Structure/function interface with sequential shortening of basal and apical components of the myocardial band

Abstract

Objective: To study the sequential shortening of Torrent-Guasp's ‘rope-heart model’ of the muscular band, and analyze the structure–function relationship of basal loop wrapping the outer right and left ventricles, around the inner helical apical loop containing reciprocal descending and ascending spiral segments. Methods: In 24 pigs (27–82 kg), temporal shortening by sonomicrometer crystals was recorded. The ECG evaluated rhythm, and Millar pressure transducers measured intraventricular pressure and dP/dt. Results: The predominant shortening sequence proceeded from right to left in basal loop, then down the descending and up the ascending apical loop segments. In muscle surrounded by the basal loop, epicardial muscle predominantly shortened before endocardial muscle. Crystal location defined underlying contractile trajectory; transverse in basal versus oblique in apical loop, subendocardial in descending and subepicardial in ascending segments. Mean shortening fraction average 18 ± 3%, with endocardial exceeding epicardial shortening by 5 ± 1%. Ascending segment crystal displacement followed descending shortening by 82 ± 23 ms, and finished 92 ± 33 ms after descending shortening stops, causing active systolic shortening to suction venous return; isovolumetric relaxation was absent. Conclusions: Shortening sequence followed the rope-like myocardial band model to contradict traditional thinking. Epicardial muscle shortened before endocardial papillary muscle despite early endocardial activation, and suction filling follows active systolic unopposed ascending segment shortening during the ‘isovolumetric relaxation’ phase.

1 Introduction

In spite of recent remarkable progress in understanding cellular myocardial function at the genetic/molecular level, the advancement in developing a comprehensive understanding of ventricular structure has been limited to microscopic scales [1,2]. These concepts are mainly based on Streeter's [3–5] two-dimensional measurements of a uniformly changing angle of orientation of myocardial microscopic fibers from epicadium to endocardium. Streeter's findings have helped us to recognize the intricate nature of local fiber structure and perhaps function. However, in order to understand cardiac function and its efficiency from the point of view of vectors of force generated by cardiac sarcomeres, one needs to develop a global three-dimensional model of ventricular structure adhering to the local and global dynamics of myocardial spatial architecture.

Streeter recognized the seminal conceptual framework of Krehl's Treibwerk representation of the figure-of-eight oblique fiber pathways, connecting inner and outer cones, and verification of this idea by the dissection techniques of Torrent-Guasp that vindicates the Treibwerk, showing a highly ordered compact structure [6]. The re-introduction of a helical rope-like heart muscle concept by Torrent-Guasp et al. [7–9] offers a radically different view thus challenging the prevailing anatomical views of myocardial structure and function. In this concept, the ventricular structure consists of a single myofiber band, with a principal fiber pathway extending from the right ventricular muscle just below the pulmonary artery, to the left ventricular muscle attached to the aorta, twisted into a transverse outer shell and then wrapped into an oblique double helical coil. In this view (Fig. 1a–c, and on video on http://www.gharib.caltech.edu/∼heart/) both the outer circumferential wrap and inner helical fiber bundle, which weave through myocardial substance, provide a preferential pathway for possible sequential contractile dynamics.

The validity of any cardiac model relies upon showing that the structure can explain physiological function when activated. Torrent-Guasp's concept, if correct, offers interesting opportunities to explain ventricular shape and mechanics based on realistic fiber shortening and dynamic rearrangement. Also, unlike the current views, it offers a unified concept of systolic ejection and the early suction phase of ‘diastolic’ cardiac function. It provides a unifying theme because it can explain both phases of cardiac function through a sequential contractile action. This is in contrast with the conventional acceptance of elastic recoil from stored potential energy during systole to explain isovolumetric relaxation [10–13]. With Torrent-Guasp's structural model, a sequential contractile wave through the preferential helical band can induce a coil-like twisting and reciprocal twisting in the opposite direction of the left ventricle and septum to cause the physiologic events of ejection and suction.

Torrent-Guasp's model, with its main emphasis on the fiber architecture and dynamics, has been criticized for not being compatible with the Purkinje network of electrical activation and its implications for the electromechanical and mechanical determinants of cardiac function. Also, the mere existence of a preferential pathway for the contractile wave has been questioned [14,15]. While the precise linking of excitation and contraction may require extensive undertaking, the latter concern over the time related contractile process can be possibly resolved through advanced MRI techniques [10,16] and, to a more limited extent, through direct mapping of contractile wave pathways by direct measurement of local myocardium mass along Torrent-Guasp's proposed helical fiber band.

The main objective of this paper is to mechanically probe the intact cardiac structure with sonomicrometer crystals [17,18], thus testing the hypothesis that the contractile wave produces fiber shortening along Torrent-Guasp's myocardial band and such contractile wave can move sequentially with its principle axis (maximum displacement orientation) aligned with local fiber bundle orientation. We used multiple sonomicrometer transducer-pairs to identify the principle angle of local fiber bundles at subendocardial/epicardial positions throughout the left and right heart.

2 Material and methods

All animals received humane care in compliance with the ‘Principles of Laboratory Animal Care’ formulated by the Institute of Laboratory Animal Resources and the ‘Guide for the Care and Use of Laboratory’ prepared by the National Institutes of Health (NIH Publication No. 86-23, revised 1985).

Twenty-four Yorkshire–Duroc pigs (27–82 kg) were premedicated (ketamine 15 mg/kg, diazepam 0.5 mg/kg intramuscularly) and anesthetized with inhaled isoflurane 1.5% (MAC 1%) throughout the operation. Support with a volume-controlled ventilator (Servo 900C, Siemens-Elema, Sweden) was started after tracheostomy and endotracheal intubation. The femoral artery and vein were cannulated and arterial blood gases measured to keep oxygen tension, carbon dioxide tension, and pH values within the normal range. A balloon-tipped catheter (Model 132F5, Baxter Healthcare Corp., Irvine, CA, USA) was advanced into the pulmonary artery through a jugular vein to measure cardiac output (thermodilution technique) and pulmonary artery pressure.

