INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Human "neuropathic" pain, which can be induced by peripheral nerve injury, devolves from aberrant spontaneous activity in sensory nerves (Kauppila 1998; Woolf and Salter 2000). At least some of this activity arises from the dorsal root ganglia (DRG) (Babbedge et al. 1996; Liu et al. 2000, 2001; Millan 1999; Wall and Devor 1983). In rats, sciatic nerve injury or section (axotomy) increases the excitability of neurons in L₄ and L₅ DRG (Kim et al. 1998; Stebbing et al. 1999; Study and Kral 1996; Zhang et al. 1997) and increases the number of action potentials (APs) fired in response to sustained depolarizing current (Abdulla and Smith 2001a). Axotomy affects "small" DRG cells more than "medium" cells, and these are affected more than "large" cells. In addition to decreasing rheobase (Study and Kral 1996), axotomy significantly increases spike height (AP amplitude) in "small" and "medium" cells (Abdulla and Smith 2001a). It also produces a significant increase in spike width (AP duration) in "small" cells (Abdulla and Smith 2001a; Kim et al. 1998; Stebbing et al. 1999).

Voltage-clamp analysis of DRG neurons associates their response to axotomy with a reduction in Ca²⁺ channel current (I_Ca) that leads to a decrease in Ca²⁺-sensitive K⁺ conductance (g_K,Ca). Delayed rectifier K⁺ current (I_K) is also attenuated (Abdulla and Smith 2001b; Baccei and Kocsis 2000; Everill and Kocsis 1999). The effects of nerve injury on Na⁺ channel currents (I_Na) are more complex. This reflects the variation of expression of different types of Na⁺ channels on DRG neurons (Cummins and Waxman 1997; Elliott and Elliott 1993; Roy and Narahashi 1992; Rush et al. 1998) as well as possible differences in the response of each channel type to axotomy (Cummins and Waxman 1997; Sleeper et al. 2000; Waxman et al. 1994). In addition, various Na⁺ channel conductances are affected in different ways depending on which type of injury is model is employed. For example, chronic constriction injury does not alter tetrodotoxin-resistant (TTX-R) or tetrodotoxin-sensitive (TTX-S) I_Na or the PN3 mRNA (also known as -SNS or Na_v1.8) (Goldin et al. 2000), which codes for a TTX-R channel (Novakovic et al. 1998). By contrast, sciatic nerve ligation and section has been reported to decrease both a slowly inactivating (TTX-R) I_Na (Cummins and Waxman 1997) and an additional "persistent" TTX-R I_Na in small (C-type) sensory neurons (Sleeper et al. 2000). Although, TTX-S I_Na density was unchanged after nerve ligation, the conductance exhibited more rapid recovery from inactivation ("repriming") (Cummins and Waxman 1997). These findings corroborate molecular biological studies that demonstrated up-regulation of the "III" (Na_v1.3) message for the rapidly-repriming TTX-S channel and down-regulation of the Na_v1.8 or "-SNS" message for TTX-R, slowly inactivating current (Black et al. 1997) as well as the Na_v1.9 or "NaN" message thought to be responsible for the "persistent" TTX-R current (Dib-Hajj et al. 2002; Sleeper et al. 2000).

In this study, we have addressed additional aspects of the changes in I_Na invoked by axotomy. We started by using criteria for identification of "small," "medium," and "large" cells that were previously established for current-clamp studies of excitability, rheobase, and AP characteristics (Abdulla and Smith 2001a). This allows for correlation between the present voltage-clamp and previous current-clamp studies. We have, for example, shown that axotomy significantly increases spike height in "small" cells but not in "large" cells. We can now ask whether these changes are reflected by a selective increase in total I_Na in the same "small" cell population, subjected to the same type of injury and identified by the same criteria. Second, because we have attempted to obtain representative recordings from all types of DRG neurons, we are able to compare the extent and the type of change induced by axotomy in different neuronal populations. We have therefore used whole cell recording techniques to examine the effects of axotomy on TTX-S I_Na and TTX-R I_Na (Caffrey et al. 1992; Elliott and Elliott 1993; Ikeda et al. 1986; Rush et al. 1998) in "small," "medium," and "large" DRG cells. In contrast to results obtained in "small" DRG neurons by Cummins and Waxman (1997) and by Sleeper et al. (2000), we found that axotomy increased TTX-R and/or TTX-S I_Na and slowed inactivation in all DRG cell types.

Sciatic nerve axotomy in rats sometimes invokes a self-mutilatory behavior known as "autotomy" (Coderre et al. 1986; Wall et al. 1979). The development of autotomy is viewed by some as an animal manifestation of human neuropathic pain (Coderre et al. 1986; Kauppila 1998; Liu et al. 2001; Mailis 1996). Although axotomy invokes pronounced changes in the properties of "small," putative nociceptive DRG neurons, little further change occurs in animals that develop autotomy. By contrast, "large" DRG cells, which are only modestly affected by axotomy, display significant alterations in their properties that correlate with the presence of autotomy (Abdulla and Smith 2001a,b). This transition in "large" cell properties may reflect the clinical observation that neuropathic pain is associated more with alterations in the properties of myelinated, nonnociceptive axons than with changes in the properties of nonmyelinated, putative nociceptive axons (Campbell et al. 1988; Liu et al. 2001; Nystrom and Hagbarth 1981). An additional aspect of our work therefore was to relate the expression of autotomy to axotomy-induced changes in Na⁺ channels in various DRG cell types.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All experimental procedures were in concordance with the recommendations of the International Association for the Study of Pain (IASP). Protocols were approved by the University of Alberta Animal Welfare Committee. This committee is responsible for maintaining standards set forth by the Canadian Council for Animal Care. Detailed methods for animal care, treatment, and surgery are described in our previous work (Abdulla and Smith 2001a,b). Briefly, 120- to 170-g male Sprague-Dawley rats were anesthetized with sodium pentobarbital (50-55 mg/kg, ip), and the sciatic nerve was sectioned proximal to its bifurcation into the tibial and the peroneal divisions. A 5- to 10-mm segment of nerve was removed to prevent regeneration. Control rats or operated rats were killed by decapitation and neurons enzymatically dissociated from L₄ and L₅ DRG. These ganglia were selected because 98% of sensory fibers in the sciatic nerve have cell bodies in the L₄ and L₅ DRG. (Swett et al. 1991). It should be noted, however, that the axons of as many as 30% of the neurons in the L₄ and L₅ DRG originate from afferent nerves other than the sciatic (Himes and Tessler 1989). Although we therefore cannot guarantee that all cells studied in the axotomy experiments actually had severed axons, statistically significant differences were found between neurons from control ganglia and axotomized ganglia. Cells were studied 2-10 h after dissociation. As in our previous studies, data were collected only from the DRG on the side of the sciatic nerve lesion. Control data were obtained from unoperated rats or nerves that had undergone nerve exploration without cutting (Abdulla and Smith 2001a,b). The morphological and functional effects of axotomy on DRG neurons start as early as 3 days and continue for more than 15 wk (Titmus and Faber 1990). We investigated effects that occur within a period ranging from 2 to 7 wk postaxotomy in age-matched control rats. In a previous study (Abdulla and Smith 2001a), we initially attempted to make a detailed time course study of the electrophysiological changes induced by axotomy. Because we found little significant difference between groups sampled at successive weekly intervals over the 2- to 7-wk postaxotomy period, data from our previous and present study have been pooled from axotomized animals 2-7 wk postoperatively.

