Fresh porcine stomachs were chosen for this study due to their structural and mechanical similarity to the human stomach (Zhao et al., 2008; Jia et al., 2015). The stomachs were taken from 41 freshly killed female pigs (~6 months and 100 kg) of a local slaughterhouse. The experimental set-up, handling, and preparation of gastric smooth muscle tissue have been described earlier (Menzel et al., 2017).

Briefly, immediately after death a predefined piece of smooth gastric tissue (30 × 20 mm) was dissected from the ventral fundus of the proximal stomach (Figure 1) and separated from the tunica mucosa. Care was taken to prevent contact of the tissue strip with gastric acid to avoid smooth muscle degeneration. Afterwards, the preparation was preserved in an icebox in aerated (95% O₂ and 5% CO₂) Krebs solution (124 mM NaCl; 5 mM KCl; 2.5 mM CaCl₂; 15 mM NaHCO₃; 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 10 mM C₆H₁₂O_6; pH 7.3 at 32°C) (Moriya and Miyazaki, 1982) at a constant temperature of 4°C for transportation. Subsequently, within 60–90 min postmortem, small tissue strips (0.66 ± 0.2 g) of 16 × 8 mm in longitudinal orientation (i.e., parallel to the direction of the gastric serosal fold, parallel with the greater curvature, Figure 1) were prepared in the laboratory from the predefined tissue sample of the fundus.

The gastric strips were mounted vertically in an Aurora 805A in-vitro muscle apparatus between an alligator clip and two flexible cannula hooks on the bottom. The clamp was attached to a dual mode lever arm system (Aurora Scientific 305C-LR, force range: 10 N, force resolution: 1 mN). An initial passive baseline force of 5–10 mN was set by the test apparatus and the tissue strip length at this force, measured between clamp and hooks with a digital sliding caliper, was defined as the slack length (L_s). The mean slack length was 13.42 ± 1.42 mm.

The tissue strips were electrically stimulated (Aurora Scientific 701C) with alternating pulses of 1 A, 100 Hz frequency, and 5 ms pulse width (van Mastrigt and Glerum, 1985; Menzel et al., 2017). All mechanical experiments were conducted at a constant temperature of 32 ± 0.1°C in aerated Krebs solution. At this temperature, the tissue samples proved very stable and able to withstand active protocols over an extended period of time (Essig et al., 1985; Tomalka et al., 2017). After an equilibration period of 30 min at L_s, the strips were stimulated isometrically for about 15 s every 5 min until a steady state force [deviation <5% of maximum isometric force (F_im)] was reached (Herrera et al., 2005).

To determine the specific biomechanical muscle properties FLR, FVR, and history-dependent effects, isometric, isotonic, and isokinetic contractions have been performed in accordance with previous studies for smooth muscle tissue (Menzel et al., 2017; Seydewitz et al., 2017) and skeletal muscles (Böl et al., 2013; Siebert et al., 2015). Fifty-four (n = 54) tissue strips were used in total for the in-vitro dynamic parameter determination (Table 1). To investigate the active and passive FLR, a series of 18–19 isometric contractions (with length increments of 0.1 L_s in ascending order), starting from an initial sample length of 0.8 L_s, have been conducted. Tissue strips were stretched up to passive forces of about 50% F_im to avoid muscle damage induced by excessive lengthening. At each length, the passive and maximum active muscle force were determined as the maximum force value and the passive force at the instant before muscle activation, respectively. Linear regression models following the equationf(x) = mx +b, withx = L/L_S, were used to fit the ascending- and descending limb of the FLR. The muscle length at F_im was defined as the optimal muscle length (L₀).

The FVR was identified by a series of about six isotonic contractions starting from L₀ against forces in the range of 0.10–0.90 F_im (Till et al., 2008) in ascending order. The FVR followed the typical hyperbolic Hill equation (Hill, 1938) f(v)=(vmax-v)/(vmax+vcurv), v < 0 for concentric contractions, with ν_max defined as maximum shortening velocity and curv=aFim (where a is the asymptote of force; Hill, 1938). A recovery phase of 7 min between the experiments has been conducted over the entire experimental protocol for determination of force-length and FVRs.

