Coronary arterial inflow is impeded and venous outflow is increased as a result of the decrease in coronary vascular volume due to cardiac contraction. We evaluated whether cardiac contraction is influenced by interfering with the changes of the coronary vascular volume over the heart cycle. Length-tension relationships were determined in Tyrode-perfused rat papillary muscle and when coronary vascular volume changes were partly inhibited by filling it with congealed gelatin or perfusing it with a high viscosity dextran buffer. Also, myocyte thickening during contraction was reduced by placing a silicon tube around the muscle. Increasing perfusion pressure from 8 to 80 cmH2O, increased developed tension by ∼40%. When compared with the low perfusion state, developed tension of the gelatin-filled vasculature was reduced to 43 ± 6% at the muscle length where the muscle generates the largest developed tension (n = 5, means ± SE). Dextran reduced developed tension to 73 ± 6% (n = 6). The silicon tube, in low perfusion state, reduced the developed tension to 83 ± 7% (n = 4) of control. Time-control and oxygen-lowering experiments show that the findings are based on mechanical effects. Thus interventions to prevent myocyte thickening reduce developed tension. We hypothesize that when myocyte thickening is prevented, intracellular pressure increases and counteracts the force produced by the contractile apparatus. We conclude that emptying of the coronary vasculature serves a physiological purpose by facilitating cardiomyocyte thickening thereby augmenting force development.
- cardiac mechanics
- intracellular pressure
cardiac contractionejects blood from the ventricular lumen to the arterial network. During the contraction, coronary vessels decrease in volume (12), which results in a decrease in arterial inflow (flow impediment) and an increase in venous outflow in systole (16, 20). Classically, the alterations in vascular volume have been viewed as merely a consequence of the compressive forces generated by the contraction of the surrounding myocardium (16). However, because the coronary vasculature is intertwined with the myocardium, we hypothesize that the emptying of the vasculature serves a physiological purpose. Coronary vascular emptying could facilitate cardiomyocyte thickening and hence shortening (because intracellular volume remains constant), thereby contributing to an increase in contractile force generation of the muscle as a whole for a given force development by the contractile elements. If this hypothesis is true, then force generation should be attenuated when emptying of the coronary vasculature is prohibited by filling the vasculature with gelatin or a highly viscous fluid. This hypothesis would also predict that if thickness changes of the muscle are limited by encasing it in a tube, contractile force is attenuated. This hypothesis cannot be tested in the intact heart, because coronary flow is essential for oxygen supply. In contrast, the perfused isolated rat papillary muscle, under carefully chosen conditions, is not dependent on coronary flow because diffusion can provide sufficient oxygen (21, 22). Therefore, we used the papillary muscle and studied the effect of filling the coronary vasculature with gelatin or high viscous perfusion fluid (dextran) to reduce vascular volume changes and the effect of an external tube to reduce muscle thickening during contraction, with the aim to relate myocyte thickening and force development.
