Sensory neuron sodium channel NaV1.8 is essential for pain at low temperatures

Abstract

Sensory acuity and motor dexterity deteriorate when human limbs cool down, but pain perception persists and cold-induced pain can become excruciating. Evolutionary pressure to enforce protective behaviour requires that damage-sensing neurons (nociceptors) continue to function at low temperatures. Here we show that this goal is achieved by endowing superficial endings of slowly conducting nociceptive fibres with the tetrodotoxin-resistant voltage-gated sodium channel (VGSC) Na(v)1.8 (ref. 2). This channel is essential for sustained excitability of nociceptors when the skin is cooled. We show that cooling excitable membranes progressively enhances the voltage-dependent slow inactivation of tetrodotoxin-sensitive VGSCs. In contrast, the inactivation properties of Na(v)1.8 are entirely cold-resistant. Moreover, low temperatures decrease the activation threshold of the sodium currents and increase the membrane resistance, augmenting the voltage change caused by any membrane current. Thus, in the cold, Na(v)1.8 remains available as the sole electrical impulse generator in nociceptors that transmits nociceptive information to the central nervous system. Consistent with this concept is the observation that Na(v)1.8-null mutant mice show negligible responses to noxious cold and mechanical stimulation at low temperatures. Our data present strong evidence for a specialized role of Na(v)1.8 in nociceptors as the critical molecule for the perception of cold pain and pain in the cold.

Full text

LETTERS
Sensory neuron sodium channel Na
v
1.8 is essential
for pain at low temperatures
Katharina Zimmermann
1
*, Andreas Leffler
2
*, Alexandru Babes
1,3
, Cruz Miguel Cendan
4
, Richard W. Carr
1
,
Jin-ichi Kobayashi
5
, Carla Nau
2
, John N. Wood
4
& Peter W. Reeh
1
Sensory acuity and motor dexterity deteriorate when human limbs
cool down, but pain perception persists and cold-induced pain can
become excruciating
1
. Evolutionary pressure to enforce protective
behaviour requires that damage-sensing neurons (nociceptors)
continue to function at low temperatures. Here we show that this
goal is achieved by endowing superficial endings of slowly conduct-
ing nociceptive fibres with the tetrodotoxin-resistant voltage-gated
sodium channel (VGSC) Na
v
1.8 (ref. 2). This channel is essential for
sustained excitability of nociceptors when the skin is cooled. We
show that cooling excitable membranes progressively enhances the
voltage-dependent slow inactivation of tetrodotoxin-sensitive
VGSCs. In contrast, the inactivation properties of Na
v
1.8 are
entirely cold-resistant. Moreover, low temperatures decrease the
activation threshold of the sodium currents and increase the mem-
brane resistance, augmenting the voltage change caused by any
membrane current. Thus, in the cold, Na
v
1.8 remains available as
the sole electrical impulse generator in nociceptors that transmits
nociceptive information to the central nervous system. Consistent
with this concept is the observation that Na
v
1.8-null mutant mice
3
show negligible responses to noxious cold and mechanical stimu-
lation at low temperatures. Our data present strong evidence for a
specialized role of Na
v
1.8 in nociceptors as the critical molecule for
the perception of cold pain and pain in the cold.
Nociceptors are peripheral sensory neurons that innervate the skin
and ‘deep’ tissues and respond to stimuli that are capable of producing
tissue damage and pain. Independent of the sensory transduction
molecules required for cold sensing
4–8
, the generation of action poten-
tials and their propagation to the central nervous system necessitate
the activity of VGSCs. Sensory neurons express several VGSC a-
subunits, with fast (for example Na
v
1.7) or slow (Na
v
1.8 and
Na
v
1.9) kinetics
2,9–11
. Whereas fast VGSCs are selectively blocked by
the puffer-fish poison tetrodotoxin (TTX), both Na
v
1.8 and Na
v
1.9
are resistant to TTX (TTXr)
2,11
.Na
v
1.8 is expressed exclusively in
sensory neurons, generates a slow-inactivating current with a high
threshold for activation, and is the only VGSC in the nociceptor
capable of generating action potentials in the presence of TTX
9,12,13
.
Using TTX as a tool to detect action potentials generated by Na
V
1.8 in
intact nociceptive terminals, we were able to define the conditions
under which this channel becomes physiologically relevant as an
impulse generator.
The isolated skin-nerve preparation
14
allows the focal application
of physical stimuli and chemicals to the cutaneous receptive field
of a single intact nociceptor and permits the measurement of the
consequent electrical events, propagated action potentials. In rat
preparations, we found that mechanocold-sensitive C-fibre (CMC)
nociceptors were blocked at 30 uC after the application of 1 mM TTX,
and as a result the nociceptors were rendered almost unexcitable when
stimulated mechanically (using forces up to 1,000 mN) or electrically
(to a maximum of 10 mA for 1 ms, applied with needle electrodes).
However, TTX-treated nociceptors still responded to cold stimulation
with action potential discharge, although with overall smaller cold
responses and thresholds at lower temperatures (Fig. 1a). Almost all
mechanosensitive C-fibres of all sensory categories in rat and mouse
skin were blocked by TTX at 30 uC, but on cooling they regained
excitability to mechanical (Fig. 1b) and electrical (Fig. 1c) stimulation
with thresholds of activation that were similar to those without TTX.