In order to investigate the impact of ramp length and velocity on history-dependent effects, isokinetic contractions (Herzog and Leonard, 1997; Menzel et al., 2017) have been applied to n = 17 tissue strips in dependence of varying ramp amplitudes at a given velocity (40% ν_max) and in dependence of different velocities at a constant length amplitude (7% L₀), respectively. The isokinetic ramps started after a period of pre-stimulation (around 14 s) under full activation until steady-state maximum isometric force, characterized by the development of a plateau, was reached. The stimulation continued for at least 14 s after the end of the ramp. Shortening and lengthening ramps for the determination of FD and FE in dependence of ramp length started at L₀ ± 5, ± 7, and ± 10% L₀, respectively, and were finished at L₀. To examine the dependency of history-effects on speed, concentric and eccentric ramps from L₀ ± 7% to L₀ with different ramp velocities (20, 40, 70% ν_max) were applied to the tissue strips. All ramp experiments have been carried out in randomized order (Figure 2). To calculate FE and FD, the difference between the redeveloped and the corresponding purely isometric force at the same length has been identified 10 s after the end of each ramp. The relaxation period during the entire history-block was 12 min to ensure an optimal recovery phase of force-generating processes (Gunst, 1986). The “cycling-protocol” by Brenner (1983) was utilized to conserve the structural and mechanical properties in maximally activated smooth gastric strips over an extended period of time as well as to reduce length-inhomogeneities of overlapping myofilaments.

To calculate force degradation, isometric reference contractions at L₀ were performed before and after each of the ramp experiments. Plastic length adaptations of smooth muscle strips by dissolution and reformation of myosin filaments (Seow, 2005), occurring at periods of hours (Wang et al., 2001; Martinez-Lemus et al., 2004) or days (Arner et al., 1984; Zeidan et al., 2000; Naghshin et al., 2003) were not considered within this study, as the time period between the reference contractions was much shorter and muscle strips were held at a constant length (L₀) during the experiments. During isokinetic ramp experiments, the isometric force in successive activations decreased at an average rate of 1.1% per activation. Data from preparations that produced an isometric force <6 N/cm² have been rejected.

The histological examination was carried out in accordance with the procedure previously described by Menzel et al. (2017) and has been realized on samples of the same tissue-region that were used for the dynamic parameter-determination in the present study. Briefly, the cryo-histological sections of smooth gastric tissue strips were stained with Picrosirius Red staining protocol (Junqueira et al., 1979) at 200% L_S, and photographed using a digital microscope (Zeiss Smartzoom 5). To examine the muscle cross-sectional areas (CSA) in the longitudinal direction (Figure 1), the average length and width of the individual sample strips (Figure 3) was determined with an image editing software (ImageJ 1.49 v, National Institutes of Health, USA). The CSA was calculated by assuming a rectangular cross-section. The muscle tension (P_im) of a smooth gastric tissue sample was determined with P_im = (F_im/CSA)/p (where p describes the percentage of the longitudinal muscle layer from the total CSA).

The length- and force-signals from the dual-mode muscle lever were recorded at 100 Hz with an A/D interface (604A, Aurora Scientific, Canada). A real-time software package (610A Dynamic Muscle Analysis, Aurora Scientific) was used for data acquisition. A program written in Matlab (The Mathworks, Inc., Nattick, MA, USA) was used for data analysis. Data were expressed as mean ± standard deviation (s.d.). For statistical analyses, normalized data were used. Force values were divided by individual F_im. Length data were divided by L_s and L₀ and velocity data were expressed in absolute values [mm/s] and normalized to optimum muscle length [L₀/s], respectively. A Kolmogorov-Smirnov Test with P > 0.05 indicated no deviation from normality. To test significant differences of history-effects in dependence of ramp amplitude and velocity, a one-way ANOVA with repeated measures was used. For homogeneous variances, post-hoc analyses were performed using the Tukey-HSD test. A significance level of P < 0.05 was used for all analyses. Statistical analyses were carried out using SPSS 23 (IBM Corp, Armonk, NY, USA).

