INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

IN THE HEART an increase in coronary perfusion results in an increase in cardiac oxygen consumption and strength of contraction (cardiac contractility). The mechanism of this so-called Gregg phenomenon (8, 10) is still unknown. Previous work has focused on two possibilities, namely, unrecognized local ischemia (7) or perfusion-induced changes in cardiac muscle length (the so-called garden-hose effect) (2).

Schouten et al. (12) concluded from isolated perfused rat papillary muscle experiments that neither ischemia nor muscle length changes were the cause of the Gregg phenomenon but that perfusion affects contractility per se, i.e., the contractile aspect of the Gregg phenomenon. We recently suggested that both aspects of the Gregg phenomenon are based on different mechanisms: the contractile aspect is related to perfusion pressure and/or (vessel wall) shear stress, whereas the oxygen consumption is related to coronary flow (5).

According to Bai et al. (3) the increase in myocardial oxygen consumption is found to be related to an increase in coronary vascular volume. This increase in coronary vascular volume is thought to produce a more rigid coronary hydraulic "skeleton," requiring greater energy expenditure (increase in myocardial oxygen consumption) for deformation during systole, which can be measured as an increase in developed left ventricular pressure (11). Indeed, perfusion-induced changes in cardiac contractility and oxygen consumption and changes in coronary vascular volume are greater during poor coronary autoregulation (3, 11, 14). Good coronary autoregulation protects the capillaries and postcapillary venules from changes in transmural pressure and shear, suggesting that the origin of the Gregg phenomenon is to be found in perfusion of the capillaries.

We therefore abolished changes in capillary perfusion by injection of plastic microspheres (15 µm). We could do this because oxygen supply in the isolated perfused papillary muscle does not depend on perfusion (12). We thus compared perfusion-induced changes in cardiac contractility before and after capillary obstruction in the isolated rat papillary muscle.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Preparations and setup. Details of the preparation and setup for studying isolated perfused rat papillary muscles have been presented earlier (12). In brief, the heart of Wistar rats (n = 7) weighing between 350 and 450 g was used. Anesthesia was induced by placing the rat inside a box containing ether vapor mixed with air. Subsequently the thorax was opened and the heart was rapidly excised and placed in a Langendorff setup. During preparation the hearts were perfused with Tyrode solution (see below) enriched with 15 mM KCl to obtain arrest. After the free wall of the right ventricle was removed, a suitable papillary muscle with the adjacent part of the septum and septal artery was removed. The excised piece of the septum was clamped under a stainless steel ring onto a Perspex plate to fix one end of the muscle and to prevent perfusate leakage via the cut edges. The tendon of the muscle was tied to a force transducer (type AE801, Mikro-Electronikk, Horten, Norway). The preparation was submerged in a muscle bath. The papillary muscle was stimulated electrically by pulses of 4-ms duration and 20-50% above threshold strength with a rate of 0.2 Hz to evoke isometric contractions. Isometric force at constant length is considered as a measure of cardiac contractility.

The septal artery was cannulated using a small (tip diameter 100-300 µm) polyethylene cannula that was connected to a pressurized reservoir via a glass capillary. The pressure difference (P) over the glass capillary was measured by a sensitive differential pressure transducer (type LX 160ID, National Semiconductor, Santa Clara, CA). Flow was found to be linearly related to P; thus after calibration P is a measure of flow. Before each experiment, the pressure drop over the perfusion system (from reservoir to cannula tip) was measured for different reservoir pressures. In this way perfusion pressure at the entrance of the coronary arterial system could be calculated. Muscle bath and pressurized reservoir were filled with identical Tyrode solutions. The Tyrode solution had the following composition (in mM): 140 NaCl, 5 KCl, 1 CaCl₂, 2 NaH₂PO₄, 1.2 MgSO₄, 10 glucose, and 5 N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid. The pH of the solution was adjusted to 7.3-7.4 at room temperature by addition of NaOH. The solution was equilibrated with 100% O₂.

The cross section of a papillary muscle, half-way between its base and tendon, is oval. Cross-sectional area was calculated from the muscle diameter measured in two perpendicular directions in unperfused condition at slack length with the use of a calibrated grid in the microscope. Muscle tension is expressed in force per cross-sectional area (mm²). An on-line video micrometer system was used to measure either segment length or muscle diameter continuously. In three of the experiments segment length and muscle diameter were measured simultaneously. For segment length measurements two small black self-adhesive markers (~0.3 × 0.05 mm) were attached to the surface of the muscle, at least 0.7 mm apart. Muscle diameter measurements were used to estimate the increase in muscle volume due to perfusion pressure changes (1).

