Variations in the cerebrospinal fluid dynamics of the American alligator (Alligator mississippiensis)

The cerebrospinal fluid (CSF) system is found in all vertebrates [1, 2]. Studies in different vertebrates have identified a number of influences, or forces, acting on the CSF including: cardiac, respiratory, vasomotion, orthostatic, ciliary, transmantle gradients, and skeletal muscle. The relative importance of these different influences remains unclear, even in humans; furthermore, the functional complexity of the CSF system is still uncertain. The CSF plays a crucial role in protecting the brain from traumatic injury [3], but also maintains and regulates the brain´s chemical environment [4, 5]. Recent breakthroughs [6, 7] in neuroscience have amplified the potential regulatory capacity of the CSF system and demonstrated that neuro-degenerative diseases, and Alzheimer’s in particular, may be due to insufficient clearance of metabolic waste by the CSF. As such, it is interesting and timely to re-investigate the CSF system and its variance across species.

The forces influencing the CSF system are best known in humans. Here, as embodied by the classic Monro-Kellie doctrine [8], the out-of-phase pulsations of the arterial and venous blood flow are compensated for by CSF pulsations [9]. Furthermore, both the respiratory and vasomotor influences on the CSF have been measured by MRI [10, 11] and confirmed by modeling [12, 13]. In humans the CSF pressure varies with orthostatic (postural) pressure, likely due to a functional coupling with orthostatic changes in venous pressure [14, 15]. Finally, transmantle gradients in humans are very small, or zero [12, 16‐18]. From a clinical perspective, it is known that alterations in CSF pressure can result in both physical displacement of the neural tissue as in cerebellar herniation [19], and functionally-disruptive compression of the neural tissue as in hydrocephalus [20] or spinal syrinx [21].

In humans the CSF in the intraventricular and subarachnoid spaces is freely exchanged through the median (Foramen of Magendie) and lateral (Foramen of Lushcka) apertures in the 4th ventricle. These apertures are not present in amphibians and reptiles [22]; in these species CSF continuity depends on exchange across the tela choroidea. The purpose of this study was to explore the CSF dynamics in the American alligator, Alligator mississippiensis, a species in which the intraventricular and subarachnoid spaces are spatially isolated. We hypothesized that the relatively restricted mobility of the CSF in Alligator, compared to the condition in humans, would alter the cerebral compliance and aspects of the CSF waveform dynamics. In alligators the ventilatory pattern is far more plastic than in humans or other mammals; alligators are capable of sustained apnea [23, 24], and switch between multiple ventilatory mechanics [25, 26]. Cerebral blood flow is more variable in alligators than in humans, due in part to the alligator's ability to actively redirect (shunt) cardiac output between the pulmonary and systemic circuits [27, 28]. Alligators, like some other reptiles, appear to lack a barostatic reflex [29, 30].

Thus, when compared to humans, alligators have a different cerebral compliance, and at least three of the known CSF influences (cerebral arterial pressure, ventilation, and orthostatic forces) are more variable in alligators. For these reasons, we hypothesized that the CSF dynamics in Alligator would be more variable than in humans, particularly with respect to the cardiac and ventilatory cycles. To explore these hypotheses we surgically implanted pressure transducers into the cranial subarachnoid space of Alligator mississippiensis and analyzed the CSF pressure traces in both the frequency and temporal domains.

The changes in CSF pressure that occur during rotation (Figs. 4d, 5a) are similar to expectations; assuming a 50 cm distance from the axes of rotation to the pressure cannula, a 45° rotation would produce a 27 mmHg pressure head. Similar postural/barostatic shifts in intracranial pressure have been documented in mammals [32]. Alligators do not exhibit a compensatory baroreflex when orthostatically challenged [29]; head-up rotations cause a significant reduction in carotid blood flow in A. mississippiensis, while head-down rotations are associated with a significant increase in carotid blood flow [30]. A lack of compensatory baroreflex, and significant alterations of blood flow patterns due to gravitational gradients, have been previously described in reptiles [33, 34]. These postural shifts in the blood distribution of A. mississippiensis would cause a corresponding change in cardiac output and cranial artery/arteriole pulsation amplitude, and would presumably shift the cranial compliance [35, 36], which could explain the differences in CSF pulsation amplitude observed during rotation (Fig. 3c).