The results of the structural observations of histologically stained tissue samples (ventral fundus of the stomach) at 200% slack length are illustrated in Figure 3. Based on the photographs of the histological sections different tissue types can be distinguished. Following Figure 3, the stomach wall can be divided into four distinct layers: (1) tunica serosa, (2) tunica muscularis in longitudinal- and (3) transversal (circumferential) orientation and (4) tunica submucosa. Note that the tunica mucosa is dissected from the tissue samples. All structures that stained yellow-orange correspond to muscle tissue, while the stained red structures correspond to collagen. In agreement with Zhao et al. (2010), the highest collagen content was found in the submucosal layer (in between the fat tissue). The collagen content likely determines the gastric wall stiffness since collagen is supposed to be the stress-bearing structure in most tissues (Fung, 1993; Zhao et al., 2010). Picrosirius staining also revealed a continuous intermeshing network of collagen throughout the entire stomach wall. The CSA of the longitudinal tissue strip is 30.01 ± 5.66 mm² (n = 53), whereas the percentage of the longitudinal muscle layer from the total CSA is 36.6 ± 2.6%.

The smooth gastric tissue samples exhibited a characteristic FLR—similar to skeletal muscles. The active isometric FLR showed clearly visible pronounced slope changes, characterized by a linear ascending and descending limb, and a bell-shaped plateau region (see Figure 4). The plateau region (>95% F_im) ranged from 0.9 to 1.1 L₀ or 2.0 to 2.4 L_S. Extrapolation of the ascending and descending limb of the FLR, see section Determination of Smooth Muscle Tissue Properties, yielded zero force at 0.65 ± 0.14 L_S or 0.29 ± 0.06 L₀ and 4.37 ± 1.00 L_S or 1.99 ± 0.45 L₀, respectively (cf. Figure 4). The maximum isometric force (F_im) at L₀ was 1141.2 ± 295.8 mN (n = 20). Longitudinal muscle strips of porcine stomach, stimulated by supramaximal AC field stimulation, attained their F_im after 13.6 ± 1.5 s (n = 20) at 32°C. Based on histological observations in section Histological Characterization, the mean maximum muscle tension (P_im) of gastric tissue samples was 10.4 ± 2.6 N/cm² in longitudinal orientation. The passive FLR of the unstimulated muscle strip was characterized by an exponential increase of force with sample length, accompanied by passive forces of 25.8 ± 13% F_im at L₀. The gastric muscle strips' ability to shorten under various loads was studied by means of a series of isotonic releases. The V_max was 1.08 ± 0.32 mm/s (corresponding to 0.04 ± 0.01 L₀/s) with a curv-factor of 0.36 ± 0.15 (n = 25). The investigated FVR of the tested stomach samples featured the typical hyperbolic shape (see Figure 5).

The investigation of history-dependent effects yielded significantly (P < 0.05) enhanced forces following stretching and depressed forces following shortening compared with the corresponding isometric forces. A summary of all history-dependent effects investigated in this study is shown in Table 2. Statistical analyses yielded significant influence of ramp length on isometric muscle force after length change. FD and FE increased almost linearly with ramp length (see Figures 6, 8A). FD is about twice as much (31.78% F_im) for the longest ramp length (10% L₀) compared to the shortest ramp length (15.52% F_im for 5% L₀; P < 0.001) (see Figure 8A). For varying ramp velocities at constant ramp length (7% L₀; n = 13) statistics yielded a significant increase in FE between slow (20% ν_max) and moderate (40% ν_max) stretch velocities (P < 0.05; cf. left columns of Figure 8B). Further increases in eccentric velocity up to 70% ν_max yielded no further change in FE. FD decreased by trend (from 25 to 23% F_im) but not significantly with increasing ramp velocities (see Figure 7; cf. right columns of Figure 8B).

