INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the last decade, mechano-insensitive C fibers (CMi units) were discovered in human skin (Schmidt et al. 1995). They clearly differ from the long known mechano-responsive C-nociceptors (CMH units) by structural and functional characteristics (Schmelz et al. 2000a,b; Schmidt et al. 2000; Weidner et al. 1999). These mechano-insensitive units have been called "sleeping nociceptors" because they are not excited even by von Frey hairs of 750 mN but become mechano-responsive on inflammation. They have been shown to play a dominant role for encoding of painful stimuli such as intracutaneous capsaicin injection and tonic pressure and for induction of primary and secondary hyperalgesia (Schmelz et al. 2000b; Schmidt et al. 2000). Furthermore, they apparently mediate the axon reflex flare by neuropeptide release from their axon terminals on stimulation (Schmelz et al. 2000a) because CMH units have too small receptive fields (Schmidt et al. 1997) to account for it.

The present study is the first quantitative assessment of the innervation territories of CMi units compared with those of mechano-responsive units. The innervation territories of the latter have already been assessed in a previous study (Schmidt et al. 1997) in which we used electrical stimuli (constant current, 10 mA, or constant voltage, 80-100 V) that turned out to be supramaximal for mechanoresponsive units but insufficient to excite all branches of mechano-insensitive fibers. For comparing the field sizes of both unit classes, we had to increase stimulus intensities in the present study. Part of this work has been published in abstract form before (Schmidt et al. 1998).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

The experiments were performed in 20 healthy human subjects, 14 men and 6 women (age range, 20-32 yr). All subjects gave their informed consent according to the Declaration of Helsinki and the study was approved by the ethics committees of the Universities of Uppsala and Erlangen/Nürnberg.

Search for C-fiber responses

Microneurography techniques were employed for recording from cutaneous C fibers in the peroneal nerve, innervating the skin of the lower leg and the dorsum of the foot and toes. The search procedure for single C fibers and the mapping of their receptive fields has been described before in detail (Schmelz et al. 1995; Schmidt et al. 1995; Torebjörk and Hallin 1974). Briefly, a tungsten microelectrode was inserted into a skin fascicle of the peroneal nerve at knee or ankle level. The innervation territory of the impaled fascicle was identified by multifiber responses of large mechano-receptive fibers to touching the skin. Single transcutaneous electrical stimuli from an insulated stimulator (0.3-Hz, 0.2-ms, 50-mA constant-current pulses, DS7A, Digitimer) were applied within the innervation territory of the impaled fascicle via a pointed steel electrode (1 mm diam), which was moved on the skin until a long-latency C-fiber response was obtained. Two tungsten microelectrodes (0.2 mm diam) were inserted intracutaneously approximately 5 mm apart at the skin site from where the C-fiber response had been elicited. Pulses of moderate intensity (0.2 ms, 1-20 mA) suprathreshold for C-fiber activation were then applied through these needles for repetitive intracutaneous stimulation. The C-unit responses were recorded on-line by a PC computer via an interface card (DAP, Microstar) using the SPIKE/SPIDI software package (Forster and Handwerker 1990). Traces triggered by the intracutaneous pulses were displayed successively from top to bottom on the computer screen.

Conduction velocity measurements

The latencies of C-fiber responses to the first electrical pulse delivered through the intracutaneous needle electrodes after a rest period of at least 2 min were used for computing the conduction velocities. The shortest distance between the stimulating electrodes in the skin and the recording electrode in the nerve was measured in millimeter. Room temperature was kept constant at 22-24°C throughout the experiments.

Classification of C units by the "marking" technique

For subclassification of C fibers, repetitious intradermal stimulation at 0.25 Hz was used. After a minute or so the latencies of C responses would stabilize at latencies characteristic for each fiber in a multifiber recording (Torebjörk and Hallin 1974). Activation of a C fiber by additional natural or electrical stimulation would result in activity dependent slowing of impulse conduction, seen as a transient increase in latency, followed by a slow recovery after the end of the stimulus (Fig. 1).

