Neuroscience Letters 592 (2015) 12–16
Contents lists available at ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
Research article
Mice and rats differ with respect to activity-dependent slowing of conduction velocity in the saphenous peripheral nerve T. Hoffmann ∗ , R. De Col, K. Messlinger, P.W. Reeh, C. Weidner Institute for Physiology and Pathophysiology, University of Erlangen-Nuremberg, Universitaetsstrasse 17, D-91054 Erlangen, Germany
h i g h l i g h t s • We assessed in mice the a published criteria for C fibre differentiation into sub-classes. • For this, we developed an in vivo electrophysiological technique in the mouse. • The criteria for differentiation into fibres sub-classes is invalid in the mouse.
a r t i c l e
i n f o
Article history: Received 19 December 2014 Received in revised form 22 February 2015 Accepted 23 February 2015 Available online 27 February 2015 Keywords: Electrophysiology Activity induced slowing In vivo In vitro Nerve fibre subclasses
a b s t r a c t We assess in mice, the electrophysiological criteria developed in humans and rats in vivo for unmyelinated (C) fibre differentiation into sub-classes, derived from the activity-induced latency increase (“slowing”) in response to electrical stimulation during 6 min at 0.25 Hz followed by 3 min at 2 Hz. Fibres are considered nociceptors if they show more than 10% slowing at 2 Hz; nociceptors are further divided into mechanosensitive (“polymodal”) and mechanoinsensitive (“silent”) ones according to a latency shift of less and more than 1% during the first minute at 0.25 Hz, respectively. Sympathetic postganglionics are recognised by 2–10% slowing at 2 Hz; units slowing less than 2% at 2 Hz remain uncategorised. For assessment of these criteria, we also developed a novel in vivo technique for recording of peripheral single-fibres in the mouse. We compared the theoretical slowing-rate discriminator criteria with experimental data obtained from mice in vivo/in vitro and rats in vitro. Out of 69 cutaneous mouse C-fibres in vitro and 19 in vivo, only 38 (67%) and 9 (47%) met the above 1% criterion, respectively; sympathetics were not identified. In contrast, out of 20 rats nerve fibres in vitro, 19 (95%) met this criterion. We conclude that (A) our novel electrophysiological technique is a practical method for examining mouse cutaneous single-fibres in vivo and (B) the published criterion for identifying silent nociceptors in rats and humans is not applicable in mice. © 2015 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Neuronal electrophysiological recordings on single-fibre level are a useful tool in the research on various neuropathies. They allow refined examination of the changes in many receptive and conductive properties under pathological conditions and moreover, enable linking specific fibre types to these conditions. This
Abbreviations: AP, action potential; C-HTMs, C high threshold mechanosensitive fibre; C-LTMs, low threshold mechanosensitive C-fibre; CMi, mechanoinsensitive C-fibre; CH, heat-sensitive C-fibre; CMC, mechanosensitive cold C-fibre; CMH, mechanosensitive heat C-fibre; CMHC, mechanosensitive heat and cold C-fibre; CMs, mechanosensitive C-fibre; RF, receptive field. ∗ Corresponding author. Tel.: +49 9131 8526730; fax: +49 9131 8522497. E-mail address: [email protected] (T. Hoffmann). http://dx.doi.org/10.1016/j.neulet.2015.02.057 0304-3940/© 2015 Elsevier Ireland Ltd. All rights reserved.
has proven significant, with various pathologies being connected to particular subsets of fibres. For instance, mechanical hypersensitivity induced by partial nerve injury has been shown to be signaled by myelinated (A) fibres, whereas specific unmyelinated (C) fibres are responsible for concomitant heat sensitisation and pain [1,2]. The discrepancies between mechanically sensitive and insensitive (C) fibres seem to be of particular importance. Mechanically, insensitive (“silent”) fibres have previously been connected to chronic pruritus and erythromelalgia in human patients [3,4]. The classification into specific neuronal fibre types by electrophysiological recordings can be made using a combination of conductive and receptive criteria. Receptive characterization could be achieved through screening for heat, cold and mechanosensitivity and a conductive index includes velocity of action potential (AP) conduction and changes in conduction velocity in response to
T. Hoffmann et al. / Neuroscience Letters 592 (2015) 12–16
repetitive electrical stimulation. Such velocity changes are utilised in the “collision technique”, first described by Iggo in the cat [5], in which activity-induced slowing of conduction velocity during repetitive electrical stimulation is used to identify discrete fibre units. Since then, distinct activity-induced changes in conduction velocity have been established as a marker for specific fibre subsets in rats and humans [6–9]. A particularly elaborated index for fibre screening through activity-induced conduction velocity changes (termed “slowing” and “speeding”) has been established in vivo by Hugh Bostock and collaborators. A specific electrical stimulation protocol was shown to be discriminative between specific peripheral nerve fibre subclasses in both rats [10] and humans [11]. The protocol entails quiescence for 3 min, followed by 6 min of electrical stimulus at 0.25 Hz, 3 min at 2 Hz and again 6 min at 0.25 Hz. The first fibre differentiation into nociceptive, sympathetic postganglionic and uncategorised nerve fibres is based on activity-induced latency increase by more than 10%, 2–10%, or less than 2%, respectively, during the 3 min electrical stimulation at 2 Hz. A further sub-division of nociceptors is based on the conduction slowing during the first minute at 0.25 Hz following the initial rest period. Fibres slowing more than 1% (rat criterion, human criterion is 1.5%) during this period are classified as mechanoinsensitive, as opposed to mechanosensitive ones which slow less than 1%. We have tried to apply these published criteria to the mouse. For this, we have also developed a novel technique for in vivo electrophysiological recording on single-fibre level in the mouse, which is an adaptation of an in vitro recording method used for neuronal recordings from rat cranial meninges [12].
13
termination of a unit in the skin was recognised, conduction velocity was determined and a marking technique, using simultaneous electrical stimulation of the terminal and the main nerve trunk was applied, to ensure recording from a single unit. The cold, heat and electrical stimulation procedures that followed were identical to the ones used for mechanosensitive fibres. 2.3. In vivo electrophysiology Mice were initially anaesthetised in an isoflurane–oxygen atmosphere (2%) to produce deep aneasthesia with absence of limb withdrawal reflexes upon pinching with forceps. Mice were then placed on their backs, continuously supplied with an isoflurane–oxygen mixture through a custom-made perspex mask and their hind legs were fixed to a perspex block using doublesided adhesive tape. Body temperature was maintained at 37 ◦ C with a feedback-controlled homoeothermic system (TKM 0902; Föhr Medical Instruments, Frankfurt, Germany) containing a rectal thermosensor (pt100) of 1 mm diameter. A small incision of approximately 2.5 mm diameter was made on the medial side of the hind limb at knee level, the appearing saphenous nerve was freed from its surrounding tissue and cut. The isoflurane concentration
2. Methods 2.1. Animals C57BL/6 mice and Wistar rats of both sexes and ranging in weight between 20–27 g and 200–300 g, respectively, were housed in group cages in a temperature-controlled environment with a 12 h light–dark cycle and were supplied with food and water ad libitum. All animal experiments were carried out in compliance with the local Animal Protection Authority. 2.2. In vitro electrophysiology Animals were sacrificed by exposure to a rising CO2 atmosphere. Single-fibre recordings from cutaneous (C) fibres of the saphenous nerve were obtained using an isolated skin–nerve preparation essentially as described previously [13–15], using a CED Micro1401 and Spike2 software for digital storage and evaluation (Cambridge Electronics, UK). Mechanosensitive receptive fields were searched for using a blunt glass rod, and the mechanical threshold was assessed using calibrated von Frey bristles. Cold stimuli (flushing with ice-cold buffer) as well as radiant heat stimulation (20 s ramp of 32–48 ◦ C delivered by a feedback-controlled halogen lamp with an 8 mm focused beam) were applied. Heat threshold was defined as the temperature at which the second spike occurred. Following sensory characterisation, the receptive field (RF) was electrically stimulated using a metal electrode as cathode, and fibres were characterised using a specific electrical protocol consisting of 3 min quiescence, 6 min at 0.25 Hz followed by 3 min at 2 Hz and again 6 min at 0.25 Hz. The activity-induced latency shifts during 2 Hz and the first minute at 0.25 Hz were used for differentiation of the fibres into sub-classes as previously published [11,16]. Mechanoinsensitive C fibres were searched for using a hand-held metal electrode delivering supramaximal electrical pulses along the nerve arborization and into the skin, as described [14]. Once the
Fig. 1. (A) Depicts the electrical protocol used for classification of fibre subsets and the activity-induced slowing profile (dot plot) of one mechanosensitive (CM, filled spheres) and one mechanoinsensitive (CMi, empty spheres) examples. Fibres with a latency shift of more than 10% (hatched line) during the 2 Hz period should be nociceptors according to slowing criterion. (B) Illustration of the differentiating criterion between mechanically sensitive and insensitive fibres. The activity-induced latency shifts of the two units shown in 1(A) during the initial part of the electrical protocol are depicted. The 1% slowing criterion during the first minute (highlighted) of stimulation at 0.25 Hz is shown as a hatched line.
