Ventricular Development, Ventriculomegaly, and Hydrocephalus in the Fetus and Newborn

Ventricular and CSF System Development ( Box 3.1 )

The primitive ventricular system begins to develop after closure of the anterior and posterior neuropores of the neural tube around 4 gestational weeks (GW). The neuroepithelium lining the inner surface of the neural tube begins to secrete embryonic CSF, dilating the telencephalic, diencephalic, and rhombencephalic vesicles (superventricles), which are destined to become the future lateral, third, and fourth ventricles, respectively. Soon thereafter, subarachnoid fluid appears around the ventral rhombencephalon. These events precede by weeks the development of the choroid plexus anlagen and opening of the fourth ventricular foramina (see Chapter 4 ) and by many months before a functioning arachnoid granulation system.

The first vesicle to expand is the rhombencephalic vesicle, which continues to do so until the 9th to 10th week, after which it shrinks in size to that of the fourth ventricle. The mesencephalic vesicle will form the future aqueduct and begins as a thin-walled structure with relatively large diameter, until growth of the mesencephalic wall reduces its caliber to that of the aqueduct. The diencephalic vesicle will form the third ventricle.

Cerebrospinal Fluid Production ( Box 3.2 )

The mechanisms for CSF production undergo significant maturational changes between the embryonic period and early infancy. The early neuroepithelium develops into an ependymal cell layer that lines the ventricular system. The ependymal layer then undergoes regional specialization (see Box 3.1 ) that includes the choroidal and nonchoroidal ependymal cells and specialized structures like the subcommissural organ at the aqueductal entrance. The earliest elements of the choroid plexus system develop in the fourth ventricle around 6 weeks of gestation, beginning in the fourth ventricle but becoming more voluminous in the lateral ventricles. By 20 weeks the choroid plexus has an adult configuration. Secretion of CSF from the choroid plexus is a two-step process, with first a passive, pressure-driven diffusion of plasma from the choroidal capillaries into the choroidal interstitial tissues, followed by active and selective transport across the choroidal epithelium into the ventricle. The highly specialized secretory choroidal epithelium is powered by densely packed mitochondria and has an apical membrane with finely regulated ion carrier proteins and a brush border of mobile beating microvilli. Therefore CSF is more than merely an ultrafiltrate of serum, and transport mechanisms regulate the amount and composition of fluid passing through. The choroidal ependymal lining forms a blood-CSF barrier, with cells joined by tight junctions containing high levels of claudin-5 transmembrane proteins.

The apical membrane of the choroidal ependyma contains proteins and enzymes, including aquaporin-1 (AQP-1), Na ⁺ /K ⁺ /Cl ⁻ cotransporter (NKCC1), and carbonic anhydrase, that regulate fluid and solute passage into the CSF. Aquaporin-1 is localized to the lateral and fourth ventricle choroid plexuses and promotes movement of CSF into the ventricles. In rodents, the NKCC1 cotransporter is responsible for up to half of CSF production, and its production is reduced by the NKCC1 inhibitor bumetanide. Carbonic anhydrase has long been known to play a role in CSF secretion, and its antagonist, acetazolamide, has been used in the treatment of hydrocephalus.

The nonchoroidal ependymal cells and the choroid plexuses of the four ventricles have different transcriptional pathways, producing CSF with different gene product and proteomic profiles, presumably to promote specific regional differentiation programs at different levels of the rostral neuroaxis. The composition of CSF may also change over development. For example, CSF formed by the embryonic neuroepithelium that fills the rostral superventricles (see earlier) differs significantly from that formed later, which contains different growth factors, proteoglycans, and extracellular matrix proteins important for early neuronal survival, particularly that of the progenitor cells in the ventricular zone.

Current understanding is that in mature humans, CSF formation occurs at a rate of 0.3 to 0.4 mL/min, with the major source (80%) being the choroid plexuses of the lateral, third, and fourth ventricles and the remaining 20% from other sources, such as the nonchoroidal ependymal cells and interstitial fluid (ISF). However, the role of the choroid plexus as the major producer of CSF has been questioned. The ependymal layer lining the ventricles appears to play a prominent role in the production of extrachoroidal CSF (see Box 3.1 ). This source of CSF is vulnerable to alterations in the ependymal lining, especially that caused by ventricular distention.

Choroid plexus function is influenced by multiple factors, such as neurotransmitters and modulators, including acetylcholine, noradrenaline, and serotonin. CSF production may also be regulated by the autonomic nervous system, but this issue remains debated. Sympathetic activation reduces CSF secretion, whereas cholinergic activation increases CSF secretion. Peak CSF production is during sleep (especially non–rapid eye movement sleep) when sympathetic tone is low and parasympathetic tone is high; autonomic pathways may contribute to the circadian rhythms of choroidal CSF secretion.