View larger version (22K): [in this window] [in a new window]   Fig. 1. The "marking method" used for mapping the electro-receptive field (eRF). A: intracutaneous needles were inserted for regular 0.25-Hz electrical stimulation. Conditioning transcutaneous electrical pulses (pen) delivered in the vicinity of the needles were interposed between regular pulses for mapping the eRF. B: regular pulses through intracutaneous needles within the receptive field every 4 s (stimulus artifacts under straight arrow) evoked responses of the unit at a stable latency. When additional electrical pulses through a surface probe were applied within the eRF, additional action potentials (left curved arrow) were generated causing delayed responses to the regular pulses and a slow recovery (right curved arrow). C: schematic drawing of an electroreceptive field. Fifty-milliampere electrical square wave pulses were applied to the skin through a pointed probe. Points where the unit was activated are marked with filled circles, and points where A was tested but not activiated are marked with open circles.

In this way C units were shown to be afferent by their responsiveness to mechanical, thermal, or chemical stimuli (Torebjörk 1974). Sympathetic C fibers could be classified by their activation related to sympathetic reflexes caused, e.g., by arousal stimuli (Hallin and Torebjörk 1974). It has been proven in a previous study that this "marking" method is sensitive enough to allow detection of just one extra impulse in a stimulated C fiber (Schmelz et al. 1995). This means that the technique is useful for classification of C units, for threshold determinations, and for mapping the extensions of the innervation territories. Due to the long conduction distances (20-50 cm), slight differences in conduction velocities caused latency separation of multiple C fibers recorded in one intraneural site. There was never any ambiguity which unit responded to a particular stimulus even if two or more units had adjacent or even overlapping innervation territories.

Determination of mechanical and heat thresholds

A set of calibrated von Frey nylon filaments (Stoelting, Chicago, IL) was used to quantify mechanical thresholds. The forces exerted by these filaments were 1.5-750 mN (0.8-18 bar; tip diameter of 0.15-0.71 mm). The lowest force that elicited a "marking" response from the receptive field of a unit was regarded as the mechanical threshold.

Heat stimuli were delivered from a halogen lamp feedback controlled by a thermocouple attached to the skin (Beck et al. 1974). Skin temperature was increased by 0.25°C/s from an adapting temperature of 32°C to tolerance level at 50-52°C. The temperature leading to activation of the unit was noted. Because the electrical stimuli were delivered every fourth second, a response to heating could possibly be overestimated by up to 1°C. In 17/20 CMi units, heat thresholds were measured in two or more spots within the eRF to assess possible variability of thermal responsiveness.

Responsiveness to chemical stimulation

Histamine (20 mC) or acetylcholine (60 mC) was applied by iontophoresis within receptive fields of CMi units using a standardized applicator, 5 mm in diameter (Magerl et al. 1990). For five CMi units, this application was repeated at different sites within the receptive field to assess variability of the response. For five CMi units, histamine was also applied outside the 50-mA eRF. The CMH units were not tested with these substances.

Mapping of mechano-receptive fields

The extensions of the mechano-receptive fields (mRFs) were mapped with a stiff von Frey hair of 750 mN. The mRFs were mapped with stimulation points spaced by about 2 mm. A skin area with a radius of at least 2.5 cm around the intracutaneous needle electrodes was searched to maintain mapping of the whole mRF in cases of discontinuity. Each point from which mechanical stimuli activated the respective C unit was marked on the skin by ink.

Mapping of electro-receptive fields

The extension of the electro-receptive fields (eRF) was mapped as described before (Schmidt et al. 1997). For this purpose, the tip of a pointed steel electrode was moistened with sodium chloride or electrode gel and gently contacted to the skin for pseudo-unipolar electrical stimulation. A metal plate (20 cm²) on the skin more than 10 cm away was used as reference electrode. The stimulation points were spaced by about 2-3 mm. A skin area with a radius of at least 2.5 cm around the intracutaneous needle electrodes was searched. After identifying an eRF in this region a radius of at least 1 cm around, the eRF was searched. Mapping was performed at 50 mA for all units. For 13/20 CMi units and 6/6 CMH units, mapping was also performed at 10 mA. In some cases, 70- and 100-mA pulses were also used, if tolerated by the subjects. Each point from which transcutaneous electrical stimuli activated the respective C unit was marked on the skin with different colors for different current strengths. A complete mapping of eRFs at different intensities could take up to 4 h. Not surprisingly, in many instances the unit was lost during the mapping, due to involuntary withdrawal movements or muscle jerks in particular at the higher stimulation intensities.

