Nervous Control of the Heartbeat in Octopus

Cephalopods have a high pressure, fully enclosed blood system, with arteries, capillaries and veins, and some large-capacity venous sinuses. This complex system has a very elaborate innervation, and large parts of the central nervous system appear to be devoted to a detailed control of the peripheral circulation (references see Young, 1971 ; Wells, 1978).

The principal pumps for the system (most of the veins show peristalsis and also help to move the blood) are the two branchial hearts, pressing the blood through the gill capillaries, and the single systemic heart into which the gills drain, supplying the rest of the body. The trio is innervated from the suboesophageal part of the brain through the two visceral nerves. The visceral nerve on each side passes into a fusiform ganglion (= first cardiac ganglion of some authors), and from there to a cardiac ganglion, attached to the branchial heart. A commissure links the two fusiform ganglia. In addition there is a chain of small ganglia along the length of each gill; these connect to the rest of the system through the cardiac ganglion on each side. An outline of the anatomy of the system is shown in Fig. 1, adapted from Young (1967). Detailed information about the nervous arrangements within the ganglia, and the innervation of the hearts is largely lacking for Octopus. Alexandrowicz (1913) gives some details and in a later paper (Alexandrowicz, 1963) describes the peculiar pulsating body to be found in the cardiac ganglion. But the most detailed information on the nervous arrangements in the ganglia associated with the hearts of cephalopods relates to Sepia (Alexandrowicz, 1960), which is unfortunate since decapods have no fusiform ganglia. The cardiac ganglia of decapods include elements that become separated into the fusiforms in the octopods.

Accounts of the function of the ganglia associated with the hearts have hitherto been limited to studies of the effect of cutting nervous connexions and/or stimulating the visceral nerves in acute preparations. For most of these the animals were nailed to boards, upside down, with the mantle split or turned inside out. Hill & Welsh (1966) have summarized this work. The conclusions reached differ in several important ways from those arising from the experiments to be reported below, which were made with free-moving, essentially intact animals.

Octopus vulgaris of between 500 and 1500 g were used. Most of the experiments were made in July and August, at Banyuls where the seawater in an open circulation runs at 21 ± 2 °C.

Two sorts of recording device were used. The first recorded pressure, through a T-piece tied into the dorsal aorta, about 5 cm downstream from the systemic heart. The second measured movement in terms of the changes in impedence between two fine wire electrodes, threaded into the heart or mantle musculature. Two such devices allowed simultaneous recordings to be made from the two branchial hearts. Details of equipment and its installation are given in Wells (1979).

Surgery was carried out under 2.5% ethanol anaesthesia.

Both visceral nerves can be cut, between the point at which the abdominal nerves branch towards the guts and the fusiform ganglia, without stopping the heartbeat. The animals sometimes seem to take longer than usual to recover from the anaesthetic to the point where respiratory movements and the heartbeat become regular, but the effect is difficult to quantify since even controls are very variable.

This first result already contradicts convention which holds that respiratory movements cease after visceral nerves section (see Hill & Welsh, 1966). This belief appears to have originated with Ransom (1884). Ransom used acute preparations, partially dissected and pegged on their backs to a board, with the mantle slit open and folded aside. In this situation he showed (1) that cutting both visceral nerves stopped respiratory movements, (2) that the hearts nevertheless continued to beat, though rather irregularly and (3) that stimulation of the proximal cut ends of the visceral nerves induced mantle movements. Since he knew (from preliminary observations of the hearts through a hole cut in the mantle) that the heart beat rate often coincided with the respiratory rhythm, and was never slower than this, he concluded that mantle movements must be maintained in response to signals from the hearts. This conclusion appears to have been accepted by all subsequent workers (references see Hill & Welsh, 1966), who may or may not have repeated Ransom’s experiments. But they were all using acute preparations. The present series of experiments shows that nervous connexions between heart and mantle are not a necessary condition for regular respiration, in unrestrained animals.

When the visceral nerves are cut, the branchial hearts continue to beat in phase with one another, preceeding the systemic heartbeat by about two-thirds of a cycle, as in control octopuses (Fig. 2). Heartbeat frequencies, mean aortic pressures and the aortic pulses of resting animals have similar values before and after the operation (Table 1). High values, similar to those found in intact octopuses, are reached in, for example, recovery from anaesthetic or following the return of an octopus from O₂ deficient to O₂ saturated seawater.

The responses to acetylcholine, 5-hydroxytryptamine, adrenaline or tyramine injected into the hearts of freemoving octopuses were qualitatively and quantitatively similar in animals tested before and after visceral nerves section, as were the responses to extracts made from the neurosecretory systems discharging into the anterior vena cava or pharyngo-ophthahnic vein (Wells & Mangold, 1979). Again very high pressures, in excess of 100 cm of water, could be generated by the hearts in the absence of any nervous connexion with the CNS.