Preparations and Setup
Details of the preparation and setup for studying isolated perfused rat papillary muscles have been presented earlier (21,22) and were slightly adapted. In brief, hearts of male Wistar rats, weighing between 350 and 500 g, were used. All experiments were performed in accordance with the “Guiding Principles in the Care and Use of Laboratory Animals” as approved by the Council of the American Physiological Society and under the regulations of the Institutional Animal Care and Use Committee. Anesthesia was used by placing the rat in a box where ether vapor was mixed with air. Subsequently the thorax was opened and the heart was excised rapidly and placed in a Langendorff setup, where the heart was perfused with Tyrode solution containing 20 mM 2,3-butanedione monoxime to stop the beating (3, 14). A papillary muscle from the right ventricle with the adjacent part of the septum and septal artery was removed. The septal part was clamped under a stainless steel ring and the muscle submerged in the muscle bath. The tendon was tied to a force transducer (Mikro-Electronikk; Horten, Norway) to measure muscle force. The whole muscle is isometric, but considerable segment shortening (∼10%) occurs within this muscle (2, 22, 23). The muscles were constantly superfused with Tyrode and kept at 27°C (except for the gelatin experiments where we used 22°C) and continuously stimulated with a rate of 0.2 Hz with field electrodes using pulses of 50 V and the duration of 4 ms. The septal artery was cannulated using a small polyethylene or glass cannula connected to a pressurized reservoir (80 cmH2O) for perfusion. In the “low perfusion” state perfusion pressure was 8 cmH2O because of the difference in height between the muscle bath and the pressure reservoir. We define this as the low perfusion state with buffer. Perfusion with a pressure of 80 cmH2O is defined as the high perfusion state. The Tyrode solution for superfusion and perfusion had the following composition (in mM): the 140 NaCl, 5 KCl, 1 CaCl2, 2 NaH2PO4, 1.2 MgSO4, 10 glucose, 5 HEPES, and 0.01 adenosine to obtain maximal vasodilation of the coronary vasculature. The pH was adjusted to 7.45 and the solution was gassed with 100% O2. The cross-sectional area of the muscle was determined in the middle of the muscle and at the muscle length where the muscle generates the largest developed tension (L max). Area was calculated from the muscle diameter measured in two perpendicular directions, assuming the cross-section is elliptic. Muscle tension is expressed as force per cross-sectional area. Area was measured atL max (mN/mm2) and used throughout the experiment.
After a stabilization period of 1 h in whichL max was also determined, length-tension relationships were generated with length steps of 1% in the range from 85% of L max (L 85) toL max. This range was chosen because it has been shown to represent the physiological range (23). Two control length-tension relationships were determined, one with low perfusion and one with high perfusion, separated by a period of 20 min at 95% L max. Subsequently, one of the following conditions was studied by means of a length-tension relationship:group 1, coronary vascular filling with gelatin; group 2, perfusion with high viscous dextran; and group 3, prevention of whole muscle thickening during contraction by an external silicone tube.
All length-tension relationships, including low perfusion and high perfusion, were performed at 22°C. The standard perfusion buffer was changed to Tyrode containing 5% (by weight) gelatin and Patent blue (Sigma) as a staining dye. The gelatin was a type A gelatin from porcine skin (G-2625, Sigma). The reservoir containing this gelatin solution was heated to 37°C, to be able to perfuse the muscle. When the gelatin reached the main supplying artery (visualized by Patent blue), muscle contractions were interrupted by stopping the pacing. When the gelatin came out of the thebesian openings, perfusion was stopped to let the gelatin congeal at 22°C (room temperature). After 15 more minutes (for congealing of the gelatin), pacing was resumed and a tension-length relationship was constructed again.
The contribution of the gelatin may depend on whether the muscle is filled at L max or at L 85(7, 19). Filled at L max, the gelatin is compressed during a tension-length relationship fromL 85 to L max, whereas filled at L85, the gelatin is stretched. For this reason, we used two subgroups of five muscles each where we filled the vasculature and let the gelatin congeal at the two extremes of the length-tension relationships, L 85 andL max.
Possible biochemical effects of gelatin on muscle contraction were tested by adding gelatin to the superfusion fluid (n = 2, at 37°C to prevent congealing). Osmolarity of the buffer containing 5% gelatin was compared with the standard buffer (030, Osmomat; Gonotec, Germany). At the end of one experiment, we heated the superfusion fluid and therefore the muscle to 37°C with liquefaction of the gelatin as a result. Gelatin was flushed out, the muscle was cooled back to 22°C, and the low perfusion length-tension relationship was measured again.
Dextran with a molecular weight of 500,000 (D-5251, Sigma) was added to the perfusion fluid until the viscosity of the buffer was increased 10 times. Perfusion was with Tyrode-dextran and superfusion with Tyrode alone. We used this high-molecular-weight dextran because it is biologically inert (27). During perfusion with dextran (also at 80 cmH2O perfusion pressure) when muscle tension had stabilized, length-tension relationships were determined (n = 6). Subsequently, in all muscles, the perfusion buffer was changed back to the standard Tyrode at high perfusion pressure to test reversibility of the intervention. To test for a possible limited oxygen supply during dextran perfusion, in one muscle Po 2 was gradually reduced (from >600 to 150 mmHg, ABL 50, Radiometer; Copenhagen, Denmark) in the superfusate and perfusion fluid (with dextran), and tension was measured.