Axotomized animals were divided into two groups: those that did not exhibit autotomy ["axotomy (no autotomy)"] and those that did ["autotomy (axotomomized)"] (Coderre et al. 1986; Wall et al. 1979). Autotomy was scored according to the scale devised by Wall et al. (1979). A score of 1 was given for the removal of one or more nails. The score was increased by 1 for injury to each distal digit, and by another 1 for injury to each proximal digit. Although the maximum score permitted under IASP guidelines is 11, all of our animals were killed before they attained a score of 8. No abnormal behavior or premature deaths occurred in the 7-wk study period. No unoperated limbs were mutilated by any of the rats. Rats exhibiting any sign of autotomy (i.e., autotomy score > 1) were assigned to the "autotomy (axotomized)" group.

Whole cell recordings (at 20-22°C) were made using a single-electrode voltage-clamp amplifier (Axoclamp 2A) in discontinuous mode as described previously (Abdulla and Smith 1997a). With low resistance patch electrodes (2-5 M), it was possible to use high switching frequencies >30 kHz with high clamp gains (8-30 mV/nA). The effectiveness of the clamp was confirmed by examining recordings of the command voltage. Recordings from cells where the voltage trace was slow to rise or distorted were discarded as were recordings from cells that exhibited "all or none" rather than progressively incremental inward current responses. Data were acquired using PCLAMP 5.5 (Axon Instruments, Foster City, CA) and analyzed using PCLAMP 6, 7, or 8. Final data records were produced using ORIGIN 5.0 or 6.1 (Microcal, Northampton, MA). Input capacitance (C_in) was calculated from the membrane time constant and input resistance (R_in) or by integration of the capacitative transient generated by a 10-mV voltage jump (for details, see Abdulla and Smith 1997).

To limit contributions from voltage-gated Ca²⁺ and K⁺ currents, I_Na was recorded in an external solution containing the following (in mM): 100 NaCl, 5 KCl, 4 MgCl₂, 10 HEPES, and 60 D-glucose, adjusted to pH 7.4 with NaOH. The internal (pipette) solution contained the following (in mM): 140 CsCl, 10 NaCl, 2 MgATP, 0.3 Na₂GTP, 2 EGTA, 10 HEPES, and 2 MgCl₂, adjusted to pH 7.2 with NaOH.

The volume of fluid in the recording dishes was about 1 ml. These were superfused with external solutions at a flow rate of 2 ml/min, allowing the exchange of bathing solution within 1 min. TTX (1 or 10 µM) was applied by superfusion. Total I_Na was recorded in response to depolarizing voltage commands from a holding potential (V_h = 90 mV) and leak subtracted by means of a p/4 protocol. Thus a series of one-fourth amplitude, reversed polarity voltage commands were applied, and the recorded currents multiplied by four and added to the recordings of I_Na. To obtain the TTX-S and TTX-R components of the current, currents persisting in the presence of 1 µM TTX were subtracted from the corresponding values of total I_Na.

Clear-cut differences in the C_in provided a criterion for classification of DRG cells. C_in was always >90 pF for "large" neurons, 70-90 pF for "medium" neurons, and <70 pF for "small" neurons.

TTX was from Research Biochemicals International (Natick, MA), and other chemicals were from Sigma (St. Louis, MO). All data are presented as means ± SD and significance of difference assessed using Student's unpaired t-test or ANOVA followed by Student-Newman-Keul's test as appropriate. In the few cases where no error bars are visible, the error bars are smaller than the symbols used to designate the data points.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Identification and definition of "large," "medium," "small-fast," and "small-slow" neurons

Neurons were first assigned to the "small," "medium," or "large" group on the basis of their C_in and according to criteria set forth in our previous studies (Abdulla and Smith 2001a,b). Total I_Na, the sum of the TTX-S and TTX-R components of the current (Roy and Narahashi 1992), was recorded using a series of depolarizing voltage commands from V_h = 90 mV. Neurons were then treated with 1 µM TTX to distinguish the TTX-S and TTX-R components of the current. In some cells, the TTX concentration was increased to 10 µM, but this never achieved any greater level of block than that seen with 1 µM TTX.

Cursory examination of the effects of TTX on I_Na in "small" cells revealed two subpopulations. In the first, >70% of the current persisted in the presence of 1 µM TTX and inactivation was slow; >7 nA of I_Na persisted after 10 ms at 10 mV. These were defined as "small-slow" cells. These cells, which made up 54% (22/41) of the "small" cell population, corresponded to "type B and C small DRG cells" as defined by Rush et al. (1998). In the second group, >70% of I_Na was blocked by TTX and inactivation was rapid and pronounced; <5 nA of I_Na persisted after 10 ms at 10 mV. These were defined as "small-fast" cells. These cells, which likely correspond to "type A and/or D small DRG cells" (Rush et al. 1998), made up 46% (19/41) of the "small" cell population.

Figure 1 shows typical recordings of I_Na from control "large," "medium," "small-fast," and "small-slow" cells (V_h = 90 mV; Fig. 1, A-D, respectively). Note the slow onset and slow inactivation of the current recorded in the "small-slow" cell. The bottom traces in Fig. 1, A-D are voltage recordings. Better voltage control was achieved in the "small" cells, but even in the "large" cell, the clamp voltage settled within 500 µS.