14
T. Hoffmann et al. / Neuroscience Letters 592 (2015) 12–16
was then reduced to 1% and mice were allowed to breathe spontaneously through the mask for the remainder of the experiment. The exposed subcutaneous tissues were kept moist throughout the experiment through occasional synthetic interstitial fluid (SIF, consisting in mM of: 107.8 NaCl, 26.2 NaHCO3, 9.64 Na-gluconate, 7.6 sucrose, 5.05 glucose, 3.48 KCl, 1.67 NaH2PO4, 1.53CaCl2 and 0.69 MgSO4) pipetting. A small fascicle of the distal nerve stump was pulled into a SIF-filled glass suction electrode with a tip diameter of approximately 20 m (see [12] for further description of the technique). Receptive fields of single nerve units were searched for and characterised using the same stimuli as in vitro applied to the dorsal (hairy) hindpaw skin. The mechanical thresholds were determined using a custom-made electromechanical stimulator [17]. This particular device allows for feedback-controlled punctate mechanical stimulation with discrete forces (0–300 mN) and for simultaneous electrical stimulation through the metallic probe tip. Threshold was determined as the force at which poking the skin with increasing force reproducibly evoked an AP in all or nothing manner (see Fig. 2). Finally, fibres were further characterized using the same discriminating electrical protocol as in in vitro experiments. 3. Results 3.1. Mice in vitro Evaluating in mice the criteria published for rats and humans for division of unmyelinated fibres into particular subclasses, we sampled in vitro a group of 69 C-fibres with receptive fields in the dorsal hindpaw skin. A total of 55 mechanically sensitive and 14 mechanically insensitive fibres were examined. Electrical search stimulus for mechanoinsensitive fibres was not employed as part of the standard protocol and the numbers reported here do not reflect actual incidence of mechanoinsensitive nerve units within the entire population of unmyelinated fibres. The mechanically responsive units were sub-classified as: 41 mechano-heat sensitive C-fibres (CMH), 10 high-threshold mechanosensitive C-fibres (C-HTM), 1 low-threshold mechanosensitive C-fibre (C-LTM), 2 mechanosensitive cold C-fibres (CMC) and 1 mechano heat and cold C-fibre (CMHC). The mechanically unresponsive fibres comprised
6 heat-sensitive C-fibres (CH) and 8 fibres which did not respond to either mechanical or heat stimulation (putative CMi units). The median mechanical threshold for the 55 cutaneuos mechanosensitive C-fibres was 8 mN (range 1–45.3 mN). Conduction velocity of the whole 69 unmyelinated fibres population was 0.51 ± 0.02 m/s. The average maximal activity-induced latency shift during 3 min at 2 Hz of the 69 cutaneous units was 50 ± 3%. The small variance of the individual latency shifts at 2 Hz electrostimulation did not allow to detect a significant difference between mechanosensitive and insensitive fibres (1-factorial ANOVA). According to the published discrimination criterion for rat and human, fibres slowing more than 10% at 2 Hz are expected to be nociceptors, fibres slowing 2–10% are considered sympathetic efferents and a classification of units slowing less than 2% is not possible [8,11,16]. Fig. 1A illustrates the slowing profiles of one mechanosensitive (and heat sensitive; CMH) and one mechanoinsensitive (CMi) units in response to the electrical stimulation protocol. The conduction latency of both fibres at 2 Hz decreased by more than 10%, and thus, they represent nociceptive units. We have re-categorized all our sampled units according to this criterion; the results are summarized in Table 1A. Out of 69 sampled units 60 slowed more than 10% during 2 Hz stimulation and were predicted nociceptors. Indeed, 47 of these units were mechanosensitive and 13 mechanoinsensitive fibres. The remaining 9 units slowed between 2 and 10% and should theoretically represent sympathetic efferent fibres. However, these 9 units were almost all nociceptors (1CMH, 4C-HTM, 1C-LTM, 2CMC, 1CH), and thus did not meet the sorting criterion. An important further sub-classification of the nociceptive fibres (i.e., those slowing more than 10% during 3 min at 2 Hz) is proposed according to the latency shift during the very first minute at 0.25 Hz. Units that slow more than 1% during this period should theoretically represent mechanoinsensitive fibres, whereas units slowing less than 1% are predicted to be mechanosensitive nociceptors (rat criterion; the human criterion is 1.5%). Fig. 1B focuses on the divergence of these early latency changes (same units as in Fig. 1A). The mechanosensitive fibre does not exceed the 1% threshold, whereas, the mechanoinsensitive fibre decelerates progressively, exceeding the threshold within less than 1 min. Out of the 55 mechanosensitive units we found 16 slowed more than
Fig. 2. (A) Description of the in vivo recording set up. Ref = reference electrode, Elect/Mech = custom made stimulator (see Section 2) able to alternate between electrical and mechanical stimulation. (B) Testing for mechanical sensitivity is illustrated by one example. Triangle-shaped stimuli were applied with increasing forces at 2 s interval (left panel titled “Mech. stim.”). Mechanically induced action potentials (highlighted in a box titled “Evoked spikes”) from this fibre appeared at a threshold of 18 mN. (C) Illustration of the marking technique for one mechanically sensitive fibre. Latency changes of electrically evoked action potentials are to be seen during simultaneous stimulations with suprathreshold mechanical pulses (highlighted by arrows).
