (green), paranodin (red, top panel), or KCNQ2 (red, bottom panel)
in teased nerves of ventral roots from Cav3.2-GFP KI mice. C-D. Adult lumbar spinal cord dorsal horn
sections from Cav3.2-GFP KI mice co-stained for GFP (green), IB4 (in red, labeling lamina II inner), and
CGRP (C, blue, labeling lamina I & IIo), or parvalbumin (D, blue, labeling lamina III-IV).

Figure S5 (related to figure 5)

Representative skin-nerve recording of a C-LTMR fiber in response to mechanical stimulation of its receptive
field in the hairy skin. Top trace is the response to a threshold stimulation in control conditions, and bottom
traces show responses in the presence of 500nM TTA-A2 to increased mechanical stimuli, illustrating the
effect of T-type channel blockade on the firing threshold of the C-LTMR fibers. The inset on the right is a
superposed magnification of isolated single action potential (AP) of traces on the left.

Figure S6 (related to figure 6)

(A) Current densities for low voltage activated (LVA) T-type calcium currents measured at -40mV from a
holding potential of -90mV in rosette A-LTMR DRG neurons from homozygous Cav3.2GFPe KI flox (KI
green bar) or from homozygous Cav3.2GFPe KI flox X heterozygote Nav1.8-Cre (SNS KO hatched green
bar). (B) Typical trace of LVA T-type calcium currents in an A-LTMR DRG neuron from a SNS KO mouse.
(C-E left) Representative phase contrast images of live DRG neurons form KI (C), SNS KO (D), and WT (E)
mice. (C-E right) Fluorescent images of the same cells showing Cav3.2-GFP membrane expression of
medium and small sized DRGs in the KI (C), of medium sized DRG in the SNS KO (D), and the lack of
labeling in the WT (E).

Figure S7 (related to figure 7)

A. Thermotaxis with the temperature gradient test is shown as the time spent by mice on each of these zones,
for the first, second and third period of the experiment (30 min each). B.C Development of mechanical (B) or
cold (C) allodynia using the Spared Nerve Injury (SNI) model in males, females, or pooled males/females
(blue, red, and black bars respectively) Cav3.2-GFP KI and Cav3.2 SNS CKO mice (open and filled bars
respectively). B. Data are shown as the D14 threshold decrease percentage calculated as the % of decrease
in the PWT between d-2 (before SNI) and d14 (after SNI). In Cav3.2 GFP KI genotype, both genders exhibited
a significant decrease in mechanical threshold 14 days post SNI surgery that is not different from the pooled
data of both genders (black versus blue versus red). Moreover male and female Cav3.2-SNS KO animals
similarly exhibited a reduced allodynia. C. Cav3.2-GFP KI mice developed an augmented score to the
acetone test, while Cav3.2-SNS KO animal did not developed any cold allodynia after SNI (d14 versus d-2,
open bars). Males and females are not significantly different each other (blue versus red) and from the pooled
groups (Figure 7 I). D. Motor coordination as assessed by measuring the latency to fall (in s) from the rotarod
wheel subjected to an accelerating protocol (0-16rpm, 3min). Cav3.2-GFP KI and Cav3.2 SNS KO mice have
both similar performances before and after SNI surgery (d-2 versus d14, black and open bars).All data
presented are mean SEM. *p<0.05 **p<0.01; ***p<0.001, NS: non significative; Mann-Whitney test, or oneway ANOVA with Fisher LSD post hoc test.

References
Allchorne, A.J., Broom, D.C., and Woolf, C.J. (2005). Detection of cold pain, cold allodynia and cold
hyperalgesia in freely behaving rats. Mol Pain 1, 36.
Chaplan, S.R., Bach, F.W., Pogrel, J.W., Chung, J.M., and Yaksh, T.L. (1994). Quantitative assessment of
tactile allodynia in the rat paw. J Neurosci Methods 53, 55-63.
Chen, C.C., Lamping, K.G., Nuno, D.W., Barresi, R., Prouty, S.J., Lavoie, J.L., Cribbs, L.L., England, S.K.,
Sigmund, C.D., Weiss, R.M., et al. (2003). Abnormal coronary function in mice deficient in alpha1H T-type
Ca2+ channels. Science 302, 1416-1418.
Decosterd, I., and Woolf, C.J. (2000). Spared nerve injury: an animal model of persistent peripheral
neuropathic pain. Pain 87, 149-158.