View larger version (38K): [in this window] [in a new window]   Fig. 1. Sample recordings of total INa and membrane voltage from a "large" cell (A), a "medium" cell (B), a "small-fast" cell (C), and a "small-slow" cell (D). Calibration (40 nA) refers to A, B, and D. Calibration (20 nA) refers to C. Calibration (2 ms) refers to A, B, and C. Calibration (5 ms) refers to D.

In our previous work (Abdulla and Smith 2001a,b), we identified a fourth population of DRG neurons that we termed AD cells. These exhibit an afterdepolarization (ADP) under current-clamp or a predominant, low voltage-activated, T-type Ca²⁺ channel current (I_Ca,T) under voltage clamp (Abdulla and Smith 1997b; Scroggs and Fox 1992; White et al. 1989). The solutions used to study Na⁺ currents precluded the identification of AD cells. Since their size fell mainly within the "medium" cell range (Abdulla and Smith 2001a,b; Scroggs and Fox 1992) it is presumed that AD cells make up some of the "medium" cell category investigated under these conditions.

Effects of axotomy and the presence of autotomy on total, peak, and total "residual" Na⁺ channel current

Figure 2, A-D, illustrate plots of total peak I_Na density versus command voltage (V_c) for "large," "medium," "small-fast," and "small-slow" neurons, respectively. All data were obtained from a holding potential (V_h) of 90 mV. Under control conditions (Fig. 2, A-D, ), the density of total leak-subtracted I_Na is similar in "large" (n = 21), "medium" (n = 17), "small-fast" (n = 19), and "small-slow" cells (n = 22).

View larger version (37K): [in this window] [in a new window]   Fig. 2. Relationship between command voltage (Vc) and peak, leak-subtracted total INa density recorded from a holding potential (Vh) of 90 mV. A-D: peak current density data for "large," "medium," "small-fast," and "small-slow" cells, respectively. E-H: "residual" current density at the end of 10-ms voltage commands for "large," "medium," "small-fast," and "small-slow" cells, respectively. The 3 lines in each panel represent currents recorded from () control neurons, () axotomized neurons (from animals that did not exhibit autotomy), and () axotomized neurons from animals that exhibited autotomy. For "large," "medium," "small-fast," and "small-slow" control cells, n = 21, 17, 19, and 22, respectively. The corresponding n's for "large," "medium," "small-fast," and "small-slow," axotomized neurons are 22, 20, 16, and 22, respectively, and 22, 22, 15, and 20, respectively, for neurons from animals that exhibited axotomy. Error bars indicate SD. In the few cases where no error bars are visible, the bars are shorter than symbols used to designate the data points. I: sample record of INa activated at 10 mV to show method of measuring peak, total current (A-D), and residual current (E-H). J: histograms to show ratios of peak to residual current (densities at 10 mV) for "large," "medium," "small-fast," and "small-slow" cells from the control, axotomy (no autotomy), and autotomy (axotomized) groups. This ratio is an index of the "apparent" inactivation of total gNa. Note marked decrease in apparent inactivation of gNa in "large cells" in the autotomy (axotomized) group; this means more current persists at the end of a 10-ms test pulse. "Small" control cells exhibit modest inactivation and further decreases in activation are seen in the axotomy and autotomy groups. Mean currents for the various experimental situations (A-H) were originally calculated and displayed with SD. Since the data presented in J are ratios of these mean currents, it is difficult to obtain an accurate estimate of error.

In "large" cells, axotomy alone produced little change in total I_Na density (Fig. 2A, ; n = 22). A pronounced increase in current density (56% larger than control at 10 mV, P < 0.0001, n = 22) was seen in "large" cells from animals that exhibited autotomy (Fig. 2A, ). By contrast, axotomy alone produced a profound increase in the density of total I_Na in "small-slow" cells (69% larger than control at 10 mV, n = 20, P < 0.0001, Fig. 2D, ) but little further increase was seen in "small-slow" neurons from animals that exhibited autotomy (n = 20, Fig. 2D, ). Similarly, axotomy alone produced a profound increase in the density of total I_Na in "small-fast" cells (74% larger than control at 10 mV, n = 16, P < 0.0001, Fig. 2C, ), but little further increase was seen in "small-fast" neurons from animals that exhibited autotomy (n = 15, Fig. 2C, ). Figure 2B shows that the effect of axotomy on "medium" cells was intermediate between the two extremes seen in "large" and "small" cells. Thus axotomy produced a 29% increase in total I_Na seen at 10 mV (n = 20, P < 0.0001, Fig. 2B, ) and a further increase was seen in currents recorded from animals that exhibited autotomy (n = 22, Fig. 2B, ).

Figure 2, E-H, shows the relationships between V_c and the total "residual" I_Na density that persisted at the end of a 10-ms command pulse (see Fig. 2I). These data and those in Fig. 2, A-D, were collected from the same cells. Axotomy has little effect on residual I_Na density in "large" cells (Fig. 2E, ) but the current density is substantially greater in "large" cells from animals that exhibited autotomy (Fig. 2E, ). In "medium," "small-fast," and "small-slow" cells (Fig. 2, F, G, and H), total residual current density was increased by axotomy and increased further in neurons from animals that exhibited autotomy. Figure 2J shows the ratios of peak to residual current for control, axotomy (no autotomy), and autotomy (axotomized) cells. Decreases in the ratio reflect decreases in apparent inactivation. The observed increases in total residual current that were seen in all cell types (Fig. 2, E-H) could however reflect alterations in I_Na inactivation per se and/or altered expression of different Na⁺ channel types after axotomy. To distinguish between these possibilities, the characteristics of I_Na were analyzed in greater detail.

TTX-R and TTX-S Na⁺ channel currents in control cells

Because the slowly inactivating TTX-R and the rapidly inactivating TTX-S components of I_Na may be differentially affected by axotomy (Cummins and Waxman 1997), we next characterized these two components of the total current in our various cell populations.