The pericardium was incised after median sternotomy and a solid-stated pressure transducer-tipped catheter (Model MPC-500, Millar Instruments Inc., Houston, TX, USA) was inserted through the apex to monitor left ventricular pressure (LVP). Regional contractility within the right and left ventricles was measured with pairs of 2 mm ultrasonic microtransducer crystals (Sonometrics, London, Ont., Canada). Each pair of crystals was oriented in order to measure contractility at a certain myocardial depth and orientation. The placement position of each crystal was made by using a 1 mm cut of the epicardium and introduction of the crystal to reach the depth selected. In the left ventricle two depths were chosen, endocardial, where the crystals were positioned transmurally to reach the inner surface via the ventricular cavity, or subepicardial, by insertion of 1 mm deep into the ventricular muscle. In the right ventricle, crystals were positioned in the endocardial wall, either in the outflow tract, or laterally by the atrioventricular groove. Aortic pressure, LVP, dP/dt, and sonomicrometer crystals data were digitally processed by specific hardware and software (Sonometrics). Velocity of sound through cardiac tissue was fixed to 1590 m/s. Sonomicrometer measurements were recorded with a sampling rate of 95.8 samples per second, a transmitter spacing of 652 μs, transmit inhibit delay of 1.81 μs, and transmit pulse length of 375 ns. Synchronicity between myocardial contractility was compared to left ventricular performance with 1 ms precision, by real-time plotting and processing of segment shortening, EKG, LVP, and dP/dt. Sequence of contraction of different segments of the heart was then established and compared with ventricular hemodynamics. All cases were performed and analyzed by the same surgeon.

3 Experimental protocol

The pattern of orientation is described in the Section 4. This dimension (angle of highest contractility relative to the long axis of the heart) was registered for both endocardial and subepicardial contraction, and compared with synchronized EKG, LVP, and dP/dt. These measurements were compared in real time with different pharmacologic changes of regional contractility, and position sites in the left and right ventricles.

4 Crystal orientation

Francisco Torrent-Guasp's model of the helical heart is presented in Fig. 1. In this model, the cardiac structure includes two simple connected loops that start at the pulmonary artery and end in the aorta. The first loop, termed basal loop, includes a horizontal basal band, which comprises right and left segments surrounding the right and left ventricles. This basal band connects to the second apical loop through a downward spiral fold that causes an oblique orientation of the band fibers that form the apical loop which consists of an endocardial or descending helical segment and an epicardial or ascending helical segment with opposing and oblique directional fiber pathways connected at the apical tip. The oblique fibers of the ascending segment complete their trajectory by bifurcating (as aberrant fibers) as shown in Fig. 2[19] to either the septal fibers coursing deeply toward the aorta, or stay superficial and jump from the anterior left ventricular surface of the right ventricle to connect to the pulmonary artery, A-V rings and posterior perimeter of the ventricular base Fig. 3 compares a silicon rubber cardiac mould [20] to the anatomic heart counterpart.

The sonomicrometer crystals were placed into the intact heart to test any relationship between sequential temporal and mechanical extent of fiber shortening between couples of crystals to model shape and timing. Their orientation was nested within and without the principal pathways comprising the suggested rope-like arrangement of the helical heart. Each of these internal and external suggested principal direction trajectories of transverse and oblique muscle mass were assumed to comprise the architectural scaffold of the intact heart by inserting couples of micrometer crystals both in the suspected pathway, and angles opposite this direction in the same site.

In order to better describe the orientation of the different couples of micrometer crystal pairs, their positions were described by depth and angle of line connecting them. To measure the angle of the line connecting the crystals in each pair, a local coordinate was used (Fig. 3); the aortic annulus was the highest point and the apex the lowest, in a North and South fashion. Following the North and South simile, contraction at 0° was considered when both crystals were oriented North to South, 90° East to West, 45° Southwest to Northeast and 135° Northwest to Southeast. The interchange of angle orientation at a specific site allowed recording of the different magnitude of contraction at the same location, but with a varied angulation. For example, an endocardial location can be recorded at 0° and 90° to analyze if these crisscross, and different fiber planes determined how forcefully paired crystals shortened relative to one another.

5 Results

5.1 Anterior wall of the left ventricle

When searching for the orientation of the micrometer crystals with highest contraction in the endocardial wall, segment shortening was most powerful at angles between 80° and 90°, reaching 17–27%, with the more extensive shortening in the apical than the basal position (Fig. 4, traces 1 and 2). In contrast, placing the crystals at 0° showed reduced contraction segment shortening (Fig. 5). The onset of contraction at this anterior wall myocardial depth precedes the systolic rise of LVP and dP/dt, and occurred between the Q and R waves of the EKG (Fig. 6). Subendocardial muscle shows two distinct rates as it shortens. First, a short and steep descent followed by a longer and less steep contraction phase. A notch was present on this curve, and divided these phases. The magnitude of this notch increased as crystals were placed closer to the papillary muscle (Figs. 7 and 8⁸). In normal hearts, the end of endocardial contraction consistently coincided with the beginning the descent phase of the left ventricular pressure, which also coincided with the appearance of the negative slope of dP/dt (Fig. 3).

At the same position (same distance from apex) subepicardial contraction averaged 12 ± 2% segment shortening (Fig. 4), when the angle of crystal placement was oriented at 150 ± 10°%, and placed approximately 60° opposite endocardial placement. Conversely, the extent of contraction was 6% at a 45° angulation, so that the 90° crisscross angulation between the subendocardial and subepicardial crystal pairs usually produced the greatest contractile force. We observed a delay of 82 ± 23 ms for the onset of subepicardial segment shortening with respect to that of the subendocardial segment. Typically, subepicardial segment shortening started at the maximum value of dP/dt, and finished 92 ± 31 ms after subendocardial contractions ended (Fig. 4).

The extent of contraction was more intense when the crystal pairs were situated near the apex in both the endocardial and epicardial sides of the left ventricle (Fig. 9). This reflected an anisotropic contractile effort as we compared basal and apical segment shortening. For example, basal contraction averaged 35 ± 5% less than apical contraction, and this general trend was consistent for both the endocardial and the epicardial muscles.

5.2 Posterior wall of the left ventricle

In contrast to the anterior wall of the left ventricle, endocardial and epicardial sides of the posterior wall presented near similar amount of contraction. Segment shortening ranged 20 ± 3% in the endocardial wall and 18 ± 3% in the epicardial one. However, the orientation of contraction was opposite in each side: better angle of contraction of the endocardial side was found at 0°, while at the epicardial side was 90° (Fig. 10). This fractional shortening variance was most evident at the epicardial site, as the transverse, or horizontal positioning showed 19% versus 7% for vertical placement. With endocardial placement the vertical crystal pair showed 23% versus 17% with transverse positioning. This continuum of contraction changed its organized pattern when the crystal touched the papillary muscle with a different fiber pathway (Figs. 7 and 8).