When the receptive field was warmed again, nociceptors returned
to the almost unexcitable state of TTX block (Fig. 1b, c; see Supple-
mentary Information and Supplementary Fig. S1).
These results indicate that the TTXr VGSC Na
v
1.8 becomes neces-
sary for the generation of action potentials at low temperatures. To test
this hypothesis we studied the influence of cooling on TTX-sensitive
(TTXs) and TTXr (namely Na
v
1.8) sodium currents in dorsal root
ganglion (DRG) neurons
13,15,16
. Cooling from 30 uCto10uC caused a
slowing of activation and inactivation kinetics and reduced peak cur-
rent amplitudes of both types of current (Fig. 2a, b). The decrease was
significantly larger for TTXs than TTXr currents when cells were held
at a membrane potential of 280 mV, which is similar to the physio-
logical membrane potential (Fig. 2c). This differential effect vanished
when neurons were held at 2120 mV (Fig. 2c), a configuration that
places sodium channels into a resting state devoid of inactivation. This
implied a differential modulation by cooling of mechanisms of volt-
age-dependent inactivation. Cooling had little effect on fast inactiva-
tion (Fig. 2d) but markedly shifted the slow inactivation of TTXs
currents towards more hyperpolarized potentials (Fig. 2e). In con-
trast, slow inactivation of TTXr currents was resistant to cooling
(Fig. 2f). Cooling also caused a shift of the voltage-dependent activa-
tion towards more hyperpolarized potentials; however, this shift was
small and similar for both TTXs and TTXr currents (Supplementary
Fig. S3). These distinct properties of TTXs and TTXr currents were
conserved in recombinant Na
v
1.7 (TTXs) and Na
v
1.8 (TTXr) chan-
nels; that is, cooling resulted in a leftward shift of slow inactivation of
Na
v
1.7 (Fig. 2g) but not of Na
v
1.8 (Fig. 2h). Consequently, the
decrease in current amplitudes on cooling was significantly stronger
for Na
v
1.7 than for Na
v
1.8 (Fig. 2i). Thus, the differential sensitivity to
cold of TTXs and TTXr currents in native neurons results from the
inherent properties of specific VGSC subunits.
Whether these mechanisms are relevant for the generation of
action potentials was first analysed by current-clamp recordings on
DRG neurons (Fig. 2j, k). At 30 uC, electrically evoked action poten-
tials in Na
v
1.8-deficient neurons required stronger currents than in
the wild type (WT; Fig. 2k, l). In addition, DRGs of both genotypes
*These authors contributed equally to this work.
1
Department of Physiology and Pathophysiology,
2
Department of Anesthesiology, Faculty of Medicine, Friedrich-Alexander University Erlangen-Nuremberg, 91054 Erlangen,
Germany.
3
Department of Animal Physiology and Biophysics, Faculty of Biology, University of Bucharest, 050095 Bucharest, Romania.
4
Department of Biology, University College
London, London WC1E 6BT, UK.
5
Department of Fixed Prosthodontics, Faculty of Dental Science, Kyushu University, Maidashi 3-1-1, Higashi-ku, Fukuoka 812-8582, Japan.
Vol 447
|
14 June 2007
|
doi:10.1038/nature05880
855
Nature
©2007
Publishing
Group

were TTX-resistant at all temperatures (data not shown). Both find-
ings are probably due to the low resting membrane potential of
cultured neurons (250 mV; Supplementary Fig. S4) at which most
TTXs channels are inactivated (Fig. 2d). These findings imply that
action potential generation at body temperature is largely resistant to
TTX, involving Na
v
1.8 and voltage-gated calcium channels
13,17
.In
contrast to the behaviour of WT neurons, even strong current injec-
tions failed to evoke action potentials in Na
v
1.8
2/2
neurons at 10 uC
(Fig. 2k, l). This suggests that action potential generation in DRG
neurons depends entirely on Na
v
1.8 at low temperatures.
TheimportanceofNa
v
1.8 as an action potential generator in noci-
ceptive terminals was analysed with the skin-nerve preparation. In con-
trast to DRGs, WT and Na
v
1.8
2/2
fibres showed comparable thresholds
for electrical (Fig. 3a) and mechanical (Fig. 3b) stimulatio n at both
30 uC and 10 uC. Cooling increased electrical thresholds in terminals
of both genotypes by about 2.5-fold. Whereas Na
v
1.8-deficient term-
inals lost excitability in the cold only in the presence of TTX, WT
terminals regained their excitability at the same electrical threshold as
in the absence of TTX (Fig. 3c, d). This implies that threshold responses
17 °C
23 °C
TTX
1 µM
100 µV
5 ms
0306090120
31
31
31
21
18
15
13
12
10
13
18
21
24
25
26
27
28
29
30
350 µA
0
30 60 90 120
16
11.4
128
512
128 mN
25 °C
2 ms
100 µV
20 s
10 s
TTX 1 µM
~5 min
11
Latenc
y
(
ms
)
4
8
256
30
20
10
0
0
100
Voltage
(µV)
a
Temperature
(ºC)
30
20
10
0
100
Voltage
(µV)
Temperature
(ºC)
b
31
Temperature (ºC)
c
31
28
24.5
Figure 1
|
TTX-blocked sensory C-fibre terminals regain responsiveness on
cooling. a, Original cold responses of a rat CMC fibre (0.4 m s
21
, 5.2 mN)
before and after treatment with TTX that rendered the unit unexcitable at
30 uC. Upper trace, temperature; lower trace, action potentials; arrow,
activation threshold. Inset: action potential shape.