The filled circles in Fig. 3, A-D, show the relationships between command voltage (V_c) and averaged, leak-subtracted, peak, TTX-S I_Na density for "large" (n = 7), "medium" (n = 7), "small-fast" (n = 6), and "small-slow" (n = 7) neurons. These were obtained by subtracting the recordings of TTX-R I_Na from the recordings of total I_Na in each cell. Figure 3, E-H, shows similar relationships for the TTX-R I_Na density in the same groups of cells. All data were obtained from V_h = 90 mV. In confirmation of previous studies, TTX-R I_Na is greatest in "small-slow" and least in "large" cells. It accounts for 6.7% of the total current in "large" cells, 8.1% in "medium" cells, 10.4% in "small-fast" cells, and 80.7% in "small-slow" cells (see Fig. 5). Also, as previously demonstrated (Elliott and Elliott 1993; Ikeda et al. 1986), maximum TTX-R I_Na occurred at relatively positive voltages. Thus maximum TTX-R currents were seen at 0 mV compared with 10 mV for TTX-S current.

View larger version (30K): [in this window] [in a new window]   Fig. 3. Relationship between command voltage (Vc) and peak, tetrodotoxin-sensitive (TTX-S), and tetrodotoxin-resistant (TTX-R) INa density recorded from a holding potential (Vh) of 90 mV. A-D: peak TTX-S current density data for "large," "medium," "small-fast," and "small-slow" cells, respectively. E-H: peak TTX-R current density data for "large," "medium," "small-fast," and "small-slow" cells. The 3 lines in each panel represent currents recorded from () control neurons, () axotomized neurons from animals that did not exhibit autotomy, and () axotomized neurons from animals that exhibited autotomy. Note that data for TTX-R INa from control "large" cells is eclipsed by the data points for axotomized cells in E. For TTX-R and TTX-S INa in "large," "medium," "small-fast," and "small-slow" control cells, n = 7, 7, 6, and 7, respectively. The corresponding n's for "large," "medium," "small-fast," and "small-slow" axotomized neurons are 6, 8, 5, and 7, respectively, and 7, 8, 6, and 7, respectively, for neurons from animals that exhibited axotomy. In some cases, error bars which indicate SD are shorter than symbols used to designate the data points.

Also, in confirmation of previous studies (Ikeda et al. 1986), TTX-S g_Na inactivated much more rapidly than TTX-R g_Na. Thus in "large," "medium," and "small-fast" cells, the peak TTX-S I_Na was 50-100 times greater than the residual current flowing at the end of a 10-ms voltage command. By contrast, the TTX-S I_Na in "small-slow" cells inactivated less, and the ratio of peak to end-of-pulse current was about 5. The time constant (_h1) for TTX-S I_Na inactivation at 10 mV was 0.45 ± 0.04 ms (n = 7) for "large" cells, 0.42 ± 0.05 ms for "medium" cells (n = 7), and 0.38 ± 0.01 (n = 6) for "small-fast" cells (Table 1). The value for "large" cells is similar to the value of 0.44-0.53 ms reported by Cummins and Waxman (1997). In most cells, however, we noted the presence of a slower and smaller component of inactivation that we termed _h2. For "large" cells, _h2 at 10 mV was 2.4 ± 0.31 ms (n = 6; Table 1). For "medium" cells, _h2 was 2.96 ± 0.16 ms (n = 6) and for "small-fast" cells was 3.54 ± 0.23 ms (n = 6). TTX-S I_Na in "small-slow" cells was too small to permit accurate curve fitting.

The time constant for TTX-R I_Na inactivation (_hR) at 0 mV was 4.51 ± 0.29 ms for "small-slow" cells (n = 7, Table 1). This is again similar to value of approximately 4.7 ms reported in identified C-cells (small cells) by Cummins and Waxman (1997). TTX-R I_Na was too small in "large," "medium," and "small-fast" cells to allow accurate determination of _hR.

Typical recordings of the TTX-S and the TTX-R components of I_Na from control "large," "medium," "small-fast," and "small-slow" neurons (at 8 mV) are shown in Fig. 4, A-D, respectively. For each cell, I_Na was recorded before and after application of 1 µM TTX. This yielded total I_Na and TTX-R I_Na. The illustrated records of TTX-S I_Na were obtained by subtraction. In some of the cells illustrated, the extracellular concentration of TTX was increased to 10 µM to confirm the TTX insensitivity of the recorded currents. The "large" cell that is illustrated in Fig. 4A exhibits only TTX-S I_Na, whereas the "small-slow" cell, illustrated in Fig. 4D, exhibits almost exclusively TTX-R I_Na. The "medium" and "small-fast" cells (Fig. 4, B and C) exhibit small amounts of TTX-R I_Na that are much less than that seen in the "small-slow" cell (Fig. 4D). These records are thus representative of the average distribution of TTX-R and TTX-S I_Na across control cells of all four types shown in the I-V plots of Fig. 3. Percentage contributions of TTX-R I_Na to the total I_Na are summarized for "small-slow," "small-fast," "medium," and "large" cells in Fig. 5.

View larger version (27K): [in this window] [in a new window]   Fig. 4. Sample recordings of TTX-S INa (at 10 mV) and TTX-R INa (at 0 mV; Vh = 90 mV). Recordings of membrane voltage omitted for clarity. A-D: currents from typical control "large," "medium," "small-fast," and "small-slow" cells. E-H: currents from typical axotomized "large," "medium," "small-fast," and "small-slow" cells from animals that did not exhibit autotomy. I-L: currents from typical "large," "medium," "small-fast," and "small-slow" cells from animals that exhibited autotomy after sciatic nerve lesion.

View larger version (34K): [in this window] [in a new window]   Fig. 5. Histogram to show percentage contribution of TTX-R INa to total INa in "large," "medium," "small-fast," and "small-slow" cells in the 3 experimental situations. Mean currents for the various experimental situations (as illustrated in Fig. 3) were originally calculated and displayed with SD. Since the data presented here were derived from the ratios of these mean currents, it is difficult to obtain an accurate estimate of error. Thus no error bars are presented.

Effects of axotomy on Na⁺ channel currents in "large" cells

Axotomy had little effect on peak TTX-S I_Na recorded at 10 mV in "large" cells. (n = 6, P > 0.15, Fig. 3A, ). This current was, however, significantly greater in animals that exhibited autotomy. In this situation, TTX-S I_Na was increased by 56% compared with control (n = 7, P < 0.001, Fig. 3A, ). Currents recorded in animals that exhibited autotomy were therefore significantly greater than those from the axotomy (no autotomy) group (P < 0.001). There was also a decrease in the rate of inactivation of TTX-S I_Na for "large" cells in the autotomy group. In this group, _h1 was twice the control value (P < 0.01, n = 6) and _h2 was increased by 60% (P < 0.05 compared with control; Table 1). Despite these changes, the rate of decay of TTX-S I_Na in "large" cells from the autotomy group is still too rapid to have much effect on total residual current seen at the end of a 10-ms command (Fig. 2E).

