Sympathetic activation by the cold pressor test does not increase the muscle force generation

26 A positive inotropic action by the sympathetic nervous system on skeletal muscles has been 27 observed and investigated in animal and in-vitro studies. This action provided a theoretical 28 basis for the putative ergogenic action of catecholamines and adrenergic agonists, although 29 there is no clear evidence of this effect in humans. The aim of this study was to investigate 30 the occurrence of inotropic effects associated to physiological sympathetic activation in 31 healthy subjects. The muscle force capacity was investigated in the tibialis anterior (n = 9 32 subjects) and in the soleus (n = 9) muscles electrically stimulated with single pulses, double 33 pulses with variable inter-spike interval (ISI: 4-1000 ms) and short pulse trains (frequency: 34 5-14 Hz) before, during and after sympathetic activation by the cold pressor test (CPT). 35 CPT significantly decreased by 10.4±7.2 % and 10.6 ± 4.4 % the force produced by single 36 and double pulse stimulation, respectively, and produced smaller decreases in the force 37 obtained by train stimulation in the tibialis anterior while no significant changes were 38 observed in either type of contraction in the soleus muscle. CPT failed to induce any 39 increase in the force capacity of the investigated muscles. The prevalent decrease in force 40 evidenced in this study support the concept that the weakening sympathetic action on type-I 41 fiber, already shown to occur in humans, prevails over the putative potentiating action. 42


INTRODUCTION 46
Sympathetic activity is known to support motor function by acting at different levels, including the cardiovascular, respiratory, and motor systems. The release of 48 catecholamines in the blood accompanies physical exercise depending on its extent and 49 duration (39), and catecholamine outflow was found to be correlated with motor 50 performance (11). In addition, administration of sympathomimetics, particularly beta2 51 adrenergic agonists, such as salbutamol, have been shown to improve motor performance in 52 different types of tasks (8,20,32,34). To explain this ergogenic effect, a specific 53 potentiating action on skeletal muscle contractility is often invoked (36). Indeed, a positive 54 inotropic effect of epinephrine (EPI) and adrenergic agonists on skeletal muscles has been 55 well documented in anesthetized animals as well as in isolated muscles and fiber bundles. 56 This effect has been found to be mediated by beta2 adrenergic receptors leading to 57 increased Ca release from the sarcoplasmatic reticulum (1,3,5,12,35). There are, 58 however, studies in which administration of EPI or beta2 agonists failed to induce 59 contractility potentiation (21) or improvement in motor performance (7,19). 60 One possible factor behind these conflicting results may be the complexity of the 61 adrenergic action, differentially affecting the contractile machinery of type-I and type-II 62 muscle fibers. A positive inotropic effect is indeed mainly exhibited by type-II fibers 63 7 EMG, torque, and VAS were concurrently sampled (12-bit A/D conversion, 2 kHz 137 sampling frequency) and stored on a PC. 138

Electrical stimulation 139
Electrical stimulation was provided by a voltage-controlled current source 140 stimulator (NoxiSTIM; JNI Biomedical A/S, Aalborg, Denmark). In study 1, the 141 stimulation of the tibialis anterior muscle was obtained by stimulating the peroneal nerve, 142 the cathode (electrode diameter: 2 cm) being placed just above the fibula neck and the 143 anode (3 x 3 cm) at the patella. In study 2, the calf muscle was stimulated by a cathode 144 electrode (diameter 2 cm) placed on the tibial nerve at the popliteal fossa. In order to 145 reduce the contribution of the gastrocnemius muscle, the knee was flexed at about 100°. 146 However, in this position the electrode nerve coupling is impaired, as compared to the 147 knee-extended position. Therefore, a custom device was fixed to the thigh and exerted an 148 adjustable pressure on the cathode electrode at the popliteal fossa in order to improve the 149 effectiveness and reliability of the nerve stimulation. The anode electrode (3 × 3 cm) was 150 placed at the patella. The anode position was adjusted in order to avoid unwanted 151 contractions of antagonist muscles during the stimulation, which was detected by 152 monitoring EMG activity on the tibialis anterior muscle. 153 For each subject the stimulation intensity evoking the maximum compound muscle 154 action potential was determined. However, in some cases, a supramaximal intensity of 155 stimulation was reported to be painful. Because it was important to avoid preventive pain-156 induced sympathetic activation related to the stimulation, in those cases we adopted the 157 maximum stimulation intensity which was non-painful.