b,TTX-blockedCMCfibre
terminalfrom a C57BL/6 mouse (0.55 m s
21
; 4 mN) respondingto mechanical
stimulation only during cooling. Upper trace, temperature; horizontal bars,
force stimuli.
c, Electrical stimulation protocol (twice threshold current, 1 ms,
0.5 s
21
) during cooling and rewarming in a C57BL/6 mouse mechanosensitive
C-fibre (0.4 m s
21
; 11.4 mN). Black traces, control; red traces, under TTX
(both at 75 mA); excitability of C-fibres recovered on cooling (less than
25 6 2 uC, n 5 9). Under TTX, action potentials showed longer latencies (see
Supplementary Fig. S2) and, if at all excitable, required much stronger current
(bottom trace). The time that passed between the bottom trace of the upper
panel and the first trace of the lower panel was 2 min.
10 ms 10 ms
2 nA
2 nA
Fraction of current
Fraction of current
–140
–100
–60
–20
20
–120
–80
–40
0
–120
–80
–40
0
Normalized current
*
*
Normalized current
*n.s.
*n.s.
Prepulse potential (mV)
Prepulse potential (mV)
10 °C
WT
Na
v
1.8
–/–
Not excitable
Threshold
current (pA)
*
10 ms10 ms
–120
mV
–80
mV
–120
mV
–80
mV
–120
mV
–80
mV
–120
mV
–80
mV
0.0
0.25
0.50
0.75
1.00
20 °C 10 °C
bc
0
0.2
0.4
0.6
0.8
1.0
0.2
0.4
0.6
0.8
1.0
0
–120
–80
–40
–120
–80
–40
0
0.2
0.4
0.6
0.8
1.0
0
80
0
–80
80
0
–80
600
0
0
–80
80
0
600
0
0.2
0.4
0.6
0.8
1.0
0
0.2
0.4
0.6
0.8
1.0
ef
a
g
d
n.s.
*
0.0
0.25
0.5
0.75
1.0
20 °C 10 °C
i
h
0
100
200
300
400
10 °C30 °C 30 °C
80
–80
Current (pA)
Voltage (mV)
jlk
Figure 2
|
Low temperature differentially affects TTXs and TTXr VGSCs.
a
, b, TTXs (a) and TTXr (b) current traces in DRGs held at 2120 mV. Red,
30 uC; black, 20 uC; blue, 10 uC.
c, Cold-induced decrease in current
amplitudes in DRGs normalized to values obtained at 30 uC (red dotted line).
Filled columns, TTXr; open columns, TTXs.
d, Steady-state fast inactivation
of TTXs (open circles) and TTXr (filled squares) currents in DRGs. Red,
30 uC; blue, 10 uC.
e–h, Steady-state slow inactivation of TTXs (e) and TTXr
(
f) currents in DRGs and of heterologously expressed Na
v
1.7 (g) and Na
v
1.8
(
h). Red, 30 uC; blue, 10 uC. i, Cold-induced reduction of current amplitudes
of Na
v
1.7 (open columns) and Na
v
1.8 (filled columns), normalized to values
obtained at 30 uC (red dotted line).
j, k, Current-clamp recordings from WT
(
j) and Na
v
1.8
2/2
(k) DRGs. Red, 30 uC; blue, 10 uC. l, Current injections
required to evoke action potentials in WT and Na
v
1.8
2/2
DRGs. Asterisk,
P , 0.001; Student’s t-test. Error bars represent s.e.m.; data in
a–l are for
n 5 7–12; see Supplementary Table S1 and Methods for details.
LETTERS NATURE
|
Vol 447
|
14 June 2007
856
Nature
©2007
Publishing
Group

of WT afferents in the cold are due entirely to Na
v
1.8, and that excit-
ability of Na
v
1.8-deficient afferents in the cold is retained only by virtue
of TTXs channels
3,16
. We assessed nociceptor function in the cold by
applying suprathreshold mechanical stimuli. In WT fibres, cooling
decreased mechanically induced discharges (Fig. 3e). This decrease
was even more pronounced in Na
v
1.8
2/2
fibres that displayed a severe
impairment of mechanonociception in the cold (Fig. 3f, and Supple-
mentary Fig. S5). Whereas noxious cold responsiveness in rat skin was
found to be partly resistant to TTX (Fig. 1a), the comparatively poor
cold responses in CMC fibres of mouse skin were retained in Na
v
1.8
2/2
(Supplementary Fig. S6). To induce more sustained responses to nox-
ious cold, we applied the TRPM8 (transient receptor potential cation
channel, subfamily M, member 8) agonist menthol
4
to the receptive
field (500 mM; Fig. 3g, h). Whereas the excitatory effect of menthol at
30 uC was similar in both genotypes (Fig. 3g, and Supplementary Fig.
S6), the menthol-sensitized cold responses observed in the WT were
almost abolished in Na
v
1.8
2/2
mice (Fig. 3g, h) and blocked by TTX
(not shown). Although transduction by cold (and menthol) is appar-
ently intact in cutaneous terminals of Na
v
1.8
2/2
neurons, action poten-
tial generation in response to noxious cold is strongly impaired. The
fact that C-fibre terminals but not DRG neurons of Na
v
1.8-deficient
mice were able to generate TTXs action potentials at 10 uCseems
contradictory. However, the depolarized membrane potential of cul-
tured DRG neurons is not representative for nerve terminals. Consi-
dering the strong voltage dependence of cold-induced slow inactivation
of TTXs currents, a more hyperpolarized resting membrane potential
would render TTXs sodium channels in terminals less susceptible to
cold. Moreover, overexpression of TTXs channels
3
and altered inactiva-
tion properties of TTXs sodium currents in Na
v
1.8-deficient neurons
16
have been proposed to compensate for the loss of Na
v
1.8.