Pulsations have been reported in the extra brain fluid of the Port Jackson shark (Heterodontus portusjacksoni) [37], but there are few additional descriptions of CSF pulsations from non-mammals. CSF pressures and pulsations have been quantified from a variety of mammalian species, including: cats [38]; rabbits [39]; dogs [40]; monkeys [41]; horses [42]; mice [43]; and humans [44]. These mammalian studies report similar CSF pressures, pulsations with similar temporal relationships to the cardiac cycle, and frequency analyses which consistently show that the cardiac cycle has the greatest contribution to CSF pressure pulsation [45].

In Alligator the CSF pulsations do not show a consistent temporal relationship to the cardiac cycle, even within an individual and while the animal is horizontal (compare Figs. 1a, 4a, b and 5d). As in mammals, the cranial arteries course along the base of the alligator brain [46, 47]. Blood flow through these vessels may be more variable in Alligator than in a mammal, owing in part to the multiple shunts and physiological regulations of cardiac outflow in Alligator [48‐50]; but these are primarily voluntary systems which would have been impeded, if not eliminated, by the anesthesia. If the CSF pulsations in Alligator are driven by arterial pulsation of some form, be it pulsations of the large arteries surrounding the brain [51], arteriole driven piston-like pulsation of the brain within the dural sac [52], or pulsations of the choroid plexus [53], how can there be such temporal variation between the cardiac cycle and the CSF pulsations?

Ventilatory movements of the mammalian thorax often produce synchronous CSF pressure pulsations [54, 55], though these pulsations are of a lower amplitude than those attributed to the cardiac cycle [45]. The mechanics of these ventilatory pulsations are uncertain; direct expansion of the thoracic spinal dura has been hypothesized [56] as has pressure influences on the cephalic venous drainage [57]. Though alligators share with mammals cyclic changes in thoracic pressure associated with ventilation [58], In some ways Alligator ventilation is different from that of mammals; alligators are intermittent breathers [24], and can use a hepatic pump or abdominal displacement to drive ventilation [25, 59]. A recent study [60] argued that during the evolution of tetrapods when axial musculature de-coupled from locomotion to drive ventilation, it also produced a novel propulsive mechanism for the CSF. Alligators were used for this study specifically because their unusual plasticity in ventilatory mechanics suggests that the functional link between ventilation and CSF propulsion would not be as developed in this group of tetrapods.

In the experiments reported herein, the alligators were actively ventilated at a frequency and tidal volume similar to what has been reported from non-anesthetized alligators [59, 61]. Although during these experiments the alligators were undergoing cyclic changes in intrathoracic pressure (like a ventilating mammal), the relationship between ventilation and CSF pulsations in the alligator is unlike anything every reported from a mammal. In some cases the CSF pulsations appear to be driven solely by the ventilatory cycle (Fig. 4b), whereas in others the ventilatory cycle appeared to interfere with the CSF pulsations (Fig. 4c). The unique relationship among the cardiac, ventilatory, and CSF pulsatory cycles of Alligator is also apparent during frequency analysis. Unlike the situation in mammals, in which the cardiac cycle makes the greatest relative and absolute contribution to CSF pulsations [45], in Alligator there was more variation (Fig. 1c) and sequences were analyzed in which the ventilatory component had a greater relative and absolute contribution than the cardiac component.

