The arguments put forward against a peristaltic pumping action of the early embryonic heart are mainly of hemodynamic nature and derive from three observations: (1) the blood flow generated by the tubular embryonic heart is pulsatile rather than peristaltic (Hu and Clark,1989). (2) The peak blood flow velocity (mm/sec) recorded in the ventricular inflow exceeds the speed of the corresponding contraction wave traveling along the wall of the inflow region (Farouhar et al., 2006; Butcher et al.,2007). (3) The relationship between the contraction frequency and the blood flow rate (ml/sec) is nonlinear and exceeds the maximum flow rate possible for a technical peristaltic pump (Farouhar et al., 2006).
An important question is whether these data really speak against a peristaltic pumping action of the early embryonic heart. The answer must be yes, if we use the fluid dynamics of a roller pump as defining criteria (Farouhar et al., 2006). This is because the pumping behavior of roller pumps is characterized by (1) the generation of continuous rather than pulsatile flow, (2) a peak flow velocity that does not exceed the speed of the compression wave, and (3) a linear relationship between the compression frequency and the flow rate (Fig. 3). If we go into biological systems, however, we will find at least one example of peristalsis, which shows flow characteristics that are qualitatively similar to that recorded in embryonic heart tubes. This example is gastric emptying. Gastric emptying of liquid meals is accomplished by two mechanisms: (1) a pressure pump, and (2) propulsive peristalsis (Kunz et al.,1998). Peristaltic emptying of the stomach generates a pulsatile transpyloric flow (Anvari et al.,1995) whose peak flow velocity is approximately 60 cm/sec (Hausken et al.,1991), and thereby exceeds the propagation speed of the propulsive peristaltic waves in the antrum, which is approximately 3 mm/sec (Kunz et al.,1998). Biological peristaltic pumps obviously seem able to generate a wider range of fluid dynamical features than technical peristaltic pumps. With respect to the embryonic heart tube, this idea is supported by recent data from a computational model, which show that peristaltic heart tubes can generate pulsatile blood flow as a consequence of the presence of endocardial cushions at the inflow and outflow of the ventricular loop (Taber et al.,2007). Considering all of the above-mentioned fluid dynamic data, we do not think that it is justified, at the present time, to conclude that the early embryonic heart does not work as a peristaltic pump. The data rather show that it does not work according to a technical standard for peristalsis (roller pumps). On the other hand, however, we must emphasize that this fact does not justify exclusion of the possibility that the tubular embryonic heart may work as a Liebau pump.
Direction of Net Flow
While going through the literature about valveless pumping, we became aware of one previously neglected aspect of fluid flow that may cause problems in considering the Liebau effect as a good candidate to explain unidirectional blood flow in the valveless embryonic heart. This aspect is the direction of the net flow generated by valveless pumps. Due to the absence of valves, the direction of peristaltic as well as Liebau effect-driven net flow is not a structurally fixed quality so that both types of valveless pumps can generate anterograde as well as retrograde net flow. The direction of peristaltic net flow corresponds to the direction of the waves of compression traveling along the length of a peristaltic pump. Thus, if the early embryonic heart should work as a peristaltic pump, we must expect that pulsating embryonic hearts regularly show a positive correlation between the direction of the traveling contraction waves and the direction of the net flow. This seems indeed to be the case. We do not know of any observation in embryonic hearts in which the normal anterograde sequence of contractions (inflow ventricular loop outflow) was associated with retrograde net flow. We have observed, however, situations in chick and mouse embryonic hearts in which we found a reverse contraction sequence that was associated with retrograde net flow from the ventricular loop toward the systemic venous sinus (data not shown). Thus, with respect to the direction of the net flow, the observations made in living embryos are in full accord with the peristaltic pump theory.