At the end of an experiment, the marks were transferred to a transparent sheet, and the contours of differently marked mRF and eRF of each unit were scanned into computer for determining the areas and circumferences by a dedicated computer program (Dept. Physiology 1, Erlangen). The quotient circumference/square root of area was calculated to indicate the degree of irregularity or roundness of the eRFs (a figure of maximal roundness, i.e., a circle, produces a quotient of 3.55, and higher values indicate increasing degrees of irregularity of the eRFs.)

The maximal diameter for each stimulation intensity was measured. Further we measured the maximal extension of the 50-mA eRF perpendicular to the maximal diameter, the maximal extension in the proximo-distal direction, and the maximal extension in the medio-lateral direction.

Location of units

Units were allocated to one of three regions of the cutaneous innervation territory of the peroneal nerve: the dorsum of the toes except the ungual phalanges and the lateral part of the little toe, the dorsum of the foot up to the malleoli, and the distal lateral half of the lower leg. The upper half of the lower leg was covered by the preamplifier equipment and could not be used for recording. Each unit was classified as belonging to one of these skin areas.

Statistics

Measures are given as median and range. The data refers to the sample of units studied for the present work. However, we have a large sample of characterized units that have already been published in studies focusing on other questions. For the general properties of C fibers, like conduction velocity or receptive thresholds, and for eRFs of CMH units, an average of all fibers studied to date is provided in addition (indicated by running average, RA). Differences were tested by the Mann-Whitney U test or the Wilcoxon test for paired samples as stated in the text. Probability levels were regarded significant at P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Sample of C units

A total of 26 C units were mapped. Six of them responded to von Frey filaments and heat radiation and were classified as CMH units with a median mechanical threshold of 31 mN (range: 14-90 mN, RA: 30 mN, 5.4-226 mN, n = 129) and a median heat threshold of 41.0°C (range: 36.5-45.8°C, RA: 41.3°C, 35.6-49.5°C, n = 141). The conduction velocity was on average 1.0 m/s (median, range: 0.85-1.07 m/s, RA: 0.97 m/s, 0.52-1.32 m/s, n = 155). Three of the six CMH units were located on the distal lower leg and three on the proximal dorsum of the foot, close to the ankle.

The other 20 units did not respond to stimulation with von Frey filaments of 750 mN and were hence classified as CMi. They were even not excited by penetration of the skin with a hypodermic needle. Conduction velocity was on average 0.82 m/s (median, range: 0.37-1.53 m/s, RA: 0.79 m/s, 0.47-1.29 m/s, n = 126). Of these, 17 units were CH because they responded to heating (up to 52°C) with a median threshold of 48.6°C (range: 41.7-52.0°C, RA: 48.0°C, 41.7-51.5°C, n = 52). Six CH units responded with more than 40 "markings" to histamine iontophoresis and belonged therefore to the subclass of "itch" units (Schmelz et al. 1997). Three CMi units did not respond to heating (or histamine) and were classified as mechano-heat insensitive C nociceptors (CMiHi). They were proven to be afferents by their pronounced activity dependent slowing clearly separating them from sympathetic efferents (Weidner et al. 1999). One of the total of 20 CMi units was located on a toe, 13 on the foot dorsum and 6 on the lower leg.

Conduction velocities and receptive thresholds were consistent with previous studies on larger samples of CMH and CMi units (Schmidt et al. 1997; Weidner et al. 1999).