In intact octopuses the beat rate is closely related to temperature and oxygen tension, slowing progressively as either falls (Wells, 1979). Both responses continue, apparently unchanged, after section of the visceral nerves.

The fact that the steady rhythm of the hearts can continue and that the hearts can respond directly to a variety of stimuli in the absence of the visceral nerves does not necessarily mean that the nerves normally play no part in the control of the heartbeat. A considerable number of experiments has been made in the past, using an isolated systemic heart-visceral nerve preparation (references see Krijgsman & Divaris, 1955). In this situation stimulation of the nerves regularly induces inhibition followed by transient increases in the frequency and amplitude of the systemic heartbeat, effects that one might expect to find reflected in the performance of the heart in vivo, and absent in octopuses with the visceral nerves cut.

The present series of experiments did not include any designed specifically to investigate inhibition of the heartbeat following visceral nerves section. But we know that cardiac arrest can occur in vivo, and there is evidence that this must at least sometimes be a direct nervous effect. The systemic heart will often miss one or more beats, if the animal is startled. Wells (1979) includes records of an octopus habituating to a hand or a white rectangle shown at intervals, and of a male animal as it approaches and copulates with a female. It is true that the animal often tenses its muscles when startled. It may flatten out into the ‘dymantic’ response (see Wells, 1978) or grip tightly to the bottom of its tank. It often ceases to make respiratory movements. Any of these responses might restrict the venous return. But there would appear to be plenty of blood in the large vena cavae to supply at least two or three systemic heartbeats, and the effects observed (both the cessation of the beat and its return) often seem to be far too abrupt and complete to permit an explanation based on strangulation of the venous return, except just possibly by contraction of the gills (which in any case are innervated via the visceral nerves). It is far more likely that sudden cardiac arrest is a direct nervous effect, ordered by the CNS through the visceral nerves.

In the other direction, enhancement rather than inhibition of the heartbeat, there is more direct evidence of visceral nerve involvement. If an animal is handled, or otherwise obliged to take exercise, the normal response is to increase pulse amplitude and (to a lesser extent) heartbeat frequency; both increases persist for a while after the animal has come to rest (Fig. 3 a). Animals with their visceral nerves cut, in contrast, often behave in the manner shown in Fig. 3b; the systemic heartbeat slows and may even stop, and the branchial heartbeat is disrupted. Table 2 summarizes the results obtained with 4 animals that were tested both before and after visceral nerves section. Reduction in beat frequency with concurrent falls in aortic pressure were seen to follow brief periods of exercise of 14 out of 18 occasions. In corresponding observations of the same animals before section of the visceral nerves short periods of exercise nearly always (in 15 out of 16 tests) produced an immediate rise in aortic pressure, and usually (in 10 out of 16 tests) a concurrent increase in beat frequency. Presumably it is undesirable for the animal to reduce, let alone to stop, the supply of blood to the arms during or after exercise and effects of the sort shown in Fig. 36 and in Table 2 represent a breakdown of a system that would determine an appropriate response to an increased oxygen demand in the intact animal.

Removal of the fusiform ganglia cuts the visceral nerves, with effects upon the response to exercise that have already been noted above. Removing the ganglia had no additional effects upon systolic or diastolic pressures or upon the frequency of the systemic heartbeat ; some typical examples, taken from animals tested both before and after the operation, are given in Table 3.

As well as being a pathway through which the visceral nerves might act, the fusiform ganglia link the two cardiac ganglia, and one might expect their removal to disrupt coordination of the branchial heartbeats. In the event, it was found that the two branchial hearts generally manage to remain in phase with each other and with the systemic heart after severance of the nervous connection between them (Fig. 4). They may, moreover, change their pattern of beat simultaneously and apparently spontaneously. The branchial hearts share a common blood supply from the anterior vena cava and in such cases it seems possible that they are responding to pressure or hormonal stimuli arising upstream of them, a situation in which the nervous link might be redundant.

In an attempt to establish whether the activities of one branchial heart can influence the other as a result of stimuli arising downstream of the bifurcation of the anterior vena cava, drugs were injected into one branchial heart through a fine cannula, while recording the activities of both.

Fig. 5 shows a typical series of results. 5-hydroxytryptamine excites and acetylcholine inhibits the branchial heartbeat ; the effect on the injected side is immediately mirrored in the activity of the ‘control’ side heart. Similar tests were made with adrenaline, tyramine, and neurosecretory materials derived from the anterior vena cava or the pharyngo-ophthalmic vein (Wells & Mangold, 1979). Ignoring instances where the dosage was subthreshold or the result uncertain, a total of 25 tests was made with 3 different octopuses. In 19 of these tests there was an immediate response by the ‘control’ side heart, closely paralleling the effect of the drug on the injected side.