Axotomy-induced changes in TTX-R I_Na were similar to those seen with TTX-S I_Na. Thus TTX-R I_Na recorded at 0 mV in axotomized "large" cells (from animals that did not display autotomy) was similar to control (n = 7, P > 0.6) and the I-V plots derived from control and axotomized cells for TTX-R I_Na superimpose exactly ( eclipses  in Fig. 3E). There was a 10-fold increase in TTX-R I_Na in "large" cells from animals that exhibited autotomy (n = 6, P < 0.001, Fig. 3E, ). This was much greater than the corresponding percentage increase in TTX-S I_Na. As might be expected, TTX-R I_Na recorded in animals that exhibited autotomy was significantly greater than that from the axotomy group (P < 0.001). It was not possible to determine whether there was a change in the rate of inactivation of TTX-R I_Na (_hr) because this current was too small in control "large" cells for accurate curve fitting. Whereas TTX-R I_Na accounted for approximately 6% of the current in control and axotomized "large" cells, this fraction was increased to approximately 24% in "large" cells from animals that exhibited autotomy (Fig. 5). It is likely therefore that the increase in residual total I_Na in cells from the autotomy group (Fig. 2E) reflects increased contribution of slowly inactivating TTX-R I_Na.

The typical recordings of TTX-R and TTX-S I_Na illustrated for the "large" axotomized cell in Fig. 4E are thus similar to those seen in the control "large" cell (Fig. 4A). By contrast, there is much more TTX-R I_Na in the "large" cell from an animal that exhibited autotomy (Fig. 4I). TTX-S I_Na is also increased and there is a clear slowing of inactivation; for the neuron illustrated in Fig. 4I, _h1 = 0.79 ms compared with control _h1 = 0.55 ms (Fig. 4A).

Effects of axotomy on Na⁺ channel currents in "medium" cells

The electrophysiological properties of "medium" cells are assumed to lay between those of "large" and "small-slow" cells (Abdulla and Smith 2001a,b). Axotomy produced a modest (26%) increase in peak TTX-S I_Na and a larger (170%) increase in TTX-R I_Na measured at 0 mV in "medium " cells (P < 0.001 and n = 8 for both, Fig. 3, B and F, ). Both currents were further increased in animals that exhibited autotomy. In this situation, TTX-S I_Na was increased by 61% (n = 8, P < 0.001, Fig. 3B, ) and TTX-R I_Na was increased sixfold (n = 8, P < 0.001, Fig. 3F, ) compared with control. Currents recorded in animals that exhibited autotomy were also significantly greater than those from the axotomy group (P < 0.001 for both TTX-S and TTX-R I_Na). Whereas TTX-R I_Na accounted for approximately 8% of the current in control "medium" cells, this fraction was increased to approximately 15% after axotomy and to approximately 23% in cells from animals that exhibited autotomy (Fig. 5).

There was also significant slowing of inactivation of TTX-S g_Na (_h1 and _h2) in "medium" cells after axotomy, and further significant increases in these parameters in animals that exhibited autotomy (Table 1). Despite this, _h1 only attained values of 0.81 ± 0.1 ms after axotomy and 1.31 ± 0.1 ms in the autotomy group. These values are too small to account for the observed increase in total residual I_Na seen at the end of a 10-ms voltage command (Fig. 2F). Since _h2 accounts for only a small part of the total inactivation, increased residual I_Na in "medium" cells from animals after axotomy and the further increase after autotomy may be attributable to increased expression of TTX-R I_Na (Fig. 3F). It was not possible to determine whether any changes occurred in _hr because TTX-R I_Na was so small in control and axotomized "medium" cells that accurate curve fitting was precluded.

As already mentioned, the "medium" cell population probably included the AD cells identified in our previous studies (Abdulla and Smith 2001a,b). The identification of AD cells was precluded because the solutions used to study I_Na contained 4 mM Mg²⁺ and no extracellular Ca²⁺. This meant that the T-type Ca²⁺ current, which is a defining characteristic of AD cells, would not have been seen under the present experimental conditions.

Typical recordings of TTX-R and TTX-S I_Na from control, axotomy (no autotomy), and autotomy (axotomized) group "medium" cells are shown in Fig. 4, B, F, and J, respectively.

Effects of axotomy on Na⁺ channel currents in "small-fast" cells

"Small-fast" cells represent a category of DRG neuron that was not identified in our previous studies (Abdulla and Smith 2001a,b), but which likely correspond to "type A and/or D" "small" DRG cells defined by Rush et al. (1998). Apart from some subtle differences in the pattern of alterations of _h1 and _h2, changes seen in "small-fast" cells after axotomy and in the presence of autotomy were similar to those in "medium" cells. Axotomy increased peak TTX-S I_Na density (at 10 mV) or TTX-R I_Na density (at 0 mV) in "small-fast" cells (by 63% and 230% respectively; Fig. 3, C and G, ; P < 0.0001 and n = 5 compared with control for both). Both currents were further increased in the autotomy group. In this situation, TTX-S I_Na was increased by 91% (n = 6, P < 0.001, Fig. 3C, ) and TTX-R I_Na was increased 5.8-fold (n = 6, P < 0.001, Fig. 3G, ) compared with control. Currents recorded in animals that exhibited autotomy were significantly greater than those from the axotomy group (P < 0.0001 for TTX-R and P < 0.001 for TTX-S I_Na). Whereas TTX-R I_Na accounted for approximately 10% of the current in control "small-fast" cells, this fraction was increased to approximately 18% after axotomy and to approximately 25% in cells from animals that exhibited autotomy (Fig. 5).

The rate of inactivation of TTX-S I_Na in "small-fast" cells was slowed after axotomy as _h1 was significantly increased, whereas _h2 was unchanged (Table 1). Little further change of _h1 was seen in the autotomy group. It was not possible to determine whether any changes occurred in _hr because TTX_R I_Na was so small in control, "small-fast" cells that accurate curve fitting was precluded.