Delfini, M.C., Mantilleri, A., Gaillard, S., Hao, J., Reynders, A., Malapert, P., Alaonso, S., Francois, A.,
Barrere, C., Seal, R.P., et al. (2013). TAFA4, a C-LTMRs-derived Chemokine-like Protein, Modulates
Injuryinduced Mechanical and Chemical Pain Hypersensitivity in Mice. Cell Reports 5, 11.
Descoeur, J., Pereira, V., Pizzoccaro, A., Francois, A., Ling, B., Maffre, V., Couette, B., Busserolles, J.,
Courteix, C., Noel, J., et al. (2011). Oxaliplatin-induced cold hypersensitivity is due to remodelling of ion
channel expression in nociceptors. EMBO Mol Med 3, 1-13.
Dittert, I., Benedikt, J., Vyklicky, L., Zimmermann, K., Reeh, P.W., and Vlachova, V. (2006). Improved
superfusion technique for rapid cooling or heating of cultured cells under patch-clamp conditions. J Neurosci
Methods 151, 178-185.
Hulse, R.P., Donaldson, L.F., and Wynick, D. (2012). Differential roles of galanin on mechanical and cooling
responses at the primary afferent nociceptor. Mol Pain 8, 41.
Li, L., Rutlin, M., Abraira, V.E., Cassidy, C., Kus, L., Gong, S., Jankowski, M.P., Luo, W., Heintz, N., Koerber,
H.R., et al. (2011). The functional organization of cutaneous low-threshold mechanosensory neurons. Cell
147, 1615-1627.
Moqrich, A., Hwang, S.W., Earley, T.J., Petrus, M.J., Murray, A.N., Spencer, K.S., Andahazy, M., Story, G.M.,
and Patapoutian, A. (2005). Impaired thermosensation in mice lacking TRPV3, a heat and camphor sensor
in the skin. Science 307, 1468-1472.
Noel, J., Zimmermann, K., Busserolles, J., Deval, E., Alloui, A., Diochot, S., Guy, N., Borsotto, M., Reeh, P.,
Eschalier, A., and Lazdunski, M. (2009). The mechano-activated K+ channels TRAAK and TREK-1 control
both warm and cold perception. EMBO J 28, 1308-1318.
Zimmermann, K., Hein, A., Hager, U., Kaczmarek, J.S., Turnquist, B.P., Clapham, D.E., and Reeh, P.W.
(2009). Phenotyping sensory nerve endings in vitro in the mouse. Nat Protoc 4, 174-196.

Normalized expression (%)

Brain

Cav3.2 probe
GFP probe

KO KI

WT

100
60
40

Cav3.2
GFP

< 250

KI

kDa
< 250

GFP

20

Tubulin

ISH - Cav3.2 / GFP

KI

Peak current (pA/pF)

100m

KI

kDa
< 250

GFP

GFP

KI

100m

WT/WT
KI/KI

-20
-40
-60
-80

-60

-40

Test potential (mV)

Franois et al, Figure S1

< 100

< 55

Normalized avaibility

80

DRG Sp.Cord

WT

100m

1.0

WT/WT
KI/KI

0.5

0.0
-100

-80

-60

Holding Potential (mV)

-40

GFP

NF200

GFP

TrkC

Merge

100m

100m

Peripherin

Franois et al, Figure S2

GFP

Parvalbumin

Merge

GFP

CGRP

Merge

GFP

TrpV1

Merge

Cav3.2 ISH

VGLUT3 GFP

Merge

Cav3.2 ISH

Tafa4 GFP

Merge

Merge

A
TAFA4 GFP

Piezo2 ISH

Merge

100m

Franois et al, Figure S3

GFP Peripherin

Merge

GFP

B
Ventral roots
(Motor)

GFP

IB4

Merge

KCNQ2

Merge

20m

GFP
Ventral roots
(Motor)

Paranodin

20m

GFP

IB4

Parvalbumin

Merge

100m

100m

CGRP

Franois et al, Figure S4

Merge

Mech. Stim.

11.2 mN
Control
20V
20V
0.2 s

11.2 mN
TTA-A2 500 nM

22.5 mN

32 mN

Franois et al, Figure S5

2 ms

LVA current (pA/pF)

60

NS

A-LTMR SNS KO
(-90 to -40mV)

40

20

20ms

A-LTMR

9pA/pF

KI

SNS KO

KI

>
>>
25m

SNS KO

WT

>

Franois et al, Figure S6

A Temperature gradient (successive periods)

n=7/7

40

30-60 min

Occupancy (%)

0-30 min

60-90 min

30

* **

20

* *

10

0
10

20

30

40

50

10

20

30

40

50

10

20

30

40

50

Floor temperature (C)

NS

C SNI cold allodynia

Response to acetone (s/2min)

B SNI mechanical allodynia

n=18

n=17

n=18

n=17

n=36

n=34

D14 PWT decrease (%)

-50

-100

***

D Motor coordination (rotarod)

(n=16/16)

Falling latency (s)

200

NS

NS

150

100

50

Ctrl

SNI

(d-2)

(d14)

Franois et al, Figure S7

NS
10

**
6

(n=8/6)

*
(n=7/7)

6
4
4
2
0

2
0

Ctrl SNI Ctrl SNI

(d-2)(d14)(d-2)(d14)

Ctrl SNI Ctrl SNI

(d-2)(d14)(d-2)(d14)

+
Cav3.2 GFP KI
Cav3.2 SNS KO