Stretch activation combats force loss from fatigue in fast-contracting mouse skeletal muscle fibers

Stretch activation (SA), the delayed increase in tension (force normalized to cross-sectional area) following a rapid stretch of at least a partially calcium (Ca²⁺)-activated muscle fiber, is a mechanical phenomenon that potentially benefits any muscle that is stretched immediately prior to shortening (Pringle, 1949; Rüegg et al., 1970; Huxley and Simmons, 1971; Steiger, 1971; Moore, 2006; Swank, 2012; Straight et al., 2019). SA is especially useful in cyclical muscle contractions to increase force during shortening, leading to increased work and power. In the extreme case of asynchronous insect flight muscle, SA takes over force level modulation from calcium, and in the heart, a moderate amount of SA helps increase contractility. In skeletal muscle, SA has previously been thought to be too low to be helpful. However, our recent work suggested that for specific situations and in certain muscle types, SA can be beneficial to skeletal muscle force production (Straight et al., 2019).

What makes SA physiologically relevant is its unique tension transient that occurs following a muscle’s rapid length increase. The tension transient displays a four-phase response (Fig. 1): an immediate tension increase (phase 1), a rapid decay (phase 2), a delayed and slower secondary increase (phase 3), and a slower secondary decay (phase 4) (Ford et al., 1977). Stretch-activated tension (F_SA) is the highest tension during phase 3, while the timing of F_SA (t₃) is measured from the initiation of the rapid fiber length change to the point in time where F_SA occurs (Fig. 1). Both F_SA and t₃ dictate SA’s physiological relevance. Within antagonistic muscle pairs that undergo repeated shortening and lengthening cycles to facilitate locomotion, like that of the indirect flight muscles found within Drosophila melanogaster, force needs to be higher during shortening than during lengthening to generate net positive work and power. SA, characterized by a delayed increase in force, can play a pivotal role in this process. When the timing of the delayed stretch-activated force aligns with muscle shortening, it synergizes with Ca²⁺-activated force (Moore, 2006). Depending on the specific muscle type, this synergy either contributes to or enables the production of positive work and power (Josephson et al., 2000).

SA has been examined in three striated muscle types: asynchronous indirect flight muscles of insects (Pringle, 1949; Pringle, 1977; Boettiger, 1957; Steiger, 1977; Wray, 1979; Josephson et al., 2000; Linari et al., 2004; Bullard and Pastore, 2011; Perz-Edwards et al., 2011; Wang et al., 2011; Swank, 2012; Glasheen et al., 2017) as well as vertebrate cardiac (Steiger, 1971; Steiger, 1977; Vemuri et al., 1999; Campbell and Chandra, 2006; Stelzer et al., 2006a; Stelzer et al., 2006b; Stelzer et al., 2006c) and skeletal (Rüegg et al., 1970; Steiger, 1977; Galler, 1994; Galler et al., 1994; Galler et al., 1997a; Galler et al., 1997b; Zhao and Swank, 2013; Zhao and Swank, 2017; Straight et al., 2019) muscles, with the F_SA response differing depending on the muscle (Moore, 2006). Indirect flight muscles, which power wing motions during flight in some insects, such as Drosophila, produce a high amount of F_SA relative to their low Ca²⁺-activated isometric tension (F_SA/F₀, where F₀ = Ca²⁺-activated isometric tension), about 200% (Loya et al., 2022), which results in SA being the main modulator of force (Abbott and Steiger, 1977; Josephson et al., 2000; Moore, 2006). Cardiac tissue has a moderate F_SA/F₀, about 25%, where SA aids calcium in modulating force levels to power cardiac output and blood ejection throughout the circulatory system (Abbott and Steiger, 1977; Steiger, 1977; Moore, 2006). Skeletal muscle has the smallest F_SA/F₀ under maximally calcium-activated conditions, around 10%, which has resulted in SA being considered physiologically irrelevant to skeletal muscle function (Abbott and Steiger, 1977; Steiger, 1977; Moore, 2006).