The typical recordings of TTX-R and TTX-S I_Na illustrated for the "small-fast" cells in Fig. 4, C and G, illustrate the increase in TTX-S and TTX-R I_Na that occur after axotomy. Additional TTX-R I_Na is illustrated in the "small-fast" cell from an animal that exhibited autotomy (Fig. 4K). TTX-S I_Na is also increased and the conductance inactivates more slowly (for the cell illustrated in Fig. 4K, _h1 = 2.2 ms compared with the control cell, Fig. 4C where _h1 = 0.54 ms).

Effects of axotomy on Na⁺ channel currents in "small-slow" cells

"Small-slow" cells likely correspond to type B and/or C small DRG cells defined by Rush et al. (1998). Axotomy doubled peak TTX-S I_Na density (at 10 mV) and increased TTX-R I_Na density (at 0 mV) in "small-slow" cells by 67% (Fig. 3, D and H, ; n = 7 and P < 0.0001 compared with control for both). TTX-S I_Na was further increased to three times control amplitude in the autotomy group (n = 7, P < 0.001, Fig. 3D, ), whereas no further change occurred in TTX-R I_Na (n = 7, P > 0.5 for autotomy compared with axotomy, Fig. 3H, ). Whereas TTX-R I_Na accounted for approximately 82% of the current in control "small-slow" cells, this fraction was unchanged (approximately 81%) after axotomy and was slightly decreased to approximately 76% in cells from animals that exhibited autotomy (Fig. 5).

The inactivation time constant for TTX-R I_Na (_hr) was significantly increased after axotomy and was increased further in the autotomy group (Table 1). Thus the increased expression of TTX-R I_Na with slowed inactivation may contribute to the increase in total residual I_Na seen in axotomized "small-slow" cells and in "small-slow" cells from animals that exhibited autotomy (Fig. 2H). The small amplitude of TTX-S I_Na in control "small-slow" cells precluded measurement of _h1 and _h2 for this group.

The typical recordings of TTX-R and TTX-S I_Na illustrated for the "small-slow" axotomized cell in Fig. 4H thus exhibit much more TTX-S and TTX-R I_Na than the control "small-slow cell" (Fig. 4D). The further increase in TTX-S I_Na seen in animals that exhibit autotomy is also clear from the neuron illustrated in Fig. 4I.

Time course of changes

Changes in the electrophysiological properties of axotomized neurons progress with time after injury (Cummins and Waxman 1997; Govrin-Lippmann and Devor 1978; Wall and Devor 1983), as does the incidence and the severity of autotomy (Abdulla and Smith 2001a; Coderre et al. 1986; Wall et al. 1979). Since all the data described above are pooled from animals 2-7 wk after axotomy, it is possible that animals assigned to the "axotomy nonautotomy" group are simply those studied at short time interval after axotomy. Similarly, those assigned to the "autotomy (axotomized)" group may be those studied at longer periods after axotomy. Thus the increase in g_Na seen in "large" cells may be a simple function of time after axotomy as opposed to a characteristic of DRG neurons from animals that are exhibiting autotomy. To address this possible bias in our data, we compared data pooled from animals at 2 and 4 wk after axotomy. In our experimental situation, 45% of animals exhibit autotomy at 2 wk after axotomy, and at 4 wk, 75% of animals exhibit autotomy (Abdulla and Smith 2001a). As shown in Fig. 6, there is a significant (P < 0.01) difference between peak total I_Na amplitude between axotomy (no autotomy) and autotomy (axotomized) groups at both 2 and 4 wk. There is, however, no difference in I_Na density for the axotomy (no autotomy) group at 2 wk and that at 4 wk. Moreover, there is no difference between the autotomy groups at the two time intervals (n = 10-12 for all groups). This suggests that the density of I_Na is related more to the presence or absence of autotomy than to the time after axotomy was carried out.

View larger version (21K): [in this window] [in a new window]   Fig. 6. Comparisons of total, peak INa density in "large" cells from the axotomy (no autotomy) and autotomy (axotomized) group at 2 and 4 wk after axotomy. Current densities recorded in axotomized cells 2 wk after axotomy (n = 12) were not significantly different from those recorded in cells 4 wk after axotomy (n = 10). Similarly, current densities in cells from animals that exhibited autotomy after 2 wk (n = 11) were not significantly different from those recorded in cells from animals that exhibited autotomy at 4 wk after axotomy (n = 11). However, at both 2 and 4 wk after axotomy, current densities recorded from animals that exhibited autotomy (dark gray bars in histogram) were greater than those recorded from animals that did not exhibit autotomy (light gray bars in histogram; P < 0.01, ANOVA, Student-Newman-Keul's test). Error bars indicate SD.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

General trends in the data

The main findings of this study are summarized in Fig. 7, A and B. We suggest that the properties of "small-fast" cells lie between those of "medium" cells and "small-slow" cells. Several trends may be identified in comparing data across a continuum of "large," "medium," "small-fast," and "small-slow" cells.

View larger version (13K): [in this window] [in a new window]   Fig. 7. A: chart to illustrate axotomy-induced percentage increase of TTX-R and TTX-S INa for the 4 cell types from animals that had been subject to sciatic nerve axotomy but which did not exhibit autotomy. Note lack of change of both currents following axotomy of "large" cells. B: chart to illustrate percentage increase of TTX-R and TTX-S INa for the 4 cell types in animals that exhibited autotomy after sciatic nerve axotomy. Note marked increase in both currents in "large" cells from animals that exhibited autotomy. Mean currents for the various experimental situations (as illustrated in Fig. 3) were originally calculated and displayed with SD. Since the data presented here were derived from the ratios of these mean currents, it is difficult to obtain an accurate estimate of error. Thus no error bars are presented.

First, TTX-S and/or TTX-R I_Na tended to increase after axotomy or in the presence of autotomy (Fig. 3). No decreases in either type of current were seen in any cell type. There was also a trend toward slowed inactivation of both TTX-R and TTX-S I_Na after axotomy and in the autotomy group (Table 1). Axotomy-induced increases in TTX-S I_Na or in Na⁺ channel gene expression have been reported in a variety of cell types including cutaneous afferent neurons in rat DRG (Rizzo et al. 1995), rat facial motoneurons (Iwahashi et al. 1994), frog sympathetic ganglion cells (Jassar et al. 1993), and in cat motoneurons (Sernagor et al. 1986; Titmus and Faber 1990).