Consistent with the data obtained in vitro was the observation that
the loss of Na
v
1.8 resulted in a very distinct phenotype in the cold-plate
test as an in vivo assay for cold nociception
6,18
. The characteristic foot-
lifting and jumping behaviour observed in WT mice was completely
absent in Na
v
1.8
2/2
mice when placed on a plate held at 0 uC (Fig. 3i).
It remained puzzling why much lower currents were required in the
cold than at 30 uC to trigger Na
v
1.8-generated action potentials in
TTX-treated WT terminals (Fig. 3c). It seems unlikely that the small
shift of the activation curve of TTXr currents in the cold accounts fully
for this marked decrease in activation threshold. We therefore studied
the electrical excitability of rat skin-nerve terminals by analysing
the electrical strength–duration relationship
19
to assess cold-induced
changes in passive and active membrane properties (see Methods).
Cooling caused an increase in the current threshold at shorter stimulus
durations and a prolongationof the chronaxy, the mostefficient stimu-
lus duration in terms of charge transfer (Fig. 4a); both effects reflect the
slowing of voltage-gated sodium channels in the cold (Supplementary
Fig. S7)
20
. In addition, the chronaxy represents the passive membrane
time constant, t
M
5 R
M
3 C
M
, and its prolongation indicates an in-
crease in membrane input resistance
19
, which is a cogent consequence
of cooling
20–22
. TTX increased both the current threshold at all stimulus
durations and the chronaxy, which is consistent with the high thresh-
old and slow kinetics of Na
v
1.8. However, cooling under TTX clearly
decreased current thresholds irrespective of stimulus duration (Fig. 4b,
and Supplementary Fig. S7). This cold-induced increase in TTXr excit-
ability is consistent with both a leftward shift of the TTXr activation
curve and an increase in input membrane resistance. The latter effect
may be of major importance in nerve endings (with their high ratio of
0105
von Frey threshold (mN)
Threshold current (µA)
N
ot excitable
N
ot excitable
N
ot excita
ble <350
N
ot excita
ble
N
ot
excitable
N
ot excitable <256
Threshold current (µA)
von Frey thresh
old (mN)
g
–10
Liftings and jumpings
Spi
kes per stimulus
WT
Na
v
1.8
–/
–
i
30 ºC 30–10 ºC
**
** *
10
30
4
8
12
16
0
20
Temperature
(°C)
0 102030
Time (s)
40 50 60 70
0
20
40
60
80
100
**
n.s.
0
5
10
15
20
25
30
35
15 20 25 30010515
Time (s)Time (s)
20 25 30
0
6
12
18
6
12
18
0
Spikes per 4 s
Spikes s
–1
h
ef
0
20
40
60
80
1
4
16
64
256
Na
v
1.8
–/–
WTWT WTNa
v
1.8
–/–
Na
v
1.8
–/–
Na
v
1.8
–/–
a WT bdc
1
4
16
64
256
0
20
40
60
80
30 ºC
10 ºC
30 ºC
10 ºC
30 ºC
10 ºC
30 ºC
10 ºC
30 ºC
10 ºC
30 ºC
10 ºC
30 ºC
10 ºC
30 ºC
10 ºC
Figure 3
|
Loss of Na
v
1.8 impairs responsiveness to noxious stimulation in
the cold. a–d, Electrical (a, c) and von Frey (b, d) thresholds of C-fibre
terminals in WT and Na
v
1.8
2/2
mice at 30 uC (open columns) and 10 uC
(filled columns). With 1 mMTTX(
c, d), WT terminals regained excitability in
the cold, whereas Na
v
1.8
2/2
terminals remained unexcitable (n 5 10).
e, f, C-fibre responses to noxious mechanical stimulation (constant pressure
of three times the von Frey threshold) of the terminals at 30 uC (open
columns) and 10 uC (filled columns) in WT (
e) and Na
v
1.8
2/2
(f) (52%
decrease at 10 uC in WT; 92% decrease in Na
v
1.8
2/2
; n 5 10; see
Supplementary Fig. S5).
g, Menthol (500 mM)-sensitized cold responses of
C-fibres in WT (open columns) and Na
v
1.8
2/2
(filled columns) (4-s bins,
two-point adjacent averaging).
h,At30uC menthol excited C-fibres from WT
(open columns) and Na
v
1.8
2/2
(filled columns) similarly (action potentials
per 2 min; see Supplementary Fig. S6b); sensitization to cold occurred only in
WT (action potentials per 60 s; n 5 7–12).
i, Cold-plate test (0 uC). Na
v
1.8
2/2
mice showed negligible response to noxious cold (n 5 9–10). Error bars
indicate s.e.m. The boxes in
b and d represent the 25th and 75th centiles; the
horizontal lines inside the boxes show the median (if not identical with the
25th centile), and the ‘whiskers’ (enhanced by asterisks) indicate minima and
maxima. Asterisk, P , 0.01; two asterisks, P , 0.001; Student’s t-test.