However, our recent skeletal muscle work has shown that F_SA/F₀ can be increased by increasing inorganic phosphate [P_i] (Straight et al., 2019) and also in insect jump muscle by expressing different myosin isoforms (Zhao and Swank, 2013; Zhao and Swank, 2017; Straight et al., 2019). In mouse soleus muscle, F_SA became larger as [P_i] increased from zero to moderate (16 mM) concentrations (Straight et al., 2019). Genetically modifying the Drosophila jump muscle, which is more similar to skeletal vertebrate muscle than Drosophila indirect flight muscle, to express an embryonic myosin isoform increased F_SA in the presence of moderate [P_i] (Zhao and Swank, 2013). Increasing [P_i] also made F_SA occur faster, as shown by a shorter t₃, in vertebrate and insect muscles (Zhao and Swank, 2013; Zhao and Swank, 2017; Straight et al., 2019). Together these findings demonstrate that in the presence of P_i, SA can elevate and quicken F_SA. This effect should improve force production in soleus skeletal muscles under conditions that cause P_i accumulation (Fitts, 2008; Debold et al., 2013).

The finding that greater [P_i] increased a type of force production is novel as this contradicts the well-known phenomena that muscle fiber force production decreases with greater [P_i] under isometric and shortening conditions (Cooke and Pates, 1985; Hibberd et al., 1985). Is this just happenstance, or might there be a physiological reason for SA to have evolved to increase with [P_i]? One muscle condition that results in P_i accumulation is muscle fatigue, which led us to suggest a new hypothesis, that SA has evolved in skeletal muscle to combat fatigue, particularly the loss of force production, and thus increase muscle endurance. However, P_i is not the only metabolic by-product that accumulates during fatiguing conditions and contributes to lower force generation, as reviewed by Kent-Braun et al. (2012). Hydrogen ions (H⁺) also accumulate quickly within the myoplasm during fatigue (Hermansen and Osnes, 1972; Metzger and Fitts, 1987; Cady et al., 1989; Broxterman et al., 2017) and inhibit the crossbridge cycle, leading to a decline in isometric force production (Allen et al., 2008; Fitts, 2008; Debold, 2016). Additionally, the prolonged accumulation of both P_i and H⁺ have been shown to disrupt Ca²⁺ sensitivity of single muscle fibers (Debold et al., 2006; Nelson and Fitts, 2014; Debold, 2016) and may impact Ca²⁺ concentration ([Ca²⁺]) of the myoplasm by reducing Ca²⁺ release from the sarcoplasmic reticulum, further decreasing force production (Westerblad and Allen, 1991). As such, muscle fatigue has been hypothesized to be multifactorial (Allen et al., 1995), having an initial phase where [P_i] and [H⁺] are increased (high calcium fatigue) (Broxterman et al., 2017) and a prolonged phase where [Ca²⁺] is additionally reduced (low calcium fatigue) (Lee et al., 1991; Allen et al., 1995). Therefore, to test our hypothesis that SA combats muscle fatigue in some skeletal fiber types, we needed to expose muscle fibers to more than just increased [P_i].

We subjected skinned single fibers from mouse soleus and extensor digitorum longus (EDL) muscles to three distinct conditions: an Active condition with normal [Ca²⁺], [P_i], and [H⁺] (pCa 4.5, 5 mM P_i, pH 7); a High [Ca²⁺] Fatigue condition with normal [Ca²⁺] but fatiguing levels of [P_i] and [H⁺] (pCa 4.5, 30 mM P_i, pH 6.2); and a Low [Ca²⁺] Fatigue with lower than normal [Ca²⁺] as well as fatiguing levels of [P_i] and [H⁺] (pCa 5.1, 30 mM P_i, pH 6.2). High [Ca²⁺] Fatigue was designed to investigate the initial phase of muscular fatigue, where [P_i] and [H⁺] are increased (Broxterman et al., 2017). Low [Ca²⁺] Fatigue was intended to understand the prolonged phase of fatigue, where [Ca²⁺] is additionally reduced, causing isometric tension decreases in the range of 30–80% compared with unfatiguing conditions (Lee et al., 1991; Allen et al., 1995), with pCa 5.1 being chosen as this was expected to reduce isometric tension around 50% from Active based on our preliminary experiments. The fatiguing metabolite levels were selected because the intramuscular [P_i] can be increased to 30 mM and above (Cady et al., 1989; Broxterman et al., 2017) while pH can be reduced to as low as 6.2 (Hermansen and Osnes, 1972; Vandenborne et al., 1993) during high intensity exercise, and previous single fiber experiments have used these concentrations (Potma et al., 1995; Debold et al., 2006; Karatzaferi et al., 2008; Nelson and Fitts, 2014). Since we previously found that the myosin isoforms alter the SA response (Zhao and Swank, 2013; Zhao and Swank, 2017; Straight et al., 2019), we determined myosin heavy chain (MHC) isoform expression (I, IIA, IIX, or IIB) for each fiber tested.