After preparation and mounting were completed, the Tyrode solution was warmed from room temperature to 27°C and used to perfuse and superfuse the muscle. At 27°C the muscle is still provided with sufficient oxygen and nutrients by diffusion alone, so that perfusion may be varied freely or even stopped (12).

Experimental protocol. After a stabilization period of ~1 h the experimental protocol (Fig. 1) was started. A tension-segment length relationship was determined at the lowest perfusion pressure (P₀ = 12.9 ± 1.0 cmH₂O, mean ± SD) applied to the muscle. At this perfusion pressure level an active tension development is equal to a nearly unperfused papillary muscle. A second tension-segment length relationship was determined at the highest perfusion pressure (P_max = 106.5 ± 31.7 cmH₂O) applied to the muscle. This perfusion pressure was chosen to obtain the maximal increase in active tension (Gregg effect) and varied between muscles (12).

View larger version (30K): [in this window] [in a new window]   Fig. 1.   Schematic representation of the experimental protocol. T/L, tension-segment length relation; Tmax, maximal developed active tension; T/P, tension-perfusion pressure relation; L, segment length; P0, lowest perfusion pressure (nearly unperfused); P1 and P2, intermediate perfusion pressures; Pmax, maximal perfusion pressure.

Muscle length was then set at ~0.9 L_max with L_max defined as the segment length at which the passive tension was equal to active tension (i.e., total tension minus passive tension) at perfusion pressure P₀ (1, 12). Maximum developed active tension (T_max) at the chosen muscle length before microsphere injection was determined at P₀ by postextrasystolic potentiation or a high Ca²⁺ concentration in the superfusion solution (3.0 mM). The two methods result in a similar value for T_max (13). In total, four perfusion pressures were applied to the isolated papillary muscle in the range between P₀ and P_max. The two remaining perfusion pressures were P₁ = 66.6 ± 26.2 cmH₂O and P₂ = 89.1 ± 28.4 cmH₂O. Perfusion pressure was increased stepwise while perfusion flow, isometric tension, segment length, and muscle diameter were continuously monitored on a thermal array recorder and also digitized and stored on disk using an Olivetti M290 personal computer with a sample rate of 180 Hz.

A bolus of plastic microspheres (fully decayed radioactive-labeled microspheres, Du Pont, 15 µm) was carefully injected distal from the glass capillary via a three-way stopcock. During injection the stainless steel ring on the septum was temporarily loosened, and perfusate leakage via the cut edges was possible. In this way we induced high-perfusion flows in the perfusion system, thereby avoiding sedimentation and allowing the microspheres to reach the muscle quickly. The septum was clamped again under the stainless steel ring, and passive force was adjusted at approximately the same level as before. The amount of microspheres injected was adjusted in such a way that flow at the intermediate perfusion pressure (P₁) decreased to <20%. After a stabilizing period of 15 min, perfusion pressure was again increased stepwise (same range as during control). F_max was determined again by high Ca²⁺ concentration in the superfusion solution at 0.9 L_max . At the end of the protocol two tension-segment length relationships were made, at P₀ and at P_max.

Analysis and statistics. The tension developed by the papillary muscle was expressed in millinewtons per square millimeter. In each experiment the passive tension, active tension (total tension minus passive tension), and T_max (maximal total tension minus passive tension) during control and after microspheres at P₀ were compared using a paired Student t-test. The relationships between active tension, muscle diameter, segment length, and perfusion pressure were normalized with respect to their values at P₀.

For each experiment the relationship between flow and perfusion pressure during control and after microsphere injection was normalized by setting flow at P_max at 100%. All normalized values are expressed as means ± SE, and all other values given are expressed as means ± SD. Each relationship made was analyzed using parametric repeated-measures of analysis of variance. P levels < 0.05 were considered to be significant.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

We performed seven experiments in which perfusion pressure was changed before (control) and after microsphere injection. The injected microspheres were present inside the microvasculature on microscopic sections. Passive and active tension developed by the papillary muscle at P₀ were not significantly affected by microsphere injection (control vs. after microspheres: passive tension 2.2 ± 2.7 vs. 2.3 ± 2.5 mN/mm², active tension 34.8 ± 19.2 vs. 26.5 ± 10.3 mN/mm²). The response to postextrasystolic potentiation or high Ca²⁺ concentration (3.0 mM) in the superfusion solution (T_max) was also not significantly different after microspheres (72.1 ± 38.4 vs. 67.3 ± 31.9 mN/mm², Table 1).