A silicone tube with an internal diameter of 0.4 mm, which was perforated with a 18-gauge needle, was placed over the string connecting the tendon with the force transducer (n = 4). After the determination of the control length-tension relationships and after pacing was interrupted, the tube was slipped over the papillary muscle. To prevent movement, the tube was secured with a tie to the tendon. Length-tension relationships with the tube in place were constructed with low and high perfusion. To show reversibility, the tube was removed and control length-tension relationships were constructed. The wall thickness was 200 μm. The permeability of silicon for oxygen (24) is about 10 times higher than the permeability for oxygen in rat heart muscle (25). With this difference in diffusion rate, this equals ∼20 μm of cardiac tissue. With a diameter of 300–400 μm for the papillary muscles this is negligible. In two muscles, with tube present, Po 2 was gradually reduced (from >600 to 150 mmHg) in the superfusate (muscle in low perfused state) to test whether oxygen supply was the limiting factor in this experiment.
Time-control experiments were performed that lasted 90 min after stabilization. The determinations of the effect of gelatin and dextran lasted 60 min (after 1 h of stabilization). The one experiment where gelatin was washed out (see Gelatin) lasted 3.5 h. Therefore, we also collected data in the control experiments at 3.5 h.
Analysis and Statistics
All length-tension data were stored with a program developed in our department. Data were sampled with a frequency of 500 Hz. Developed tension was calculated as systolic tension minus diastolic tension. Diastolic tensions were normalized with respect to diastolic tension atL max in the muscle in low perfused state as 100%. Developed tensions were normalized with the tension atL max of the muscle in low perfused state as 100%.
All values are expressed as means ± SE. The effects of changing vascular mechanics were evaluated by ANOVA at all measured lengths. For the statistics, the raw data were employed, whereas for clarity in the figures, normalized values are shown. P values <0.05 obtained during post hoc testing were statistically significant.
Diastolic and developed length-tension relationships after filling the vasculature with gelatin are compared with the muscles during the low and high perfusion states with Tyrode solution. The two conditions during filling (L 85 andL max group) are presented separately. The diastolic tension length-relationships (Fig.1, A and B) with low perfusion, high perfusion, and after filling the vasculature with gelatin were not different. Developed tension atL max during low perfusion state with buffer was 76 ± 9 mN/mm2 for muscles in theL 85 group and 69 ± 9 mN/mm2for muscles in the L max group. At all muscle lengths, all individual muscles showed an increased developed tension when changing from low to high perfusion state with buffer (Gregg phenomenon) and a decreased developed tension due to congealed gelatin in the vasculature. In the L 85 group measured atL max, increased perfusion significantly changed developed tension to 143 ± 15% of the muscle during low perfusion with buffer. Gelatin-filled muscles atL 85 reduced the developed tension to 43 ± 6% of the muscle in low perfused state (Fig. 1 D). There was no statistical difference in the reduction in the developed tensions of the muscles filled at L 85 and those filled atL max (Fig. 1, D and E).
Direct biochemical effects on muscle function were tested by adding gelatin to the superfusion buffer. To prevent congealing, length-tension relationships were measured at 37°C before and after superfusion with gelatin. There was no change in developed tension even after 30 min, which was the exposure time to gelatin in the perfusion experiments. Osmolarity of the gelatin buffer (322 ± 4 mosmol) did not differ from the osmolarity of the Tyrode buffer (317 ± 9 mosmol).
Immediately after the gelatin reaches the muscle, developed tension is reduced to about 65% of the tension before gelatin perfusion. The final reduction of developed tension is maximal when the gelatin is congealed. These observations rule out a time effect. In one experiment with gelatin in the vasculature, we heated the muscle to 37°C with liquefaction of the gelatin as result. Gelatin was flushed out and subsequently the muscle was cooled to 22°C. The developed tension at L max during low perfusion with buffer returned to 80% of its control value, whereas with the gelatin-filled vasculature, the force had been reduced to 54% of control. Time control experiments show that after 3.5 h (length of experiment where gelatin was washed out) the developed tension was significantly reduced to 79 ± 6% (n = 5 muscles) of the developed tension at the start of the experiments. Thus filling with gelatin does not have an additional effect on tension compared with the time effect.