Second, the changes promoted by axotomy were usually expressed more clearly in cells from the animals that exhibited autotomy. This trend has been noted in our previous experiments on axotomy-induced changes in excitability (Abdulla and Smith 2001a) and in Ca²⁺ and K⁺ channel function (Abdulla and Smith 2001b).

Third, we have previously noted that the presence of autotomy correlates with a shift in the properties of "large" cells (Abdulla and Smith 2001a,b). This generalization was supported by the observation that axotomy per se failed to affect TTX-S I_Na in "large" cells (Figs. 3A and 7A), but in animals where autotomy occurred, the current was 56% larger than control (Figs. 3A and 7B). This pattern was exaggerated for TTX-R I_Na (Fig. 7A). Although this current was not increased in "large" cells by axotomy alone (Figs. 3E and 7A), the amplitude of TTX-R I_Na was 900% of its control value in "large" cells from animals that exhibited autotomy (Figs. 3E and 7B). Autotomy therefore seems to correlate with a change in properties of nonnociceptive cells. This suggestion helps to explain the observation that destruction of peripheral C-fibers with capsaicin fails to prevent the autotomy induced by a prior nerve injury (Nagy et al. 1986). It is also consistent with the finding that tactile allodynia can be produced after nerve injury in the absence of C-fiber activation (Liu et al. 2000). Last, it fits with the clinical observation that large myelinated afferents seem to signal the mechanical hyperalgesia associated with nerve injury (Campbell et al. 1988). It has been suggested that nerve injury promotes sprouting of A-fibers, which normally signal innocuous information, into the substantia gelatinosa of the spinal cord (Woolf et al. 1992). Since the substantia gelatinosa is a major site of processing nociceptive information, normally innocuous or spontaneous A-fiber information may be perceived as pain. Moreover, persistent activity in A-fibers may contribute to the establishment of dorsal horn sensitization, which is an essential feature of neuropathic pain (Liu et al. 2001). Our findings thus complement the idea that changes in the properties of "large" DRG neurons, which exhibit brief, noninflected action potentials, are responsible for both the onset of autotomy and the induction of chronic pain (Liu et al. 2001).

Last, changes in the types of channels expressed may reflect an injury-induced de-differentiation process (Cummins and Waxman 1997; Kuno et al. 1974; Waxman et al. 1994). In other words, "large" cells tend to acquire the characteristics of "small" cells and vice versa. This is especially clear for the changes accompanying autotomy that are summarized in Fig. 7B. The aforementioned 900% increase in TTX-R I_Na seen in "large" cells implies that they start to acquire a Na⁺ channel phenotype that is characteristic of "small-slow" cells. Although there is a relatively modest increase (67%; Figs. 3H and 7B) in TTX-R I_Na in "small-slow" cells, there is a 200% increase in TTX-S I_Na (Figs. 3D and 7B). "Small-slow" cells may thus start to acquire the Na⁺ channel phenotype that is characteristic of "large" cells. In terms of percentage contribution of TTX-R I_Na to the total current, there is a net increase in TTX-R I_Na in "large" cells (6.6% in controls compared with 23.6% in the autotomy group, Fig. 5) and a slight loss of TTX-R I_Na in "small-slow" cells (82.3% in controls compared with 76.3% in the autotomy group, Fig. 5).

Relationship to changes in total residual I_Na

The increases in total residual I_Na at the end of a 10-ms command that accompany axotomy or the presence of autotomy (Fig. 2, EH) likely reflect the overall increase in expression of TTX-R I_Na in all cell types (Fig. 7, A and B) and its slowed inactivation (Table 1). Although inactivation of TTX-R I_Na is clearly slowed in "small-slow" cells (Table 1), it was not feasible to determine whether similar changes occurred in "large," "medium," or "small-fast" cells. This was because the current was too small in the control populations to allow measurement of the control values for _hr. The increase in _h1 for TTX-S I_Na is unlikely to have contributed to increases in residual current seen in any of the cell types either after axotomy or in animals that exhibited autotomy (Fig. 2, E-H). Although the value of _h1 in "large" cells from the autotomy group was twice that from the controls (Table 1), the slowed current would have decayed to zero well before the end of a 10-ms pulse. A similar situation obtains for "medium" cells; here _h1 was threefold greater in the autotomy group than in controls, yet the maximum value obtained (1.31 ± 0.1 ms) was still too small to have much effect on total residual I_Na (Fig. 2F).

Relationship to previous current-clamp findings

Axotomy-induced increases in I_Na and slowed inactivation correspond to the increase in spike width and height, the increase in excitability, and the decrease in AP threshold noted in current-clamp studies of DRG neurons (Abdulla and Smith 2001a; Gallego and Eyzaguirre 1978; Gurtu and Smith 1988; Kim et al. 1998; Stebbing et al. 1999; Study and Kral 1996). The correlation between changes in Na⁺ currents and changes in spike shape, rheobase, and excitability is clearly apparent across the whole sensory neuron population. Thus axotomy alone has little or no effect on I_Na in "large" cells (Fig. 7A), and this correlates with its general lack of effect on spike height and spike width in these cells (Abdulla and Smith 2001a). By contrast, axotomy has much greater effects on I_Na in both "small-fast" and "small-slow" cells (Fig. 7A), and this correlates well with axotomy-induced increases in spike height and width and lowering of rheobase in all "small cells" (Abdulla and Smith 2001a). Moreover, the presence of autotomy coincides with an increase in spike height and width in "large" cells and a reduction in rheobase (Abdulla and Smith 2001a); this is mirrored by changes in I_Na under this circumstance in "large" cells.