Receptive fields of CMH units

eRFs OF CMH UNITS AT 10 mA. The eRFs of CMH units mapped with 10 mA had median areas of 1.95 cm² (range: 0.84-2.86 cm²; Fig. 2), which is not statistically different (Wilcoxon test) from the mRFs tested with 750 mN von Frey filaments (median: 1.87 cm², range: 0.78-2.81 cm²). They had distinct borders. Moving the stimulus probe 1 mm across the border drawn with pen on the skin was usually enough to reproducibly activate a unit or not. The borders of the 10 mA eRFs corresponded exactly to the 750-mN mRFs for most of the circumference of the CMH units. However, for parts of the circumference the 10-mA eRFs slightly exceeded the 750-mN mRF in four units. No CMH unit was activated by 750 mN outside the 10-mA eRF. Five of six CMH units exhibited one continuous 10-mA eRF. The remaining unit had two distinct 10-mA eRFs separated by 2 mm. The 10-mA eRFs and 750-mN mRFs for the six CMH units are illustrated in Fig. 2A.

View larger version (20K): [in this window] [in a new window]   Fig. 2. Electro- and mechano-receptive fields of mechano-heat-responsive (CMH) fibers. A: eRFs of 6 CMH units in response to 10-mA stimulation (black and white) and mechanoreceptive fields (mRFs) in response to 750 mN von Frey stimulation (black). For 2 units, the eRFs and mRFs coincided exactly. For 4 units, minor parts at the borders of the eRFs were not responsive to 750-mN von Frey stimulation. B: eRF of the same 6 mechano-responsive C units as shown in A mapped with electrical pulses of 10 mA (black) and 50 mA (crosshatched). Two units were also mapped with 70 mA (hatched) and 1 with 100 mA (white). Receptive fields located on the lower leg and foot are shown separately (top and bottom). The 50-mA fields are not much larger than the 10-mA fields in contrast to the mechano-insensitive units depicted in Fig. 3. The 10 mA eRFs for these mechano-responsive units are (in contrast to the mechano-insensitive units) mainly located near the center of the 50-mA receptive fields, suggesting that the 50-mA fields for these units may represent current spread rather than true anatomical receptive fields.

HIGH-INTENSITY eRFs OF CMH UNITS. Mapping the 6/6 CMH units with 50-mA electrical pulses produced slightly larger eRFs as compared with 10 mA (Fig. 2). For units with irregular 10-mA eRFs, the gaps and indentations of the eRFs tended to be filled out. The median area of the 50-mA fields increased to 3.08 cm² (range 1.15-3.66 cm²) and the median diameter from 24.4 to 28.7 mm. Increasing the stimulation intensity to 70 and 100 mA tended to produce a small concentric growth of the eRFs (Figs. 2B and 4).

Receptive fields of CMi units

eRFs OF CMi UNITS AT 10 mA. Stimulation at 10 mA in 13/20 CMi units produced small eRFs in all units tested. The median area was 0.35 cm² (range: 0.1-1.74 cm²). The units were excited at one or several small spots (median 3, range: 1-7, Fig. 3). Chemical or thermal excitability could regularly be found outside the 10 mA eRF. If the 10-mA eRF consisted of multiple spots, their closest borders had a median distance of 5 mm (range: 3-45 mm) for the two closest spots and of 25 mm (range: 3-45 mm) for two most distant spots.

View larger version (51K): [in this window] [in a new window]   Fig. 3. Electroreceptive fields of mechano-insensitive C-fibers. Electroreceptive fields of heat-responsive (CH) and -unresponsive (CMiHi, marked) mechano-insensitive C fibers on the lower leg, the foot, and the toes are depicted proximal side up. The 50-mA (crosshatched, tested for all units) fields were much larger than the 10-mA (black) fields but 70 (hatched) and 100 mA (white) produced relatively little further increase in the electroreceptive fields, suggesting that 50, 70, or 100 mA reflects the true anatomical receptive field. Note that even for 100 mA, no proximal elongation of the eRF indicative of the axonal stem is visible. There were no major differences in electroreceptive fields based on location (leg, foot, or toes) or histamine responsiveness (His). Ten-milliampere mapping stimuli produced an eRF in the units testedunits lacking a black area in the figure were not tested at 10 mA. Units lacking a 70- or 100-mA field in the figure were not tested with these intensities.