Isoflurane causes respiratory and cardiac depression in reptiles, presumably through increasing vagal tone [62]. Reptiles appear to have the same autonomic control over cardiorespiratory interactions as mammals [63]. The increased vagal tone produced by the isoflurane disrupted some of these controls: though three different ventilatory rates (apnea, 8 breaths per minute, and 24 breaths per minute) were applied to A. mississippiensis, there were no corresponding changes in heart rate (Fig. 1c). Previous workers have argued that ventilatory movements were transferred to the CSF by intrathoracic pressure changing venous drainage such that the venous pressure of the dural sinuses would change [64]. An increase in intracranial venous pressure would lead to an increase in intracranial CSF pressure, but would not disrupt the CSF pulsations; rotating the alligators head down caused the venous blood to pool in the head [30], and was associated with an increase in CSF pressure (Figs. 4d, 5a) but also resulted in an increase in CSF pulsation amplitude (Fig. 3c). Frequency analysis demonstrated that when ventilatory movements were held constant, but the alligator was rotated head-up (so arterial pulsations were reduced) the relative contribution of the cardiac peak within the CSF pulsation was decreased while the ventilatory peaks were increased (Fig. 6b).

Accordingly, we follow [56] and hypothesize that in Alligator mississippiensis ventilatory movements result in displacement of the spinal dura adequate to create CSF pulsations. The capacity of the spinal dura to create these pulsations is demonstrated by the activation of the myodural bridge (Fig. 5b). The myodural bridge only has a contact area of approximately 3.5 cm with the spinal dura (and thus is a small fraction of the thoracic length), yet twitch stimulation causes a significant decrease in CSF pressure even though the dura was estimated to be displaced only 0.1 mm [60]. We do not propose that the myodural bridge contracts rhythmically, only that the influence of the myodural bridge on CSF pressure suggests that a rhythmic thoracic dural displacement could well produce the type of CSF pulsations recorded in Alligator mississippiensis.

The relationship between between the ventilatory cycle and CSF pulsations in the alligator may also relate to the unusual long duration of the CSF pulsations reported in this study. The CSF pulses in A. mississippiensis had a frequency of approximately 0.3 Hz (Figs. 1b, 3a), which is far lower than the ~ 1.3 Hz of human CSF pulsations [45, 65], but reflects the slow heart rate of Alligator when anesthetized with isoflurane [66]. In mammals as heart rate increases there are corresponding increases in CSF pulsation frequency and decreases in CSF pulse duration; in mice, where the resting heart rate is approximately 600 bpm, [43] recorded CSF pulsations with a frequency of around 6 Hz, some 20 × that of the alligator. In the alligator an increase in heart rate increases the frequency of the pulsations, but the duration of each CSF pulsation seems to hit a minimum at around 1 s. This can lead to CSF pulsations originating from the diastolic slope of earlier pulsations (Fig. 3b). These summated curves precluded quantifying the duration of these CSF pulsations. We are treating 1.0 s as the approximate minimum duration of a CSF pulsation in Alligator because we see discrete pulsation of slightly greater than 1.0 s (Fig. 3c) but did not record any discrete pulsations with durations less than 1.0 s. This minimum duration is very close to the inhalatory duration of the ventilatory cycle (Fig. 4b), which is presumably when the spinal dura would be displaced and a CSF pulsation produced. This hypothesis that in Alligator the ventilatory contributions create a minimum duration to the CSF pulsations is supported by the observation that only a portion of "summated" CSF curves (Fig. 4b) increase in amplitude during head-down rotations. Nevertheless, the presence of an (apparent) minimum duration to the CSF pulsations, even during apnea, indicates that other factors are influencing the variation in CSF pulse duration.

Studies of the bulk flow of CSF between the cranial and spinal sub-arachnoid spaces of mammals have demonstrated pulsatile flow, but the pulsations are quite different to those reported herein from Alligator. In humans, and other mammals, these CSF flow pulses are of much shorter duration, typically one second or less [67], compared to the 3 s duration in alligators (Fig. 6c). Furthermore, the mammalian flow pulses are typically bi-directional [68], though asymmetry in the two flow directions may yield net directional flow [69]. The flow velocities recorded in Alligator were unidirectional; though the CSF slowed down at the end of the pulse it did not change direction. These Alligator flow velocities were recorded from the subarachnoid space dorsal to the spinal cord; there is no evidence that all of the spinal CSF fluid was flowing unidirectionally. Similar "large scale" unidirectional CSF flow pulses have been reported from zebra fish embryos [70, 71] and previous studies on human CSF flow have wrestled with flow velocities and patterns that seem "discordant" with potential reservoir volumes [72].