Despite these differences in the vigor of shortening between endo- and epicardial sites in the posterior inferior LV wall, the starting point of contraction was similar, shown in Fig. 10, a finding different from the ∼80 ms time delay in initiating shortening by crisscross sites of crystal placement evident on the anterior call, shown in Fig. 3.

5.3 Right ventricle

Two main types of contraction are found in the right ventricular free wall, depending upon site of crystal placement. In the right ventricular outflow tract beneath the pulmonary artery, contraction pattern with an orientation of 150 ± 10° started later than the lateral wall, and is similar to the epicardial side of the anterior wall of the left ventricle (Fig. 11a). This fiber orientation matches the oblique direction seen by cardiac unraveling (Fig. 11b). In the rest of the free wall of the right ventricle, contraction is oriented between 100 ± 5°. In the lateral free wall, close to the atrioventricular junction, contraction started at the time of LV anterior shortening, and was more prominent when the crystals were placed in a transverse (shortening fraction 23 ± 1% at 100°), compared to a vertical direction (SF% 10 ± 1% at 0°). When crystal pairs were placed at 45°, contraction was 6% (Fig. 12a and b).

5.4 Sequence of contraction

Contraction starts initially in the lateral free wall of the right ventricle, and simultaneously in the endocardial side of the antero-septal wall of the left ventricle. This observation is evident in Fig. 13, showing that the initial contraction slope is most vigorous in the endocardial segment of the left ventricle. The initiation of this early contraction corresponded with the Q wave of the EKG and initial slow LVP rise, as pressure rise remained below 15 mmHg. The sequence of events continues after a delay of 10 ± 5 ms with the contraction of the posterior wall of the left ventricle. The contraction of right and posterior left ventricular, and anterior endocardial wall segments began before contraction of the rest of the myocardium resulted in the rapid ascent in ventricular pressure recording that exceeded aortic diastolic pressure. Consequently, the early steep segmental shortening of the right free wall, posterior LV and endocardial regions (Fig. 13) occurred while there was no contraction detected in the subepicardial sites of the segments located in the subepicardial anterior wall of the left ventricle, as reflected by Fig. 14 in the right ventricular outflow tract.

Contraction at epicardial segments began 82 ± 23 ms after the initial muscle contraction (free wall of the right ventricle and the endocardial wall of the left ventricle), and corresponded to the peak of the positive dP/dt wave, the S wave on the QRS complex of the EKG signal (Fig. 6), and most importantly, the steep rise (equivalent to the peak dP/dt) in left ventricular pressure for ejection of blood. For the rest of systole during ejection, contraction was present in all segments of the heart resulting in ‘co-contraction’ of both endocardial and epicardial fibers. The initiation of contraction in the subepicardial fibers was maximal in a 90° opposite direction from endocardial ones, and coincided with a reduction of the slope of contraction of the subendocardial fibers (Fig. 4). Conversely, the contraction of the epicardial fibers did not change the slope of segmental contraction of either the lateral free wall of the right ventricle, or the posterior free wall of the left ventricle, as shown in Figs. 11–14.

We evaluated the sequential temporal and spatial relationships between the endocardium, posterior inferior LV wall, and epicardium, shown in Fig. 15 by comparing sites with the angle of maximum displacement between pairs of crystals in these regions. This time-related course of shortening started by displacement of endocardial muscle, followed ∼10 ms later by the posterior inferior wall, with initiation ∼80 ms later in epicardial muscle. The anterior LV muscle crystal angles, recording the maximal extent of crystal displacement, were in a superimposed crisscross (80° and 150°) pattern for deep and superficial, while a similar extent of the posterior LV wall maximal shortening occurred in deep and superficial muscle at 90° and 0°, with the same starting time in deep and superficial sites as seen in Fig. 10. Consequently, we could use this evaluation of the superimposed crisscross endo- and epicardial pathways to distinguish between the posterior LV and anterior LV wall, because the anterior wall showed an earlier endocardial starting point and ∼80 ms delay in epicardial shortening. Conversely, the posterior wall deep and superficial fibers started together, about ∼10 ms later (Fig. 15), with the absence of any time delay (Fig. 10) in the endocardial and epicardial regions.

As shown in Fig. 4, the first regions to stop shortening were the segments that started first, located in the endocardial side of the antero-septal left ventricle wall, and the free wall of the right ventricle (Fig. 14), and then 10 ms later by posterior free left ventricle wall. The end of contraction for these segments coincided with the end of the flat peak region in the left ventricular pressure tracing, corresponding to the onset of the negative dP/dt (Fig. 4). However, the epicardial component of left anterior wall and right ventricular outflow tract fibers continued their contraction phase for 92 ± 33 ms after the RV free wall, posterior LV, and endocardial LV segments ceased their contraction phase.

This extended time interval during which the epicardial segment that continues its contraction corresponds to the so-called ‘LV isovolumetric relaxation’ phase. In addition, this time interval, also, overlaps the LV's rapid pressure decrease or otherwise known as the ‘suction’ phase (Fig. 4).

Analysis of the pressure recordings, coupled with simultaneous analysis of regional recordings of contraction in endocardial, epicardial, lateral right ventricular and posterior left ventricular segments, shows a linkage between the acceleration and deceleration phases of developed pressure (i.e., the slope velocity of the pressure recording) and regions of contraction. During the initial rapid acceleration, or isovolumetric contraction of left ventricular pressure, all segments shortened simultaneously, whereas during the later deceleration of LV pressure, only the subepicardial segment was actively shortening. Consequently, there was no interval of isovolumetric relaxation.

The prolonged contraction of the anterior wall epicardial fibers during the cessation of endocardial shortening was associated with a reversal, or upward slope of the endocardial crystal tracing as shown in Fig. 4. Simultaneously, the basal segments of the free posterior wall of the left ventricle presented an increased distance between crystals (Figs. 14 and 16¹⁶), resulting in a rapid reversal of the slope of the crystal recording during endocardial noncontraction or relaxation (i.e., zero contraction). This separation or widening between crystals reached the point of maximum fiber stretch (i.e., separation between crystals), only surpassed by the added stretch due to ventricular filling by atrial contraction, as shown in Fig. 4.