All papillary muscles showed a Gregg phenomenon before microspheres were injected, that is, an increase in isometric tension with an increase in perfusion. An example is shown in Fig. 2, in which perfusion pressure changed from 14 to 35 cmH₂O (indicated by arrow) and flow (Fig. 2B) also increased stepwise. In all muscles, flow increased linearly with perfusion pressure. Muscle diameter increased significantly with perfusion pressure (Fig. 2C), but diastolic segment length was unaffected (Fig. 2D). The averaged pressure-flow relation before microspheres (Fig. 3) shows the absence of autoregulation of the coronary vascular bed. After microsphere injection, flows were significantly reduced in comparison with control, and at the lower perfusion pressures flow did not increase with pressure. A significant increase in coronary flow was found only at the highest perfusion pressure (106.5 ± 31.7 cmH₂O) applied. The average results of the relationships between normalized active tension, muscle diameter, and segment length with perfusion pressure are shown in Fig. 4. With constant segment length, muscle volume changes could be determined from cross-sectional area changes. The muscle volume increased significantly with 16.2 ± 9.3% by P₀ to P_max.

View larger version (37K): [in this window] [in a new window]   Fig. 2.   Typical recording of effect of increase in perfusion pressure (arrow) from 14 to 35 cmH2O on force (A and E), flow (B and F), muscle diameter (C and G), and segment length (D and H) before (left) and after microsphere injection (right).

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Averaged relation (means ± SE) between perfusion pressure and flow before () or after microsphere injection (). Flow values were normalized to flow at Pmax during control conditions. * Significantly different from flow at P0; # significantly different from flow after microspheres at similar perfusion pressure.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Averaged effect (means ± SE) of perfusion pressure and normalized active tension (at 0.9 Lmax; A), normalized muscle diameter (B), and normalized segment length (C). All values are normalized with respect to control values at P0. , Before microsphere; , after microspheres. * Significantly different from P0 value; # significantly different from value after microspheres at similar perfusion pressure. Lmax, segment length at which passive tension was equal to active tension at P0.

Figure 2, right, shows an example of the effect of an increase in perfusion pressure on force (Fig. 2E), flow (Fig. 2F), muscle diameter (Fig. 2G), and segment length (Fig. 2H) after microsphere injection (same experiment as in Fig. 2, left). The increase in flow after a change in perfusion pressure is significantly decreased compared with before microsphere injection. At intermediate perfusion pressure (P₁ = 66.6 ± 26.2 cmH₂O), the averaged flow before and after microsphere injection is 17.6 ± 5.4 and 3.2 ± 1.3 µl/min, respectively. After microsphere injection, none of the muscles showed an increase in active tension with an increase in perfusion pressure. At the two highest perfusion pressures even a decrease in active tension was seen. In Fig. 4, the averaged data are presented, and at P₂ and P_max a significant decrease in active tension was found. At P_max the decrease in tension was accompanied by a significant increase in muscle diameter (Fig. 4B), leading to a muscle volume increase of 7.5 ± 6.3% at this pressure. In all experiments in which a decrease in tension and an increase in muscle diameter was seen from P₀ to P_max, active tension increased and muscle diameter decreased again when perfusion pressure was returned to P₀. This shows that the changes in active tension and muscle diameter were reversible. The decrease in tension was not due to muscle length changes because segment length was unaffected by a change in perfusion pressure in all muscles (Fig. 4C, n = 3).

In Fig. 5 active tension-segment length relationships before and after microsphere injection in one typical experiment are shown at P₀ and at P_max. The relationships at P₀ before and after microspheres are very close together. A perfusion-induced increase in cardiac contractility is seen before microsphere injection, whereas a perfusion-induced decrease in cardiac contractility is seen after microspheres.

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Active tension-segment length relationships at unperfused (P0 = 12 cmH2O, circles) and at maximal applied perfusion pressure (Pmax = 82 cmH2O, squares) before (open symbols) and after (filled symbols) microsphere injection in 1 typical experiment.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The aim of our study was to determine whether the contractile aspect of the Gregg phenomenon, i.e., the perfusion-induced increase in force in the rat papillary muscle, is related to capillary perfusion. We showed that an increase in perfusion induced an increase in active tension under control conditions. After microsphere injection, capillary perfusion was blocked and the perfusion-induced increase in active tension was abolished. These findings are in agreement with the hypothesis that the Gregg phenomenon is related to changes in capillary perfusion (5). In the maximally dilated bed changes in perfusion pressure result in changes in capillary pressure and shear (5). After capillary perfusion was blocked by injection of plastic microspheres (15 µm) no effect or even a decrease in active tension, at high perfusion pressure, was found.