8
The types of stimulation performed were single pulses, doublets, and pulse trains. 160 The stimulation pattern for single stimuli and doublets consisted of a sequence of 22 161 alternated single and paired pulses separated by 1-s interval. The paired pulses (doublets) 162 had an inter-spike interval ranging between 4 and 1000 ms (4, 8, 12, 15, 20, 30, 50, 75, 100, 163 125, 150, 175, 200, 225, 250, 300, 400, 500, 750, 1000 ms) according to a protocol 164 adopted in previous studies (17,18). This sequence of pulses was followed by 4 pulse 165 trains of 5 s in duration, separated by 5-s intervals. The pulse trains had frequencies 5, 8,166 10, and 12 Hz (Fig. 1B). 167 The set of stimulations was repeated seven times, corresponding to three control 168 conditions (C1, C2, C3), one condition of sympathetic activation (CPT), and three recovery 169 conditions (P1, P2, P3) (Fig. 1A). Each recording condition was separated by 5-min 170 intervals. 171 In the CPT condition, the left hand was immersed in iced water (3-4°C), stirred by a 172 peristaltic pump, for 4 min. The subjects could withdraw the hand from the water if the 173 pain became unbearable, in which case the data were excluded from the analysis. In study 174 2 only, in one of the control conditions the left hand was immersed in water at 32-35 °C 175 (neutral condition) and the sequence of the control and neutral conditions was randomized. 176 Systolic and diastolic blood pressures were measured during the control condition 177 just after C2, immediately after the CPT and during the recovery just after P2 (see black 178 dots in Fig 1A).
computed from the average of 21 single twitch torques, following the first stimulus. The 184 twitch torque elicited by the first stimulus was excluded because it was systematically 185 smaller than all the others. The same parameters were extracted from the doublet 186 stimulation for the second stimulus in each pair of stimuli. The PA value was identified as 187 the maximum torque increase (with respect to the pre-stimulation level) reached after the 188 stimulating pulse. Since no changes in nerve conduction velocity were expected, TTP was 189 more conveniently computed as the interval between the stimulation pulse and the time 190 instant corresponding to the torque peak, rather than between the onset of torque 191 development and the torque peak. HRT was computed as the interval between the torque 192 peak and the instant in which the torque was reduced to half its peak value. 193 From the pulse trains, the average torque and the amplitude of torque oscillations 194 were extracted. The average torque was computed by averaging the torque signal over the 195 last 1-s of stimulation during the pulse train. The amplitude of torque oscillation was 196 obtained as the peak-to-peak amplitude of the torque signal, as average value over the last 197 1-s of stimulation. These values were normalized with respect to the average of all 198 conditions before averaging over subjects. 199

Statistical analysis 200
For both experiments, non-parametric statistical analysis was adopted because the 201 normality tests failed for some of the analyzed variables (diastolic blood pressure in control 202 condition, experiment 1 and 2). The Kruskal-Wallis analysis of variance (ANOVA) and 203 Mann-Whitney U-test were used to compare blood pressure changes and VAS score in the 204 two studies. One way ANOVA for repeated measures was used to assess an effect of 205 condition (C1, C2, C3, CPT, P1, P2, P3) on the measured variables. When ANOVA was 206 significant (P < 0.05), pair-wise comparisons were tested by the Newman-Keuls post-hoc 207 test. Values are presented as mean and SD in the text and as mean and standard error of the 208 mean (SE) in the figures. CPT evoked a persistent painful sensation that outlasted the duration of the test. 219 The peak VAS score was 4.5 ± 1.9 (range: 2.7-8.7). The painful sensation vanished in all 220 subjects before P2. 221 CPT produced an increase in diastolic blood pressure from 73.8 ± 6.0 to 89.0 ± 9.2 222 mmHg (P <0.01) and systolic blood pressure from 109.0 ± 9.7 to 127.6 ± 11.0 mmHg (P < 223 0.01). Both variables returned to control values when reassessed, after P2 condition 224 (diastolic: 75.7 ± 5.6 mmHg; systolic: 113.2 ± 8.2 mmHg) 225 The effect of CPT on the torque twitch evoked by single electrical stimuli is 226 exemplified by the recordings from a representative subject (Fig. 1A). Group effects across 227 all subjects on AMP, TTP and HRT are shown in the bar diagrams in Fig. 2B-D. We 228 observed a slight increase in AMP during the three control conditions: C3 being higher then 229 C1 by 8.2 ± 9.1 % (P<0.01), possibly due to post contraction potentiation mechanisms. 