Mean maximum tension of active smooth muscle from porcine stomach is 10.92 ± 2.85 N/cm². This value is about 2–3 times higher compared to muscle tension values of porcine bladder tissue (2.5–6.0 N/cm² van Mastrigt and Glerum, 1985; Menzel et al., 2017), but about half as much when compared to smooth muscle tissue of stomachs from small mammals such as guinea pigs (19.2 N/cm² Moriya and Miyazaki, 1982). The mean passive forces of about 26% F_im at L₀ are in accordance with previous findings on smooth muscle tissue (Gordon and Siegman, 1971; Siegman et al., 2013; Menzel et al., 2017).

In agreement with the FLR from porcine urinary bladder (Menzel et al., 2017), rabbit taenia coli (Gordon and Siegman, 1971), and rat arterial vessel (Mulvany and Warshaw, 1979), the force-length dependency determined in the present study exhibited an obvious linearity of the ascending and the descending limb (see Figure 4). Furthermore, the shape of the active FLR of smooth muscles is generally similar to that of striated skeletal muscles (Gordon et al., 1966; Herlihy and Murphy, 1973)—implicating a dependence of muscle force with regards to myofilament overlap and muscle length. Considering this, active force production in smooth muscles is roughly based on the mechanisms described by the sliding filament and cross-bridge theories (Huxley and Hanson, 1954; Huxley and Niedergerke, 1954; Gordon et al., 1966) for skeletal muscles, although this is much more thoroughly understood in smooth muscles (Gordon and Siegman, 1971; Arner and Malmqvist, 1998; Siegman et al., 2013). However, there are distinct differences in the underlying microstructure of smooth muscles. Actin filaments are connected to dense bodies (Somlyo et al., 1973) and thus there is no Z-disc as in striated muscles. In contrast to the perfect straight alignment of actin and myosin filaments in skeletal muscles, ultrastructural studies of smooth muscles demonstrate a quite randomly orientated arrangement of myofilaments under activation, accompanied by cellular twisting (corkscrew-like) during active shortening (Fay and Delise, 1973; Bond and Somlyo, 1982; Warshaw et al., 1987).

These structural characteristics might explain specific differences in the muscle properties between smooth and striated muscles. Smooth muscles exhibit no slope change at the ascending limb of the FLR (Herrera et al., 2005), which is typical for striated muscles (Rode and Siebert, 2009; Winters et al., 2011) and muscle fibers (Stephenson, 2003; Tomalka et al., 2017) at sarcomere lengths of about 1.7 μm. In striated muscles this slope change is attributed to the myosin filament sliding through the Z-disc (Rode et al., 2016). The absence of this slope change in smooth muscles (Figure 4) might be partially explained by the side-polar myosin filaments in smooth muscles (Herrera et al., 2005) compared to bipolar myosin filaments in striated muscles (Craig and Megerman, 1977). However, the structural understanding of the entire FLR in smooth muscles is incomplete so far (Siegman et al., 2013).

The maximal shortening length of smooth gastric muscle tissue from pigs investigated in this study is 0.65 ± 0.14 L_S (corresponding to 0.29 ± 0.06 L₀), which is in accordance with findings of other smooth muscle studies (Gordon and Siegman, 1971; Mulvany and Warshaw, 1979; Siegman et al., 2013). The descending limb is characterized by a linear force decrease in proportion to increasing length. This yielded zero forces at 4.37 ± 1.0 L_S or 1.99 ± 0.45 L₀, accompanied by non-myofilament overlap. These values are slightly higher than those for such smooth muscles as rat arterial vessel (1.82 L₀) (Mulvany and Warshaw, 1979) or rabbit taenia coli muscles (~1.9 L₀) (Siegman et al., 2013). For lengthening skeletal muscles, active force production is limited to about 1.6 L₀ (Gordon et al., 1966; Stephenson, 2003). Differences might be due to the functional necessity of gastric tissue to withstand high distension (Korossis et al., 2009; Jia et al., 2015). The complex microstructure of smooth muscles comprises a loose, irregular myofilament arrangement held in a mesh, whereas intermediate filaments appear to link the dense bodies in a cytoskeletal network (Fay and Delise, 1973; Somlyo et al., 1973; Mulvany and Warshaw, 1979; Bond and Somlyo, 1982). These filaments, which have a more structural rather than a contractile function, are assumed to have a role in force transmission and mechanical stability (Arner and Malmqvist, 1998; Tortora and Nielsen, 2013). Additionally, smooth muscles are able to actively contract in response to stretching, followed by a reduction in tension within a short period of time (Hill, 1926; Gordon and Siegman, 1971; Tortora and Nielsen, 2013). This stress-relaxation response allows smooth muscles to undergo great length changes while still preserving the ability to contract efficiently (Siegman et al., 1976; Tortora and Nielsen, 2013). Furthermore, side-polar myosin filaments could help to explain the ability of smooth muscles to shorten by large amounts (Xu et al., 1996).