Similar tests were made after cutting the fusiform to cardiac ganglion connectives on one or both sides in 3 octopuses, 2 of them from the trio already tested pre-operationally. In 15 out of 24 tests, the response now appeared only on the injected side (Fig. 6a). In 9 instances, there was a response on the ‘control’ side. In 6 (and perhaps 8) of these cases the control side response was noticeably delayed (Fig. 6b). Since there is a coronary circulation, fed by arteries that branch from the anterior aorta, downstream of the point of injection, it is likely that these delayed effects are blood-bom rather than nervously relayed.

It seems reasonable to conclude that nervous signals from one branchial heart can reach the other only through the fusiform ganglia and that signals along this pathway could help to keep the branchial heartbeats in time in the intact octopus.

Johansen and Martin (1962) have produced evidence for rhythmic contractions of the gills in O. dofleini which (assuming the same to occur in O. vulgaris) might contribute to the filling of the systemic heart. One might reasonably guess that severing the connexions between the gill ganglia and the hearts would produce some change in the output of the systemic heart. But no obvious consequent changes were seen in the present series of experiments.

In contrast to the above, where very considerable damage scarcely affects the rhythm of the hearts, interference with the cardiac ganglia is seriously disruptive. If a cardiac ganglion is removed, the regular beat of the corresponding branchial heart generally becomes disorganised (Fig. 7a).

In one animal the nerves from the cardiac ganglia to the gill hearts were all cut, leaving the rest of the system intact. The effect of this operation was very similar to that of removing the cardiac ganglia. Fig. 8 shows the performance of the systemic and one branchial heart before and after the operation; the second branchial heart is presumably behaving in the same way. The consequences of malfunction by the pair of gill hearts show in the record of aortic pressure; the systemic heart beats irregularly and more slowly than usual. The form of the systemic beat, however, remains essentially normal, just as it does after removal of one or both of the cardiac ganglia.

In the course of these experiments it was often observed that while removal or disconnexion of a cardiac ganglion generally disrupted the beat of the corresponding gill heart, as in Figs. 7(a) and 8(b), the effect was by no means invariable. From time to time and sometimes for periods of several minutes (and perhaps for hours, though continuous recordings were never made for so long) a gill heart without a ganglion would settle down to beat steadily, in time with its opposite number and in phase with the systemic. This could occur after removal of both cardiac ganglia (Fig. 7b) and even in animals from which the fusiform ganglia had been removed as well as the cardiacs. Four such animals survived (in some cases after several successive operations and many tests) overnight, so there is little doubt that the hearts system can function, at least sufficiently well to keep the resting animal alive, in the absence of all the heart ganglia. The maximum pressures and frequencies shown in these circumstances can approach the resting values seen in intact animals (Table 4). But there seems to be little or no capacity for a further rise ; on the few occasions when attempts were made to induce such animals to move about there was either no change, or a sudden disorganization of the heartbeat, with aortic pressures and systemic frequencies collapsing towards zero. Fig. 11 includes the only instance in which there was any indication of a (small) sustained rise in pressure following a period of exercise.

The heartbeat of an intact Octopus vulgaris is temperature sensitive, with a Q10 of about 3 over the range 7-27 °C. The hearts are also sensitive to the oxygen content of the seawater bathing them, slowing progressively as this declines; at about 2.5 ppm the beats become irregular, and may stop (Wells, 1979).

These responses continue after removal of the cardiac ganglia. Fig. 9 shows records of the branchial heartbeats of an animal with the cardiac ganglion removed from the LHS gill heart, in waters at 15 and 22 °C. The beat is notably slower at 15 °C and although it is a little irregular in the deganglionated heart, it is quite plain that the two are contracting at the same basic rate at each temperature.

Fig. 10 shows the effect of transferring an octopus from seawater at 6.8 ppm O₂ to seawater at 2.7 ppm O₂ and then returning it to the O₂ saturated water. The systemic heartbeat slows within a few seconds of transfer, and picks up again when returned to the oxygen-rich seawater. This animal had both cardiac ganglia removed.

In recovery from anaesthesia respiratory movements generally return before there is any sign of a regular beat by the hearts. As the O₂ content falls so the hearts slow, to the point where they may finally stop while respiratory movements continue (Wells, 1979). The systemic heartbeat, at least, may stop ‘spontaneously’ for periods of a few seconds (O. vulgaris, Wells, 1979), or for many minutes (O. dofleini, Johansen & Martin 1962); in most cases respiratory movements continue regularly throughout the period of cardiac arrest.