Spectral frequency analysis (Figs. 1c, 6b) was used to segregate the cardiac and ventilatory components of the CSF pulsation, but the resulting power spectra were consistently different from those reported from mammalian studies. Power spectral analyses of mammalian CSF pulsations reveal a large number of higher-order harmonics [45, 65], which have been interpreted in terms of skull resonance and system compliance [45, 65, 73, 74]. In this analysis of Alligator, spectral peaks above a 3rd order harmonic were never recovered. As the spectral analysis of the EKG waveform shows (Fig. 6a), this was not due to an inherent limitation of the recording/analysis protocol, but seems to reflect the nature of the CSF pulsations themselves. Setting aside the significance of these low energy components of the pulsation, we hypothesize that their absence may reflect the nature of the reptilian meninges. Unlike the situation in mammals, in Alligator the dura is not fused to the skull [75], and there appears to be more mobility between the meningeal layers. As such the meninges in Alligator may be more compliant than those of mammals, which could effectively dampen some of the CSF pulsation.

Perhaps the most intriguing finding to come out of this study is that the CSF pulsations "differ" between the skull and spinal canal, despite the continuity of the CSF and subarachnoid spaces. When the "same" CSF pulsations are recorded from the cranial and spinal CSF there is a clear synchrony (Fig. 5c). Spectral analysis reveals a slightly different profile between the two, perhaps indicative of the large venous sinus in the epidural space of the spinal canal [76]. The differences between the cranial and spinal CSF pressure curves could be a manifestation of compliance [77, 78] or reflexional differences; similar differences have been previously reported from human studies [79]. Closer examination shows that the CSF pulsations have a temporal offset of approximately 200 ms, with the spinal pulses lagging behind the cranial pulses. This temporal offset means there is also a pressure offset (of up to 1.5 mm Hg); if the curves are artificially synchronized the pressure difference is less than 0.5 mm Hg. The cranial and spinal pressure transducers were only 12 cm apart; sound wave propagating through fluid would cover 12 cm in roughly 0.08 ms. An earlier study comparing intracranial and lumbar pressure recordings in humans [80] also reported a temporal offset greater than sound wave propagation, and noted that the delay was decreased at higher pressures, presumably due to changes in compliance. Similar temporal offsets between cranial and spinal CSF pulsations have been found in humans [81], suggesting that there is a real, previously unexplained, delay or temporal shift that occurs as part of the CSF transition from the skull to the spinal canal.

All vertebrates have CSF systems involving pulsatile and net flow influenced by cardiac, respiratory, vasomotion and orthostatic forces the relative importance of which are currently debated, particularly in light of the CSF´s role in brain clearance. In Alligator mississippiensis CSF pulsations have amplitudes of approximately 5 mmHg and durations around 3 s (Figs. 1a, 3a), and the bulk flow of CSF around the spinal cord includes clear unidirectional flow that is maintained for a similar duration (Fig. 6c). Furthermore, the respiratory pressure gradients can be of both longer duration and larger in amplitude than the cardiac-related pressures (Figs. 1c, 3b), and the separation between pulsatile and bulk CSF flow may be diminished or eliminated. Only further comparative studies will demonstrate if the large influence of ventilation on the CSF, and the long duration CSF pulsations and CSF flow profiles are a unique attribute of A. mississippiensis, or a shared reptilian characteristic. The current study demonstrated that all of the cardiac, respiratory, vasomotion and orthostatic forces are significant players of CSF dynamics in alligators, and that the link between the pulsatile and bulk CSF flow regimens may be more amenable to experimental exploration and manipulation in this species.