Diastolic and developed length-tension relationships are shown in Fig. 1, C and F, respectively. The diastolic length-tension relationships of the dextran-perfused muscles were not different from the muscles during the low perfusion with buffer and high perfusion with buffer except at L max, where a statistically significant difference was reached. Compared with the muscle using low perfusion buffer (developed tension, 71 ± 10 mN/mm2 at L max), all individual muscles showed (Fig. 1 F), at all muscle lengths, a significantly increased developed tension using the high perfusion with buffer and a significantly decreased developed tension during perfusion with dextran. During the high perfusion state with buffer and with dextran, perfusion developed tension at L max was increased to 140 ± 9% and decreased to 73 ± 6%, respectively.
Immediately after dextran reached the muscle, developed tension was reduced to about 80% of the tension before dextran perfusion. Within a few minutes after the start of the dextran perfusion the developed tension was stable. These observations rule out an effect of time.
During low perfusion with buffer after rinsing out the dextran, developed tension at L max was 59 ± 9 mN/mm2. This tension was not significantly different from the developed tension in low perfusion with buffer measured before dextran perfusion.
No difference in osmolarity was found between dextran and Tyrode buffer. Change in colloid osmotic pressure was also negligible (<3 mmHg). Oxygen tension was reduced (from >600 to 150 mmHg) in one muscle during dextran perfusion. The developed tension atL max differed <3% (not significant) within this range of oxygen tensions.
Diastolic length-tension relationships (Fig.2, top) did not differ under the four conditions, i.e., with low and high perfusion both with and without the silicone tube. In both high perfusion state and low perfusion state, developed tension was decreased when a silicone tube was present. Before application of the tube, the developed tensions atL max were 78 ± 11 and 98 ± 11 mN/mm2 (126%) with low and high perfusion states, respectively. After placement of the tube, atL max, developed tension was 65 ± 6 mN/mm2 (83%) in the low perfused state and 85 ± 5 mN/mm2 (109%) in the high perfusion state. With the tube present, developed tension was always smaller than in the unrestricted muscle when high perfusion and low perfusion conditions were compared.
Diastolic tensions of muscles during the high perfusion state in were slightly reduced when the tube was around the muscle. Reducing the Po 2 gradually to 150 mmHg with the tube present had no significant effect on length-tension relationships. The developed tension at L max differed less than 3% (not significant) over this range of oxygen tensions.
These results show that the systolic function of the papillary muscle does not decrease significantly within the time span of the dextran and gelatin experiments (60 min). We performed control experiments and over a time period of 90 min and within this time span developed tension reduced to 83% (n = 5), which was not significantly different.
The major findings are that when emptying of the coronary vasculature is impeded during contraction (gelatin, dextran) or when muscle diameter increase is restricted (tube), tension development is decreased at all muscle lengths. Because diastolic length-tension relationships were not affected and developed tension was decreased, the implication is that systolic tension was reduced at all lengths. Thus limiting myocyte thickening during systole results in a decrease in developed tension. This means that emptying of the coronary vasculature facilitates externalization of force generated by the contractile apparatus during cardiac contraction.
We explain these findings using the hypothesis that the tension measured (Tm) equals the difference between the tension generated by the contractile apparatus (Tca) minus intracellular pressure (Pic). All variables can be expressed in the same units (with 1 mN/mm2 = 1 kPa = 7.5 mmHg). Equation 1The effect of the coronary vasculature is shown schematically in Fig. 3. The myocyte shortens during contraction, which leads to a rise in intracellular pressure (17). Because of the incompressible nature of the cytosol, cell thickness increases with cell volume constant. Thus when the myocyte shortens and is able to thicken, intracellular pressure will remain low. When restricted, intracellular pressure will become large, and shortening is reduced. When the muscle is held isometrically and no shortening occurs, no intramyocardial pressure develops and force equals that of the contractile apparatus. For our hypothesis, it is necessary that muscle shortening occurs within the papillary muscle. Although we use muscle isometric contractions, there are regions with obvious segment shortening. As a result, some segments will lengthen. The measured tension is equal for every cross section, but total tension is determined by the strongest segments. Thus the strong regions generate the force and they will shorten and thicken. The weaker regions lengthen and get thinner. During dextran perfusion, filling with gelatin, and external tube, only the strongest shortening segments are hindered and therefore have an effect on developed tension.
