Understanding muscle function during perturbed in vivo locomotion using a muscle avatar approach

Extensor digitorum longus (EDL) muscles were surgically exposed, and 4-0 silk sutures were tied in square knots at the distal and proximal muscle-tendon junctions. Tendons were cut outside the suture knots, and muscles were removed for ex vivo experiments. Extracted EDL muscles (n = 6) were attached to a dual-mode muscle lever system (Aurora Scientific, Inc., Series 300B, Aurora, ON, Canada) via tightened slip knots in the suture at the ends of the muscle. During experiments, the muscles were bathed in a 21°C, 7.4 pH Krebs–Henseleit solution containing (in mmol l–1): NaCl (118); KCl (4.75); MgSO4 (1.18); NaHCO3 (24.8); KH2PO4 (1.18); CaCl2 (2.54); and glucose (10.0). The bath was aerated with a 95% O2/ 5% CO2 gas mixture. The muscle was suspended between two platinum electrodes which delivered 1 ms square-wave pulses at tetanic supramaximal stimulation (80 mV, 200 Hz) while finding optimal length (L0). Submaximal stimulation (45 mV, 90 Hz) was used during the experimental protocol to more closely emulate in vivo activation, and to decrease fatigue during experimental trials (James et al., 1995). To investigate the function of the guinea fowl LG during these different types of strides, we used LG strain trajectories from four strides of a guinea fowl’s instrumented leg in ex vivo EDL work loop experiments: one stride in which the leg stepped down from the obstacle onto the treadmill (Fig. 1A; ‘Down’, yellow line); two in which the instrumented leg stepped up from the treadmill onto the obstacle (Fig. 1A; ‘Up1’ purple line and ‘Up2’ red line); and one in which no obstacle was present (Fig. 1A, ‘No obstacle’, blue line. We included two different Up trajectories because ‘Up1’ had a slight stretch in the strain trajectory (Fig. 1A; ‘Up1’, purple line), whereas ‘Up2’ did not (Fig. 1A; ‘Up2’, red line). Each of the strain trajectories was used in ex vivo EDL work loops at 2 Hz. For each muscle, we also included passive and active sinusoidal work loops at the same strain amplitude and frequency. Two work loop cycles were performed for each active strain trajectory, but only one cycle was performed for the passive sinusoidal strain trajectory. Activation patterns for the ex vivo EDL work loops were based on measured in vivo guinea fowl EMG, which typically starts at the longest muscle length at the onset of leg retraction (Daley and Biewener, 2011). All five strain trajectories were tested using three different activation patterns (Normal, Late and Long) at the same submaximal activation level (45 mV, 90 Hz). On average, submaximal activation produced 91% of the maximum isometric force observed during supramaximal activation. At activation levels lower than 45 mV and 90 Hz, the force in sinusoidal work loops was unfused. The onset and duration of the ‘Normal’ activation pattern (Fig. 2A) used in the ex vivo EDL work loops was based on observed electromyographic (EMG) data from the guinea fowl LG for the ‘Up1’ step. We accounted for the differing delay between activation and force onset, measured at ~25 ms in vivo (Daley and Biewener, 2011) versus only 4-5 ms in ex vivo experiments, by activating EDL muscles 20 ms later than observed in vivo. Thus, for ‘Normal’activation, stimulation began 20 ms after the longest muscle length and continued for a duration of 115 ms as observed in vivo during the ‘Up1’ step (Daley and Biewener, 2011). For ‘Late’ activation, (Fig. 2B), stimulation started at the onset of increased muscle shortening velocity, on average 32.5 ms later than for ‘Normal’ activation, for the same duration of 115 ms. ‘Long’ activation (Fig. 2C) spanned the other two, as the muscle was activated at the same time as ‘Normal’ activation, but was deactivated at the same time as ‘Late’ activation, for an average duration of 148.4 ms. The order in which strain trajectories and activation patterns were performed was randomized for each muscle. In total, 15 experiments were performed on each muscle, including all combinations of 5 strain trajectories and 3 activation patterns.