The relationship between contraction of endo- and epicardial segments was examined by use of infusions of inotropes or β-blockers (Fig. 17). The time delay between the start of contraction in the endocardial and epicardial muscles of the anterior wall of the left ventricle decreased to 26 ± 7 ms when dopamine at 10 μg/(kg min) was given. Simultaneously the extent of shortening increased from 25.7 to 29.1% in the endocardial wall, and heart rate rose from 88 to 112 beats per minute to confirm the inotropic stimulation. In contrast, propranolol, a negative inotrope, prolonged the time delay between initiation of contraction in endocardial, and epicardial segments to 121 ± 20 ms, reduced the extent of shortening to 19%, and slowed the heart rate to 78 beats per minute.

This pharmacologic prolongation, induced by propranolol, increased phase separation of the onset of contraction between the endocardial and epicardial segments. There was also an associated prolongation of the duration of the endocardial contraction. Thus, the interval or hiatus of separation shortened between the ends of endocardial contraction with ongoing epicardial contraction. The phasic time frame for end of contraction of endocardial and epicardial fibers occurred at a closer interval (in contrast to the prolongation existing during the rapid ascent of the left ventricular pressure curve), so that there was prolongation of the slope of the rapid descent of the left ventricular pressure curve.

5.5 Negligible movement of the apex

When a pair of crystals was positioned close to one another at the point of the apex (Fig. 18), we found irregular, but noneffective contraction. The selected but noneffective shortening sites were adjacent and 1 cm apart, to include locations either toward or away from the septum, or in deep or superficial placement to evaluate endocardial and epicardial fibers. This positioning could not measure rotation of the apex.

5.6 Abnormal contractile movement of apex, with more distal fiber placement

A different pattern appeared when one crystal was placed into the tip of apex, with the other inserted into muscle (at a greater distance of 2–3 cm from the apex) that had previously shown a contractile pattern (Fig. 19). The second muscular crystal placement sites (distant from apex) evaluated deep subendocardial and more superficial epicardial positions.

These tracings, defined in Fig. 20a, showed absence of the expected endocardial pattern sequence consistently present in the mid wall (Figs. 4 and 19). The normal starting point that previously coincided with the beginning of ventricular pressure rise, shifted to the (∼80 ms) starting point of the epicardial fibers. However, this endocardial contraction ended normally, with termination ∼90 ms before the expected completion of the epicardial fiber contraction. Similarly, the expected epicardial contractile pattern sequence reflected reverse mirroring of the endocardial pattern. Contraction began ∼80 ms earlier, resembling initiation of endocardial placed crystals in other studies, with a normal completion point, typically ∼90 ms later than the stoppage point of endocardial crystals. Consequently, a paradox exists when an apex tip becomes the one placement point and normally contracting muscle becomes the other; the endocardial contractile pattern becomes ∼80 ms shorter, and the epicardial pattern becomes ∼90 ms longer than in expected sequences without an apical tip starting site.

6 Discussion

The intent of this study was to use sonomicrometer crystals with high temporal and spatial resolutions [21,22] to determine if the timing and maximum extent of shortening between crystal probes relates to the principal fiber pathway orientation suggested by Torrent-Guasp's helical heart model [8,20,23]. We explored the relationship of how these two dimensional devices could identify function in the underlying three dimensional muscle, which comprises an oblique double helix called the apical loop, that is surrounded by a transverse basal loop acting like a buttress covering the double spiral endocardial and epicardial apical components.

The crystal dimension gauges provide a local view of a global concept, exploring all cardiac regions, by probing how the maximal extent of fiber shortening in preferential pathways is governed by the orientation between couples of these crystals. These local barometers do not measure either thickening [21,24], twisting [10,11], torsion [12,25,26], cross fiber shearing forces [21,22], or inception of the calcium trigger of contraction [27,28]. These local shortening measurements are influenced by transmural shearing forces, yet they do reflect how deformation from position orientation receives substantial interactions from neighboring fibers. Consequently, each change in time-related regional function comprises how three-dimensional spatial architecture alters function within the ventricular wall. The consequence of unraveling the functional implications of a novel architectural pattern, is that, if correct, a linking will exist during conventional hemodynamic measurements, between the electrical signal, formation of upslope and downslope of ventricular pressure waves, and rate of maximal acceleration and deceleration of these standard signals.

Distinction between each of the varied factors that influence the term ‘contractility’ is not the intent of this manuscript, since no effort was made to measure deformation [24,29], as it influences strain of the cross fiber or transmural shearing forces [22,29] that may result in a motion that may not be aligned with local myofibers. However, such measurements were made by others [30], where findings of shear stress and torsion that is maximal in the endocardial and approximately twice [16,30,31] that of the epicardial region, is matched and consistent with the maximal extent of shortening during our comparisons of displacement between couples of crystals. Additionally, our observation of the anisotropic shortening showing the heterogeneous contractile action of both endocardial and epicardial regions, that increases toward the apex, compares favorably with tagging-related MRI reports of deformation [30].

Consequently, despite the limitations of the crystals used for sonomicrometric measurements, these local records take maximal advantage of their temporal and spatial resolution that we recognize is also influenced by the overlying cross fiber strain and shearing forces of the inner and outer halves of the ventricular wall. The point of maximal directional shortening or widening of opposing crystal pairs defines the composite geometric action of the either overlying or underlying ventricular muscle mass containing natural cross fiber connections. The centerpiece of this shortening observation is that the location where opposing crystals show maximum reduction of distance between them is at right angles to their angle of placement [22]. We suspect that the perpendicular angular shortening of a spring within each principal fiber pathway may become a clear example of this action, whereby limited shortening follows analyzing the cross-sectional site, with a major vector developing along the principal direction.

The study tested several factors, and these included defining the correct angle orientation between crystals at each site to determine maximal displacement. Each ventricular site was selected to determine if the principal fiber course shown in the unraveled heart described in Figs. 1, 21 and 22²², and on video on http://www.gharib.caltech.edu/∼heart/, could, when refolded into the intact muscle, reflect a consistency with (a) known anatomy, (b) the recognized course of the Purkinje electrical cable system, and (c) simultaneously follow the novel model of the ventricular band. After pairs of probes within a spatial configuration that selected the point of maximum displacement at regions within basal and apical loops were tested, we then evaluated if a sequence of time and position related force displacement was consistent with progression of the contractile wave around the unrolled rope by a trajectory moving from right to left in the basal loop, and then down the descending and up the ascending segments of the apical loop.