We cannot say that the same holds true for the Liebau pump theory. In fact, the picture obtained from the data reported in the literature is an ambiguous one. Liebau's initial observations in physical models suggested that the direction of net flow generated by his valveless pumps depended primarily on the position of the active site of compression with respect to the ends of the flexible tube (Liebau,1954a,b,1955a,b,1968,1970). The net flow was shown to run from the longer toward the shorter passive portion of the flexible tube (Figs. 3, 4). Corresponding observations were reported by several other researchers (Moser et al.,1998; Kilner,2005; Vogel,2007; Bringley et al.,2008). If we assume that this relationship is characteristic for Liebau pumps, we should be surprised to note that Forouhar and coworkers (2006) see the tubular embryonic heart as a Liebau pump whose active compression site lies at the inflow region, because such a configuration would be expected to generate retrograde instead of anterograde net flow (Fig. 7). It should be noted here, however, that the situation is not as clear as it may appear on the first view. This is because we also found several papers in which the net flow was described as running from the shorter toward the longer passive portion of a Liebau pump (Ottesen,2003; Kenner,2004; Hickerson et al.,2005; Hickerson and Gharib,2006), which is a relationship that would fit with the idea that the early embryonic heart may work as a Liebau pump (Fig. 8). How can this discrepancy be explained? Physical experiments as well as mathematical simulations have shown that the direction of Liebau effect-driven net flow depends not only on the position of the active compression site but additionally depends on several other factors of which the compression frequency seems to be the most important one (Bredow,1968; Ottesen,2003; Kenner,2004; Kilner,2005; Hickerson et al.,2005; Hickerson and Gharib,2006; Jung,2007; Timmermann and Ottesen,2009). In fact, changing the frequency of compression can reverse the direction of Liebau effect-driven net flow in a seemingly unpredictable manner and such reversals have been observed even at frequencies that were in the range of the steadily increasing beating rate of the tubular embryonic heart. It thus might be possible that the above-mentioned differences in the direction of Liebau effect-driven flow may reflect differences in the experimental settings. It is the unstable and seemingly unpredictable behavior of Liebau pumps that makes it difficult to consider the Liebau effect as a good candidate to explain unidirectional blood flow in the valveless embryonic heart.
Figure 7. This schematic figure shows the relationship between the position of the active compression site and the direction of net flow as reported by Liebau and several other researchers. Net flow is said to run from the longer toward the shorter passive portion of the flexible tube. Note that, if the embryonic heart tube is seen as a Liebau pump in which the active compression site lies at the inflow region, the net flow is expected to run from the arterial toward the venous pole (B).
Download figure to PowerPoint
Figure 8. This schematic figure shows the relationship between the position of the active compression site and the direction of net flow as reported by Ottesen (2003) and several other researchers. Net flow is said to run from the shorter toward the longer passive portion of the flexible tube. Note that, if the embryonic heart tube is seen as a Liebau pump in which the active compression site lies at the inflow region, the net flow is expected to run from the venous toward the arterial pole (B).
Download figure to PowerPoint
Solid (Morphological) Dynamics
Peristaltic and Liebau pumps show striking differences in the dynamics of its wall motion that might be used to identify the nature of early cardiac pumping function. The action of peristaltic pumps is characterized by the generation of nonstationary areas of active compression traveling along the length of the tubular pump. In biological peristaltic pumps, these traveling compression waves result from the sequential action of multiple segments of active muscular contractions. Classic Liebau pumps, on the other hand, are flexible tubes that possess only a single, stationary center of active compressions. This active motor center lies at an asymmetric position along its length (Figs. 4, 5) and is said to generate bidirectional elastic waves traveling along the passive upstream and downstream sections of the pump. Thus, if the tubular embryonic heart were a classic Liebau pump, we expect to find only a single center of active myocardial contractions at an asymmetric position along its length. The rest of the heart tube should consist of passive segments. If, on the other hand, the tubular embryonic heart were a biological peristaltic pump, we expect to find multiple segments of active myocardial contractions along the whole length of the organ that normally act in proximodistal order (inflow ventricular loop outflow) to generate unidirectionally propagating waves of active contractions.