T. Hoffmann et al. / Neuroscience Letters 592 (2015) 12–16 Table 1 Implementation of the fibre classification criteria in mice and rats. The tables describe the quantities of mechanosensitive (CM), insensitive (CH/CMi), sympathetic nerve units (Symp. postgl.) and group sum (n) in mice (A in vitro, B in vivo) and rats in vitro (C); the fibres were sorted according to criteria established for rats and humans and derived from their activity-dependent slowing of conduction velocity upon two different electrical stimulation protocols (upper 3 and lower 2 rows). Mismatch rates were determined as the number of units violating the criteria. A. CM
CMi/CH
Symp. postgl.
n= Slow > 10% Slow = 2–10% Slow < 2% Mismatch
55 47 8 0 8
14 13 1 0 1
0 0 0 0
Slow > 1% Slow < 1% Mismatch
16 31 16
9 4 4
CM
CMi/CH
Symp. postgl.
n= Slow > 10% Slow = 2–10% Slow < 2% Mismatch
13 11 2 0 2
6 5 1 0 1
0 0 0 0
Slow > 1% Slow < 1% Mismatch
4 8 3
1 4 4
CM
CMi/CH
Symp. postgl.
n= Slow > 10% Slow = 2–10% Slow < 2% Mismatch
11 11 0 0 0
9 9 0 0 0
0 0 0 0
Slow > 1% Slow < 1% Mismatch
1 10 1
9 0 0
n
69 60 9 0 9(=13%) 25 35 20(=33%)
B.
n
19 16 3 0 3(=15%) 4 12 7(=44%)
C.
n
20 20 0 0 0(=0%) 10 10 1(=5%)
1% during the first minute at 0.25 Hz and should theoretically be mechanoinsensitive, and 31 slowed less than 1% during this period and should theoretically be mechanosensitive (Table 1A). From the 14 mechanoinsensitive fibres, 9 slowed more than 1%, and thus, met the criterion, but 4 slowed less than 1% and violated the criterion. Out of 60 C-fibres meeting, the first criterion (>10% at 2 Hz) 20 (33%) violated the second criterion, 16 slowing more than 1% at 0.25 Hz despite being mechanosensitive and 4 slowing less than 1% despite being mechanoinsensitive. This considerable mismatch (n = 20, 33%) was the same whether the rat (1%) or human (1.5) criterion was applied. 3.2. Mice in vivo There are three possibilities to argue about this significant gap between the theoretical expectation and our empirical results. First, the differentiation criterion was established for rats and humans in vivo and does not apply in vitro. Second, the criterion is applicable to rats and humans but not to mice. A third option is that the criterion could not be verified due to technical differences. To decide between these possibilities, we developed a novel in vivo suction electrode technique for recording from single peripheral nerve fibres in anaesthetised mice (Fig. 2A). The preparation does not require tracheal intubation and is rapidly set up. The method enables full electrical as well as sensory characterisation of cutaneous single-fibres and sufficient signal-to-noise ratio for
15
several experimental hours per day (see Fig. 2B for raw records). A total of 19 cutaneous unmyelinated C-fibres were examined. In all fibres, all test protocols could be fully executed and no fibre was excluded from the analysis. Averaged conduction velocity of the 19 unmyelinated fibres was 0.65 ± 0.05 m/s. Out of 19 tested C-fibres, 6 were experimentally classified as mechanoinsensitive (CMi) and 13 as mechanosensitive (CMs), with an average mechanical threshold of 18 ± 4 mN. Fig. 2B presents increasing mechanical stimuli and responses of one mechanosensitive unit. Time-locked mechanically evoked action potentials occurred upon repetitive (0.5 Hz) stimulation by 18 mN. The marking technique [17] ensured the identical origin of electrically and mechanically evoked APs (Fig. 2C). The evaluation of the activity-induced slowing during 3 min at 2 Hz for all 19 fibres provided a mean latency shift of 44 ± 7%, comparable to that measured in the cutaneous fibres in vitro. As in in vitro, mechanosensitive and insensitive fibres could not be differentiated by their degree of slowing (n.s., 1-factorial ANOVA). The in vivo results are summarised in Table 1B. Out of 19 sampled units, 13 were experimentally classified as mechanosensitive and 6 as mechnoinsensitive. 11 units out of the 13 mechanosensitive fibres, and 5 units out of 6 mechanoinsensitive fibres slowed more than 10% during the 2 Hz stimulation period and were theoretically predicted to be nociceptors. Two units, out of the mechanosensitive fibres and 1 unit of the mechanoinsensitive fibres slowed between 2% and 10% and should thus theoretically represent sympathetic efferents. Out of the 11 mechanosensitive units, we found 4 slowed more than 1% during the first minute at 0.25 Hz and should theoretically be mechanoinsensitive, and 8 slowed less than 1% and should theoretically be (and were in fact) mechanosensitive. From the 5 mechanoinsensitive fibres, 1 slowed more than 1% and thus, met the criterion, but 4 slowed less than 1% and violated the criterion. Out of 16C-fibres meeting, the first criterion (>10% at 2 Hz) 7 (44%) violated the second criterion, 4 slowing more than 1% at 0.25 Hz despite being mechanosensitive and 4 slowing less than 1% despite being mechanoinsensitive. This considerable mismatch (n = 7, 44%) was the same whether the rat (1%) or human (1.5) criterion was applied. Mismatch rates in vivo corresponded to those in vitro, indicating that the invalidity of the classification criterion in our mouse experiments does not originate from methodological differences between in vitro and in vivo recordings.
3.3. Rats in vitro We next evaluated the possibility that the differentiation criterion is valid in humans and rats but not in mice. For this, we sampled 20 cutaneous single-fibres in rats in vitro. The averaged conduction velocity of these fibers was 0.43 ± 0.00 m/s. These fibres were sub-classified as: 5CMH, 2C-LTM, 4C-HTM, 5CH and 4 fibres which did neither respond to mechanical nor to heat stimuli (CMi). The median mechanical threshold for the 11 cutaneuos mechanosensitive fibres was 11.4 mN (range 2–90.5 mN). All 20 units slowed more than 10% during the 2 Hz stimulation period and were theoretically predicted to be nociceptors. The averaged activity-induced latency shift during 3 min at 2 Hz for the 20 cutaneous units was 47 ± 4%. Once more, mechanosensitive and insensitive fibres could not be differentiated by their degree of slowing (n.s., 1-factorial ANOVA). Except for one CMH unit which slowed more than 1% during the first minute at 0.25 Hz and should have been mechanoinsensitive according to the classification criterion, all fibres matched the predicted classifications (n = 1, 5% mismatch, Table 1C). The activity-induced slowing of this single mismatching unit during the first minute at 0.25 Hz was 1.2%, extremely close to the 1% rat boundary and within the 1.5% human limit.
16
T. Hoffmann et al. / Neuroscience Letters 592 (2015) 12–16
4. Discussion Nerve fibres could be divided into subclasses in a tree-like manner according to various sensory and conductive properties. Mechanical sensitivity in nociceptors is of particular importance for specific pathological states. In this respect, mechanical hyperalgesia such as appearing during inflammation, nerve injury, and osteoarthrosis is promoted by mechanical sensitization of “silent nociceptors” (CMi and CH), but not of “normal” (polymodal) nociceptors in the skin [18,19]. A simple electrophysiological discriminator such as the activity-induced slowing rate criteria available for rats and humans but not mice could support the challenging search for cutaneous CMi and CH fibres. We have failed to reproduce the electrophysiological classification criteria previously established for humans and rats in vivo in the mouse. The fact that the criteria of enhanced activity-induced slowing could not be applied to CMi/CH in the mouse both in vivo and in vitro but was confirmed in rats in vitro (except for one borderline outlier) points to a species specificity of the criteria rather than to methodological discrepancies between in vivo and in vitro recordings. This implies essential discrepancies in biophysical properties of the peripheral axonal membrane in mice vs. rats and humans. The voltage-gated sodium channels Nav 1.7, Nav 1.8 and Nav 1.9 as well as the hyperpolarization-activated cyclic nucleotide-gated (HCN) channels have previously been shown to influence AP initiation and propagation [20–24]. Differential neuronal expression of these channels in specific fibre types of mice in comparison to humans and rats could be responsible for the different activity-induced latency slowing rates in these species [25]. In our attempt to test the nerve fibre sub-classification by activity-induced slowing rates in mice, we evaluated mechanically insensitive C-fibres. We cannot not rule out the possibility that some of the mechanoinsensitive fibres we recorded from were postganglionic sympathetic nerve fibres or, alternatively, axons in passage which were cut during skin dissection in vitro (“cut ends”). This could have caused a systematic error of limited extent, inflating the mismatch of the more/less than 1% criterion. The mismatch was even greater in vivo where “cut ends” are not to be feared. In addition, most of the predicted sympathetic efferents (2–10%) were in fact mechanosensitive, or at least heat sensitive, and thus violated the preliminary criterion (slowing more/less than 10%). In conclusion, our electrophysiological technique is a valid method to record in vivo from nerve single-fibres in the mouse. Furthermore, the previously published criteria differentiating between nociceptive and sympathetic and between mechanically sensitive and insensitive peripheral nerve C-fibres are invalid in the mouse and not suitable to identify silent nociceptors. This should prove important when integrating mouse models in electrophysiological studies on peripheral nociception and primary hyperalgesia. Acknowledgments The authors would like to thank Birgit Vogler for excellent technical assistance. This article was funded by theDeutsche Forschungsgemeinschaft, grant RE 704/2-1 (AOBJ) 56874. References [1] J.N. Campbell, S.N. Raja, R.A. Meyer, S.E. Mackinnon, Myelinated afferents signal the hyperalgesia associated with nerve injury, Pain 32 (1988) 89–94.