Clearly, any deviation from this expected sequence reflects either (a) an error in the model or (b) failure to take the conduction system into account. We determined if an anisotropic increment in contractile force followed the elliptical shape of the ventricular chamber, and also evaluated if these sequential contractions relate to interaction of calcium dynamics with the myocardial mass confronted with either positive or negative inotropic influences by dopamine or propanalol. Finally, we defined how (a) the apex itself, that visibly rotates, contributes to normal contractile force since it forms the tip or vortex of the double helix, and (b) how the apical tip of the ventricular helical coil becomes influenced by the principal fiber direction actions of the distant endocardial or epicardial components whose overlapping forms the helix producing this conical tip.

6.1 Sonomicrometer evidence for the myocardial band

6.1.1 Identification for the proper contraction angle

The angle defining the maximum contraction between crystal gauges conformed closely with the main fiber pathway evident on the helical model. In every instance the maximal extent of contraction was in a direction perpendicular to evident macroscopic direction. For example, epicardial fiber shortening was greatest 12% at 135° angle, with crystals moving closer together at this angulation, with markedly diminished contraction of 6% at 45°. The more extensive endocardial vs. epicardial shortening (20% vs 12%) shown by crystal sonomicrometer of the anterior LV free wall introduces data consistent with measurements in MRI recordings [30,31]. Furthermore, the extent of shortening diminished in each area as crystal placement was made at off angles to the probed site of maximal force of shortening selected from the model. Of course, the points of both maximal and minimal crystal displacement varied in different regions, as we segmentally probed the intact heart to search for the underlying scaffolding that may uncover principal and preferential fiber pathways.

6.2 Anterior left ventricle

In the anterior LV free wall, the crisscross nature defining the most active sites supported the suspicion of dual underlying descending and ascending segments. Of critical importance is the time-related difference in both appearance and end of contraction, showing a ∼80 ms delay between the origin of endocardial versus epicardial shortening. The counterpoint to these observations is a similar ∼90 ms delay between end of endocardial and ongoing epicardial contraction. The timing and spatial shortening measurements find a functional counterpart in LV pressure tracings and dP/dt events related to rapid ventricular filling described in Fig. 6, and summarized in the Section 5. These crystal observations are consistent with the ventricular rotational movements reported by McDonald [32], suggesting association with early activation and contraction of endocardium, with later contraction and persistence of epicardial contraction.

Aside from the temporal and maximum displacement for contraction, codified by the downslope of the crystal recordings of the deep and superficial muscle, an upslope (or widening between endocardial crystals) developed during the interval between cessation of deep shortening of ongoing ascending shortening. This stretching is consistent with the slight epicardial lengthening observed by Rademakers [10] during what he termed ‘isovolumetric relaxation’. We suspect Rademaker's epicadial observation reflects how the widened endocardial region causes transmural transmission to overlying surface muscle.

6.3 Circumferential studies

One alternate avenue included crystal placement around the circumferential perimeter of the left ventricle, where we did not search for the principal pathway directly, but probed a cross-sectional region to try to search for interaction between sites, in a manner similar to transverse heart sections evaluated by MRI [33] or echocardiography [34]. We uncovered one site that showed an area of negligible overall contraction, but developed an augmentation of the notch in the slope of the endocardial recordings made throughout the anterior left ventricle. This region matched the point of contraction of the papillary muscle proven at autopsy (Figs. 6 and 7) where papillary muscle fibers move in a different direction compared to free wall fibers. This structural finding thus accounts for the sudden loss or the expected normally graded contraction seen in Figs. 7 and 9. Simultaneously, normal contraction persisted in the overlying epicardial muscle where visible fibers are more transverse [35].

6.4 Posterior wall

Placement of crystals in the posterior wall defined the absence of two overlying layers, due to early and similar origin time of contraction. In this posterior location, the inner descending segment wraps around to form the apex and rise to become the ascending segments shown in Figs. 1 and 11, thereby removing the concept of two layer super-imposition (as exists in the anterior LV free wall) to provide a functional equivalent to the single principal pathway evident in the myocardial band dissection (Fig. 1). A tracing of the structural sequence of ventricular band allowed simultaneous recording of shortening in the descending, posterior, and ascending regions (shown in Fig. 15) where a simultaneous timing sequence of contraction of the wrapped band became evident with the earliest shortening in the descending segment, slightly later in the posterior muscle, and latest in the ascending segment. Further study of the posterior wall confirmed the predominant maximal contraction in the transverse direction, thereby following the evident visible fiber direction, with lesser contraction when the vertical angle was used (shown in Fig. 10). This transverse vector identifies the predominant pathway of the superimposed overlapping fibers, the posterior wall [as shown in the fiber strand peel off technique (SPOT)], and ‘shingle roof’ undercrossing of subjacent fibers described by Lunkenheimer and co-workers [23,36]. Such transverse principal direction of contractile shortening in the inferior posterior LV provides a functional equivalent to the principal circumferential structural fibers described by Greenbaum [37] in anatomic postmortem dissections. Perhaps the crisscross extent of displacement reflects the spring-like twisting within a single predominant pathway, thus contributing to the deformation seen by MRI and regional studies.

6.5 Right ventricle

The difference in timing and angle of temporal and spatial shortening in the RV free wall and outflow tract precisely followed the model; the lateral RV wall shortened simultaneously with the descending segment of the left ventricle, and RV outflow tract shortening was consistent with the ascending segment of the apical loop. The circumferential nature of fiber direction in the lateral free wall was also confirmed by the more effective shortening within the transverse predominant fiber course in the lateral RV free wall of the RV, with lesser shortening evident after vertical crystal placement as shown in Fig. 12a. This early shortening pattern would favor a constrictive motion during the isometric phase of systole that precedes ventricular ejection.

We confirmed evidence of the aberrant fibers [9] running over the RV outflow tract, originating from the LV oblique ascending segment, by finding a delay in the beginning of the shortening signal In the RV anterior and lateral walls. The anterior RV delay precisely paralleled the later origin at the LV free wall shortening, shown in Fig. 12a, with anatomic confirmation defined in Figs. 2 and 11b showing to the oblique course of the aberrant fibers that jump onto the right ventricle from the anterior left ventricle. These fibers then follow a similar course within the deep and superficial trajectories along the pulmonary artery outflow tract. The oblique nature of the fiber orientation would be consistent with a later twisting motion during ejection.