Soleus and EDL muscles were removed after euthanasia from adult (12–15 wk) female C57BL/6NJ mice (n = 8; The Jackson Laboratory). All mice used in this study served as controls for other experiments approved by the University of Massachusetts Institutional Animal Care and Use Committee and were raised under standard housing conditions at room temperature and supplied with breeder’s chow ad libitum. Muscles were immediately placed into cold (4°C) dissecting solution (20 mM N,N-bis[2-hydroxyethyl]-2-aminoethanesulfonic acid, 5 mM EGTA, 5 mM MgATP, 1 mM free Mg²⁺, 1 mM dithiothreitol, and 0.25 mM P_i) with an ionic strength of 175 mEq, pH 7.0, and at pCa 8.0. Muscles were tied to glass rods and placed in skinning solution (170 mM potassium propionate, 10 mM imidazole, 5 mM EGTA, 2.5 mM MgCl₂, 2.5 mM Na₂H₂ATP, protease inhibitor [Roche] with pH 7.0 and pCa 8.0), a low-[Ca²⁺] solution that begins the permeabilization of the muscle fibers’ external membrane (Wood et al., 1975), for 24 h at 4°C. Addition of glycerol, which protects the fibers from freezing and aids in the permeabilization of the muscle fibers’ external membrane, was accomplished by transferring the fiber bundles to storage solution (identical to skinning solution, but with 1 mM sodium azide and without protease inhibitor) and stepwise adding increasing amounts of glycerol to the solution: 10% vol/vol glycerol for 2 h, 25% glycerol vol/vol for 2 h, and 50% vol/vol glycerol for 2 h. Thereafter, muscles were stored in 50% vol/vol glycerol, including the protease inhibitor, at 20°C until single fibers were isolated for mechanical measurements, which occurred within 4 weeks of the dissection.

On the day of the experiment, soleus and EDL muscles were placed in dissecting solution, and a section of tissue from each muscle was removed with scissors to create a muscle fiber bundle containing ∼100 fibers. Muscle fiber bundles were moved to a dissecting solution with 1% Triton X-100 for 30 min at 4°C to remove more of the sarcolemma and sarcoplasmic reticulum (Stienen, 2000), then transferred back to dissecting solution, cut ∼1 mm in length, and single muscle fibers were manually separated using forceps. Aluminum T-clips were connected to both ends of the fiber. The top and side fiber diameters were measured at three positions along the length of the fiber in dissecting solution using the scale bar superimposed within the eye piece of a stereomicroscope to estimate the height-to-width ratio. This ratio is later used to calculate the side diameter when the fiber is on our apparatus, as we can more accurately measure top diameter. Individual muscle fibers with T-clips were demembranated a second time (dissection solution with 1% vol/vol Triton X-100) for 30 min at 4°C to ensure removal of the sarcolemma and sarcoplasmic reticulum.