200
6,000
a
b
4,000
2,000
0
6,000
4,000
2,000
0
150
100
10 ºC
30 ºC
20 ºC
10 ºC
Threshold current (µA)Threshold charge (µC)
50
0
0.1 1 10 0.1 1
201002010
Time (ms)
Stimulus width (ms)
0 –5.6
–6.0–0.46–4.07
10
300
200
100
0
Figure 4
|
Cooling increases TTXr excitability in nociceptive terminals.
Electrical stimulus strength–duration measurements (upper panels) from a
rat CMC-fibre terminal (0.47 m s
21
; 5.2 mN) at different temperatures
19
,
without TTX (
a) and in the presence of 1 mM TTX (b). Lower panels:
rheobase current and chronaxy were determined by the gradient and zero-
charge intercept, respectively, of the regression of charge (Q) on stimulus
duration. In the absence of TTX (
a), cooling increased current thresholds at
shorter stimulus durations (less than chronaxy) but decreased the rheobase
and prolonged the chronaxy. In
b, TTX increased current thresholds and
chronaxy, but cooling under TTX primarily decreased the rheobase and
decreased rather than increased the chronaxy (see Supplementary Fig. S7).
NATURE
|
Vol 447
|
14 June 2007 LETTERS
857
Nature
©2007
Publishing
Group

surface area to volume), because it will decrease the current required to
activate Na
v
1.8 by augmenting the voltage change across the mem-
brane evoked by any depolarizing current
21,22
. The increase in input
membrane resistance is consistent with physical laws and with cold-
induced closure of certain potassium channels that have an established
role in cold nociception
23,24
. Finally, a cold-induced block of the
sodium–potassium (ATP-ase) pump
25
may also contribute to the
increase in TTXr excitability.
Mammalian sodium channels are evolutionarily the most recent
of the voltage-gated channels; however, they represent the best-
characterized family. VGSCs are encoded by a family of genes that have
been highly conserved throughout evolution, which reflects their func-
tion in the regulation of excitability
10,26
. The fact that Na
v
1.8 inactiva-
tion does not increase with cooling is likely to represent a conserved
feature, vitally important for cold-blooded animals (poikilotherms),
that has enabled homeotherms to detect and avoid tissue-damaging
levels of cold. It may also explain the exquisite tissue specificity of
Na
v
1.8 expression in mammals, in which it is expressed only in noci-
ceptive neurons. Several findings imply a role for Na
v
1.8 in inflammat-
ory and neuropathic pain
15,27,28
. One role for Na
v
1.8 in these conditions
is to maintain the excitability of nociceptive neurons during repetitive
firing and sustained depolarization that leads to progressive inactiva-
tion of TTXs sodium channels
17
. The high threshold for voltage-
dependent fast inactivation of Na
v
1.8 is the distinguishing feature that
renders it tolerant of tonic depolarization
28
. At dangerously low tem-
peratures, however, the cold resistance of slow inactivation of Na
v
1.8
seems to be a decisive property making Na
v
1.8 an essential moleculefor
nociception in the cold and for cold pain.
METHODS SUMMARY
Animals. Adult Wistar rats and littermates of WT and Na
v
1.8-deficient mice
3
(continuously backcrossed to C57BL/6/J since 1999) were used.
Behavioural assays. A cold-plate analgesia meter was used to assess noxious cold
sensitivity on the plantar surface of the paw as described
6
.
DRG cell culture. Adult mice were killed by CO
2
inhalation. DRGs from all
spinal levels were removed and cultured as described
29
. Recordings were made
within 24 h in culture.
Transfection procedures. Rat Na
v
1.8 was transiently transfected in the DRG
neuroblastoma hybridoma celline ND7/23, and human Na
v
1.7 was transiently
transfected in HEK 293T cells as described
29
. Recordings were made within 3–4 days.
Electrophysiology: single-fibre recordings. The isolated skin-saphenous-nerve
preparation and single-fibre recording technique were used as described
14
.
Protocols for electrical stimulus strength–duration measurements were adapted
19
.
Electrophysiology: patch-clamp recordings. Whole-cell recordings were per-
formed with an Axopatch 200B amplifier and pClamp 8.2 software (Axon
Instruments) at defined temperatures
30
. Current-clamp recordings were made
by using DRG neurons with a resting membrane potential more negative than
240 mV. Voltage-clamp recordings were performed on DRG neurons from
WT mice (18.0 6 1.0 pF, n 5 39) and Na
v
1.8
2/2
mice (16.9 6 1.0 pF, n 5 50;
P 5 0.46, Student’s t-test). TTXr currents were studied in the presence of
250 nM TTX. Cells with an increase in passive leak current during cooling were
discarded. Sodium channels of DRG neurons or heterologously expressed Na
v
1.8
and Na
v
1.7 were studied as described
29
. Values of all experiments are presented as
means 6 s.e.m.; statistical comparisons were calculated with Student’s t-test,
significance symbols are defined in figure legends.
Full Methods and any associated references are available in the online version of
the paper at www.nature.com/nature.
Received 19 March; accepted 25 April 2007.
1. Solomon, S. & Stearns, C. R. On the role of the weather in the deaths of R. F. Scott
and his companions. Proc. Natl Acad. Sci. USA 96, 13012
–
13016 (1999).
2. Akopian, A. N., Sivilotti, L. & Wood, J. N. A tetrodotoxin-resistant voltage-gated
sodium channel expressed by sensory neurons. Nature 379, 257
–
262 (1996).