6.6 Basal loop during late ventricular isovolumetric contraction

Additional functional information was provided by widening of the right and posterior LV free wall basal loop crystals (Figs. 14 and 16) during the hiatus of ongoing ascending segment shortening responsible for suction generated rapid ventricular filling. This widening between outer rim basal loop crystals implies a stretching of noncontracting muscle during the presumed reciprocal clockwise direction twisting, as the ventricle prepares to suck blood into its cavity during isovolumetric systolic contraction. This stretch and widening between crystals is also clear in Fig. 4, where we presume the noncontracting stiff endocardial fibers are pulled apart during the active ongoing contraction during continuing ascending segment shortening.

6.7 The conical ventricle

The anisotropic quality of heterogeneous contraction of different regions is evident from MRI recordings [16,30], where the force vector of contraction increases as segments are followed from base to apex. Our crystal recordings looked at both endocardial and epicardial regions, and confirmed this improved shortening as we moved towards the elliptical chamber apex, with 35% reduced contraction when basal endocardial segment bases were compared to the near apical records (shown in Fig. 9). The dynamic change in MRI tracings compare the base and apex to define increased strain and deformation as the radius of curvature narrows and the extent of intracavitary thickening increases towards the tip of the conical chamber [22,24].

Our studies did not measure clockwise or counterclockwise rotation but the sequence of initial contraction of the right, then the left circumferential basal loop regions would produce the slight clockwise rotation, described initially by prior vector analysis by Coghlan and coworkers [38], recent MRI studies [33], and characterized as ‘cocking of the heart’ in the comprehensive reports by Ingels and co-workers [26,39,40]. This clockwise basal loop motion precedes the predominant counter clockwise twist responsible for ventricular shortening during ejection. Our analysis indicates that the shortening caused by subsequent endocardial and then co-contraction of both descending and ascending segments of the oblique helical apical loop will shorten the heart by pulling downward the attached basal loop. This counterclockwise pathway is evident on MRI [10,11] and other twisting analysis defining torsion of the macroscopic heart [41] as it twists during ventricular ejection.

6.8 The apex

Visualization of the apex shows that it rotates, as it is formed by the vortex of the criss-crossing descending and ascending segments. Anatomically, it is thin at cross-sectional observation [9,42]. Placement of a crystal at the apex with connection to the second of these couples at adjacent deep of superficial muscle showed only an irregular contraction (shown in Fig. 18). These movements were consistent, thus occurring with and without the Millar catheter in the apex, to thereby exclude a traumatic cause for these movements. Moreover, the contour of this contraction was ineffective, and the pattern did not reflect tracings gathered from the other segments. We could only interpret this record to confirm existence of some motion that varied markedly from the consistent and coordinated contractile scheme evident from all three angular patterns in nonapical regions (Fig. 4).

This regional evaluation of only the apex, rather than how it interacts with other distal regions, may follow prior studies of apical rotation done by placing only one isolated probe that is restricted to studying the tip region alone [43,44]. We amended this isolation by selecting two points, but limited our analysis to closely adjacent regions. These local measures defined irregular shortening patterns in the tip, and set the stage to determine how the tip of the helix interacts with the segments that form the apical tip. These include the inside-out wrap of the functional descending segment, and the outside-in wrap of the ascending segments [45–47].

An unusual pattern was seen in Fig. 20a when one crystal was placed into the apical tip, and then connected to a second distal crystal attached to functional endocardial or epicardial muscle. These functional observations contradict expected patterns from nontip connected endocardial or epicardial recordings shown in Fig. 4. Each apical tracing showed insights of how the nonefficient contractile apex became influenced by attachment to the descending and ascending segments whose inside-out and outside-in vortex forms the cone tip. The apical recording connected to the active descending segment started at ∼80 ms later time than epicardium initiation, and ended normally, approximately ∼90 ms before completion of the ascending segment shortening. Conversely, the apical tip to the ascending segment started ∼80 ms early, at precisely the time of expected early start of the descending segment, but ended normally ∼90 ms after the descending segment stopped.

We suspect how these tracings illustrate the helical coil of the apex, formed by these segments, and influenced by reciprocal shortening of the two components that create the ventricular tip (as shown in Fig. 20b). Perhaps these contractile changes reflect the transition point, where opposing contractile forces nullify each other, and produce the evident rotation. With the apex and descending segment, rotation was clockwise due to the pull of the ascending segment on the apex, with the normal completion due to the correct position of the distal endocardial crystal. Conversely, with the apex and ascending segment, the early start is due to the early pull of the descending segment in a counterclockwise direction, coupled with a normal ending point, ∼80 ms beyond expected descending segment stoppage of contraction. We believe the apex thus rotates due to how the opposing segments effect the helix, and this pattern is only visible when one of the crystals is placed into the rotating, but noneffective contracting apex, while the other is in a correct position in either deep or superficial descending or ascending muscle can be shortened considerably.

6.9 Implications of myocardial band analysis

The first concept that is forthcoming from the data analysis is that systolic contraction exists both during the phase of ejection, and during the time-related period previously termed as isovolumetric relaxation. This interval coincides with the rapid deceleration of ventricular pressure thought previously to reflect elastic recoil from isovolumetric relaxation related to potential energy stored during the systolic contraction [10,48–52] but is produced by ongoing contraction of the epicardial muscle forming ascending segment. This muscle continues to shorten beyond the end contraction phase of the endocardial muscle forming the descending segment. Consequently, a contractile phase exists throughout LV pressure acceleration during systolic ejection as well as during the early LV pressure deceleration that precedes the time of rapid filling.

Review of the pressure, dP/dt, and sonomicrometer tracings in Figs. 4 and 15 defines three contractile phases responsible for (a) isovolumetric contraction prior to ejection, (b) contraction and co-contraction for ejection, and (c) contraction for early rapid filling. These muscular shortenings for systolic action during both ejection and rapid filling will involve a shortening wave in each of the segments of the basal and apical loops, but our findings contradict (a) the concept of longitudinal progression along the sequential four segments of the band, i.e., right to left in basal loop, then descending followed by ascending in the apical loop [7,9,17,18], and (b) the prevailing notion that elastic recoil from stored potential energy from systole is responsible for the isovolumetric relaxation that generates suction for rapid filling [10].