A custom-built muscle mechanics apparatus was used, as previously described (Miller et al., 2010). Single fibers were transferred to the experimental apparatus containing relaxing solution (same as dissecting solution, but with 5 mM P_i) (as resting P_i levels are around 5 mM [Pathare et al., 2005; Kemp et al., 2007]), 15 mM creatine phosphate, and 300 μl/ml of creatine phosphokinase at 15°C and mounted to the experimental apparatus by sliding the holes in the T-clips over the hooks attached to a piezo actuator linear motor (P-841.10; Physik Instrumente) and an Akers strain gauge (AE-801; SensorOne). Sarcomere length was adjusted to 2.5 µm, and baseline tension was recorded, transferred to pre-activating solution (same as relaxing, except EGTA is decreased to 0.5 mM) for 30 s, then transferred to active solution (same as relaxing, but at pCa 4.5), where an initial Ca²⁺ activation occurs at 15°C. Following this initial 15°C activation, the fiber was returned to relaxing solution where temperature was set to 25°C. To stay within the force plateau of the force-length relationship, ∼2.5–2.65 μm (Gordon et al., 1966; Chen et al., 2016), sarcomere length was set to 2.65 µm, and cross-sectional area was determined by measuring the width of the fiber on the apparatus’ compound microscope and estimating the height using the previously determined height-to-width ratio, which presumes the fiber’s cross-sectional area is elliptical. The fiber was slackened completely, the force gauge was zeroed, and the fiber was pulled back to its original sarcomere length of 2.65 µm, and the force was allowed to equilibrate for 1 min before undergoing isometric tension and step experiments. Fiber images were obtained while on the experimental apparatus using a video camera (Point Grey, FLIR Integrated Imaging Solutions, Inc.), and sarcomere length was determined using ImageJ (Schneider et al., 2012) to perform a fast-Fourier transformation of the visible light and dark sarcomere patterns.

Step experiments were performed to induce SA (Straight et al., 2019). Following force equilibration, relaxed isometric tension was measured and a relaxed tension transient was collected by stretching the fiber by 0.5% of its optimal muscle length over a period of 6 ms. The fiber was held at this stretched length for 2.5 s and returned to its original length over a period of 6 ms. This rapid stretch-hold-release pattern is referred to as a “trapezoidal perturbation,” as the length trace has a trapezoidal pattern. The fiber was placed into pre-activating solution for 30 s, then into Low [Ca²⁺] Fatigue solution (same as active but 30 mM P_i, pH of 6.2, and pCa 5.1) with isometric tension measured after a plateau was reached (Low [Ca²⁺] Fatigue F₀) and a step experiment performed (0.5% over 6 ms, 2.5-s hold, 6-ms return). Isometric tension (F₀) and step experiments were subsequently performed at High [Ca²⁺] Fatigue (same as Low [Ca²⁺] Fatigue but pCa 4.5) and the Active condition (active solution). Fibers were exposed to experimental solutions from high to low [P_i] so that isometric tension starts low and increases throughout the protocol to prolong fiber stability. To measure fiber stability, tension was examined pre- and postlength step in each of the three experimental conditions. Postlength step tension was within 10% of pre-length tension for all fibers, and therefore no fibers were excluded. If sarcomere length shortened during the experiment, which happened occasionally, fibers were placed back into relaxing solution, where sarcomere length was stretched back to 2.65 µm, slacked, and the force gauged zeroed, re-lengthened, and allowed to equilibrate for 1 min. Relaxed tension was remeasured, fiber placed into pre-activating solution for 30 s, then placed in active solution where isometric tension and step experiments were conducted for a second time.

The fiber was deemed to not have SA if all three of the following criteria were met: (1) phase 3 was not apparent upon visual inspection, (2) the three-exponential equation resulted in identical r₃ and r₄ values, and (3) the double-exponential equation equally or better fit the tension traces compared with the three-exponential fit. If the criteria listed above were met, then the fiber was rejected for displaying a phase 3 response, and SA was considered to not be present. For fibers that displayed a SA response (were better fit by the three-exponential term equation than the two-exponential term equation), a running mean of 2 ms was applied to the entire tension transient, with the range of phase 3 being manually segmented within the Shiny App. The maximal tension value within the phase 3 range (F_SA) was automatically identified, while the time when F_SA occurred (t₃) was manually calculated by subtracting the time point of phase 3 minus the time point of the start of phase 1 (Fig. 1).

