Cerebrospinal Fluid Physiology

Published on 26/03/2015 by admin

Last modified 26/03/2015

Print this page

This article have been viewed 2331 times

CHAPTER 188 Cerebrospinal Fluid Physiology

Under normal physiologic conditions, most of the cerebrospinal fluid (CSF) is secreted by the choroid plexus and flows through the ventricular system to emerge from the fourth ventricle. The CSF then traverses the subarachnoid space (SAS) and drains from the central nervous system (CNS) through unidirectional open channels as a result of a hydrostatic pressure difference. This chapter examines the clinical and experimental evidence for these conclusions. First reviewed is the CSF physiology of the developing fetus because this sets the stage for what is to follow.

Cerebrospinal Fluid Physiology of the Developing Fetus

Cerebrospinal Fluid Pathways

Development of the ventricular system begins with closure of the neural groove to form a neural tube. Fluid is present within the neural tube even before the choroid plexus anlage appears. This fluid serves as structural support for the neural tube, as well as a pathway for diffusion of metabolites before the formation of blood vessels. In the small thin-walled fetal brain, fluid movement is characterized by a lack of communication between the ventricles and the meningeal fluid spaces. Ciliary action inside the ventricles produces directional streaming, whereas fluid mixing aids in diffusing substances from the outer surface of the brain via the extracellular spaces (ECSs) through the wall of the neural tube into the ventricular cavity and vice versa.

The mechanism of ventricular formation appears to be well conserved in vertebrates, with ventricular shape being determined by adjacent cellular proliferation. The initial ventricular fluid is not dependent on the presence of blood vessels. The CSF within the ventricles contains hormones, proteoglycans, and ions and its composition varying with time and from site to site within the ventricles, depending on adjacent parenchymal development.¹ Ventricular enlargement is present in early development but steadily decreases to its size at term by approximately 30 weeks’ gestation.^2,³

The mesenchyme surrounding the brain thins out in a definite, organized pattern to form the pia-arachnoid membrane, the cisterns, and the SAS. The residual mesenchyme forms the trabecular meshwork of the arachnoid. Ultrasonography has largely been supplanted by magnetic resonance imaging (MRI) for study of the ventricular system and SAS in the fetus.² The width of the SAS in early development is fairly large until 32 to 34 weeks’ gestation, when it declines to its size at term. The volume of the SAS is related to CSF formation, CSF absorption, and fetal CNS development. An enlarged SAS can indicate chromosomal abnormalities, infection, and CNS underdevelopment. Enlargement of the cisterna magna (CM) can be seen with trisomy 18 and 21, Dandy-Walker malformation, cerebellar hypoplasia, and posterior fossa cysts.² A large CM by itself can be an isolated finding and compatible with normal development. A small CM can also be seen in patients with Chiari malformations.

The SAS and its configuration are virtually complete at birth.⁴ The SAS develops independently of choroid plexus CSF secretion and does not require the presence of CSF circulation. There is no movement of fluid out of the ventricular system during early development of the SAS.⁵ The outlets to the fourth ventricle are covered with a membrane, even after the choroid plexus begins to secrete CSF. This membrane does not appear to impair outflow of CSF from the ventricles because drainage occurs via intracellular pores in the membrane.⁶ The membrane subsequently becomes progressively attenuated, and larger and larger holes develop until it is no longer present.

Resistance to outflow from the ventricles increases as gestation progresses, but it does not change to any degree after birth.⁷ Resistance to CSF drainage in turn is the end product of differentiation of the cells that make up the pathways. Glycoconjugates appear to influence development of the matrix of the drainage pathways and to determine the degree of resistance.⁸ Presumedly, impaired function or absence of normal glycoconjugates could lead to increased resistance, which if significant, would result in hydrocephalus.