230 However, AMP was significantly reduced by 10.4±7.2 % (P<0.01) during CPT with respect 231 C3. A gradual recovery of twitch amplitude was observed in the recovery conditions, with 232 P3 being significantly different from CPT and matching the value of C3. Conversely, no 233 significant changes were instead observed in the time course of TTP (116±13 ms in C1, 234 The analysis of the twitch torque produced by the second of two spikes 236 administered with variable ISI is shown in Figure 3. For ISI smaller than 20 ms, the two 237 twitches are fused together and the amplitude is almost independent of the ISI. With 238 increasing ISI above 20 ms, the two twitches begin to split and the amplitude of the second 239 one starts to fall (Fig. 3A). Above 300-400 ms the second twitch is completely separated 240 from the first one and its characteristics tend to approach the characteristics of the single 241 twitch described in Fig. 2. With ISI >30ms the two twitches are only partly fused and the 242 peak amplitude (detected after the second pulse) starts to decrease. The TTP and HRT also 243 exhibit a clear dependency on ISI. 244 The CPT influenced the amplitude of the peak torque produced by the second pulse 245 at all ISIs tested (thick line in Fig. 3B). The differences were significant for the average 246 peak amplitude of the first three doublets (ISI= 4, 6, 8 ms) which was reduced by 10.6 ± 4.4 247 % during the CPT with respect to C3 and was significantly different from all other 248 conditions (P<0.01, except vs. C1: P<0.05 ). TTP and HRT were not influenced by CPT

12
The contractions evoked by burst stimulations were analyzed in terms of the torque 251 reached at the end of the burst and of the amplitude of torque oscillations. On average, the 252 torque developed during CPT was lower than that developed in all other conditions for each 253 of the stimulation frequencies employed, however the significance level was not reached. 254 The amplitude of torque oscillation during burst stimulation at 5 Hz exhibited a similar 255 time course as the torque twitch amplitude, i.e., a slight increasing trend between C1 and 256 C3 (10 ± 12 %) and between P1 and P3 (9 ± 13 %) but with a decrease between C3 and 257 CPT (6 ± 13 %). Similar but less marked changes were observed at 8, 10 and 12 Hz 258 although none of these changes reached statistical significance. 259

Experiment 2 -Stimulation of the tibial nerve 260
One subject had to be excluded because of instability of the force recording and two 261 other subjects were discarded because of the presence of a H-reflex in response to the 262 electrical stimulation. The results are described for the 9 remaining subjects. 263 Electrical stimulation of the tibial nerve appeared to be relatively more painful than 264 stimulation of the peroneal nerve and, in order to avoid pain sensations associated to the 265 electrical stimulation, the intensity often had to be reduced below the one producing the 266 maximum M-wave (89 ± 10.3 %, range: 76 -106 %). 267 Hand immersion in water at neutral temperature did not evoke a pain sensation 268 (VAS= 0 in all subjects) while CPT evoked similar effects to those described for 269 experiment 1. The VAS score peaked at 5.6 ± 3.1 (range: 2.3-9.2) during the test and 270 returned to 0 in all subjects at P2. Diastolic blood pressure rose from 73.8 ± 6.0 to 89.0 ± 271 9.2 mmHg (P <0.01) and systolic blood pressure from 109.0 ± 9.7 to 127.6 ± 11.0 mmHg 272 (P < 0.01). Both variables returned to control values when reassessed, after P2 condition condition (average of C1-C3) was slower than in Experiment 1, as observed both for TTP 278 (124 ± 19 ms vs. 105 ± 11 ms; P<0.01) and for HRT (99 ± 23 ms vs. 71 ± 18 ms; P< 0.05). 279 In one of the control conditions the left hand was immersed in water at neutral temperature 280 and on average, the twitch parameters did not depend on the control condition. CPT did not 281 influence the twitch amplitude ( Also for this stimulation paradigm the parameters did not depend on the condition. As 287 observed for the single twitch, HRT was slightly but not significantly reduced at ISI < 50 288 ms (Fig. 5C). Absence of systematic changes in the time course of the contraction is also 289 confirmed by the absence of changes in the pattern of summation of the double-twitches, as 290 indicated by the curves in Fig. 5B. 291 The torque developed by burst stimulation was also unaffected by CPT both in 292 terms of average torque developed and of amplitude of torque oscillation. 293 Physiological sympathetic activation by CPT did not produce a potentiation of the 296 contraction in any of the tested muscles. Conversely a significant decrease in twitch 297 amplitude was observed in TA while only a trend towards twitch shortening was observed 298 in the calf muscles. This set of results indicates that a weakening rather than a potentiating 299 effect has been induced by sympathetic activation. 300