When all currents are considered, it is actually quite difficult to explain why axotomy increases spike width (Abdulla and Smith 2001a). In "medium" or "small" cells, suppression of Ca²⁺ current with Cd²⁺ or noradrenaline decreases spike width, whereas increasing Ca²⁺ influx with BAYK 8644 increases spike width (Abdulla and Smith 1997a). This is consistent with the idea that Ca²⁺ current underlies the "hump" or "shoulder" on the repolarizing phase of the AP (Ikeda et al. 1986). It would be predicted therefore that axotomy-induced attenuation of Ca²⁺ influx (Abdulla and Smith 2001b; Baccei and Kocsis 2000) would attenuate the shoulder and thereby shorten rather than lengthen the AP. In preliminary modeling studies of APs in "small" DRG neurons, we find that AP shape can only be predicted if it is assumed that the shoulder of the AP depends not only on Ca²⁺ influx, but also on the persistence of g_Na (P. S. Pennefather, P. A. Smith, and F. A. Abdulla, unpublished observations). It will be recalled that 80% of I_Na in "small-slow" cells is TTX-R (Fig. 5) and the underlying conductance inactivates slowly (Table 1), yielding a large persistent I_Na in these cells (Fig. 2H). We find from the model that a slight slowing of inactivation, as would be seen after axotomy, results in increased spike width even when g_Ca is reduced. It is likely therefore that the increased spike width seen after axotomy reflects slowed inactivation of TTX-R g_Na.

Relationship to other studies

Since as many as 30% of neurons in L₄ and L₅ DRG have afferent fibers in nerves other than the sciatic (Himes and Tessler 1989), it is likely that some of the cells studied in the axotomy or autotomy groups did not actually have severed axons. Despite this, the SDs on the data shown in Figs. 2, A-H, and 3, A-H, seem quite small. We have no simple explanation for this lack of variability. One possibility is that the axotomy-induced changes do not reflect effects of axon interruption per se but may rather reflect exposure of surviving axons to a environment in which Wallarian degeneration is occurring (Wu et al. 2001).

Our findings also appear to differ from those reported by Cummins and Waxman (1997) and by Sleeper et al. (2000). These authors reported that axotomy of identified nociceptive C-type neurons resulted in down-regulation of TTX-R I_Na (Na_v1.8 or SNS gene product) and of persistent TTX-R I_Na (Na_v1.9 or NaN gene product) and the appearance of a "rapidly repriming" embryonic TTX-S, -III type Na⁺ channel (Na_v1.3). This led to an overall increase in the rate of inactivation of the total I_Na and the appearance of a g_Na in axotomized C-neurons that readily recovered from inactivation. There is no simple explanation for the difference between these findings and our observation of slowed inactivation and increased expression of TTX-R I_Na after axotomy. Differences in the methodology and in the interpretation and presentation of the data between our studies and those of Cummins and Waxman (1997) and by Sleeper et al. (2000) should therefore be noted.

First, the experimental lesion was different; while we used a simple nerve cut, Cummins and Waxman (1997) and Sleeper et al. (2000) used ligation, cutting, and insertion of the proximal nerve cut into a silicone cuff. If more inflammation was induced in the model we used, this could contribute to the differences in Na⁺ channel expression seen in the two studies. This is because inflammation and inflammatory mediators can increase Na⁺ channel expression (Baker and Wood 2001; Gould et al. 1998; Khasar et al. 1998; Tanaka et al. 1998; Waxman et al. 1999, 2000).

Second, our data were collected from freshly isolated cells, whereas the other two studies used control and axotomized neurons that were maintained in culture for 12-24 h before study. Although the practice of using short-term cultures of control and axotomized DRG neurons (Cummins and Waxman 1997) has certain advantages in that the cells studied have an opportunity to "recover" from the acute dissociation procedure, culturing may invoke other changes. We have shown, for example, that total I_Na density in bullfrog sympathetic B-neurons doubles within 24 h of culture (Lei et al. 2001). This may reflect accumulation in the cell body of newly synthesized TTX-S and TTX-R Na⁺ channels destined for translocation into the axon (Novakovic et al. 1998). The expression of TTX-R SNS Na⁺ channels in DRG is driven by nerve growth factor (NGF) (Black et al. 1997; Dib-Hajj et al. 1998) and axotomy decreases SNS-type sodium channel mRNA (for TTX-R I_Na) in sensory neurons (Okuse et al. 1997; Waxman et al. 1994). The loss of target-derived NGF when neurons are axotomized in vivo (Bongenhielm et al. 2000) would therefore restrict the accumulation of TTX-R channels when neurons from axotomized animals are placed in culture. By contrast, nonaxotomized, cultured cells would have had recent access to target-derived NGF in vivo so that they would accumulate more channels than axotomized neurons in culture. This might account for the observed reduction in TTX-R I_Na in DRG cells cultured from axotomized animals (Cummins and Waxman 1997). We suggest that the increase in TTX-R I_Na and TTX-S I_Na seen in our experiments may simply reflect increased accumulation of both types of Na⁺ channels in the cell body as a result of in vivo axotomy. This effect may reflect the loss of axons in vivo and perhaps inflammation at the site of injury that would tend to up-regulate Na⁺ channel expression (Baker and Wood 2001; Gould et al. 1998; Khasar et al. 1998; Tanaka et al. 1998; Waxman et al. 1999, 2000).

In summary, it may be supposed that three factors control TTX-R and TTX-S I_Na expression in DRG cells after axotomy: loss of retrograde supply of NGF, channel accumulation in the cell body, and up-regulation of channel expression by inflammatory mediators (Gold et al. 1996). Whereas our procedures emphasized channel accumulation in the cell body, perhaps as a consequence of inflammation-induced up-regulation of Na⁺ channels, those employed by Cummins and Waxman (1997) and Sleeper et al. (2000) may have emphasized the loss of effect of target-derived NGF. Recent evidence suggests however that the up-regulation of TTX-R I_Na seen with our procedures may be more relevant to the etiology of neuropathic pain. This is because selective "knock-down" of TTX-R I_Na expression with specific Na_V1.8 antisense oligonucleotides has an inhibitory effect on nerve-injury induced neuropathic pain phenomena in rats (Lai et al. 2002).

The essential lesson from this and our previous voltage-clamp study (Abdulla and Smith, 2001b) is that axotomy-induced changes in ion channel properties, increased I_Na and decreased I_K and I_Ca, are all in a direction that would tend to increase excitability. We have now completed a study of axotomy-induced changes in excitability (Abdulla and Smith 2001a). Since the same populations of neurons were used in those two and in the present study and all neurons studied were subjected to the same type of injury, we are now in a position to relate the current-clamp findings to the voltage-clamp studies. This will be the subject of future computer modeling studies in which we will attempt to account for the alterations in the repetitive discharge characteristics of axotomized neurons in terms of alterations in ion channel properties.