Following single-fiber mechanical analysis, single fibers were placed in 30 μl loading buffer, heated for 2 min at 65°C, and stored at −80°C until determination of MHC isoform composition by SDS-PAGE could be performed to identify fiber types. Fiber type gels were run as previously described (Miller et al., 2010) with the following modifications. Resolving gel was comprised of 8% acrylamide/bis-30% glycerol (wt/vol), and gels were run at 70 V for 1 h followed by 275 V for 26 h.

Data are expressed as mean ± SE, and differences were considered significant at P ≤ 0.05. As there were multiple observations within the same mouse, main effects were determined using a repeated measures linear mixed model (Kuznetsova et al., 2017), with a random effect to account for grouping of observations within mice and a repeated effect to account for the between-mice factor (i.e., Active group versus High [Ca²⁺] Fatigue versus Low [Ca²⁺] Fatigue). Tukey’s post-hoc tests were performed to determine pairwise differences between the three-experimental conditions. Within group differences for t₃ were determined using a linear mixed model. Analysis of covariance was performed on F_SA versus F₀ plots to determine differences between slopes for the three conditions. All statistical analyses and graphs were conducted using custom written R scripts (v4.2.3, R Core Team, 2021).

Because the type of myosin alters the SA response (Zhao and Swank, 2013; Zhao and Swank, 2017; Straight et al., 2019), MHC isoform expression (I, IIA, IIX, or IIB) of each fiber tested was determined. 89 total fibers from the soleus (n = 53) and EDL (n = 36) muscles from female C57BL/6NJ mice (biological replicate, N = 8) were collected and classified: MHC I (n = 14, all from soleus), MHC IIA (n = 32, 31 from soleus), MHC IIAX (n = 7, 4 from EDL), MHC IIX (n = 11, 7 from EDL), MHC IIXB (n = 6, all from EDL), and MHC IIB (n = 19, all from EDL). Due to the small sample size (technical replicate, n < 10) of MHC IIAX and IIXB, analysis was conducted only on the four pure MHC isoforms: the slow-contracting MHC I and the fast-contracting MHC IIA, IIX, and IIB fibers.

The stretch-activated tension response (F_SA) was greater in fibers with fast-contracting MHC isoforms while in active and fatiguing solutions (Fig. 1, Fig. 2 A, and Table 1), supporting our hypothesis that MHC isoform plays an important role in setting F_SA (Zhao and Swank, 2013; Zhao and Swank, 2017; Straight et al., 2019). MHC I fibers displayed minimal F_SA values under Active conditions and no measurable F_SA under either fatiguing condition. MHC IIA fibers’ F_SA increased going from Active to High [Ca²⁺] Fatigue and had similar F_SA values in Active and Low [Ca²⁺] Fatigue solutions. MHC IIX fibers’ F_SA was not statistically different in Active, High [Ca²⁺] Fatigue, and Low [Ca²⁺] Fatigue solutions. MHC IIB fibers’ F_SA was the same in Active and High [Ca²⁺] Fatigue solutions, and F_SA decreased in Low [Ca²⁺] Fatigue solutions. The number of MHC IIA fibers producing a noticeable F_SA peak dropped from Active (n = 32, 100%) to High [Ca²⁺] Fatigue (n = 30, 94%) to Low [Ca²⁺] Fatigue solutions (n = 8, 25%), while all MHC IIX and IIB fibers displayed an F_SA peak in all three solutions. Thus, the fast-contracting MHC isoforms displayed a more robust SA response to fatigue compared with slow-contracting.

Although the SA response was different based upon MHC isoform, the isometric tension (F₀) response was similar for all MHC isoforms. F₀ decreased from Active to High [Ca²⁺] Fatigue to Low [Ca²⁺] Fatigue solutions for all four MHC isoforms, except for High [Ca²⁺] Fatigue to Low [Ca²⁺] Fatigue in MHC I fibers (Fig. 2 B and Table 1).