The curv-factor (0.36 ± 0.15) observed in this study is within the range (0.1–0.5) reported for smooth- (Moriya and Miyazaki, 1985; van Mastrigt, 2002; Menzel et al., 2017) and skeletal muscles (Siebert et al., 2015), respectively. Maximum shortening velocity (0.04 ± 0.01 L₀/s) is in the lower range of values from 0.03– 0.6 L₀/s reported for other mammalian smooth muscles (Gordon and Siegman, 1971; Kong and Stephens, 1983; Moriya and Miyazaki, 1985; van Mastrigt, 2002; Menzel et al., 2017). Compared to values reported for skeletal muscles (ν_max: 3–7 L₀/s, Ranatunga, 1984; Siebert et al., 2008; Gollapudi and Lin, 2013), maximum shortening velocities of smooth muscles were one to two orders of magnitude smaller (Barany, 1967).

The behavior of the smooth gastric tissue strips investigated in this study is generally in agreement with history-effects reported for other smooth- (Gunst, 1986; van Asselt et al., 2007; Menzel et al., 2017) and skeletal muscles (Rassier and Herzog, 2004; Siebert et al., 2015). Whereas, the evidence of FD and FE has been shown for other canine and porcine smooth muscles in previous studies (Gunst, 1986 trachea; van Asselt et al., 2007; Menzel et al., 2017 urinary bladder), this study represents the first approach investigating smooth gastric muscle tissue. Findings reveal a linear dependency of history-effects (FE and FD) with regards to their ramp amplitude (Table 2, Figure 6) similar to findings on skeletal muscles (Abbott and Aubert, 1952; Edman et al., 1982). Maximum values of FE (up to 16% F_im) and FD (up to 32% F_im) exceed previous findings on smooth muscles (12 and 18% F_im for FE and FD, respectively, Menzel et al., 2017).

In accordance with previous investigations (Menzel et al., 2017), ramp velocity was found to have no significant influence on FD and FE, except for differences between lengthening contractions with 20 and 40% ν_max (P < 0.05; Table 2). The investigated influence of ramp velocity on FE (at low ramp velocities) might be influenced by the experimental protocol and the contractile properties of smooth gastric muscle tissue. Low ramp velocities require long stimulus durations resulting in a decrease in muscle force induced by fatigue—accompanied by an additional flattening of the force trace—yielding to significantly reduced FE for slow eccentric ramps (Table 2, Figure 8B). Thus, results of FE for low velocities should be considered with caution. In accordance with findings by Menzel et al. (2017) on smooth muscle tissue as well as on skeletal muscles (Abbott and Aubert, 1952), FD decreased by trend (but not significantly) with increasing ramp velocity.