Animal studies suggest that the specialized membrane systems of the choroid plexus can be altered by pathophysiological influences . In a rodent model, hydrocephalus decreased AQP-1 expression in the apical membrane of the choroidal ependyma, potentially as a compensatory response to decrease CSF production. In another model, NKCC1 was activated by inflammatory pathways triggered by intraventricular hemorrhage leading to hydrocephalus, which may be prevented by bumetanide.

Cerebrospinal Fluid Flow (see Box 3.2 )

Several different mechanisms promote CSF movement through the central nervous system, and the dominant mechanisms vary across regions. Broadly, these mechanisms include different pressure-based forces and motile ependymal cilia . Synchronized motile cilia are key to moving CSF along the ventricular surface. In the human fetal brain, the differentiation of ependymal cells into ciliated cells occurs almost exclusively in the embryonic period. Different types of motile ciliated cells are spatially distributed and direct CSF flow in specific directions. Cilia-driven CSF flow moves CSF along the ventricular wall lining, presumably distributing neuropeptides and trophic factors to specific targets at that level of the neuroaxis. Normal synchronized ciliary beating appears to be important for ventricular development and might ultimately influence cortical development through its support of the adjacent subependymal ventricular zone. Specifically, the motile cilia are thought to distribute trophic factors across the ventricular surface, thereby supporting neural stem cell proliferation and subsequent neuronal migration from the ventricular zone. What remains unclear is whether and to what extent ciliary movement plays a role in CSF bulk flow between ventricles. In conditions with defective cilia motility (ciliopathies), hydrocephalus is common. Acquired hydrocephalus due to conditions such as intraventricular hemorrhage and infection is likely in part mediated by ciliary dysfunction.

The ependyma has become recognized as an important secretory “organ” during brain development and might play a play an important early role in hydrocephalus . Specifically, the ependymal denudation associated with hydrocephalus has been shown to precede development of hydrocephalus and/or Chiari II formation and therefore may not be the result of hydrocephalus as is commonly thought. Murine models suggest that a primary ependymal abnormality may play a causative role in hydrocephalus. In the normal brain, CSF is secreted continuously, albeit with circadian fluctuation. Such secretion is modulated on the vascular aspect of the choroidal cells by autonomic innervation, with sympathetic inhibition and cholinergic stimulation of CSF production. On the ventricular aspect of the choroidal cells, receptors for neurotransmitters and modulators such as dopamine, serotonin, and melatonin (which are normally present in the CSF) have been identified and likely play a regulatory role in CSF production.

Another region of specialized ependyma, the subcommissural organ-Reissner fiber complex, is located at the rostral entrance of the aqueduct, an area where the CSF pathways are at their narrowest. The subcommissural organ-Reissner fiber complex is a transient maturational expansion of the ependyma that is thought to play important roles in the early developing ventricular system and in the circulation and reabsorption of CSF. In rodent models, dysfunction of this complex has been implicated in aqueductal stenosis, an important cause of congenital hydrocephalus in humans.

Animal studies suggest that pressure differentials of various forms play a role in bulk CSF movement, including tonic, phasic, and pulsatile pressure changes (see Box 3.2 ). Tonic pressure changes resulting from CSF production at the choroid plexus and ependyma provide a background pressure upon which other forces operate. Phasic pressure changes emanate from respiratory, abdominal, postural, and movement changes and drive CSF between ventricles and through the ventricular system in either direction. For example, CSF in the ventricles moves rostral with deep inspiration and caudal with deep expiration; changes in CSF flow through the aqueduct have been described with respiration. Pulsatile forces of cardiac origin drive CSF-ISF flow through the penetrating arterioles and the glymphatic system (see later) ( Figs. 3.2 and 3.3 ); cardiac pulsations of the choroid plexus probably play less of a role than previously believed.

Impaired CSF flow through the extraaxial spaces of the convexities can result from hemorrhage, infection, and other malformations, such as the obstructed SAS in cobblestone lissencephaly, and cause hydrocephalus.

Cerebrospinal Fluid Clearance (see Box 3.2 )

The different pathways for CSF absorption have also come under scrutiny in recent years, especially as they relate to the immature brain. In the mature brain the arachnoid villi, especially those in the major dural venous sinuses and, to a lesser extent, around the brainstem and spinal nerve roots, are considered by many to play the dominant role in removing CSF from intracranial compartment. However, the early development of the arachnoid villi is delayed relative to CSF production, and they are unlikely to play a significant role in CSF clearance from the immature brain. Development of the arachnoid granulations begins around 26 GW, initially as dural depressions filled with arachnoidal cell clusters, which penetrate the dural fibers toward the venous wall, eventually entering the lumen around 35 GW as endothelial protrusions (arachnoid villi). After 39 to 40 GW their number and complexity begin to accelerate but only become fully functional around 18 months of age. The late development of the normal arachnoid villi makes it difficult to implicate them in pathways leading to VM in the fetus and newborn.