The combination of stretch-activated and isometric tension, F_SA + F₀ (Fig. 2 C), is an important value for normalizing F_SA because the [F_SA/(F_SA + F₀)] ratio represents how much SA is contributing to total tension generation. F_SA + F₀ decreased from Active to High [Ca²⁺] Fatigue to Low [Ca²⁺] Fatigue solutions for all three MHC isoforms, except for Active to High [Ca²⁺] Fatigue in MHC IIX fibers (P = 0.058, Table 2).

Two important indicators of how physiologically relevant an F_SA value is to muscle tension generation are the ratio of F_SA to isometric tension production, F_SA/F₀, and the contribution of stretch-activated tension to total tension generation, [F_SA/(F_SA + F₀)]. MHC II isoforms’ F_SA/F₀ ratios significantly increased (76–148%) under fatiguing conditions compared with active conditions (Fig. 3 and Table 1). [F_SA/(F_SA + F₀)] showed a similar pattern, increasing from Active to High [Ca²⁺] Fatigue to Low [Ca²⁺] Fatigue in fast-contracting fibers (58–114%). MHC I fibers’ ratios were <5% under active conditions and not obtainable under fatigue conditions since no visible F_SA peak occurred. Overall, that the MHC II fibers F_SA contribution to total tension increased to as much as 30% (Fig. 3 B) shows that under fatigue conditions SA plays a physiologically relevant role in tension generation.

The effect of isometric tension on the SA response was evaluated by plotting F_SA versus F₀. Under active conditions, F_SA showed little to no response with increasing F₀ in MHC I fibers, while F_SA increased similarly in MHC IIA, IIX, and IIB fibers with F₀ (Fig. 4 A). These results show that the effect F₀ has on F_SA is different between the slow- and fast-contracting fiber types. Under fatigued conditions, F_SA also increased similarly in MHC IIA, IIX, and IIB fibers with F₀ (Fig. 4, B and C), with the fatigued slopes being higher than active (both P < 0.01) but not different from one another (P = 0.10). By putting all three conditions on the same graph (Fig. 4 D), our findings indicate that, at a given F₀, the fast-contracting fibers produce a larger F_SA under both High [Ca²⁺] Fatigue and Low [Ca²⁺] Fatigue compared with Active solutions. These results also show that fast-contracting fibers produce greater F_SA as F₀ increases for active and fatigue conditions, meaning that fibers that have the greatest tension production capabilities will benefit the most from SA.

The time from the start of lengthening to when the F_SA peak occurred (t₃) remained relatively consistent in active compared with fatiguing conditions within fibers expressing the same MHC isoform. The exceptions were t₃ being 25% shorter in High [Ca²⁺] Fatigue compared with Active in MHC IIA fibers and t₃ being 23% longer in Low [Ca²⁺] Fatigue compared with Active and High [Ca²⁺] Fatigue solutions in MHC IIB fibers (Fig. 5 and Table 1). Comparing fiber types while in Active solutions, MHC I t₃ was longer than all other isoforms, MHC IIA was greater than MHC IIB, and MHC IIX was longer than MHC IIB. Under fatiguing conditions, t₃ was greater in MHC IIA and IIX fibers compared with MHC IIB.

The goal of this study was to investigate whether SA improves muscle performance, particularly force generation, during simulated conditions of muscle fatigue. A main characteristic of fatigue in muscle fibers is lower isometric muscle force production, as observed in many studies where muscles are repetitively activated while held isometric (Godt and Nosek, 1989; Lee et al., 1991; Westerblad and Allen, 1993; Potma et al., 1995; Karatzaferi et al., 2008; Nelson and Fitts, 2014). A way to combat this force decay would be beneficial for fibers that are repetitively used over long periods of time. From our work and others on insect asynchronous flight, muscle we know that high levels of SA can occur independent of calcium cycling (Wang et al., 2011) or at lower calcium levels (Linari et al., 2004) where isometric force is reduced. We also recently observed in both flies and mice that, counterintuitively, increased [P_i] increased F_SA, the opposite of what is observed for F₀. Together these observations led us to theorize that SA might have evolved in some muscle types to increase endurance. Therefore, we replicated several of the known conditions that occur during fatigue to test how SA responds.