HIGH-INTENSITY eRFs OF CMi UNITS. Stimulation with 50-mA pulses in 20/20 CMi units revealed fairly large irregular eRFs often with small "satellites" outside the main territory. The median area was 5.34 cm² (range: 1.1-14.2 cm²) and the median diameter was 46.6 mm (range: 21.0-78.6). This is larger than for CMH units (median area: 3.1 cm², median diameter: 28.7 mm, U test, P = 0.0013).

View larger version (24K): [in this window] [in a new window]   Fig. 4. Summary of electro-receptive field size of CMH and CMi fibers. Raw data (thin lines) and medians ± quartile (thick lines) are depicted for CMi (n = 20) and CMH (n = 6) units. Note that the areas of the CMi units grew considerably when the current intensity was increased from 10 to 50 mA, whereas the areas of the CMH units changed only little with increasing current intensity.

Shape of 50-mA eRFs of CMH and CMi units

The median circumference/square root area quotient for the 50-mA eRFs of CMH units was 5.0 (range: 4.35-5.82), indicating a more rounded shape than for CMi units with a median quotient of 7.2 (range: 5.1-10.8). The latter were thus significantly more irregular in shape (U test, P = 0.002). There was no difference in 50-mA eRF area between CMi units on the lower leg (median area: 5.91 cm², n = 6) or foot (median area: 5.08 cm², n = 14). This is in contrast to the proximo-distal size diminution in CMH units as shown previously (Schmidt et al. 1997). The median diameters of 50-mA eRFs of CMi units were also similar (48 mm) on the foot and the lower leg. The maximal diameters were about 50% larger than the perpendicular diameters. The direction of the maximal extent seemed to be randomly orientated (Table 1).

Discrepancy between 10- and 50-mA eRFs in CMi and CMH units

The eRFs of CMi units grow in an irregular and discontinuous manner when increasing the stimulus intensity from 10 to 50 mA and the most distant part of a 50-mA field could be quite remote from its closest 10-mA field. In contrast, the eRFs of CMH units grow little and in a more regular concentric manner. To illustrate this numerically, the maximal distance from the outer border of the 50-mA eRF to the nearest outer border of the 10-mA eRF was measured for all the 13 CMi and 6 CMH units that were mapped with both these intensities. We named this measure the "10-50 eRF distance." For CMi units, the median 10- to 50-eRF distance was 22.7 mm (range: 4.6-33.2 mm). For CMH units, the corresponding median distance was 3.4 mm (range: 2.4-5.7 mm). The difference is statistically significant (U test, P = 0.0009). Even after normalizing these data by dividing 10- to 50-eRF distance with the mean diameter of the 50-mA eRFs to compensate for the difference in size between CMi and CMH units, the difference remained highly significant (U test, P = 0.007).

Variability of chemical and heat responsiveness inside the eRFs

Temperature thresholds were measured at multiple locations within the eRFs of all CH units except three. In some units, activation thresholds to heat and chemical responsiveness showed considerable variations even inside the eRFs of a single unit as shown in Fig. 5 (e.g., the 2 lower units). Maximal differences between heat thresholds were 42.7 versus 52 and 41.7 versus 50.2. The local temperature thresholds and responsiveness to chemicals were not clearly related to whether the units were excited by 10-mA or only by 50-mA electrical pulses through a pointed probe at the location where heat or chemical stimulation was applied.

View larger version (45K): [in this window] [in a new window]   Fig. 5. Responsiveness to natural stimulation inside and outside the electro-receptive field of CMi fibers. Six examples of electroreceptive fields of CMi units are depicted using the same method to indicate responsiveness to electrical stimuli as in Fig. 3. , temperature thresholds at tested locations within the eRFs or responsiveness to histamine and acetylcholine when applied at multiple sites inside or outside the eRFs.

Histamine iontophoresis applied outside the 50-mA eRFs for five CMi units was negative.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The most important findings in this paper are the differences in size and shape of innervation territories between CMH and CMi fibers, the latter being larger and more irregular.

Reliability and validity of the mapping procedure

Innervation territories of individual C units were mapped by transcutaneous electrical stimulation delivered from a pointed probe. In all experiments, the marking technique was employed to make sure that responses of an individual unit were not confused with those of other fibers having overlapping innervation territories. To address the question whether the eRF as assessed here corresponded with the distribution of receptive nerve endings mechanical, heat or chemical stimuli were applied within the electrically mapped innervation territories.