[2] H.E. Torebjork, R.H. LaMotte, C.J. Robinson, Peripheral neural correlates of magnitude of cutaneous pain and hyperalgesia: simultaneous recordings in humans of sensory judgments of pain and evoked responses in nociceptors with C-fibers, J. Neurophysiol. 51 (1984) 325–339. [3] K. Orstavik, C. Weidner, R. Schmidt, M. Schmelz, M. Hilliges, E. Jorum, H. Handwerker, E. Torebjork, Pathological C-fibres in patients with a chronic painful condition, Brain 126 (2003) 567–578. [4] M. Schmelz, M. Hilliges, R. Schmidt, K. Orstavik, C. Vahlquist, C. Weidner, H.O. Handwerker, H.E. Torebjork, Active itch fibers in chronic pruritus, Neurology 61 (2003) 564–566. [5] A. Iggo, The electrophysiological identification of single nerve fibres, with particular reference to the slowest-conducting vagal afferent fibres in the cat, J. Physiol. 142 (1958) 110–126. [6] J.G. Thalhammer, S.A. Raymond, F.A. Popitz-Bergez, G.R. Strichartz, Modality-dependent modulation of conduction by impulse activity in functionally characterized single cutaneous afferents in the rat, Somatosens. Mot. Res. 11 (1994) 243–257. [7] M. Schmelz, C. Forster, R. Schmidt, M. Ringkamp, H.O. Handwerker, H.E. Torebjork, Delayed responses to electrical stimuli reflect C-fiber responsiveness in human microneurography, Exp. Brain Res. 104 (1995) 331–336. [8] M.D. Gee, B. Lynn, B. Cotsell, Activity-dependent slowing of conduction velocity provides a method for identifying different functional classes of C-fibre in the rat saphenous nerve, Neuroscience 73 (1996) 667–675. [9] H.E. Torebjork, R.G. Hallin, Identification of afferent C units in intact human skin nerves, Brain Res. 67 (1974) 387–403. [10] A. George, J. Serra, X. Navarro, H. Bostock, Velocity recovery cycles of single C fibres innervating rat skin, J. Physiol. 578 (2007) 213–232. [11] H. Bostock, M. Campero, J. Serra, J. Ochoa, Velocity recovery cycles of C fibres innervating human skin, J. Physiol. 553 (2003) 649–663. [12] R. De Col, K. Messlinger, R.W. Carr, Conduction velocity is regulated by sodium channel inactivation in unmyelinated axons innervating the rat cranial meninges, J. Physiol. 586 (2008) 1089–1103. [13] P.W. Reeh, Sensory receptors in mammalian skin in an in vitro preparation, Neurosci. Lett. 66 (1986) 141–146. [14] M. Kress, M. Koltzenburg, P.W. Reeh, H.O. Handwerker, Responsiveness and functional attributes of electrically localized terminals of cutaneous C-fibers in vivo and in vitro, J. Neurophysiol. 68 (1992) 581–595. [15] K. Zimmermann, A. Hein, U. Hager, J.S. Kaczmarek, B.P. Turnquist, D.E. Clapham, P.W. Reeh, Phenotyping sensory nerve endings in vitro in the mouse, Nat. Protoc. 4 (2009) 174–196. [16] A. George, J. Serra, X. Navarro, H. Bostock, Velocity recovery cycles of single C fibres innervating rat skin, J. Physiol. 578 (2007) 213–232. [17] C.R. De, K. Messlinger, R.W. Carr, Repetitive activity slows axonal conduction velocity and concomitantly increases mechanical activation threshold in single axons of the rat cranial dura, J. Physiol. 590 (2012) 725–736. [18] T. Schlegel, S.K. Sauer, H.O. Handwerker, P.W. Reeh, Responsiveness of C-fiber nociceptors to punctate force-controlled stimuli in isolated rat skin: lack of modulation by inflammatory mediators and flurbiprofen, Neurosci. Lett. 361 (2004) 163–167. [19] M. Schmelz, R. Schmidt, M. Ringkamp, H.O. Handwerker, H.E. Torebjork, Sensitization of insensitive branches of C nociceptors in human skin, J. Physiol. 480 (Pt 2) (1994) 389–394. [20] M. Renganathan, T.R. Cummins, S.G. Waxman, Contribution of Na(v) 1.8 sodium channels to action potential electrogenesis in DRG neurons, J. Neurophysiol. 86 (2001) 629–640. [21] S. Tate, S. Benn, C. Hick, D. Trezise, V. John, R.J. Mannion, M. Costigan, C. Plumpton, D. Grose, Z. Gladwell, G. Kendall, K. Dale, C. Bountra, C.J. Woolf, Two sodium channels contribute to the TTX-R sodium current in primary sensory neurons, Nat. Neurosci. 1 (1998) 653–655. [22] E.C. Emery, G.T. Young, P.A. McNaughton, HCN2 ion channels: an emerging role as the pacemakers of pain, Trends Pharmacol. Sci. 33 (2012) 456–463. [23] S. Schnorr, M., Eberhardt, K., Kistner, H., Rajab, J., Kasser, A., Hess, P., Reeh, A., Ludwig, S. Herrmann, HCN2 channels account for mechanical (but not heat) hyperalgesia during long-standing inflammation, Pain (2014). [24] E. Leipold, L. Liebmann, G.C. Korenke, T. Heinrich, S. Giesselmann, J. Baets, M. Ebbinghaus, R.O. Goral, T. Stodberg, J.C. Hennings, M. Bergmann, J. Altmuller, H. Thiele, A. Wetzel, P. Nurnberg, V. Timmerman, J.P. De, R. Blum, H.G. Schaible, J. Weis, S.H. Heinemann, C.A. Hubner, I. Kurth, A de novo gain-of-function mutation in SCN11A causes loss of pain perception, Nat. Genet. 45 (2013) 1399–1404. [25] M.E. Petersson, O. Obreja, A. Lampert, R.W. Carr, M. Schmelz, E. Fransen, Differential axonal conduction patterns of mechano-sensitive and mechano-insensitive nociceptors–a combined experimental and modelling study, PLoS One 9 (2014) e103556.