It has been held that the reverse does not occur, that if the animal stops breathing, the heart stops (Johansen & Martin, 1962, O. dofleini). This was certainly found to be the case in most instances obesrved during the present study ; cardiac arrest was typically abrupt, which would imply nervous inhibition of the heartbeat. It would plainly be economic for the animal to stop pumping blood through the gills while the water in the mantle remains unchanged.

The cardio-inhibitory reflex would be a further function for the visceral nerves and it should cease if these are cut. In experiments following visceral nerves section mantle movements were, unfortunately, rarely recorded and a sequence including a halt to these was observed only once, on an occasion when the animal was chased down its tank by the experimenter. This sequence is shown in Fig. 11 (a) ; heartbeats continued throughout.

Fig. 11(b) shows a sequence in which both heartbeat and respiratory movements were irregular in an animal with its visceral nerves intact, apparently at rest in its tank. Respiratory movements ceased, or were at least very much reduced, on several occasions, without any obvious corresponding effect upon the systemic heartbeat. The matter plainly requires some further study.

When all the nervous ganglia associated with the hearts are removed or disconnected, the three hearts of Octopus will still beat. They normally remain in phase, within the two branchial hearts contracting together, followed by the beat of the systemic. Pressures and frequencies can approach the values found in intact resting octopuses. The beat frequency in such cases is temperature and oxygen dependent, falling with a decline in either, just as it does in animals having the full compliment of nerve ganglia present and connected to the central nervous system.

Studies of the performance of the hearts of intact animals under a variety of conditions show that while the mean aortic pressure and pulse amplitude can vary enormously, heartbeat frequency is very stable (Wells, 1979; Wells & Mangold, 1979). This implies a system based on a pacemaker or pacemakers rather than one triggered reflexedly by the filling of the hearts, a conclusion that would eliminate the systemic ventricle as the source of the cardiac rhythm, since the systemic ventricle plainly varies in frequency, slowing down if its blood supply is limited. One must suppose either that the beat originates at the level of the branchial hearts, or that the pacemaker for the whole system lies outside the trio of hearts and their associated ganglia, perhaps within the walls of the vena cava.

Arguing against the latter possibility are two observations. One is that the isolated branchial heart continues to beat in vitro (and that, unlike the isolated systemic heart, no pressure-head of blood or seawater is required to cause it to do so ; Ransom 1884). The second is Alexandrowicz’s (1965) report that each cardiac ganglion contains a vesicle that pulsates with the same frequency as the branchial heart, and slightly in anticipation of the main muscular contraction; ‘When the ganglion is cut out with its neighbouring organs its activity as a rule lasts longer than that of the branchial heart.’ Taken together with the results of the present series of experiments which show the very disruptive effects of removal or disconnexion of the cardiac ganglia from the branchial hearts, these observations indicate the cardiac ganglion as the likely site of the main cardiac pacemaker on each side. The fact that the two branchial hearts regularly contract together, even after severing the nerves that connect the two cardiac ganglia through the fusiform ganglia, implies that the two branchial heart pacemakers are normally brought into phase by events upstream.

It has already been pointed out that the picture of the origin and control of cephalopod heartbeats that emerges from the present study differs radically from that to be found, for example, in the ‘Physiology of Molluscs’, where Hill & Welsh (1966) have summarized a literature based mainly on the behaviour of acute, partially dissected, preparations. From these, it has generally been accepted, following Ransom (1884), that the ventricle is the source of the beat, and that some connexion with the central nervous system is essential if a steady beat is to be maintained. In these acute preparations, moreover, it seems that the gill hearts cease to beat if the connexions between the fusiform and cardiac ganglia are severed, a result quite contrary to that now reported from free-moving and largely intact animals. The very unnatural conditions of much of the earlier experimentation make comparison of the two sets of results difficult. In particular it should be noted that restriction of mantle movements (for acute preparations it is split and pinned back, or folded inside out) may interfere with the heartbeat, through what appears to be an inhibitory reflex. This, and a number of other features (notably the means by which the two gill hearts tend to fall into phase after severance of all nervous connexions) of the hearts control system will require further investigation. What does seem quite plain even at this stage in the investigation is that much of the long-established ‘conventional wisdom’ about the control of cephalopod (and possibly other molluscan) heartbeats may be wrong and should now be re-examined under conditions more nearly resembling the state of affairs in the intact animals.

This work was aided by grants from the Royal Society (Scientific Investigations Program, and the European Scientific Exchange Program) in 1975 and 1977. I should also like to thank the staff and Directors of the Stazione Zoologica, Naples and the Laboratoire Arago, Banyuls, for their hospitality and the use of their facilities.