3. Akopian, A. N. et al. The tetrodotoxin-resistant sodium channel SNS has a
specialized function in pain pathways. Nature Neurosci. 2, 541
–
548 (1999).
4. Peier, A. M. et al. A TRP channel that senses cold stimuli and menthol. Cell 108,
705
–
715 (2002).
5. McKemy, D. D., Neuhausser, W. M. & Julius, D. Identification of a cold receptor reveals
a general role for TRP channels in thermosensation. Nature 416, 52
–
58 (2002).
6. Kwan, K. Y. et al. TRPA1 contributes to cold, mechanical, and chemical nociception
but is not essential for hair-cell transduction. Neuron 50, 277
–
289 (2006).
7. Story, G. M. et al. ANKTM1, a TRP-like channel expressed in nociceptive neurons,
is activated by cold temperatures. Cell 112, 819
–
829 (2003).
8. Babes, A., Zorzon, D. & Reid, G. Two populations of cold-sensitive neurons in rat
dorsal root ganglia and their modulation by nerve growth factor. Eur. J. Neurosci.
20, 2276
–
2282 (2004).
9. Brock, J. A., McLachlan, E. M. & Belmonte, C. Tetrodotoxin-resistant impulses in
single nociceptor nerve terminals in guinea-pig cornea. J. Physiol. (Lond.) 512,
211
–
217 (1998).
10. Catterall, W. A., Goldin, A. L. & Waxman, S. G. International Union of
Pharmacology. XLVII. Nomenclature and structure
–
function relationships of
voltage-gated sodium channels. Pharmacol. Rev. 57, 397
–
409 (2005).
11. Dib-Hajj, S., Black, J. A., Cummins, T. R. & Waxman, S. G. NaN/Nav1.9: a sodium
channel with unique properties. Trends Neurosci. 25, 253
–
259 (2002).
12. Strassman, A. M. & Raymond, S. A. Electrophysiological evidence for
tetrodotoxin-resistant sodium channels in slowly conducting dural sensory fibers.
J. Neurophysiol. 81, 413
–
424 (1999).
13. Blair, N. T. & Bean, B. P. Roles of tetrodotoxin (TTX)-sensitive Na
1
current, TTX-
resistant Na
1
current, and Ca
21
current in the action potentials of nociceptive
sensory neurons. J. Neurosci. 22, 10277
–
10290 (2002).
14. Reeh, P. W. Sensory receptors in mammalian skin in an in vitro preparation.
Neurosci. Lett. 66, 141
–
146 (1986).
15. Wood, J. N., Boorman, J. P., Okuse, K. & Baker, M. D. Voltage-gated sodium
channels and pain pathways. J. Neurobiol. 61, 55
–
71 (2004).
16. Matsutomi, T., Nakamoto, C., Zheng, T., Kakimura, J. & Ogata, N. Multiple types of
Na
1
currents mediate action potential electrogenesis in small neurons of mouse
dorsal root ganglia. Pflu¨gers Arch. 453, 83
–
96 (2006).
17. Renganathan, M., Cummins, T. R. & Waxman, S. G. Contribution of Na
v
1.8 sodium
channels to action potential electrogenesis in DRG neurons. J. Neurophysiol. 86,
629
–
640 (2001).
18. Bautista, D. M. et al. TRPA1 mediates the inflammatory actions of environmental
irritants and proalgesic agents. Cell 124, 1269
–
1282 (2006).
19. Bostock, H. The strength
–
duration relationship for excitation of myelinated nerve:
computed dependenceonmembraneparam eters.J. Physiol. (Lond.) 341, 59
–
74 (1983).
20. Hodkin, A. L. & Katz, B. The effect of temperature on the electrical activity of the
giant axon of the squid. J. Physiol. (Lond.) 109, 240
–
249 (1949).
21. Thompson, S. M., Masukawa, L. M. & Prince, D. A. Temperature dependence of
intrinsic membrane properties and synaptic potentials in hippocampal CA1
neurons in vitro. J. Neurosci. 5, 817
–
824 (1985).
22. Volgushev, M., Vidyasagar, T. R., Chistiakova, M., Yousef, T. & Eysel, U. T.
Membrane properties and spike generation in rat visual cortical cells during
reversible cooling. J. Physiol. (Lond.) 522, 59
–
76 (2000).
23. Reid, G. & Flonta, M. Cold transduction by inhibition of a background potassium
conductance in rat primary sensory neurones. Neurosci. Lett. 297, 171
–
174 (2001).
24. Viana, F., de la Pena, E. & Belmonte, C. Specificity of cold thermotransduction is
determined by differential ionic channel expression. Nature Neurosci. 5, 254
–
260
(2002).
25. Pierau, F. K., Torrey, P. & Carpenter, D. O. Mammalian cold receptor afferents: role of
an electrogenic sodium pump in sensory transduction. Brain Res. 73, 156
–
160 (1974).
26. Goldin, A. L. Evolution of voltage-gated Na
1
channels. J. Exp. Biol. 205, 575
–
584
(2002).
27. Matthews, E. A., Wood, J. N. & Dickenson, A. H. Nav 1.8-null mice show stimulus-
dependent deficits in spinal neuronal activity. Mol. Pain 2, 5 (2006).