Crystal shortening follows changes in wall motion without reference to electrical events, and analysis of these temporal and spatial findings that comprise the data must include insight into (a) connections of the electric cable system for the Purkinje neural penetration distribution, (b) myocardial spread of impulses to fibers in relation to muscle thickness along the myocardial band, and (c) a concept of matching the electric and mechanical interaction along the visible sequential muscular segments progressing through the unfolded, then rewrapped band. This matching of the electrically defined depolarization and repolarization must then be correlated with the evidence of ongoing contraction during the period of repolarization, since a contractile effort takes place during the entire QRS and T waves on the EKG. Clearly, an explosion of electricity comprises the QRS pattern, but ongoing contraction is evident during the electrical interval of repolarization as the T wave appears, so that a more complete understanding of excitation–contraction linking is essential.

The genesis of ventricular contraction during the isovolumetric contraction phase of systole shows that shortening begins immediately after the Q wave on the EKG, and involves both left ventricular endocardial the right lateral free wall fibers, to be followed ∼10 ms later with LV posterior contraction in the distribution of the basal loop. These three segments contract before the rapid acceleration of LV pressure for ejection. The predominant force becomes circumferential compression, thereby explaining the ∼25% narrowing of the mitral valve annulus during the isovolumetric contraction that precedes ventricular emptying, as defined by Shah and co-workers [53]. Contraction of the epicardial segment follows ∼80 ms later, and correlates with the rapid acceleration of pressure, the peak of the positive dP/dt and the end of the QRS complex.

Consequently, isovolumetric contraction involves only three segments, while the rapid ascent of pressure for ejection involves all four segments, bringing into play the later contribution of the epicardial or ascending segment. Based upon the direction of crystals, we think that the RV and posterior LV basal loop shortenings are dominant at the initiation of contraction, and cause compression and narrowing of the chamber. The oblique squeeze of the endocardial segment initiates the twisting responsible for ejection with shortening of the cavity, due to an initiation of the torsion-like counterclockwise twisting, or wringing of a towel or with a wine press, as suggested by Borrelli [41] in 1660, and now codified by noninvasive MRI recordings with tagging studies [10,11,33]. The downward movement of the ventricle during this later phase of ejection implies that the endocardial or descending segment muscle carries the dominant force during contraction for ejection.

With consideration of only the rope model, this contractile phase sequence defeats Torrent-Guasp's concept of progression of contraction sequentially along the visible myocardial band [8,9,17,24,38,54], since endocardial contraction would be expected to begin after the left posterior LV contraction. There must be a matching with nerve and muscle anatomy, since the earliest muscle contact of the Purkinje penetration is via the massive fan type neural distribution described by Tawara [55] that directly touches endocardial fibers. Sodi-Pallares and Calder [56] and Lewis et al. [57,58] have described this early activation of endocardial muscle that is translated into immediate contraction.

We suspect there is a delay between the initial excitation of endocardial fibers in a direct neural myocyte connection, and the later contraction of fibers in the thicker endocardial muscle mass that does not have Purkinje cells touching each fiber, and thus must get its excitation by transmission via a matrix transmission [59]. We know from the studies by Delhaas et al. [60] that transient ischemia does not alter the QRS interval during reperfusion, but slows the transmural contraction scheme, presumably by slowing the conduction pathway through the muscle. Further testing of this matrix concept is needed in the normal heart to define the relationship of spread of the impulse for contraction of the wall.

The early rapid downslope of the crystal record may define how the obliquely oriented fibers have a more forceful contraction, as postulated by Sallin [61], while the lesser slope in the free wall of the RV may reflect reduced force by the transverse fibers, whose maximal circumferential course is codified by the alignment of transverse crystal angles during RV recording (Fig. 11a and 12a). These mechanical observations may help explain why the predominant fiber angle orientation may allow the 15% shortening found in isolated fibers without scaffolding connections, to change as they exist in the wrapped heart with scaffolded spatial architecture, to show more forceful shortening (i.e., 60%) when predominant fiber orientation is oblique, rather than transverse (i.e., 30%), as exists in the predominant pathways of fibers comprising the two segments of the apical and basal loops [61,62].

Another gap in conventional thinking is the evidence of early contraction of the entire base of the heart before ejection occurs. This separation between basal contraction before papillary muscle shortening for rapid ejection starts was reported by Armour and Randall [63] in 1970, who confirmed the work of Roy and Adami [64] in 1890. Our data supports the concept of Armour and Randall that the base of the heart forms a stiff outer shell around the bulk of myocardium (really the endocardial and epicardial or descending and ascending regions of the apical loop) responsible for ejection. Armour documented the later stimulation of the base, as also shown by Sodi-Pallares and Calder [56], who confirmed a ‘mismatch’ in excitation–contraction coupling.

We have recently [17,18,59] considered this concept, and our observations support the work of Robb and Robb [65], that correlates the axial flow spread of the impulse along the fiber bundles, with more rapid transmission along thinner than thicker fibers. Consequently, the more slender RV and LV free wall basal loop segments shorten more quickly than the thicker papillary muscle and outer LV wall. We suspect that excitation impulse spreads through a matrix system [59] along these fine fibrous sheaths that extend from the termination of the Purkinje conduction fibers to escape and interweave with muscle fibers beyond the transition between conduction tissue and the remaining working myocardium.

These observations thereby also contradict the concept of transmission of the contraction wave impulse from the inner to the outer wall, since the direct penetration of the Purkinje fibers extends only to inner wall muscle. Clearly, conduction along cell–cell via low resistance gap junctions, at a maximum velocity of 0.3 mm/ms could not occur along the 80 mm descending segment, and 130 mm ascending segment, since 433 ms would be needed in the epicardium to complete activation [59]. The beginning of excitation of most fibers that enter into contraction is only ∼105 ms, and only ∼80 ms in this study. Most likely the impulse transmission spreads along spiral pathways, as suggested by Taccardi et al. [66,67] to explain the spread of activation beyond the superficial Purkinje network.