Segment shortening of ∼10% within a muscle-isometric contraction is normal and described in the literature (21, 23). This segment shortening also leads to systolic flow impediment. Schouten et al. (22) show in rat papillary muscle that arterial inflow is reduced during contraction and also that venous outflow is increased during systole. Thus during contraction of the papillary muscle coronary vascular volume is decreased.
The generated intracellular pressure induced by cardiomyocyte shortening has until now not been reported in literature. Rabbany et al. (17) used isolated skeletal muscle cells of the giant barnacle as a model for the cardiomyocyte. Both isotonic and isometric contractions were studied. The isometrically contracting myocytes with peak force about 60 mN/mm2 did not develop intracellular pressure (Fig. 4 of Ref.17). During 20% shortening the intracellular pressure increase was about 7 kPa (7 mN/mm2). This approximated the difference in peak force between isometric and isotonic contraction, which was about 10%, i.e., about 6 mN/mm2. The increase in intracellular pressure of 7 kPa is equivalent to ∼52 mmHg, which should provide sufficient lateral force to displace coronary microcirculatory and venous volume during systole.
An increase in perfusion pressure has several effects on tension generation, this is called the Gregg phenomenon, i.e., with an increase in perfusion pressure developed tension increases in a maximally dilated coronary vasculature (1, 9, 12, 13).
Several factors play a role. First, an increase in perfusion pressure could increase systolic stiffness (5), leading to a more isometric contraction and thus an increase in developed tension. Second, increased perfusion results in increased intracellular calcium, also leading to increased tension (8). Third, increased vascular filling facilitates myocyte thickening. With a more filled vasculature, especially the myocytes in the core of the papillary muscle can increase their diameter during contraction more easily than without perfusion. The myocytes in the outside layers are probably less limited in their increase in diameter during contraction when the vasculature cannot be squeezed.
Previously, experiments have been performed after blockade of the microvasculature of the isolated papillary muscle with 15-μm microspheres (10). After a few contractions, blocking the small arterioles leads to emptying of the distal microcirculation and veins (1). When perfusion pressure is now increased the arteries and arterioles larger than 15 μm are more difficult to empty and, based on our hypothesis, developed muscle tension should be smaller, which is indeed the case. Thus, when the emptying of the vasculature is hindered, as with blocking the vasculature with microspheres, the effects of perfusion (see above) are overruled by the limitation imposed on myocyte thickening.
We conclude that hindered vascular emptying has a different effect on tension generation than perfusion pressure. A limitation of emptying reduces tension and perfusion increases tension. In control conditions with normal Tyrode buffer perfusion, the emptying and perfusion effects may both be present. Bai et al. (4) and Bian et al. (5) reported that an increase of perfusion pressure even to 180 mmHg had no negative effect on contractility, suggesting perfusion effects to be dominant.
During experimental conditions, which hinder vascular emptying using gelatin, dextran, or silicon tube, the only variable is the degree of limitation of myocyte thickening during contraction. Perfusion pressure is not changed thus cannot explain any of the reported differences. Figure 4 shows the effect of the severity of restriction on myocyte thickening on developed tension (at L max) as percentage of the low perfusion with buffer. When volume reduction of the coronary vasculature was prohibited during contraction (gelatin) reduction in developed tension was largest. When volume reduction of the vasculature was impaired by high viscous perfusion fluid (dextran), reduction of developed tension was less than with gelatin filling.