Choroid Plexus

The choroid plexus of the third and fourth ventricles arises from invaginations in the roof plate, whereas the choroid plexus of the lateral ventricles arises from the choroidal fissure of the developing telencephalon. The choroid plexus consists of an epithelium covering a stromal core. The stromal core, or tela choroidea, is derived from mesenchyme, whereas the epithelium arises from neural tube spongioblasts lining the ventricles. The epithelium is initially pseudostratified but is subsequently transformed into a single layer of cuboidal cells. During development, the choroid plexus forms lobules, which in turn become fronds covered with microvilli. This process markedly increases the surface area of the choroid plexus while reducing the proportional volume that the choroid plexus occupies within the ventricular system. The microvilli become progressively more convoluted, which may relate to secretory activity. In humans, as in animals, the fourth ventricular choroid plexus is the first to develop. However, the greatest choroid plexus bulk resides within the lateral ventricles and is attached to the medial ventricular walls, where it is suppled by branches of the anterior and posterior choroidal arteries. The remaining choroid plexus hangs from the roof of the third and fourth ventricles and is supplied by branches of the medial posterior choroidal artery and the anterior inferior and posterior inferior cerebellar arteries, respectively. The choroidal veins drain mainly into the internal cerebral vein, a part of the deep venous or galenic system.

Various genetic factors that result in malformation or abnormal function of the choroid plexus have been described in animal models but not thus far in humans.⁹ Tight junctions between the epithelial cells of the choroid plexus become operative early in development and limit the free passage of proteins (Fig. 188-1). Synthesis or transport of proteins (or both) into and out of the CSF by the choroid plexus appears to influence neurogenesis.^10,¹¹ When developed, the choroid plexus, in addition to secreting CSF, also performs regulatory functions such as transporting various substances out of the CSF, neutralizing substances that could be harmful to the CNS, and helping maintain the homeostasis needed for normal CNS function.

FIGURE 188-1 Diagrammatic representation of the choroid plexus. The capillary endothelium is of the attenuated fenestrated type, which allows an ultrafiltrate of plasma to reach the basal side of the epithelial cells. Tight junctions at the apical or ventricular side of the epithelial cells restrict molecular movement and constitute the blood-CSF interface.

The locations of the lateral, third, and fourth ventricular choroid plexus may also play a unique role in CNS development inasmuch as trophic substances can be added or removed from CSF at specific sites. The CSF in the lateral ventricles is in continuity with the germinal matrix, the main site for cortical cell proliferation and subsequent migration, whereas the CSF in association with the choroid plexus of the fourth ventricle may more likely influence structures in the basal cisterns.¹²

Development of the Cerebrospinal Fluid–Blood-Brain Barriers

The earliest CSF is most likely an ultrafiltrate of plasma. With development of the CSF-blood-brain barriers, CSF becomes a secretion. The concept that these barriers are less developed in the fetus and infant is based on observations that blood-borne dyes stain the immature brain more extensively, that the concentration of protein in CSF is higher in the newborn, and that metabolites and various solutes enter more readily and reach higher concentrations in the fetal brain than in the adult brain.¹³

However, many problems exist in trying to determine the permeability of the CSF-blood-brain barriers in an immature brain because there are many different and continual changes occurring during gestation. The very early embryonic brain has no blood vessels, so exchanges with blood in vessels external to the CNS must occur indirectly. The main route of transfer appears to be from blood into CSF and then into the parenchyma, and it seems likely that the choroid plexus and early CSF have important roles in nutritional supply to the developing brain.¹⁴

After vascularization begins, the initial low density of blood vessels increases steadily with gestation. The number of blood vessels and cerebral blood flow increase in relation to metabolic needs. The ECS appears to be larger at certain stages, so there is less restriction to free diffusion. In addition, an increase in ventricular volume takes place. This increases the volume distribution, which in turn decreases CSF/plasma concentration ratios for small, passively defusing molecules and may not reflect changes in permeability.¹⁵ The volume of the brain is initially small relative to its surface area, and the amount of CSF is proportionally higher. The kinetics of CSF circulation is different in the fetus such that removal of dyes and other markers of permeability, such as inulin, is not the same as in the adult CNS.

Finally, the concentration of various substances in the CNS depends on many different active transport mechanisms, which mature independently. It has been shown that the developing choroid plexus distinguishes between different types of albumin.¹⁶ All these factors produce alterations that make it difficult to assess any changes in the passive permeability characteristics of the CSF-blood-brain barriers.¹⁷ Despite these complexities, experimental models have been designed and are able to answer some of the questions regarding the development of barriers particular to the CNS.