28. Rush, A. M. et al. A single sodium channel mutation produces hyper- or
hypoexcitability in different types of neurons. Proc. Natl Acad. Sci. USA 103,
8245
–
8250 (2006).
29. Leffler, A., Reiprich, A., Mohapatra, D. P. & Nau, C. Use-dependent block by
lidocaine but not amitriptyline is more pronounced in tetrodotoxin (TTX)-
resistant Na
v
1.8 than in TTX-sensitive Na
1
channels. J. Pharmacol. Exp. Ther. 320,
354
–
364 (2007).
30. Dittert,I. et al. Improvedsuperfusion technique for rapidcooling or heatingof cultured
cells under patch-clamp conditions. J. Neurosci. Methods 151, 178
–
185 (2006).
Supplementary Information is linked to the online version of the paper at
www.nature.com/nature.
Acknowledgements We thank H. Bostock and O. Krishtal for discussions. This
work was supported by the Wellcome Trust, the MRC, SEUI/MEC, the German
Research Foundation and the Humboldt Foundation.
Author Contributions J.K. made the decisive discovery that TTX-blocked rat CMC
fibres fired in response to noxious cold stimulation. A.L. performed the
voltage-clamp recordings, A.B. the current-clamp recordings, and K.Z. and P.W.R.
the skin-nerve recordings. J.N.W. and C.M.C. provided the mice and conducted
behavioural experiments. C.N. provided heterologously expressed Na
v
1.7 and
Na
v
1.8. R.W.C. wrote a script for Spike2 enabling the modified excitability testing.
Author Information Reprints and permissions information is available at
www.nature.com/reprints. The authors declare no competing financial interests.
Correspondence and requests for materials should be addressed to K.Z.
(zimmermann@physiologie1.uni-erlangen.de).
LETTERS NATURE
|
Vol 447
|
14 June 2007
858
Nature
©2007
Publishing
Group

METHODS
Animals. Wistar rats (150–200 g) and littermates of WT and Na
v
1.8-deficient
mice (either sex)
3
continuously backcrossed to C57BL/6/J since 1999 (25–32 g),
were used.
Behavioural assays. A cold-plate analgesia meter (IITC Life Science), main-
tained at 0 6 0.5 uC, was used to assess noxious cold sensitivity of the plantar
surface of the hind paws of the mice
6
. Nocifensive responses are given as means
from two separate countings of the number of paw lifts during 5 min (60-min
interval).
Single-fibre electrophysiology. We used the technique based on isolated skin-
saphenous-nerve preparation and single-fibre recording
14
. To apply solutions of
defined temperature, we designed a flow-through heater connected to a roller
pump. A stainless steel cannula was coiled with an insulated resistive wire
(10 V m
21
) over a length of 10 cm; the wire was connected to a computer-
controlled power supply. The cold stimulus (60 s) consisted of a dynamic phase
(30–12 uC) and a subsequent static phase (12–10 uC). The criterion for assigning
cold responsiveness to a fibre (CMC) was a discharge of at least three spikes and/
or sensitization to cooling after the application of 500 mM menthol. Menthol was
considered effective when the cold response at least doubled, when the unit
discharged on menthol application at 30 uC, or when a cold-insensitive unit
developed responsiveness to cold. For temperature control, TTX superfusion
and menthol superfusion, the receptive field was isolated with a metal ring
(volume 500 ml) and continuously perfused at a rate of 10 ml min
21
. Noxious
mechanical stimulation was performed with a gravity-driven von Frey type
stimulus of threefold the threshold strength. DAPSYS (http://www.dapsys.net)
was used to record and analyse the data.
Electrical stimulus strength–duration measurements. A high-impedance (9–
12-MV) needle electrode and a constant-current stimulus isolator (WPI) were
used to perform excitability testing at the most sensitive spot of the receptive field
(threshold current less than 50 mA at 1 ms), applying electrical current pulses of
variable strength and duration to the nerve ending. For each duration of current
pulse the threshold current to evoke an action potential was determined by using
the method of limits. The strength–duration relationship obtained for C-fibre
excitation was converted to threshold-stimulus charge transfer (Q 5 It) to use
the linear function of charge (Q) versus stimulus duration (t) for the determina-
tion of two key parameters of excitability: rheobase current and chronaxy
19
.
Rheobase is the minimal electric current of infinite duration that triggers an
action potential; chronaxy is the stimulus duration for which a current of double
the rheobase strength needs to be applied to evoke an action potential. The
regression of threshold charge on stimulus duration was used to estimate rheo-
base (gradient) and the x-axis intercept at zero charge was used to estimate
chronaxy. The effects of cold and TTX on C-fibre excitability are described in
terms of changes in chronaxy and rheobase. Spike2 (CED) was used to record
and analyse the data.
Cell culture. Adult mice were killed by CO
2
inhalation. DRGs from all spinal
levels were removed and cultured as described
30
. Recordings were made within
24 h in culture.
Complementary DNA and transient transfection. Rat Na
v
1.8 cDNA was tran-
siently transfected in the DRG neuroblastoma hybridoma cell line ND7/23.
Human Na
v
1.7 cDNA was transiently transfected in HEK 293T cells. ND7/23
and HEK 293T cells were maintained in DMEM (Gibco), supplemented with
10,000 U ml
21
penicillin/streptomycin (Gibco), 1 M HEPES (Gibco), 10% heat-
inactivated FBS (HyClone) and 0.3 M taurine (Sigma) at 37 uC and 5% CO
2
. All
cells were transfected by using the calcium phosphate precipitation method and
included a reporter plasmid (1 mg of CD8-pih3m). Transfected cells were used
for experiments within 3–4 days.