Contents lists available at ScienceDirect
Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet
Research article
Mice and rats differ with respect to activity-dependent slowing of conduction velocity in the saphenous peripheral nerve T. Hoffmann ∗ , R. De Col, K. Messlinger, P.W. Reeh, C. Weidner Institute for Physiology and Pathophysiology, University of Erlangen-Nuremberg, Universitaetsstrasse 17, D-91054 Erlangen, Germany
h i g h l i g h t s • We assessed in mice the a published criteria for C fibre differentiation into sub-classes. • For this, we developed an in vivo electrophysiological technique in the mouse. • The criteria for differentiation into fibres sub-classes is invalid in the mouse.
a r t i c l e
i n f o
Article history: Received 19 December 2014 Received in revised form 22 February 2015 Accepted 23 February 2015 Available online 27 February 2015 Keywords: Electrophysiology Activity induced slowing In vivo In vitro Nerve fibre subclasses
a b s t r a c t We assess in mice, the electrophysiological criteria developed in humans and rats in vivo for unmyelinated (C) fibre differentiation into sub-classes, derived from the activity-induced latency increase (“slowing”) in response to electrical stimulation during 6 min at 0.25 Hz followed by 3 min at 2 Hz. Fibres are considered nociceptors if they show more than 10% slowing at 2 Hz; nociceptors are further divided into mechanosensitive (“polymodal”) and mechanoinsensitive (“silent”) ones according to a latency shift of less and more than 1% during the first minute at 0.25 Hz, respectively. Sympathetic postganglionics are recognised by 2–10% slowing at 2 Hz; units slowing less than 2% at 2 Hz remain uncategorised. For assessment of these criteria, we also developed a novel in vivo technique for recording of peripheral single-fibres in the mouse. We compared the theoretical slowing-rate discriminator criteria with experimental data obtained from mice in vivo/in vitro and rats in vitro. Out of 69 cutaneous mouse C-fibres in vitro and 19 in vivo, only 38 (67%) and 9 (47%) met the above 1% criterion, respectively; sympathetics were not identified. In contrast, out of 20 rats nerve fibres in vitro, 19 (95%) met this criterion. We conclude that (A) our novel electrophysiological technique is a practical method for examining mouse cutaneous single-fibres in vivo and (B) the published criterion for identifying silent nociceptors in rats and humans is not applicable in mice. © 2015 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Neuronal electrophysiological recordings on single-fibre level are a useful tool in the research on various neuropathies. They allow refined examination of the changes in many receptive and conductive properties under pathological conditions and moreover, enable linking specific fibre types to these conditions. This
Abbreviations: AP, action potential; C-HTMs, C high threshold mechanosensitive fibre; C-LTMs, low threshold mechanosensitive C-fibre; CMi, mechanoinsensitive C-fibre; CH, heat-sensitive C-fibre; CMC, mechanosensitive cold C-fibre; CMH, mechanosensitive heat C-fibre; CMHC, mechanosensitive heat and cold C-fibre; CMs, mechanosensitive C-fibre; RF, receptive field. ∗ Corresponding author. Tel.: +49 9131 8526730; fax: +49 9131 8522497. E-mail address: [email protected] (T. Hoffmann). http://dx.doi.org/10.1016/j.neulet.2015.02.057 0304-3940/© 2015 Elsevier Ireland Ltd. All rights reserved.
has proven significant, with various pathologies being connected to particular subsets of fibres. For instance, mechanical hypersensitivity induced by partial nerve injury has been shown to be signaled by myelinated (A) fibres, whereas specific unmyelinated (C) fibres are responsible for concomitant heat sensitisation and pain [1,2]. The discrepancies between mechanically sensitive and insensitive (C) fibres seem to be of particular importance. Mechanically, insensitive (“silent”) fibres have previously been connected to chronic pruritus and erythromelalgia in human patients [3,4]. The classification into specific neuronal fibre types by electrophysiological recordings can be made using a combination of conductive and receptive criteria. Receptive characterization could be achieved through screening for heat, cold and mechanosensitivity and a conductive index includes velocity of action potential (AP) conduction and changes in conduction velocity in response to
T. Hoffmann et al. / Neuroscience Letters 592 (2015) 12–16
repetitive electrical stimulation. Such velocity changes are utilised in the “collision technique”, first described by Iggo in the cat [5], in which activity-induced slowing of conduction velocity during repetitive electrical stimulation is used to identify discrete fibre units. Since then, distinct activity-induced changes in conduction velocity have been established as a marker for specific fibre subsets in rats and humans [6–9]. A particularly elaborated index for fibre screening through activity-induced conduction velocity changes (termed “slowing” and “speeding”) has been established in vivo by Hugh Bostock and collaborators. A specific electrical stimulation protocol was shown to be discriminative between specific peripheral nerve fibre subclasses in both rats [10] and humans [11]. The protocol entails quiescence for 3 min, followed by 6 min of electrical stimulus at 0.25 Hz, 3 min at 2 Hz and again 6 min at 0.25 Hz. The first fibre differentiation into nociceptive, sympathetic postganglionic and uncategorised nerve fibres is based on activity-induced latency increase by more than 10%, 2–10%, or less than 2%, respectively, during the 3 min electrical stimulation at 2 Hz. A further sub-division of nociceptors is based on the conduction slowing during the first minute at 0.25 Hz following the initial rest period. Fibres slowing more than 1% (rat criterion, human criterion is 1.5%) during this period are classified as mechanoinsensitive, as opposed to mechanosensitive ones which slow less than 1%. We have tried to apply these published criteria to the mouse. For this, we have also developed a novel technique for in vivo electrophysiological recording on single-fibre level in the mouse, which is an adaptation of an in vitro recording method used for neuronal recordings from rat cranial meninges [12].
13
termination of a unit in the skin was recognised, conduction velocity was determined and a marking technique, using simultaneous electrical stimulation of the terminal and the main nerve trunk was applied, to ensure recording from a single unit. The cold, heat and electrical stimulation procedures that followed were identical to the ones used for mechanosensitive fibres. 2.3. In vivo electrophysiology Mice were initially anaesthetised in an isoflurane–oxygen atmosphere (2%) to produce deep aneasthesia with absence of limb withdrawal reflexes upon pinching with forceps. Mice were then placed on their backs, continuously supplied with an isoflurane–oxygen mixture through a custom-made perspex mask and their hind legs were fixed to a perspex block using doublesided adhesive tape. Body temperature was maintained at 37 ◦ C with a feedback-controlled homoeothermic system (TKM 0902; Föhr Medical Instruments, Frankfurt, Germany) containing a rectal thermosensor (pt100) of 1 mm diameter. A small incision of approximately 2.5 mm diameter was made on the medial side of the hind limb at knee level, the appearing saphenous nerve was freed from its surrounding tissue and cut. The isoflurane concentration
2. Methods 2.1. Animals C57BL/6 mice and Wistar rats of both sexes and ranging in weight between 20–27 g and 200–300 g, respectively, were housed in group cages in a temperature-controlled environment with a 12 h light–dark cycle and were supplied with food and water ad libitum. All animal experiments were carried out in compliance with the local Animal Protection Authority. 2.2. In vitro electrophysiology Animals were sacrificed by exposure to a rising CO2 atmosphere. Single-fibre recordings from cutaneous (C) fibres of the saphenous nerve were obtained using an isolated skin–nerve preparation essentially as described previously [13–15], using a CED Micro1401 and Spike2 software for digital storage and evaluation (Cambridge Electronics, UK). Mechanosensitive receptive fields were searched for using a blunt glass rod, and the mechanical threshold was assessed using calibrated von Frey bristles. Cold stimuli (flushing with ice-cold buffer) as well as radiant heat stimulation (20 s ramp of 32–48 ◦ C delivered by a feedback-controlled halogen lamp with an 8 mm focused beam) were applied. Heat threshold was defined as the temperature at which the second spike occurred. Following sensory characterisation, the receptive field (RF) was electrically stimulated using a metal electrode as cathode, and fibres were characterised using a specific electrical protocol consisting of 3 min quiescence, 6 min at 0.25 Hz followed by 3 min at 2 Hz and again 6 min at 0.25 Hz. The activity-induced latency shifts during 2 Hz and the first minute at 0.25 Hz were used for differentiation of the fibres into sub-classes as previously published [11,16]. Mechanoinsensitive C fibres were searched for using a hand-held metal electrode delivering supramaximal electrical pulses along the nerve arborization and into the skin, as described [14]. Once the
Fig. 1. (A) Depicts the electrical protocol used for classification of fibre subsets and the activity-induced slowing profile (dot plot) of one mechanosensitive (CM, filled spheres) and one mechanoinsensitive (CMi, empty spheres) examples. Fibres with a latency shift of more than 10% (hatched line) during the 2 Hz period should be nociceptors according to slowing criterion. (B) Illustration of the differentiating criterion between mechanically sensitive and insensitive fibres. The activity-induced latency shifts of the two units shown in 1(A) during the initial part of the electrical protocol are depicted. The 1% slowing criterion during the first minute (highlighted) of stimulation at 0.25 Hz is shown as a hatched line.