Based on high-resolution in vivo imaging data, Forouhar et al. (2006) have calculated that embryonic zebrafish heart tubes seem to possess only a single center of active myocardial contractions at its inflow (see Fig. 2 in Forouhar et al.,2006) and regard this finding as a supporting evidence for their theory that the embryonic vertebrate heart works as a Liebau pump. A possibly corresponding observation has, subsequently been made in the tubular heart of bird embryos by another research group (Jenkins et al.,2007). This group analyzed the morphological dynamics of pulsating embryonic quail hearts using an ultrahigh-speed optical coherence tomography imaging system, which facilitated visualization of morphological dynamics that could not be achieved at low temporal resolution. The study uncovered striking differences in the dynamics of wall motion between the proximal (inflow) and distal (outflow) portions of the heart tube. The cardiac inflow was identified as a region of fast contraction and fast propagation of the contractile waves (“kickstart” center), whereas the portions downstream to the inflow were identified as regions of decreasing contraction speed and decreasing wave propagation speed. These data were interpreted as suggesting “… active contraction at the beginning of the tube initiating a wave which then propagate more passively through the rest of the tube” (Jenkins et al.,2007). Although this study has not definitively determined the presence of only a single site of active myocardial contraction in the tubular embryonic quail heart, it is tempting to speculate that the presence of a leading kickstart center might support the Liebau pump theory of early cardiac pumping function.
On the other hand, however, we have to consider that regional differences in the wall motion of tubular embryonic hearts have previously been documented by several other researchers and, thus are not really new findings (Boucek et al.,1959; Castenholz and Flórez-Cossio,1972; Steding and Seidl,1990; Sarre et al.,2006; Vennemann et al.,2006). Furthermore, the presence of a leading kickstart region at the cardiac inflow must not necessarily mean that the tubular embryonic heart owns only a single segment of active myocardial contractions. In fact, there is ample evidence (1) that each segment of the tubular heart of higher vertebrate embryos (inflow, ventricular loop, outflow) is capable of active myocardial contraction; and (2) that the mechanical waves traveling along the wall of the pulsating heart tube of higher vertebrate embryos result from active, sequential contractions of the embryonic cardiac segments. The historically oldest evidence comes from in vitro experiments. If an explanted heart tube from a chick embryo is dissected into multiple pieces, each of these physically isolated pieces will undergo active contractions in a segment-specific frequency, which is high in pieces derived from proximal (inflow) and low in pieces derived from distal (outflow) heart segments (Fano and Badano,1890; Johnstone,1924; Barry,1942). Corresponding observations were made in intact embryonic chick heart tubes in which the segments were electrically isolated from each other by the placement of ligatures around the tube (Pickering,1893; Johnstone,1924). Optical mapping of the electrical activity of the myocardium of the tubular embryonic chick heart has uncovered spatiotemporal sequences of periodic electrical activation that correspond to the propagation pattern of its traveling contractile waves (Chuck et al.,2004; Sedmera et al.,2004). In view of the fact that myocardial contraction is coupled to electric activation, this observation may be regarded as a strong evidence for an active nature of the traveling contractile waves. One might reply that electric activation of an embryonic myocardial segment downstream to the cardiac inflow region must not necessarily mean that it will respond with active contraction, because the segment might be in an immature, precontractile state of electromechanical uncoupling (Hirota et al.,1983; Kamino,1991). However, why should such a segment then start active contraction if it is experimentally isolated from the rest of the heart (see above)? Moreover, during the formation of the embryonic heart tube, the inflow segment becomes added to the venous pole of the heart only after the ventricular segment has already started rhythmic beating (Patten,1949), what makes it difficult to believe that the cardiac segments downstream to the inflow are in a passive state in the fully formed embryonic heart tube. A final evidence for the active nature of the traveling contractile waves is the observation that fully formed embryonic heart tubes can develop partial AV-blocks similar to those found in mature four-chambered hearts (Pickering,1893; Lewis,1923; Paff et al.,1964). If the traveling waves were passively propagating elastic waves initiated by active myocardial contractions of the cardiac inflow segment, we would expect that every contraction of the inflow segment should initiate an elastic wave traveling along the passive downstream segments of the heart. Such a 1:1 relationship is found in the normally beating heart tube. In heart tubes with partial AV-block, however, the relationship between atrial (inflow) and subsequent ventricular contractions can be 2:1 or 3:1 (Pickering,1893; Lewis,1923; Paff et al.,1964). This finding conflicts with the idea of passively propagating elastic waves but can be easily explained by conduction disturbances in the electric activation of the myocardium.