The endocardial very rapid shortening phase becomes less steep with the onset of shortening of the oppositely directed epicardial or ascending segment. Consequently, there is then ‘co-contraction’ of all areas (including the basal loop right ventricle and posterior LV, and endocardial and epicardial, or descending and ascending segments of the apical loop) during ejection. The crystals when paced in a crisscross manner defined maximal shortening force along principal fiber pathways but did not measure the torsion confirmed by MRI [10] and in isolated segment recordings by Ingels et al. [26] and Beyar et al. [48,49]. We suspect the change in slope of the crystal tracing during co-contraction reflects the oblique and clockwise counterforce and dominant counterclockwise twist of the descending segment (shown in Fig. 23) that becomes opposed by the reciprocal twisting of the ascending segment in a different direction. This reciprocal force may become unleashed when there is cessation of descending segment active contraction, so the reciprocally twisting ascending segment can then predominate and produce the clockwise twisting so characteristic of abrupt lengthening during rapid ventricular filling described in the following sections.

The final phase of contraction relates to the hiatus between loss of contraction in the RV, posterior LV and endocardial or descending segment records, and ongoing contraction in the epicardial or ascending segment. This time interval is linked to the deceleration of LV pressure and the onset of the negative phase of dP/dt recording. Consequently, there is an active period of epicardial or ascending contraction during the period previously termed isovolumetric relaxation. A more precise term is late isovolumetric contraction [17,18]. Fig. 23 displays the theme for dominant interaction between endocardial and epicardial segments and directional displacement.

This active and ongoing epicardial systolic muscular shortening after active endocardial or descending systolic shortening has stopped, specifically contradicts the concept that the rapid descent of pressure, and negative dP/dt is from recoil of stored potential energy [48,68–72], since systole persists [17,18]. This contractile cause differs from the interaction between the contracting myofilaments and elastic collagen surrounding the muscle fibers [48]. As a proposed mechanism for lengthening, we suspect that despite completion of active endocardial contraction in the descending segment, tension persists in this muscle, allowing it to act as a fulcrum [17,18] for ascent of the ventricle during the rapid filling phase [73]. We believe this ongoing twist of epicardial muscle, occurring in an opposite direction from the endocardium or descending region, makes this contractile effort govern the rapid clockwise rotation seen by MRI [33] during lengthening during rapid filling.

This active contractile role was also suggested by Rademakers and co-workers [10,11] by MRI studies, defining that 50% of filling develops during this time frame, and accentuation of speed and rate untwisting (or reciprocal twisting in a reverse direction) for rapid filling by can be increased by inotropic drug infusion. Brutsaert and co-workers [27,28,74] further amended the infrastructure for rapid filling by a suggesting contractile phase of systole. Our findings tested this concept with sonomicrometer crystals, and characterized the role of calcium dynamics in this process by infusion with dopamine or propranolol. The negative inotropic effect of propranolol, aside from reducing pressure and heart rate, widened the onset of contraction of the epicardium versus the contraction of the endocardium to thereby slow ejection, while simultaneously narrowing the hiatus for rapid filling, by decreasing the time frame for otherwise unbridled epicardial contraction while the remainder of the chamber muscle was not contracting. The result was a delay or prolongation of the downslope of LV pressure and less negative dP/dt. Clearly, further advancement of the stoppage of the endocardial contraction relative to the ongoing epicardial shortening will increasingly compromise the contractile forces responsible for rapid filling and derail the mechanisms for suction, so that pressure, rather that muscle motion now becomes the principal filling determinant. The consequence is that these tracings introduce a contractile mechanism that both contributes to suction filling in the normal heart, and more importantly implies that when this action becomes disrupted, a contractile cause for diastolic dysfunction may prevail.

Evidence for this prolongation of late systole in hearts with diastolic dysfunction is evident in studies of stunning (Fig. 24) after ischemia [75–77], hypertrophy during aortic stenosis [78], posttransplant dilatation [79], and tachycardia-induced cardiomyopathy [40]. Our findings show that this is an active phenomena, related to loss of the critical gap between stoppage of shortening in the circumferential and endocardial region, and ongoing contraction of the epicardial or ascending segment. The pertinent MRI findings that would support this concept is that there is loss of the normal action of rapid ascent of the muscle, so that vigorous shortening and rapid lengthening becomes replaced by a lesser length distance of movement, and slower return to the normal diastolic filling position.

6.10 The integrated cardiac spiral

The reciprocal spiral pattern is heralded by the macroscopic double helical cardiac form defining the ascending and descending loops, and interface with the collagen weave-like structure forming the connective tissue scaffold for this muscle mass [80]. A similar dual helicoid configuration exists microscopically in myosin, actin, tropomyosin, and confocally in the coil arrangement in calcium distribution within a single cell, as well as the coil arrangement within a single calcium ion [81]. The helical cardiac formation is well known anatomically, with Senec describing the internal helix (shown in Fig. 25a) from his text in 1760 [82,83], Pettigrew describing the apex (shown in Fig. 25b) to contain reciprocal spirals with the epicardium going from without to within, and the endocardium going from within to without [82]. This spatial analysis also provides a functional confirmation to the observations of Lower [84], in 1669, who described muscle fibers of the inner wall that ran opposite the fibers of the outer wall, and that contraction can be compared ‘… to the wringing of a cloth to squeeze out the water’. Our functional analysis shows that the lengthening is due to the reciprocal twist of the outer wall fibers that continue to contract after the inner wall fibers stop in the normal heart.

We suspect the heart borrows a trick from nature, where the reciprocal spiral horns of a ram or eland, gain internal strength from the internal spring-like coil arrangement. For these reasons, the arrangement shown in Fig. 26a and b [17,18,85] may reflect the cardiac spiral within spiral arrangement. The spatial arrangements of these coils may also help explain why a muscle can contract and shorten, yet lengthen its longitudinal position when we consider the ascending segment. This role of contraction of a spring within a spring, like coils within coils, connected in a helical manner introduces some insight into how a muscle can both contract and lengthen. Review of Fig. 26b shows how the noncontracting ascending segment may become stretched during the beginning of descending segment contraction and how by subsequent wringing can both shorten the distance between individual springs, yet lengthen due to its spatial helical orientation.

This spatial configuration parallels the structure of DNA where this recipe for an interweaving spiral configuration is planned. If this concept is correct, the natural framework for our search into understanding has been mapped before our arrival, and our task is find the new helical pathways required for us to understand novel mechanisms to explain our functional observations in the living heart.

References