14
T. Hoffmann et al. / Neuroscience Letters 592 (2015) 12–16
was then reduced to 1% and mice were allowed to breathe spontaneously through the mask for the remainder of the experiment. The exposed subcutaneous tissues were kept moist throughout the experiment through occasional synthetic interstitial fluid (SIF, consisting in mM of: 107.8 NaCl, 26.2 NaHCO3, 9.64 Na-gluconate, 7.6 sucrose, 5.05 glucose, 3.48 KCl, 1.67 NaH2PO4, 1.53CaCl2 and 0.69 MgSO4) pipetting. A small fascicle of the distal nerve stump was pulled into a SIF-filled glass suction electrode with a tip diameter of approximately 20 m (see [12] for further description of the technique). Receptive fields of single nerve units were searched for and characterised using the same stimuli as in vitro applied to the dorsal (hairy) hindpaw skin. The mechanical thresholds were determined using a custom-made electromechanical stimulator [17]. This particular device allows for feedback-controlled punctate mechanical stimulation with discrete forces (0–300 mN) and for simultaneous electrical stimulation through the metallic probe tip. Threshold was determined as the force at which poking the skin with increasing force reproducibly evoked an AP in all or nothing manner (see Fig. 2). Finally, fibres were further characterized using the same discriminating electrical protocol as in in vitro experiments. 3. Results 3.1. Mice in vitro Evaluating in mice the criteria published for rats and humans for division of unmyelinated fibres into particular subclasses, we sampled in vitro a group of 69 C-fibres with receptive fields in the dorsal hindpaw skin. A total of 55 mechanically sensitive and 14 mechanically insensitive fibres were examined. Electrical search stimulus for mechanoinsensitive fibres was not employed as part of the standard protocol and the numbers reported here do not reflect actual incidence of mechanoinsensitive nerve units within the entire population of unmyelinated fibres. The mechanically responsive units were sub-classified as: 41 mechano-heat sensitive C-fibres (CMH), 10 high-threshold mechanosensitive C-fibres (C-HTM), 1 low-threshold mechanosensitive C-fibre (C-LTM), 2 mechanosensitive cold C-fibres (CMC) and 1 mechano heat and cold C-fibre (CMHC). The mechanically unresponsive fibres comprised
6 heat-sensitive C-fibres (CH) and 8 fibres which did not respond to either mechanical or heat stimulation (putative CMi units). The median mechanical threshold for the 55 cutaneuos mechanosensitive C-fibres was 8 mN (range 1–45.3 mN). Conduction velocity of the whole 69 unmyelinated fibres population was 0.51 ± 0.02 m/s. The average maximal activity-induced latency shift during 3 min at 2 Hz of the 69 cutaneous units was 50 ± 3%. The small variance of the individual latency shifts at 2 Hz electrostimulation did not allow to detect a significant difference between mechanosensitive and insensitive fibres (1-factorial ANOVA). According to the published discrimination criterion for rat and human, fibres slowing more than 10% at 2 Hz are expected to be nociceptors, fibres slowing 2–10% are considered sympathetic efferents and a classification of units slowing less than 2% is not possible [8,11,16]. Fig. 1A illustrates the slowing profiles of one mechanosensitive (and heat sensitive; CMH) and one mechanoinsensitive (CMi) units in response to the electrical stimulation protocol. The conduction latency of both fibres at 2 Hz decreased by more than 10%, and thus, they represent nociceptive units. We have re-categorized all our sampled units according to this criterion; the results are summarized in Table 1A. Out of 69 sampled units 60 slowed more than 10% during 2 Hz stimulation and were predicted nociceptors. Indeed, 47 of these units were mechanosensitive and 13 mechanoinsensitive fibres. The remaining 9 units slowed between 2 and 10% and should theoretically represent sympathetic efferent fibres. However, these 9 units were almost all nociceptors (1CMH, 4C-HTM, 1C-LTM, 2CMC, 1CH), and thus did not meet the sorting criterion. An important further sub-classification of the nociceptive fibres (i.e., those slowing more than 10% during 3 min at 2 Hz) is proposed according to the latency shift during the very first minute at 0.25 Hz. Units that slow more than 1% during this period should theoretically represent mechanoinsensitive fibres, whereas units slowing less than 1% are predicted to be mechanosensitive nociceptors (rat criterion; the human criterion is 1.5%). Fig. 1B focuses on the divergence of these early latency changes (same units as in Fig. 1A). The mechanosensitive fibre does not exceed the 1% threshold, whereas, the mechanoinsensitive fibre decelerates progressively, exceeding the threshold within less than 1 min. Out of the 55 mechanosensitive units we found 16 slowed more than
Fig. 2. (A) Description of the in vivo recording set up. Ref = reference electrode, Elect/Mech = custom made stimulator (see Section 2) able to alternate between electrical and mechanical stimulation. (B) Testing for mechanical sensitivity is illustrated by one example. Triangle-shaped stimuli were applied with increasing forces at 2 s interval (left panel titled “Mech. stim.”). Mechanically induced action potentials (highlighted in a box titled “Evoked spikes”) from this fibre appeared at a threshold of 18 mN. (C) Illustration of the marking technique for one mechanically sensitive fibre. Latency changes of electrically evoked action potentials are to be seen during simultaneous stimulations with suprathreshold mechanical pulses (highlighted by arrows).