Whole-cell patch-clamp recordings. Recordings were performed at defined
temperatures
30
with an Axopatch 200B amplifier and pClamp 8.2 (Axon
Instruments). Patch-clamp pipettes from borosilicate capillary glass (TW150F-
3; WPI) were heat-polished to a resistance of 1–3 MV. Current-clamp recordings
were made from small and medium-sized DRG neurons with a resting mem-
brane potential of at least 240 mV. The extracellular solution contained (in
mM): NaCl 140, KCl 4, CaCl
2
2, MgCl
2
1, glucose 5, HEPES 10 (adjusted to
pH 7.4 with NaOH). The pipette (intracellular) solution contained (in mM): KCl
140, EGTA 0.5, HEPES 5, MgATP 3 (adjusted to pH 7.3 with NaOH). TTX was
applied to the extracellular solution at a concentration of 250 nM. Voltage-clamp
recordings were made on DRG neurons of WT mice (18.0 6 1.0 pF, n 5 39) and
Na
v
1.8
2/2
mice (16.9 6 1.0 pF, n 5 50; P 5 0.46, Student’s t-test). TTXr cur-
rents were studied in WT DRG neurons in the presence of 250 nM TTX; TTXs
currents were studied in Na
v
1.8
2/2
DRGs.
The extracellular solution for DRG recordings contained (in mM): NaCl 40,
choline chloride 100, KCl 3, MgCl
2
1, CaCl
2
1, HEPES 10 (pH 7.4). The intra-
cellular solution contained (in mM) CsF 140, EGTA 1, NaCl 10, HEPES 10
(pH 7.4). Cells with increase in passive leak current during cooling were
discarded. The extracellular solution for recordings from ND7/23 and HEK
293T cells contained (in mM): NaCl 65, choline chloride 85, CaCl
2
2, HEPES
10 (pH 7.4). The intracellular solution contained (in mM): NaF 100, NaCl 30,
EGTA 10, HEPES 10 (pH 7.4). Transfection-positive cells were identified with
immunobeads (anti-CD-8 Dynabeads; Dynal A.S.). Cells with an initial seal of
more than 1 GV and a leak current of less than 500 pA throughout the recording
were used for the analysis. Voltage errors were minimized by using 70–80%
series resistance compensation. The capacitance artefact was cancelled by
using the computer-controlled circuitry of the patch-clamp amplifier. Linear
leak subtraction was applied to all voltage-clamp recordings (using resistance
estimates from four hyperpolarizing pulses applied before the depolarizing
test potential). Membrane currents were filtered at 5 kHz and sampled at
20 kHz. Data were analysed with Clampex 8.2 and Origin 6.0 (Microcal
Software). To calculate the midpoint of activation, the peak conductance g
m
was estimated from the equation g
m
5 I
Na
/(E
m
2 E
rev
), where I
Na
is the peak
current, E
m
is the corresponding voltage and E
rev
is the estimated reversal
potential. The data were least-squares fitted by using the Boltzmann equation
g
m
/g
max
5 1/{1 1 exp[(E
0.5
2 E)/k
E
]}, where g
max
is the maximum conductance,
E
0.5
is the voltage at which g/g
max
5 0.5, and k
E
is the slope factor. To obtain the
midpoint of steady-state fast and slow inactivation, test-pulse-evoked peak cur-
rents were measured, normalized and plotted against the conditioning prepulse
potential. The data were least-squares fitted by using the Boltzmann equation
y 5 1/{1 1 exp[(E
pp
2 h
0.5
)/k
h
]}, where h
0.5
is the voltage at which y 5 0.5, and
k
h
is the slope factor.
Voltage-clamp protocols for DRGs. Cold-induced reduction was examined on
sodium currents activated by test pulses to 210 mV (TTXs) or 0 mV (TTXr)
(V
h
52120 mV or 80 mV). For steady-state fast inactivation, 50-ms prepulses
from 2140 mV (TTXs) or 2120 mV (TTXr) to 120 mV in steps of 10 mV were
used, followed by a 50-ms test pulse to 210 mV (TTXs) or 0 mV (TTXr)
(V
h
52120 mV or 80 mV). For steady-state slow inactivation, 30-s prepulses
from 2120 mV to 0 mV in steps of 10 mV were used, followed by a 100-ms pulse
to 2120 mV to remove fast inactivation, and then by a 50-ms test pulse to
210 mV (TTXs) or 0 mV (TTXr) (V
h
52120 mV).
Voltage-clamp protocols for ND7/23 and HEK 293T cells. Cold-induced
reduction was examined on sodium currents activated by test pulses to
150 mV (V
h
52120 mV or 80 mV). For steady-state fast inactivation, 50-ms
prepulses from 2140 mV (TTXs) or 2120 mV (TTXr) to 120 mV in steps of
10 mV were used, followed by a 50-ms test pulse to 150 mV (V
h
52120 mV).
For steady-state slow inactivation, 30-s prepulses from 2120 mV to 210 mV in
steps of 10 mV were used, followed by a 100-ms pulse to 2120 mV to remove
fast inactivation, and then by a 50-ms test pulse to 150 mV (TTXr)
(V
h
52120 mV).
doi:10.1038/nature05880
Nature
©2007
Publishing
Group