To date, there is an intensive debate about mechanisms and functions of history-dependent effects in skeletal muscles (Herzog et al., 2008; Campbell and Campbell, 2011; Siebert and Rode, 2014; Hessel et al., 2017). Titin, a non-cross-bridge, semi-active structure —is increasingly recognized as an important protein that contributes to active force production during and following eccentric contractions (Herzog et al., 2016). While titin does only exist in striated muscles, a molecule having similar functional and structural titin-like characteristics, named smitin, occurs in smooth muscles (Kim and Keller, 2002). Therefore, a transfer of currently discussed various mechanisms, that provide some possible explanations with regards to history-dependent effects in skeletal muscles, is likely (Rode et al., 2009; Nishikawa et al., 2012; Schappacher-Tilp et al., 2015; Heidlauf et al., 2016). However, further experimental and modeling evidence is required to demonstrate conclusive explanatory approaches of underlying mechanisms of history-dependent effects of force production in smooth muscle tissue (see Supplementary Material for further information).

The investigated tissue samples were dissected out of a predefined section of the fundus (see Figure 1, black rectangle), which is mainly serving as food reservoir. Hence, when the proximal stomach is filled due to ingestion, accompanied by a large expansion of surrounding tissue, history-dependent effects, observed experimentally in this study, might have physiological relevance in order to suit gastric motility. The motility of the fundus (proximal stomach) is characterized by almost tonic contractions induced by permanent but irregular muscle activity (de Wever et al., 1978; Azpiroz and Malagelada, 1984; Notivol et al., 1995). More specifically, in between meals (interdigestive phase) the fundus maintains a high basal muscle tone (a state of continuous partial contraction). Upon food intake, the muscle tone of the proximal stomach suppresses, but will never be switched off completely (Notivol et al., 1995; Schwizer et al., 2002; Tack et al., 2002; Janssen et al., 2011). This enables gastric accommodation and enhances the storage capacity of the stomach by increasing the compliance of the fundic stomach muscle (Villanova et al., 1997; Kindt and Tack, 2006). These findings suggest an eccentric contraction behavior of the fundus during ingestion, even though at suppressed gastric tone (Azpiroz and Malagelada, 1984; Janssen et al., 2011). Accordingly, we expect enhanced forces following active muscle stretch which might subsequently support gastric emptying. This potentially explains the functional relevance of significantly enhanced forces during and after lengthening as observed in this study. Hence, a protective and supportive function of FE prior to gastric emptying to avoid excessive distension of the stomach might represent specific adaptations to gastric functionality. Furthermore, FE might counteract rapid extension of gastric tissue due to acceleration of gastric contents in impact situations e.g., at ground contact during jumping or tumbling. Additionally, the large working range of about 3.7 L_S, high muscle tension of smooth gastric tissue, as well as the ability to withstand large eccentric forces, enables the fundus to operate suitably during filling.

FD might occur following food consumption (postprandial phase), when the fundus propels the gastric content distally by a tonic concentric contraction, going along with considerably dimensional changes accompanied by slow muscle shortening (Schulze-Delrieu et al., 1998; Farré and Tack, 2013). Anyway, high FD values seem to be functionally counterproductive for this region and could be seen as an unwanted by-product of FE (Rode et al., 2009). Consequently, the observed muscle properties of the fundus muscle tissue might represent functional adaptations to cover prevailing conditions. However, to verify the hypothesized relation between contractile properties and stomach function, structural reasons for history-dependent effects in smooth muscles as well as potential mechanisms of adaptation have to be analyzed in prospective studies.

Concluding, the findings provided by the current research support the idea of a holistic reflection of stomach structure and function, extending well-known mechanisms and processes (such as distinct signal pathways, various triggered mechanoreceptors) by history-dependent effects. A key role in explaining the enhanced and depressed total force-responses during active stretching and shortening, respectively, might represent the titin-like structure smitin. This semi-active element might behave like a spring during fundic distension and emptying, thus energy may be stored and recoiled as reported for skeletal muscles in stretch-shortening cycles (Roberts and Azizi, 2011), respectively. Consequently, an elegant, energy saving, elastic mechanism similar to the bouncing gait in vertebrate locomotion (Cavagna et al., 1977; Roberts and Azizi, 2011; Roberts, 2016) might enable, or at least enhance, efficient stomach function. The results presented here enable new insights into stomach function and might facilitate development and validation of realistic 3D muscle models of hollow organs like the stomach.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.