Our results suggest that for fast-contracting MHC II fibers, SA counters the effects of fatigue, specifically increased [P_i] and H⁺ with or without impaired calcium cycling. High and Low [Ca²⁺] Fatigue conditions either increased F_SA (MHC IIA) or caused no change in F_SA (MHC IIX and IIB) compared with Active conditions, while F₀ decreased, except for a decrease in F_SA for MHC IIB fibers in Low [Ca²⁺] Fatigue. This finding is highlighted when looking at the high values of F_SA normalized to isometric tension or total muscle tension under fatigue conditions, which shows that the magnitude of F_SA is enough to be relevant to muscle function. For example, F_SA/F₀ in MHC IIA fibers more than doubled under Low [Ca²⁺] Fatigue conditions compared with active conditions, with F_SA generating about 30% of total muscle tension under Low [Ca²⁺] Fatigue. Thus, as isometric tension is declining, 45% for IIA, F_SA is partially compensating for loss of F₀ due to fatigue. These results support our idea that SA has evolved to combat the decrease in force generation that occurs with fatigue in MHC II fibers. In contrast, MHC I fibers showed very low F_SA under active conditions, and the F_SA peak was not even visible under either fatigue condition, suggesting that in these fibers, SA is not helpful for improving force during extended locomotion.

That fast-contracting MHC II fibers developed a SA-based mechanism to help reverse fatigue may have occurred because fatigue happens more rapidly in these fibers compared with MHC I fibers (Allen et al., 2008). Muscles that predominately consist of slow-contracting fibers (∼80% MHC I) also produce high [P_i] (>30 mM) and [H⁺] (pH 6.4) ions when maximally fatigued (Weiner et al., 1990; Kent-Braun, 1999), indicating that slow-contracting fibers also significantly increase these ions during fatigue,; this just occurs more slowly than type II fibers (Vandenborne et al., 1993; Mizuno et al., 1994). After the initial phase, where [P_i] and [H⁺] are increased, both animal (Allen et al., 2008) and human (Olsson et al., 2020) studies show impaired [Ca²⁺] release from the sarcoplasmic reticulum contributing to impaired force production. Overall, slow- and fast-contracting fibers both appear to significantly build up [P_i] and [H⁺] ions and show impaired [Ca²⁺] release with fatigue, but these changes occur much faster in MHC II fibers, which may have caused them to develop a robust SA response to counteract fatigue.

Higher levels of F_SA under fatigue conditions are not the only requirement needed for SA to be beneficial for skeletal muscle function. Other factors, such as the time when F_SA occurs in a contraction cycle, need to be examined to determine if SA is physiologically relevant to skeletal muscle during fatigue. For SA to be helpful for muscle work and power generation, it cannot occur too early in the lengthening cycle, or F_SA will occur during muscle lengthening instead of shortening, where increased force increases net work and net power. We propose that the ideal time for the triggering of SA to occur to help increase force during shortening and power generation in synchronous muscle types would be in the later portion of the lengthening portion of a muscle contraction cycle. For most of lengthening, force needs to be low, or increased resistance to lengthening occurs, which results in loss of work and power. However, as the end of lengthening nears, force-generating mechanisms need to be initiated because force generation is not instantaneous, and force needs to be high during shortening to produce work and power. So as calcium starts increasing (SA does not occur unless at least a low amount of calcium is released [Wang et al., 2011]), plus the remainder of lengthening (stretch) occurs, these should together trigger the SA response so that F_SA occurs during shortening. This means that only a small length increase is all that should be needed to trigger SA, which is one reason we used a small stretch amplitude in this study.

Our previous research on the SA response under varying [P_i] conditions (Straight et al., 2019) evaluated the physiological relevance of in vitro SA kinetics in the context of in vivo sarcomere length changes during mouse locomotion. By overlaying the 16 mM [P_i] fit from a soleus muscle fiber onto sarcomere length change patterns recorded from a mouse running at 8 Hz (James et al., 1995), we observed that the increase in stretch-activated force coincided with muscle shortening. When comparing the t₃ values of soleus MHC IIA fibers presented in this manuscript to the same 8-Hz sarcomere length changes, we again note that the timing of the peak of phase 3 (∼30 ms from the start of the imposed stretch) aligns with muscle shortening.