Transfer of CSF across the arachnoid villi is sensitive to the pressure gradient between the CSF and the venous circulation. Unlike choroid plexus CSF production, which decreases with increased CSF pressure, the arachnoid villi respond to increased CSF pressure by distending, thereby increasing their exchange surface area and CSF transfer to the venous system. Conversely, increased pressure in the venous compartment (e.g., due to changes in jugular venous pressure, respiration, abdominal pressure, arousal state, physical activity, and posture) decreases CSF flow into the venous system. This mechanism has been implicated in the excessive CSF accumulation (primarily extraaxial) in children with certain forms of congenital heart disease. In addition, the site of major CSF drainage might change; for example, in an upright position, CSF absorption increases across the lower spinal arachnoid villi.

Recent animal studies have provided new insights into CSF dynamics after its egress from the ventricular system, although unresolved issues persist. One major conceptual advance is that of a glymphatic system (see Box 3.2 ). Unlike other organ systems, the brain lacks an intrinsic lymphatic system, although such a system has recently been identified in the meninges. The glymphatic system is thought to assume this role by linking (astro)glial pathways to the meningeal lymphatics through perivascular fluid pathways (see Figs. 3.2 and 3.3 ). The concept of a glymphatic system is based on several recent discoveries, including (1) the discovery of the meningeal lymphatic system, (2) the demonstration of significant CSF flow along perivascular spaces into and out of the brain parenchyma, and (3) the role of a gliovascular unit at the interface between capillaries and the astrocytic end-feet that bridges the perivascular inflow and outflow systems. The potential role of the glymphatic system in the cerebral hydrodynamics of the fetus and newborn has generated growing interest in recent years, driven by the prevalence of fetal VM and the fact that the arachnoid villi only begin to function after birth and remain immature for many months (see earlier).

Despite the early seminal work of Cserr and colleagues, showing convective (“advective”) bulk-flow CSF clearance (as opposed to diffusion) along perivascular pathways, it took almost 4 decades for the importance of this perivascular fluid pathway system (see Figs. 3.2 and 3.3 ) to become appreciated. This delayed recognition was due to the fact that in earlier electron microscopy studies the perivascular spaces collapsed postmortem. More recently, in vivo imaging studies in live rodents demonstrated perivascular spaces much larger than previously appreciated, making resistance-based computational models more plausible. In addition, the loose fibrous matrix of the perivascular space provides a low resistance pathway for the convective CSF flow along the periarterial spaces. More recent studies have shown that CSF moves along Virchow-Robin spaces of penetrating arterioles in the direction of blood flow, propelled by arterial pulsations, to the precapillary level (see Fig. 3.3 ). Several observations support the notion of a “perivascular pump,” including the close association between the flow velocities and pulsatility of peristaltic CSF flow, blood flow, and the cardiac cycle (see Fig. 3.3 ). These features suggest that the CSF driving force is mediated through the vessel wall, by cardiac pulsation and smooth muscle recoil. A background platform of tonic CSF pressure drives the CSF into the lower pressure in the Virchow-Robin spaces, after which perivascular flow is driven primarily by arterial pulsations (see Fig. 3.3 ). The efficiency of this perivascular pump is influenced by changes in blood pressure, which may influence the pulse waves; for example, elevated baseline blood pressure reduces the superimposed arteriolar wall excursion with each arterial pulsation. Forward flow in the perivascular spaces is dependent on the fast upstroke in arterial flow with each pulse wave, followed by a slow downstroke. The perivascular pump becomes less efficient in acute arterial hypertension because the fast upstroke is followed by an initial fast downstroke and then a slow downstroke, slowing down antegrade flow and increasing retrograde flow. Increasing blood pressure leaves the arterial diameter unchanged but changes the pulsations of the arterial wall, increasing backflow and thereby reducing net flow in the PVS. How the increase in intracranial pressure, and its effect on pressure gradients (e.g., in hydrocephalus), might change this system is not clear but is open to speculation. The glymphatic system is mainly active during sleep and is largely inactive during wakefulness.