The isolated perfused rat papillary muscle, in the chosen conditions, is provided with sufficient oxygen and nutrients by diffusion alone so that the perfusion can be varied freely or even stopped so that injection of 15-µm microspheres does not affect availability of oxygen. In the myocardium all capillaries originate from arterioles <15 µm in diameter or from other capillaries (4). After microsphere injection flow did not decrease completely to zero (Fig. 3) because microspheres may not have blocked all the arterioles completely, i.e., some perfusion fluid may pass between them and the arteriolar wall. After microsphere injection a small diameter change remained, suggesting that arteries and arterioles are still subjected to perfusion pressure but that the capillaries are not, and capillary transmural pressure changes are abolished. Injection of plastic microspheres did not significantly affect the contractile properties of the isolated perfused papillary muscle. In some of the experiments there is a decline in tension but in others there is even a small increase at P₀. This decline could partially be explained by the small reduction in passive force developed by the muscle, which indicates a small decrease in muscle length. In these experiments many microspheres were injected and some damage to the muscle may have occurred. We did not exclude these experiments because an increase in active tension by adding extra Ca²⁺ to the superfusion solution still occurred, indicating functional contractile elements. The applied perfusion pressures at which the maximal effect of perfusion pressure on active tension was found differed considerably between the individual experiments. This may result from differences in the vascular tree and thus in vascular resistance between the excised piece of the septum and the papillary muscle, resulting in different relationships between perfusion pressure and papillary muscle flow.

After plastic microsphere injection the capillary (and venous) compartments of the vascular bed are no longer influenced by changes in perfusion pressure; i.e., capillary flow, shear stress, and transmural pressure changes do not take place any more. Perfusion pressure alterations also do not change coronary arterial/arteriolar flow and shear stress any more but transmural pressure in the arteries and arterioles still varies so that coronary vascular volume can change and small muscle volume changes follow. After microsphere injection the Gregg phenomenon is abolished while muscle volume increased a little (Fig. 4B); thus we can conclude that the Gregg phenomenon (at least its contractile aspect; 5) cannot result from changes in transmural pressure and coronary vascular volume at the arterial and arteriolar level. The Gregg phenomenon is thus to be related to either shear stress (or flow) at all vascular levels or capillary transmural pressure and its related changes in coronary vascular volume. In a previous study we concluded that the involvement of transmural pressure and thus the coronary vascular volume on the contractile aspect of the Gregg phenomenon is very unlikely (5). In the isolated perfused rat heart the Gregg phenomenon was found to be still present after the arterial and arteriolar endothelium was made dysfunctional (6), making shear stress an unlikely factor. In the same study it was also concluded that the Gregg phenomenon is not related to wash-in and washout effects of inotropic agents that could be related to changes in coronary flow (5). Most of these inotropic agents are produced and/or released at arterial and arteriolar levels. These two findings indicate that shear stress at the arteriolar level probably plays a minor role. Thus the combination of our previous work (5, 6) with the present findings strongly suggests that shear stress at the capillary level is the main factor involved in the Gregg phenomenon.

Although our present results strongly suggest that the mechanism of the Gregg phenomenon is to be found at capillary level and is probably related to capillary shear, the precise mechanism is still unclear. Changes in capillary perfusion are present when autoregulation is absent and bring about changes in washout of substances. Washout of a negative inotropic factor could explain the rather slow response of tension to abrupt changes in perfusion observed in our study (Fig. 2) because the removal of such an agent requires diffusion and washout and therefore takes time. Figueredo et al. (9) suggested that perfusion-induced changes in cardiac contractility are mediated by changes in inorganic phosphate (P_i) levels. Small changes in P_i are known to alter the relationship between cytosolic free Ca²⁺ and developed pressure.

The slow increase in muscle volume is not necessarily coupled to the slow increase in force (Fig. 4B) but is related to the development of edema. The time constant of force and volume change may therefore be different. After microsphere injection capillary perfusion is blocked and edema formation is less (i.e., diameter increase is smaller; Fig. 4), and removal of a negative inotropic substance after the blockade does not take place anymore.

After injection of microspheres a significant decrease in active tension (negative inotropic effect) was found at the two highest perfusion pressures applied to the muscle (Fig. 4A). This decrease in active tension was reversible and not related to changes in segment length because these were unaffected (Fig. 4C). We can only guess about the mechanisms involved in this decrease in active tension. The mechanism of the perfusion pressure-induced decrease in active tension may be located in arterial and arteriolar vessels >15 µm in diameter because perfusion of the capillaries was blocked by microspheres. This then suggests a negative effect of arterial/arteriolar distending pressure on active tension because flow changes in these vessels were not present.

We conclude that the contractile aspect of the Gregg phenomenon is related to capillary perfusion. It is most likely that perfusion-induced changes in shear stress at the capillary level are involved.