The CNS barriers are indeed more permeable in the fetus, but this greater permeability does not relate to the tightness of the junctions at either the brain capillary endothelium ¹⁸ or the choroid plexus epithelium.^19,²⁰ Indeed, the intracellular junctions at these locations are well formed very early in fetal development and do not differ significantly from those in adults. The degree of permeability relates instead to the size of the intracellular channels, and it is a decrease in the size and perhaps the number of these “pores” that tightens the barrier.^11,²¹

For the barrier to tighten, however, it is necessary for astrocytes to be present. It has been shown that capillaries of CNS origin grown outside the CNS lose their normal barrier properties whereas non-CNS capillaries grown in the CNS acquire the appropriate characteristics.^22,²³ Additional evidence to support the contention of the inductive influence of astrocytes is the loss of normal capillary barrier function in the mature brain at the site of tumors.

For the brain to have its protected environment, there must be a barrier with the equivalent of tight junctions at the arachnoidal membrane comparable to those present in the capillary endothelium and the choroid plexus epithelium. The timing of completion of the arachnoidal barrier in the fetus is unknown, but it may coincide with the development of tight junctions in the blood vessels and choroid plexus (Fig. 188-2).

FIGURE 188-2 Diagrammatic representations of the central nervous system. A, Blood-CSF interface at the choroidal epithelium. B, Blood-brain interface at the capillary endothelium. C, Blood-CSF interface at the arachnoid membrane. D, It is the three interfaces depicted in A, B, and C that provide the specialized environment of the central nervous system.

The protein content of the CSF of fetuses, premature infants, and infants has been studied extensively.²⁴ Total protein levels reach a peak concentration at 20 weeks’ gestation and then fall steadily. For full-term infants younger than 2 months, it is normal to find a protein level of up to 100 mg/mL. Premature infants have an even higher protein level. The concentration of protein in CSF varies with conceptual age but not with birth weight or with postnatal life span, thus indicating that maturation of the barrier, as reflected by a decline in protein, is not influenced by the timing of birth. The protein level represents a steady state at the time of sampling and is dependent on multiple factors, including the CSF secretion rate, volume of distribution, CSF circulation, and absorption rates of macromolecules. The CSF acts as a sink to clear macromolecules. Consequently, the protein level is not dependent solely on the degree of barrier permeability at the time of sampling. The proteins do not appear to emanate from the brain side of the barrier because they have an electrophoretic pattern similar to that of plasma.¹¹

Although cited as an indication of barrier immaturity in the neonatal CNS, focal deposition of bilirubin in the basal ganglia (i.e., kernicterus) does not reflect increased permeability to this substance. The conjugated form of bilirubin is not lipid soluble and does not cross the neonatal CNS to any significant degree. However, the unconjugated lipid-soluble form of bilirubin easily crosses into the CNS, as is also true in adults. That the unconjugated lipid-soluble form does not usually enter the brain is due to the fact that it is bound to plasma proteins. It is only when the binding capacity of the plasma proteins is exceeded that the free, lipid-soluble bilirubin enters the CNS. Overloading of the binding capacity of plasma proteins may result from reduced numbers of available binding sites on the plasma proteins secondary to competition from drugs or lower blood pH. Why bilirubin should selectively affect some regions more than others and why the developing brain is more sensitive to this substance are not known.

Formation and Absorption of Cerebrospinal Fluid

Formation of CSF is dependent on a number of transporters and enzymes, the most important of which appear to be carbonic anhydrase, sodium-potassium adenosine triphosphatase (Na⁺,K⁺-ATPase), and aquaporin-1. Low levels of carbonic anhydrase and Na⁺,K⁺-ATPase are present early in fetal development, but whether these enzymes are functional when they first appear is not known.^21,²⁵ Their presence coupled with aquaporin-1 water channels and concomitant enlargement of the ventricles does, however, lend support to at least some degree of CSF formation fairly early in fetal development.¹⁴

The relationship of CSF formation to brain maturation has been studied in several animal species but not in humans.²⁶ The data indicate that CSF production increases at a rate greater than can be accounted for by the corresponding increase in choroid plexus weight or brain weight. Thus, the increased production of CSF may reflect maturation of the enzyme systems involved. Oversecretion of CSF does occur in the presence of a choroid plexus papilloma. In utero, hydrocephalus secondary to the presence of a choroid plexus papilloma has been demonstrated as well.²⁷