Potentiation vs. weakening 301
In-vitro studies (5, 12) have elucidated that adrenergic agonists may modulate the 302 contractile machinery of skeletal muscles in two ways: i) by increasing the reuptake of 303 Ca++ in the sarcoplasmatic reticulum, thus shortening the twitch duration (positive 304 lusitropism) and resulting in a weakening effect -this mechanisms being present in type-I 305 fibers only -and ii) by augmenting the release of Ca++ from the SR, thus producing a 306 twitch of bigger amplitude, which is a potentiation of the contraction -this mechanisms is 307 present in both fiber types although it has been observed mostly in fast-twitch muscles (1, 308 3, 4). These classic studies, performed on animal models, already pointed out that higher 309 doses of EPI or β 2 -agonist had to be administered to elicit a potentiating effect in fast-310 twitch muscles with respect to the dose required to elicit the weakening effect in slow-311 twitch muscles (1, 3, 4). This difference can be partly attributed to the fact that type-I fibers 312 have a higher density of adrenergic receptor than type-II (15, 22). On the other hand, it is a 313 widely held view that the sympathetic nervous system potentiates the contraction of skeletal 314 muscles (9,11,33,36). This idea fits well with the other actions that the sympathetic 315 nervous system exerts, particularly on the cardiovascular system, to support intense muscle 316 work, and is appropriate in a context of fight or flight. However, it must be emphasized that 317 no human study currently evidenced the occurrence of sympathetic-mediated potentiation 318 of skeletal muscles. Moreover many animal and in-vitro studies that report catecholamine-319 mediated potentiation refer to muscles that were previously fatigued (3, 23) or to muscle 320 fibers immersed in a iperkalemic medium (13). The force potentiation of fatigued muscles, 321 also called "anti-fatigue" effect, is based on the recovery of cell excitability by EPI-induced 322 potentiation of the Na/K pump (3, 13, 29), and does not mediate the positive inotropic 323 effect observed in resting fast-twitch muscles (1, 3, 4). 324 In the present study, CPT failed to induce any potentiation in either TA or calf 325 muscles, although the same stressor was adequate to induce the weakening effect in low 326 threshold, presumably type-I, motor units of the TA (28). This supports the concept that the 327 positive inotropic effect has a higher threshold of activation than the weakening effect. 328 These results also support and integrate the only investigation in humans about the effects 329 of exogenous (not spontaneously released, as in the present study) EPI on muscle 330 contractility by Marsden & Meadows (21). Although their interest was mostly focused on 331 the tremor-genic action of EPI, the authors evidenced a weakening adrenergic effect in both 332 the calf muscles (5 subjects) and the adductor pollicis (3 subjects). In particular, they 333 showed a reduction in HRT of the twitch force in the calf muscles (~15%), no significant 334 effect on the twitch force of adductor pollicis but a decrease in the force of subtetanic 335 contraction (10Hz stimulation), as we did observe for TA. 336 The protocol adopted in the present study included electrical stimulation by paired 337 stimuli at varying inter-spike interval within the range 4-1000 ms. This stimulation pattern 338 was previously employed for the investigation of the velocity recovery function of muscle 339 fibers (18) and of twitch summation (17). It was adopted in the present study for two 340 reasons: 1) the response to the doublet at short ISI is stronger than the single twitch and thus provides an improved signal-to-noise ratio for the detection of changes in muscle 342 contractility; 2) possible increase/decrease in twitch duration, resulting in 343 increased/decreased twitch fusion, would have been evidenced by rightward/leftward shift 344 of the torque amplitude vs. ISI curve in this stimulation pattern. 345 The reduction in single twitch amplitude in TA was confirmed by the reduction in 346 the response to paired stimuli (4<ISI<30, Fig. 3B) as well as by a reduction in the torque 347 developed by burst stimulation at the different frequencies. This supports the interpretation 348 that sympathetic activation by CPT produced a weakening effect. In fact, in many animal 349 studies a marked decrease of the twitch amplitude was observed in response to EPI 350 injection, as a consequence of the lusitropic effect occurring in type-I fibers (1, 3, 4). It is 351 possible that, this latter effect was masked in the present study, due to the co-activation of 352 unresponsive or differently-responding type-II fibers in TA. A decrease in HRT was instead 353 observed in response to CPT-induced sympathetic activation in our previous study where 354 single, low-threshold, presumably type-I motor units were investigated (28) while an 355 increased HRT was observed in the TA of healthy subjects in response to blockade of β-356 adrenergic receptors (2). 