Skinned myocytes can provide some evidence for our hypothesis. When cells are skinned, they are no longer isovolumic during contraction. This should lead to an increase in measured force because of the reduced generation of intracellular pressure. Brandt et al. (6) found that the steady-state tension produced by maximally activated skinned cardiac myocytes with high calcium was much larger than the developed tension in intact muscle fibers. Van der Linden et al. (25) have made a finite element model that applies to force generation of skeletal muscles even at slack length. It was shown that force development in skinned fibers was still present at lengths smaller than the slack length of the intact fibers, suggesting a contribution of intracellular pressure in intact cells. Most studies, however, report a reduced developed tension after skinning. This can be explained by a reduced myofilament Ca2+ sensitivity in skinned muscle (11) and a loss of functional proteins (15).
We used rat papillary muscle for different reasons. First, oxygen supply by diffusion is sufficient without perfusion (21,22), whereas isolated or in vivo hearts when subjected to such an intervention would have a reduced performance. Second, it is rather easy to obtain a perfused rat papillary muscle. Third, myocytes and vasculature are running in parallel (18); in whole hearts, the architecture of cardiomyocytes and coronaries is much more complex, but we believe that the present results are applicable to the whole heart because squeezing of the vasculature during systole also occurs in the whole heart (12, 16, 28).
The length-tension relationships we made were fromL 85 to L max, which is within the in vivo working range of heart muscle (23).L max did not change during the interventions and over the time course of the experiments, indicating that sarcomere length was also not changed. Our time control results prove that the reduced developed tension due to gelatin and dextran are not due to time effects. This was corroborated by the one experiment where gelatin was removed and by the washout of dextran. Also, we observed an immediate effect of gelatin and dextran perfusion ruling out a time effect. Also, with gelatin in the vasculature, developed tension was not reduced within 60 min after tension had stabilized; thus accumulation of waste products within this time span is not expected. With dextran there is still perfusion, and, therefore, supply of metabolic fuels and removal of waste products is not stopped. Also, length-tension relationships (low perfusion and high perfusion) after the silicone tube experiments showed no difference with the ones made before the silicone tube was shifted over the muscle. We reduced oxygen tension from >600 to 150 mmHg and showed no decrease in tension development. These results and those published earlier (22) show that even a reduction of oxygen tension to 25% has no effect on the papillary muscles without internal perfusion (diameter 0.3–0.4 mm).
Ideally the internal diameter of the tube should approximate the muscle diameter in diastole but limit the increase in muscle diameter during shortening. However, we found that most papillary muscles were considerably triangular. Therefore, it was not possible to standardize the amount of limitation of the increase in diameter by the tube, but the effect was sufficient to allow detection of a blunted force development. Because of the constant diameter of the tube, one would expect the tension reduction by the tube to become less at an increased muscle length. This is confirmed by our findings. Thus when we compared the muscles in low perfused state with and without the tube, we found at L 85 a reduction of 38% in developed tension but only a 13% reduction at L max. Similarly, for the high perfusion state, we observed a reduction of 32% at L85 and 13% at L max.
Osmolarity of gelatin and dextran was also tested. Because of the high molecular weight (500,000) of dextran, relatively few molecules were needed to increase viscosity. No change in osmolarity was measured with dextran or gelatin added to the Tyrode buffer.
Direct biochemical effects on muscle function are not present, and, therefore, the effects are purely mechanical. This is corroborated by the experiments with the tube where mechanical limitation in thickening gave similar results as with gelatin and dextran in the vasculature.
In conclusion, impediment of vascular emptying results in blunting of developed force, which is mediated by impaired cardiomyocyte thickening. These findings support the hypothesis that vascular emptying is not merely a consequence of the cardiac contraction but rather acts to facilitate muscle function.
This work was supported by The Netherlands Organization for Scientific Research (NWO) Project Grant 902-16-175.
Address for reprint requests and other correspondence: N. Westerhof, Laboratory for Physiology, Faculty of Medicine, Institute for Cardiovascular Research, Free Univ., van der Boechorststraat 7, 1081 BT Amsterdam, The Netherlands (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2001 the American Physiological Society