To summarize the above-mentioned data on the myocardial dynamics of early embryonic hearts, we can state that there is ample evidence that the traveling mechanical waves seen in the beating embryonic heart tube of higher vertebrates result from the sequential action of multiple segments of active myocardial contractions along the whole length of the organ. This fact is in accord with the traditional theory of a peristaltic embryonic heart action but strongly speaks against the possibility that the early embryonic heart works as a classic Liebau pump, because such a pump has only a single center of active compressions at an asymmetric position along its length. We should note here, however, that the above-mentioned evidence against the Liebau pump theory comes from higher vertebrate (chick) embryos. Thus, we cannot completely exclude the possibility that the tubular embryonic hearts of lower vertebrates such as the zebrafish may work as classic Liebau pumps as suggested by Forouhar and coworkers (2006). Furthermore, we should note that Liebau's research did not exclusively focus on physical pump models with only a single center of active compressions. He also reported on a few experiments in physical models with two, sequentially acting compression sites suggesting that the tubular hearts of higher vertebrate embryos may work as “bi-” or “multi-segmental” Liebau pumps (Liebau,1955a,1970). For unknown reasons, however, research on the Liebau effect has exclusively focused on mono-segmental pump models, while the bi-segmental models did not receive much recognition during the past 50 years. We, therefore, have no data about the characteristics of such pumps that might be used to identify them in biological systems.
Our evaluation of the functional significance of the dynamics of early cardiac wall motion should not be confined to myocardial dynamics, only. This is because the wall structure of tubular embryonic hearts differs considerably from that of the simple technical tubes traditionally used in physical models of Liebau pumps. The latter have only a single-layered flexible wall while the former have a multilayered wall of which the myocardium forms only the relatively thin outer layer. The mass of the multilayered embryonic heart wall consists of a thick layer of extracellular matrix—the cardiac jelly (Davis,1924)—interposed between the outer myocardium and inner endocardium (Fig. 1). Several data suggest that the cardiac jelly may play important roles in early cardiac pumping function: (1) The systolic narrowing of tubular embryonic hearts ends with complete occlusion of the lumen of the contracting heart segment (Patten et al.,1948; McQuinn et al.,2007; Männer et al.,2008,2009). Using geometrical analyses, Barry (1948) has shown that, under physiological conditions (e.g., ∼20% systolic shortening in circumferential length of the myocardium), the lumen of a tubular embryonic heart cannot be occluded at end-systole without the presence of a layer of cardiac jelly whose thickness must be ∼45% of the radius of the end-diastolic lumen. Because lumen occlusion is regarded as a requirement for effective peristalsis, he concluded that the presence of a thick cardiac jelly layer is a requisite for effective pumping function if the tubular embryonic heart is to propel the blood with a peristaltic-like pumping mechanism. (2) The cardiac jelly shows an uneven spatial distribution, which forces the open lumen of the heart tube to acquire an elliptic cross-section (McQuinn et al.,2007; Männer et al.,2008). This finding might be interesting because mathematical analyses have shown that tubes of elliptic cross-section have a higher mechanical efficiency of peristaltic pumping compared with tubes of circular cross-section (Usha and Rao,1995; Taber and Perruchio,2000). (3) There is evidence that the cardiac jelly is a resilient component of the embryonic heart wall, which becomes deformed during systole and springs back to its original shape during diastolic relaxation and, thereby, might suck the blood into the inflow segment of the heart tube (Barry,1948; Männer et al.,2009). In view of all these data, it may appear that the wall