357 In the soleus muscle, the reduction in HRT observed on the single twitch was also 358 observed in response to close paired stimuli (4≤ISI≥30, Fig. 5D), although the effect was 359 probably too weak to reach statistical significance and to produce appreciable changes in 360 the amplitude-vs-ISI curve as well as in the burst contractions. 361 Electrical stimulation does not allow selective recruitment of type II muscle fibers. 362 Therefore the possibility exists that a potentiating effect occurring in type-II fibers was 363 canceled by concomitant weakening effects in type-I fibers when the muscle is composed of a balanced proportion of the two types of fibers. In fact, Bowman and Zaimis (4) 365 observed clear cut potentiation in the fast-twitch tibialis muscle and marked weakening in 366 the slow-twitch soleus muscle of the cat, intravenously injected with EPI, while minor 367 effects were observed in plantaris and gastrocnemius muscles characterized by a more 368 balanced fiber-type composition. In humans, both TA and soleus muscles have a 369 preponderance of type-I fibers, so the possibility cannot be excluded that potentiation 370 effects have been canceled by weakening effects occurring in these fibers. On the other 371 hand, selective activation of type II fibers is also unlikely to occur in voluntary contractions 372 since the orderly recruitment of motor units according to the size principle predicts that 373 type-I motor units are recruited first (14). On this basis, the possible ergogenic action of 374 catecholamines would anyway hardly become functionally meaningful, given that most 375 human skeletal muscles have a large percentage of type-I fibers. 376 A possible complication in the interpretation of our results is that greater 377 sympathetic activation may be required for observing potentiating effects on muscle fiber 378 contractility than that provided by CPT. This opens for a potential functional role of 379 potentiation of contractility at higher activation levels of the sympathetic system. However, 380 CPT, which provokes a consistent increase in arterial blood pressure and in plasma 381 catecholamines (31), is a stimulus which is already quite difficult to sustain: VAS pain 382 scores up to 9.2 were reported in the present study while in previous studies some subject 383 could not tolerate the pain level and interrupted the test before completion (28). 384 Nevertheless, the involvement of the sympathoadrenal axis in the stress response is 385 stressor-dependent (27,29,31) and it cannot be excluded that the ergogenic effect can be blood flow to resting limb muscles (37). Reduction in the blood supply was shown to 389 decrease the muscle force capacity in fatigued and resting muscles (25,38). Although it is 390 unlikely that a small reduction in blood flow (20% in the study of Wray et al (37) First of all the maximum force capacity of the muscle is not impaired by the 431 lusitropic effect (3), although an increased driving frequency would be necessary to attain 432 the same force (28). Secondly, the possibility to produce rapid alternating movements, as those required in fight and flight, should be improved by faster muscle relaxation (29). On 434 this basis it is not surprising that significant increases in performance after administration 435 of beta2-agonists has been reported almost only for short-lasting and rapid tasks, such as 436 the wingate test (8,20,32,34). 437 Besides the generalized sympathetic activation that characterizes the fight-or-flight 438 response, sympathetic outflow is known to be highly differentiated to different organs and 439 tissues depending on the context or stimulus according to the so-called autonomic 440 "signature", which also concerns the balance between sympatho-neural and sympatho-441 adrenal pathways (24,27,30). Thus, sympathetic modulation of muscle contractility should 442 also be expected to occur in other situations activating the sympatho-adrenal axis. In this 443 respect and in support of the current view, it is interesting to mention the anecdotal reports 444 of back and leg muscle weakness during states of fear and anxiety as well as in response to 445 adrenaline infusion (3). 446 Nerve 36: 190-196, 2007. 511 19.