T. Hoffmann et al. / Neuroscience Letters 592 (2015) 12–16 Table 1 Implementation of the fibre classification criteria in mice and rats. The tables describe the quantities of mechanosensitive (CM), insensitive (CH/CMi), sympathetic nerve units (Symp. postgl.) and group sum (n) in mice (A in vitro, B in vivo) and rats in vitro (C); the fibres were sorted according to criteria established for rats and humans and derived from their activity-dependent slowing of conduction velocity upon two different electrical stimulation protocols (upper 3 and lower 2 rows). Mismatch rates were determined as the number of units violating the criteria. A. CM
CMi/CH
Symp. postgl.
n= Slow > 10% Slow = 2–10% Slow < 2% Mismatch
55 47 8 0 8
14 13 1 0 1
0 0 0 0
Slow > 1% Slow < 1% Mismatch
16 31 16
9 4 4
CM
CMi/CH
Symp. postgl.
n= Slow > 10% Slow = 2–10% Slow < 2% Mismatch
13 11 2 0 2
6 5 1 0 1
0 0 0 0
Slow > 1% Slow < 1% Mismatch
4 8 3
1 4 4
CM
CMi/CH
Symp. postgl.
n= Slow > 10% Slow = 2–10% Slow < 2% Mismatch
11 11 0 0 0
9 9 0 0 0
0 0 0 0
Slow > 1% Slow < 1% Mismatch
1 10 1
9 0 0
n
69 60 9 0 9(=13%) 25 35 20(=33%)
B.
n
19 16 3 0 3(=15%) 4 12 7(=44%)
C.
n
20 20 0 0 0(=0%) 10 10 1(=5%)
1% during the first minute at 0.25 Hz and should theoretically be mechanoinsensitive, and 31 slowed less than 1% during this period and should theoretically be mechanosensitive (Table 1A). From the 14 mechanoinsensitive fibres, 9 slowed more than 1%, and thus, met the criterion, but 4 slowed less than 1% and violated the criterion. Out of 60 C-fibres meeting, the first criterion (>10% at 2 Hz) 20 (33%) violated the second criterion, 16 slowing more than 1% at 0.25 Hz despite being mechanosensitive and 4 slowing less than 1% despite being mechanoinsensitive. This considerable mismatch (n = 20, 33%) was the same whether the rat (1%) or human (1.5) criterion was applied. 3.2. Mice in vivo There are three possibilities to argue about this significant gap between the theoretical expectation and our empirical results. First, the differentiation criterion was established for rats and humans in vivo and does not apply in vitro. Second, the criterion is applicable to rats and humans but not to mice. A third option is that the criterion could not be verified due to technical differences. To decide between these possibilities, we developed a novel in vivo suction electrode technique for recording from single peripheral nerve fibres in anaesthetised mice (Fig. 2A). The preparation does not require tracheal intubation and is rapidly set up. The method enables full electrical as well as sensory characterisation of cutaneous single-fibres and sufficient signal-to-noise ratio for
15
several experimental hours per day (see Fig. 2B for raw records). A total of 19 cutaneous unmyelinated C-fibres were examined. In all fibres, all test protocols could be fully executed and no fibre was excluded from the analysis. Averaged conduction velocity of the 19 unmyelinated fibres was 0.65 ± 0.05 m/s. Out of 19 tested C-fibres, 6 were experimentally classified as mechanoinsensitive (CMi) and 13 as mechanosensitive (CMs), with an average mechanical threshold of 18 ± 4 mN. Fig. 2B presents increasing mechanical stimuli and responses of one mechanosensitive unit. Time-locked mechanically evoked action potentials occurred upon repetitive (0.5 Hz) stimulation by 18 mN. The marking technique [17] ensured the identical origin of electrically and mechanically evoked APs (Fig. 2C). The evaluation of the activity-induced slowing during 3 min at 2 Hz for all 19 fibres provided a mean latency shift of 44 ± 7%, comparable to that measured in the cutaneous fibres in vitro. As in in vitro, mechanosensitive and insensitive fibres could not be differentiated by their degree of slowing (n.s., 1-factorial ANOVA). The in vivo results are summarised in Table 1B. Out of 19 sampled units, 13 were experimentally classified as mechanosensitive and 6 as mechnoinsensitive. 11 units out of the 13 mechanosensitive fibres, and 5 units out of 6 mechanoinsensitive fibres slowed more than 10% during the 2 Hz stimulation period and were theoretically predicted to be nociceptors. Two units, out of the mechanosensitive fibres and 1 unit of the mechanoinsensitive fibres slowed between 2% and 10% and should thus theoretically represent sympathetic efferents. Out of the 11 mechanosensitive units, we found 4 slowed more than 1% during the first minute at 0.25 Hz and should theoretically be mechanoinsensitive, and 8 slowed less than 1% and should theoretically be (and were in fact) mechanosensitive. From the 5 mechanoinsensitive fibres, 1 slowed more than 1% and thus, met the criterion, but 4 slowed less than 1% and violated the criterion. Out of 16C-fibres meeting, the first criterion (>10% at 2 Hz) 7 (44%) violated the second criterion, 4 slowing more than 1% at 0.25 Hz despite being mechanosensitive and 4 slowing less than 1% despite being mechanoinsensitive. This considerable mismatch (n = 7, 44%) was the same whether the rat (1%) or human (1.5) criterion was applied. Mismatch rates in vivo corresponded to those in vitro, indicating that the invalidity of the classification criterion in our mouse experiments does not originate from methodological differences between in vitro and in vivo recordings.
3.3. Rats in vitro We next evaluated the possibility that the differentiation criterion is valid in humans and rats but not in mice. For this, we sampled 20 cutaneous single-fibres in rats in vitro. The averaged conduction velocity of these fibers was 0.43 ± 0.00 m/s. These fibres were sub-classified as: 5CMH, 2C-LTM, 4C-HTM, 5CH and 4 fibres which did neither respond to mechanical nor to heat stimuli (CMi). The median mechanical threshold for the 11 cutaneuos mechanosensitive fibres was 11.4 mN (range 2–90.5 mN). All 20 units slowed more than 10% during the 2 Hz stimulation period and were theoretically predicted to be nociceptors. The averaged activity-induced latency shift during 3 min at 2 Hz for the 20 cutaneous units was 47 ± 4%. Once more, mechanosensitive and insensitive fibres could not be differentiated by their degree of slowing (n.s., 1-factorial ANOVA). Except for one CMH unit which slowed more than 1% during the first minute at 0.25 Hz and should have been mechanoinsensitive according to the classification criterion, all fibres matched the predicted classifications (n = 1, 5% mismatch, Table 1C). The activity-induced slowing of this single mismatching unit during the first minute at 0.25 Hz was 1.2%, extremely close to the 1% rat boundary and within the 1.5% human limit.
16
T. Hoffmann et al. / Neuroscience Letters 592 (2015) 12–16
4. Discussion Nerve fibres could be divided into subclasses in a tree-like manner according to various sensory and conductive properties. Mechanical sensitivity in nociceptors is of particular importance for specific pathological states. In this respect, mechanical hyperalgesia such as appearing during inflammation, nerve injury, and osteoarthrosis is promoted by mechanical sensitization of “silent nociceptors” (CMi and CH), but not of “normal” (polymodal) nociceptors in the skin [18,19]. A simple electrophysiological discriminator such as the activity-induced slowing rate criteria available for rats and humans but not mice could support the challenging search for cutaneous CMi and CH fibres. We have failed to reproduce the electrophysiological classification criteria previously established for humans and rats in vivo in the mouse. The fact that the criteria of enhanced activity-induced slowing could not be applied to CMi/CH in the mouse both in vivo and in vitro but was confirmed in rats in vitro (except for one borderline outlier) points to a species specificity of the criteria rather than to methodological discrepancies between in vivo and in vitro recordings. This implies essential discrepancies in biophysical properties of the peripheral axonal membrane in mice vs. rats and humans. The voltage-gated sodium channels Nav 1.7, Nav 1.8 and Nav 1.9 as well as the hyperpolarization-activated cyclic nucleotide-gated (HCN) channels have previously been shown to influence AP initiation and propagation [20–24]. Differential neuronal expression of these channels in specific fibre types of mice in comparison to humans and rats could be responsible for the different activity-induced latency slowing rates in these species [25]. In our attempt to test the nerve fibre sub-classification by activity-induced slowing rates in mice, we evaluated mechanically insensitive C-fibres. We cannot not rule out the possibility that some of the mechanoinsensitive fibres we recorded from were postganglionic sympathetic nerve fibres or, alternatively, axons in passage which were cut during skin dissection in vitro (“cut ends”). This could have caused a systematic error of limited extent, inflating the mismatch of the more/less than 1% criterion. The mismatch was even greater in vivo where “cut ends” are not to be feared. In addition, most of the predicted sympathetic efferents (2–10%) were in fact mechanosensitive, or at least heat sensitive, and thus violated the preliminary criterion (slowing more/less than 10%). In conclusion, our electrophysiological technique is a valid method to record in vivo from nerve single-fibres in the mouse. Furthermore, the previously published criteria differentiating between nociceptive and sympathetic and between mechanically sensitive and insensitive peripheral nerve C-fibres are invalid in the mouse and not suitable to identify silent nociceptors. This should prove important when integrating mouse models in electrophysiological studies on peripheral nociception and primary hyperalgesia. Acknowledgments The authors would like to thank Birgit Vogler for excellent technical assistance. This article was funded by theDeutsche Forschungsgemeinschaft, grant RE 704/2-1 (AOBJ) 56874. References [1] J.N. Campbell, S.N. Raja, R.A. Meyer, S.E. Mackinnon, Myelinated afferents signal the hyperalgesia associated with nerve injury, Pain 32 (1988) 89–94.