本实验对趾长伸肌（Extensor digitorum longus, EDL）进行外科暴露，以4-0号丝线在肌-腱连接部的远端和近端分别系上方结。随后在丝线结外侧切断肌腱，取出肌肉用于离体实验。提取的趾长伸肌样本（n=6）通过肌肉末端丝线的紧扎滑结，连接至双模肌肉杠杆系统（Aurora Scientific公司，300B系列，加拿大安大略省奥罗拉市）。实验过程中，肌肉浸泡于21℃、pH 7.4的克-亨氏液（Krebs–Henseleit solution）中，其成分（单位：mmol·L⁻¹）为：NaCl 118、KCl 4.75、MgSO₄ 1.18、NaHCO₃ 24.8、KH₂PO₄ 1.18、CaCl₂ 2.54及葡萄糖10.0。灌流槽持续通入95% O₂与5% CO₂的混合气体进行曝气。肌肉悬挂于两枚铂电极之间，电极可输出1 ms方波脉冲，在寻找最优肌长（L₀）时采用强直最大刺激（80 mV，200 Hz）。实验方案中采用亚最大刺激（45 mV，90 Hz），以更贴近体内激活状态，并降低实验过程中的肌肉疲劳（James等，1995）。为探究珠鸡LG肌在不同步态中的功能，本实验在离体趾长伸肌功环实验中，采用了来自4段经仪器监测的珠鸡腿部步态的LG肌应变轨迹：1段为腿部从障碍物迈步至跑台的步态（图1A；"Down"，黄色线）；2段为腿部从跑台迈步至障碍物的步态（图1A；"Up1"，紫色线与"Up2"，红色线）；以及1段无障碍物的平地步态（图1A；"No obstacle"，蓝色线）。本研究纳入2种不同的上步轨迹，原因在于"Up1"的应变轨迹存在轻微牵张（图1A；"Up1"，紫色线），而"Up2"则无此特征（图1A；"Up2"，红色线）。所有应变轨迹均以2 Hz的频率应用于离体趾长伸肌功环实验。针对每块肌肉，实验还包含了相同应变幅度与频率下的被动与主动正弦功环。每条主动应变轨迹对应2次功环循环，而被动正弦应变轨迹仅进行1次循环。离体趾长伸肌功环的激活模式基于实测的珠鸡LG肌体内肌电图（electromyographic, EMG）数据，该肌电信号通常在腿部回缩起始时，于肌肉最长长度处开始激活（Daley与Biewener，2011）。所有5种应变轨迹均采用3种不同的激活模式（Normal、Late与Long）进行测试，且激活水平保持一致的亚最大刺激参数（45 mV，90 Hz）。平均而言，亚最大刺激可产生强直最大刺激下观测到的最大等长肌力的91%。当激活参数低于45 mV与90 Hz时，正弦功环中的肌力未发生融合。"Normal"激活模式（图2A）的起始时刻与持续时长，基于"Up1"步态下的珠鸡LG肌实测肌电数据确定。考虑到体内与离体实验中激活至肌力起始的延迟存在差异——体内延迟约为25 ms（Daley与Biewener，2011），而离体实验仅为4~5 ms——本实验将趾长伸肌的激活时刻较体内实测值延后20 ms。因此，"Normal"激活的刺激起始于肌肉最长长度出现后的20 ms，并按照"Up1"步态下的体内实测时长持续激活115 ms（Daley与Biewener，2011）。"Late"激活模式（图2B）的刺激起始于肌肉缩短速度加快的时刻，较"Normal"激活模式平均延后32.5 ms，持续时长同样为115 ms。"Long"激活模式（图2C）则介于前两者之间：其激活起始时刻与"Normal"模式一致，但停用时刻与"Late"模式相同，平均持续时长为148.4 ms。每块肌肉的应变轨迹与激活模式的执行顺序均进行随机化处理。单块肌肉总计开展15次实验，涵盖5种应变轨迹与3种激活模式的全部组合。