In CMH units, the eRFs were compared with the mechanical receptive fields (mRFs) mapped with von Frey hairs suprathreshold for all mechano-responsive C fibers (750 mN). Mechanical stimuli affect stretch activated membrane channels in the C-fiber terminals, and the mRFs represent the distribution of those channels in the terminal arborization of the units.

The substrate stimulated by transcutaneous electrical stimulation is not as obvious as that of mechanical stimulation. Whereas mechanical, thermal, and chemical stimuli excite the nerve fiber at its receptive terminals, an electrical stimulus might activate it further proximally. However, we did not observe elongations of eRFs in the proximal direction which would have been the case if the parent axon would have been activated. In monkey, Meyer et al. (1991) found such elongation as a correlate of the parent axon at high stimulus intensities, whereas a sudden distal threshold decrease indicated the border of the innervation territory. In awake humans, the current necessary to excite the parent axon seems to exceed the tolerance level probably due to the thick insulating subcutis.

Still, there is a possibility of lateral spread of the electrical current such that remote nerve terminals are excited by high-intensity stimulation. Indeed, this seemed to be the case for CMH units that are known to have low electrical thresholds (Weidner et al. 1999). The concentric expansion of the eRFs on increasing the stimulus intensity from 10 to 50 mA as documented here seems to reflect such current spread. Comparing the eRFs with the mRFs mapped with a 750-mN von Frey hair revealed a good correspondence for the 10-mA eRF and the 750-mN mRF, whereas the expansion at 50 mA was always outside the mRF. Therefore for CMH units that have fairly low electrical and mechanical thresholds (Weidner et al. 1999), both mapping procedures might lead to a slight overestimation of the respective receptive fields, mechanical probing by exciting remote endings through lateral stretch, and transcutaneous electrical stimuli by current spread.

The situation is completely different for CMi units. They have much higher transcutaneous electrical thresholds than CMHs. In a previous study, we found an average threshold of 60 mA (0.2 ms pulses) for CMi units and 4 mA for CMH units when an iontophoresis probe, 5 mm in diameter, was used for stimulation (Weidner et al. 1999). In this study, 10 mA applied through a pointed probe revealed small scattered eRFs that could not reflect the real size of the innervation territories because thermal or chemical stimuli could excite the units from areas outside. Therefore 50 mA was used for mapping. This revealed a considerable expansion of the eRFs that now might better reflect the real size of the innervation territories because thermal and chemical stimulation regularly excited the units from within this area. Increase of the stimulus intensity to 70 or 100 mA induced only minimal further extensions, possibly not due to current spread because of its nonconcentric growth. In addition, 10-mA spots were found directly adjacent (2 mm) to spots from which 100 mA did not elicit any response of the recorded fiber. This also reflects that a gradual increase of current does not induce a gradual increase of size as suggested by the current spread hypothesis but that the areas recruited at very high stimulus intensities represent extremely high-threshold branches of the terminal arborization. In other words, mapping with 50 mA probably leads to an underestimation of the real size of the innervation territories for the mechano-insensitive units. The statistically significant difference between the slightly overestimated 10-mA eRF areas of CMH units (median: 1.95 cm²) and the underestimated 50-mA eRFs of CMi (median: 5.34 cm²; U test, P < 0.01) reveals that CMi units have much wider branches in human skin than CMH C nociceptors. This difference becomes even more striking (median: 1.01 vs. 5.34 cm², U test, P < 0.0001) when comparing the present CMi material with a large sample of CMH units published previously (Schmidt et al. 1997).