structure of early embryonic hearts is optimized for peristaltic-like blood transport and that the presence of cardiac jelly may be seen as a supporting evidence for the peristaltic pump theory. However, recent studies on Liebau pumps have shown that this must not be the case. Inspired by the multilayered structure of embryonic heart tubes, engineers have constructed a multilayered Liebau pump in which a thick gelatinous layer acts as cardiac jelly (Loumes et al.,2008). The thick gelatinous layer was found to amplify the elastic waves responsible for Liebau effect-driven pumping, so that only small excitations of the active compression side (10% external radius) were needed to produce significant flow (Loumes et al.,2008). The presence of a thick layer of cardiac jelly obviously seems to be advantageous for the pumping function of both kinds of tubular hearts; those working as peristaltic pumps (Barry,1948) as well as those working as Liebau pumps (Loumes et al.,2008). Therefore, the presence of a thick layer of cardiac jelly cannot be regarded as supporting evidence for either of the two competing theories of early cardiac pumping mechanism.
Periodic changes in diameter of the lumen of valveless tubular pumps are final consequences of pumping-related wall motions and the patterns of these changes show striking differences between peristaltic and Liebau pumps. The lumen of Liebau pumps, for example, becomes occluded only at the site(s) of active compressions (Hickerson et al.,2005; Hickerson and Gharib,2006), whereas ideal peristaltic pumps show traveling waves of lumen occlusion caused by the unidirectional propagating waves of compression. Due to the presence of cardiac jelly, the systolic narrowing of a given segment of the tubular embryonic chick heart ends with complete occlusion of its lumen (Patten et al.,1948; Barry,1948; McQuinn et al.,2007; Männer et al.,2008,2009). The traveling contraction waves of the tubular embryonic chick heart thus are accompanied by corresponding waves of lumen occlusion during the initial phase of its pumping activity when all embryonic cardiac segments downstream to the primitive atrium have a thick layer of cardiac jelly (Fig. 9). This picture fits very well with the pattern expected for a peristaltic pump but differs from the pattern expected for a classic mono-segmental Liebau pump.
Figure 9. Peristaltic-like movement of a wave of heart lumen occlusion in an embryonic chick heart tube (stage 13 according to Hamburger and Hamilton,1951) visualized in vivo by high-resolution optical coherence tomography (own data). A: Whole heart tube in diastolic state. B: Occlusion of the cardiac inflow segment. C: Occlusion of the mid-ventricular region. D: Occlusion of the proximal portion of the outflow segment. E: Occlusion of the distal portion of the outflow segment. Arrows point to the myocardial surface of the contracting heart segments. i, inflow; o, outflow; v, ventricle.
Download figure to PowerPoint
The peristaltic-like pattern of lumen occlusion is found until developmental stages when the embryonic ventricles start formation of a trabeculated muscle layer (Männer et al.,2009). The formation of myocardial trabeculations is accompanied by a remarkable thinning of the cardiac jelly in the apical portions of the embryonic ventricles, so that their lumen now is no longer occluded at end-systole. The ventricular inflow (atrioventricular region) and the outflow segment of the heart tube, on the other hand, retain a thick layer of cardiac jelly and, therefore, still show lumen occlusion during contraction of their myocardial mantles. These embryonic cardiac segments now act as functional heart valves because they prevent backflow of blood during the cardiac cycle (Patten et al.,1948). Their cardiac jelly layers soon become populated with endocardium-derived mesenchyme and undergo remodeling to form so-called endocardial cushions, which contribute to the formation of the definitive heart valves. In view of the valvular action of its atrioventricular and outflow segments, the question arises whether the tubular embryonic heart should be regarded as a valveless pump during advanced developmental stages.