[2] H.E. Torebjork, R.H. LaMotte, C.J. Robinson, Peripheral neural correlates of magnitude of cutaneous pain and hyperalgesia: simultaneous recordings in humans of sensory judgments of pain and evoked responses in nociceptors with C-fibers, J. Neurophysiol. 51 (1984) 325–339. [3] K. Orstavik, C. Weidner, R. Schmidt, M. Schmelz, M. Hilliges, E. Jorum, H. Handwerker, E. Torebjork, Pathological C-fibres in patients with a chronic painful condition, Brain 126 (2003) 567–578. [4] M. Schmelz, M. Hilliges, R. Schmidt, K. Orstavik, C. Vahlquist, C. Weidner, H.O. Handwerker, H.E. Torebjork, Active itch fibers in chronic pruritus, Neurology 61 (2003) 564–566. [5] A. Iggo, The electrophysiological identification of single nerve fibres, with particular reference to the slowest-conducting vagal afferent fibres in the cat, J. Physiol. 142 (1958) 110–126. [6] J.G. Thalhammer, S.A. Raymond, F.A. Popitz-Bergez, G.R. Strichartz, Modality-dependent modulation of conduction by impulse activity in functionally characterized single cutaneous afferents in the rat, Somatosens. Mot. Res. 11 (1994) 243–257. [7] M. Schmelz, C. Forster, R. Schmidt, M. Ringkamp, H.O. Handwerker, H.E. Torebjork, Delayed responses to electrical stimuli reflect C-fiber responsiveness in human microneurography, Exp. Brain Res. 104 (1995) 331–336. [8] M.D. Gee, B. Lynn, B. Cotsell, Activity-dependent slowing of conduction velocity provides a method for identifying different functional classes of C-fibre in the rat saphenous nerve, Neuroscience 73 (1996) 667–675. [9] H.E. Torebjork, R.G. Hallin, Identification of afferent C units in intact human skin nerves, Brain Res. 67 (1974) 387–403. [10] A. George, J. Serra, X. Navarro, H. Bostock, Velocity recovery cycles of single C fibres innervating rat skin, J. Physiol. 578 (2007) 213–232. [11] H. Bostock, M. Campero, J. Serra, J. Ochoa, Velocity recovery cycles of C fibres innervating human skin, J. Physiol. 553 (2003) 649–663. [12] R. De Col, K. Messlinger, R.W. Carr, Conduction velocity is regulated by sodium channel inactivation in unmyelinated axons innervating the rat cranial meninges, J. Physiol. 586 (2008) 1089–1103. [13] P.W. Reeh, Sensory receptors in mammalian skin in an in vitro preparation, Neurosci. Lett. 66 (1986) 141–146. [14] M. Kress, M. Koltzenburg, P.W. Reeh, H.O. Handwerker, Responsiveness and functional attributes of electrically localized terminals of cutaneous C-fibers in vivo and in vitro, J. Neurophysiol. 68 (1992) 581–595. [15] K. Zimmermann, A. Hein, U. Hager, J.S. Kaczmarek, B.P. Turnquist, D.E. Clapham, P.W. Reeh, Phenotyping sensory nerve endings in vitro in the mouse, Nat. Protoc. 4 (2009) 174–196. [16] A. George, J. Serra, X. Navarro, H. Bostock, Velocity recovery cycles of single C fibres innervating rat skin, J. Physiol. 578 (2007) 213–232. [17] C.R. De, K. Messlinger, R.W. Carr, Repetitive activity slows axonal conduction velocity and concomitantly increases mechanical activation threshold in single axons of the rat cranial dura, J. Physiol. 590 (2012) 725–736. [18] T. Schlegel, S.K. Sauer, H.O. Handwerker, P.W. Reeh, Responsiveness of C-fiber nociceptors to punctate force-controlled stimuli in isolated rat skin: lack of modulation by inflammatory mediators and flurbiprofen, Neurosci. Lett. 361 (2004) 163–167. [19] M. Schmelz, R. Schmidt, M. Ringkamp, H.O. Handwerker, H.E. Torebjork, Sensitization of insensitive branches of C nociceptors in human skin, J. Physiol. 480 (Pt 2) (1994) 389–394. [20] M. Renganathan, T.R. Cummins, S.G. Waxman, Contribution of Na(v) 1.8 sodium channels to action potential electrogenesis in DRG neurons, J. Neurophysiol. 86 (2001) 629–640. [21] S. Tate, S. Benn, C. Hick, D. Trezise, V. John, R.J. Mannion, M. Costigan, C. Plumpton, D. Grose, Z. Gladwell, G. Kendall, K. Dale, C. Bountra, C.J. Woolf, Two sodium channels contribute to the TTX-R sodium current in primary sensory neurons, Nat. Neurosci. 1 (1998) 653–655. [22] E.C. Emery, G.T. Young, P.A. McNaughton, HCN2 ion channels: an emerging role as the pacemakers of pain, Trends Pharmacol. Sci. 33 (2012) 456–463. [23] S. Schnorr, M., Eberhardt, K., Kistner, H., Rajab, J., Kasser, A., Hess, P., Reeh, A., Ludwig, S. Herrmann, HCN2 channels account for mechanical (but not heat) hyperalgesia during long-standing inflammation, Pain (2014). [24] E. Leipold, L. Liebmann, G.C. Korenke, T. Heinrich, S. Giesselmann, J. Baets, M. Ebbinghaus, R.O. Goral, T. Stodberg, J.C. Hennings, M. Bergmann, J. Altmuller, H. Thiele, A. Wetzel, P. Nurnberg, V. Timmerman, J.P. De, R. Blum, H.G. Schaible, J. Weis, S.H. Heinemann, C.A. Hubner, I. Kurth, A de novo gain-of-function mutation in SCN11A causes loss of pain perception, Nat. Genet. 45 (2013) 1399–1404. [25] M.E. Petersson, O. Obreja, A. Lampert, R.W. Carr, M. Schmelz, E. Fransen, Differential axonal conduction patterns of mechano-sensitive and mechano-insensitive nociceptors–a combined experimental and modelling study, PLoS One 9 (2014) e103556.