Within the brain parenchyma, astrocytic end-feet surround all arterioles, capillaries, and venules. Through this gliovascular unit CSF enters the brain ISF where it admixes with the ISF before passage into the perivenular spaces and egress to the surface of the brain. Perivascular astrocytic end-feet membranes are enriched in AQP-4 water channels, which shuttle CSF into the interstitium and through the neuropil. Astrocytes play a critical role in the homeostatic regulation of the neuronal, especially synaptic, microenvironment, where the composition of the ISF may undergo dramatic changes during neuronal activity, including glutamate and potassium reuptake at the synapse; failure of this system may profoundly influence neuronal function and cause neurotoxicity. Finally, the CSF-ISF enters the meningeal lymphatic system through which it drains ultimately into the jugular venous system. Additional routes of communication between the SAS and meningeal lymphatic system include the perivenous and perineural spaces of the cranial and spinal nerve roots. The glymphatic system has garnered much attention for its potential role in removal of injurious waste products from the interstitial space implicated in neurodegenerative conditions of aging. In addition, the glymphatic system is likely involved in the transport of other important substances, including glucose, lipids, amino acids, and neurotransmitters.

Aquaporins (AQPs) are distributed at key fluid interfaces in the brain, including the blood-CSF and blood-brain barriers, as well as the pial and ependymal surfaces, and play a prominent role in cerebral hydrodynamics. On the other hand, AQP-4 is densely expressed in the astrocytic end-feet of the gliovascular unit. These roles of AQPs in regulating brain fluid homeostasis have made them potential targets for therapeutic agents.

An unresolved question is whether the glymphatic system acts solely as a waste removal system or whether it is capable of effecting significant changes in overall volumetric CSF changes. Whether these functions are separate or part of a combined functional system remains unclear.

The sleep-wake cycle has a major influence on cerebral hydrodynamics. During the awake and active sleep states the increased sympathetic nervous activity decreases choroid plexus CSF production. Norepinephrine plays several potential roles, including a direct inhibitory action on choroid epithelial cells in wakefulness, enhanced CSF production during quiet sleep, and suppression of the glymphatic system during wakefulness. The glymphatic system is dramatically activated during sleep and suppressed during wakefulness. This difference suggests that sleep is important for the removal of brain waste products.

Diagnosis and Classification ( Box 3.3 )

Fetal ultrasound is the standard prenatal screening modality for identifying fetal VM and has become an important tool for monitoring ventricular size, configuration, and its progression over gestation. By convention, fetal ventricular size is measured by US as the diameter of the lateral ventricular atrium on an axial view at the level of the thalami or the glomus of the choroid plexus ( Fig. 3.4 ). Using this approach, the ventricular size remains relatively stable at around 6 to 7 mm during normal development between 14 weeks and term. A widely used criterion for diagnosing VM is ventricular diameter of ≥10 mm (>4 standard deviations above the mean) at the axial level discussed earlier. It is important to realize that this 10 mm cutoff criterion is based on a statistical norm and that ventricular size above 10 mm should serve only as an indicator that further investigation is warranted. Another ultrasonographic indicator of VM is the “dangling choroid” sign ( Fig. 3.5 ). Fetal US may also provide other helpful information, including detection of (1) intracranial hemorrhage, (2) the increased echogenicity of the ventricular lining when ependymitis results from intraventricular hemorrhage or infection, and (3) larger malformations. In addition, fetal US is useful for monitoring progression of VM over gestation and may help distinguish hydrocephalic from nonhydrocephalic VM. Specifically, large ventricles and a large head circumference suggest fetal hydrocephalus, whereas large ventricles and a small head are more suggestive of ex vacuo VM. Furthermore, in cases of hydrocephalic VM the configuration of the ventricular system may provide evidence of where the CSF outflow is impeded. For example, lateral and third VM with a normal fourth ventricle suggests aqueductal stenosis, whereas tetraventricular distention suggests a fourth ventricular outflow problem. In cases where VM (usually not severe) is accompanied by increased extraaxial CSF, a communicating or “nonobstructive” mechanism of VM may be inferred.

Although US may detect major parenchymal anomalies, the development of an informed prognostic opinion and management plan usually requires other modalities, such as magnetic resonance imaging (MRI), genetic studies, and studies for infectious conditions. Informed prognostication in cases of VM (see later) is based on more than ventricular size. Once VM has been diagnosed by US, the recommended next steps should include a comprehensive diagnostic workup that includes TORCH serology (see Chapter 38 ), karyotype and microarray, and serology for fetal infections associated with VM of which cytomegalovirus, toxoplasmosis and lymphocytic choriomeningitis virus constitute the vast majority, as well as serial fetal US to exclude progression, and a fetal MRI. The sensitivity of brain imaging techniques for detecting additional brain anomalies in VM increases from fetal US to prenatal MRI, with postnatal MRI being the most sensitive. When present, VM is more common in males, by a ratio of around 1:2.5.