Inhomogeneity of sensory properties within the eRF of CMi units

The mRF and 10-mA eRF of CMH units are usually highly congruent. Comparable responsiveness to natural stimulation is not easily tested in CMi units. Only part of the CMi units respond to a certain chemical stimulus or to heating, and the unresponsiveness might be attributed to a lack of receptor proteins or to a lack of terminal branches at the examined position. Several CH units, proven to be heat sensitive in at least one spot, were tested for their heat thresholds in various other spots within their eRF (Fig. 5). Heat thresholds were astonishingly variable within the eRF. For a few units, repetitive histamine or acetylcholine stimulation yielded a comparable variability. The responsiveness to these stimuli was apparently not related to whether the stimulus was applied within the 10 or 50 mA part of the eRFs. This suggests that CH units may have inhomogenous receptive properties of their endings within different parts of their innervation territories rather than only variation in depth or dimensions of the terminal branches.

In five of the CMi units, histamine iontophoresis was applied outside the 50-mA eRF. None of these units was activated. This suggests that there are no receptive endings outside the eRFs of CMi units, although it is possible that some CMi units may have branches with extremely high electrical thresholds which are not excited even by 50- to 100-mA pulses.

Organization of innervation territories

Innervation territories of several square centimeters can only be explained by extensive arborization of the nerve terminals. Anatomical evidence for branched nerve terminals that might even supply distinct anatomical structures has been found in rats and cats (Heppelmann et al. 1990; Pierau et al. 1982; Taylor and Pierau 1982). Distinct steps in response latency were found as functional evidence for branched A and C fibers in animal and human recordings (McMahon and Wall 1987; Peng et al. 1999; Ringkamp et al. 1998; Torebjörk and Hallin 1974). Our findings indicate that such branches tend to be concentrated on a confluent rounded area for CMH fibers, whereas for CMi fibers, the terminal arborization is widely distributed. Thus the CMi units had very irregular contours of their eRFs and often showed satellites outside the main innervation territory. In addition many of the CMi units had discrete spots of fairly low-threshold eRFs often eccentrically located within the high-threshold regions possibly reflecting inhomogeneity in diameter or depth of the individual terminals.

Functional implications of different innervation territories

Recently we have shown that afferent cutaneous C fibers fall into two clearly different classes that in intact skin can be distinguished by their responsiveness to mechanical stimulation with von Frey hairs, the prior with a threshold of less than 200 mN, the latter not even excitable by 750 mN or insertion of hypodermic needles. Under pathophysiological conditions, this difference may be blurred because CMi units often become responsive to mechanical stimuli in inflamed skin (Schmidt et al. 1995). Since mechanical thresholds of CMH units do not decrease in inflammation sensitization of CMi units may be the underlying mechanism for primary mechanical hyperalgesia (Schmelz et al. 1996). Alternatively, suprathreshold responses could be enhanced also in CMH unit as shown before (Andrew and Greenspan 1999).

Mechanical responsiveness is not the only difference between the two classes of C nociceptors. Whereas discharge patterns of CMH units do not explain the intensity and time course of pain in response to capsaicin injection or tonic pressure, discharge of CMi units closely matches the perceived pain (Schmelz et al. 2000b; Schmidt et al. 2000). Also the contribution to neurogenic flare responses is different. It is mediated by CMi units in human skin, whereas there is only a minor contribution of CMH units to the axon reflex flare (Schmelz et al. 2000a). The extension of the axon reflex flare induced by capsaicin injection, high-intensity electrical stimulation or histamine of up to 15 cm in diameter can only be explained by large innervation territories of up to 7.9 cm as shown for CMi units here. A subgroup of CMi units is particularly sensitive to histamine and probably mediates itch sensations (Schmelz et al. 2000a). Six CH units responding to histamine (marked with "His" in Fig. 3) were presented in this paper.

It might be speculated that the small sizes of the innervation territories of CMH units and their proximo-distal decay would allow for more precise stimulus localization than the input from CMi units with large territories. This notion is supported by the finding, that localization of noxious events is fairly precise on the foot and the dorsum of the hand with a pure C-fiber input (Jørum et al. 1989; Koltzenburg et al. 1993), whereas the two point discrimination for histamine induced itch is poor (Wahlgren and Ekblom 1996).

In conclusion, innervation territories of CMi nociceptors are more irregular and much larger than the territories of CMH units. This spatial difference adds to the functional dichotomy between the two major types of C nociceptors in human skin: mechano-insensitive versus mechano-responsive.