Multiple muscles initially undergo stretch during activation and then immediately shorten during human walking and running based on studies using combined computational and experimental approaches (Arnold and Delp, 2011; Arnold et al., 2013). During walking, 10 of the 17 muscles examined showed this behavior of being activated during stretching, typically starting during a short stretch or near the end of a longer stretch, and then remaining activated during shortening (Arnold and Delp, 2011), indicating that SA could be occurring in these muscles. Additionally, in vivo work in individual human lower limb muscles, specifically the vastus lateralis (Bohm et al., 2018) and soleus (Lai et al., 2015), indicate that muscle length changes may be small due to large tendon length changes during muscle activation. Altogether, these findings show that the appropriate conditions for SA to occur exist in multiple muscles in vivo during normal gait and that in vivo length changes may be small, making our experimental approach physiologically relevant.

Our prior work exploring the effects of P_i on the SA response proposed a P_i-dependent mechanism for SA in skeletal muscle (Straight et al., 2019). The mechanism is an extension of the proposed mechanisms for the well-known effect of P_i decreasing isometric force, which is that P_i rebinds to myosin-ADP, driving crossbridge from the strongly to weakly bound state (Hibberd et al., 1985; Kawai, 1987; Cooke et al., 1988; Millar and Homsher, 1990; Regnier et al., 1995; Takagi et al., 2004). Our myosin-based mechanism for SA suggests that, under conditions of moderate P_i concentration and mechanical stretch, certain strongly bound postpowerstroke crossbridges rebind with P_i and are forced backward into the weakly bound pre-powerstroke state. This transition from strong-to-weak crossbridge binding generates a population of crossbridges capable of undergoing an additional powerstroke without hydrolyzing a secondary ATP molecule, thereby producing stretch-activated force (phase 3). Several studies support the hypothesis that both elevated [P_i] levels (Hibberd et al., 1985; Kawai, 1987; Cooke et al., 1988; Pate and Cooke, 1989; Dantzig et al., 1992; Millar and Homsher, 1992; Takagi et al., 2004) and mechanical stretch (Lombardi and Piazzesi, 1990; Martyn et al., 2002; Colombini et al., 2007; Mansfield et al., 2012) can promote the transition of myosin from strongly to weakly bound states. This mechanism is likely contributing to our current findings that SA increases under fatigue conditions, but given all the additional changes that we have made here to simulate fatigue, increased H⁺, decreased calcium, and increased temperature compared with our last study, additional work is needed to determine all the mechanisms contributing to the changes to SA from fatigue.

This study provides novel insights into mammalian skeletal muscle SA behavior in active and fatigued muscle. Our data suggest that SA serves as a significant modulator of tension production under active, but especially fatigue conditions, in fibers containing MHC II isoforms. This leads us to propose a new mechanism, SA, for reducing the decline in force generation during prolonged locomotion, in other words, increasing muscle endurance. This mechanism would be relevant for moderate to fast locomotion, such as walking, swimming, or running, that are powered by muscles that cyclically lengthen and shorten.

All data and R scripts used for this study can be found on GitHub at https://github.com/19PWoods/Masters.

Henk L. Granzier served as editor.

We would like to thank Dr. Jesse Mager and his laboratory in the Department of Veterinary and Animal Sciences at the University of Massachusetts at Amherst, MA, USA for providing the animals used in this study, as well as Dr. Brent Scott in the Department of Kinesiology at the University of Massachusetts at Amherst, MA, USA, for aiding in the creation of the R scripts used to analyze the data.

Supported by National Institutes of Health (NIH) grant AR083916 to D.M. Swank

Author contributions: P.C. Woods: conceptualization, data curation, formal analysis, investigation, methodology, software, validation, visualization, and writing—original draft, review, and editing. D.M. Swank: conceptualization, funding acquisition, methodology, project administration, supervision, validation, visualization, and writing—review and editing. M.S. Miller: conceptualization, methodology, project administration, resources, supervision, validation, visualization, and writing—original draft, review, and editing.