No studies exist regarding CSF absorption in fetal animals or humans. Arachnoid villi, visible only microscopically, and arachnoid granulations, visible with the unaided eye, have long been thought to be the site of CSF absorption and are not found in the fetus. These structures begin to be present at birth and increase in size and number with age, with villi becoming arachnoid granulations. If they are not present, how is CSF absorbed? Although no fetal studies have been performed, neonatal animal studies confirm CSF absorption via the extracranial lymphatics, with drainage into the venous sinuses through the arachnoidal structures being secondary.²⁸

Hydrocephalus

Normal dilation of the ventricular system is needed for the cells of the germinal matrix to multiply and migrate to form the normal cortical architecture. Various factors are found to be altered with hydrocephalus, but it is difficult to determine cause and effect.^29,³⁰ One study has shown that CSF taken from hydrocephalic animals in and of itself can inhibit neurogenesis but does not alter cell migration from the germinal matrix.³¹ Experimental models of hydrocephalus are discussed in Chapter 189.

Currently, the only genetic defect directly linked to fetal hydrocephalus is the X-linked L1-NCAM mutation; however, the CNS is severely altered as well.^32,³³ The ventricular dilation seen with extensive structural CNS abnormalities may have other different underlying genetic causations.

Simpson and colleagues attempted to delineate the dynamics of fetal intracranial pressure in utero at the time of therapeutic abortion.³⁴ Although their studies suffered from having only seven patients with diverse CNS malformations, they could not correlate the type of CNS lesions with the intracranial pressure found in fetuses with an excessive amount of CSF. In hydrocephalic fetuses, an attempt has been made to correlate ventricular size with velocity waveforms of pulsed Doppler recordings of cerebral blood flow, but no correlation has been found.^35,³⁶

In utero imaging studies now make it possible to detect developmental abnormalities, mass lesions, and evidence of infection and hemorrhage, any of which can result in hydrocephalus. The antenatal diagnosis of hydrocephalus may influence the timing of delivery, the mode of delivery, and the possibility of terminating the pregnancy. In a situation that is fairly analogous to in utero hydrocephalus, a retrospective analysis was undertaken to assess the efficacy of aggressive surgical management of progressive hydrocephalus in preterm neonates with intracranial hemorrhage. The overwhelming factor in determining the outcome of this patient group was the extent of intracranial hemorrhage and parenchymal damage. The degree of hydrocephalus and aggressive treatment of it were not significant.³⁷ There is no evidence that early control of hydrocephalus significantly improves neurological function because functional outcome is determined by the underlying insult to the CNS rather than the hydrocephalus.

Blood-Brain Barrier

In 1885, Ehrlich demonstrated that many dyes injected into the systemic circulation of laboratory animals stain virtually all the organs in the body except the brain and spinal cord.³⁸ Ehrlich’s disciple Goldman continued these experiments and showed that intravenously administrated trypan blue fails to stain the CNS and CSF,^39,⁴⁰ although the choroid plexuses and meninges were stained. In his second paper, he demonstrated that interventricularly injected trypan blue rapidly stains brain parenchyma. Thus, it was concluded that there is a barrier between blood and the brain and that this barrier could be circumvented by direct injection of dye into CSF (Fig. 188-3).

FIGURE 188-3 Diagrammatic representation of Goldman’s experiments in which trypan blue was injected either into the bloodstream (A) or into CSF (B). The fact that dye injected into the bloodstream did not enter CSF or the brain whereas dye injected into CSF fully penetrated the brain gave rise to the first concept of a blood-brain barrier. In view of the extensive metabolic activity that occurs at the endothelial cell level between blood and the brain, this region is better described as an interface rather than simply as a barrier.

Electron microscopic cytochemical studies from the late 1960s duplicated Goldman’s first and second experiments.^41,⁴² Horseradish peroxidase (HRP) injected intravenously was found in the lumen of brain microvessels, but no further movement of the label beyond the endothelial membrane was observed. When HRP was injected interventricularly, it readily infused across the ependyma and along the basement membranes of capillaries, but it did not enter blood via the endothelial membrane. These experiments established the anatomic concept of the blood-brain barrier and suggested that the endothelial membrane restricts free exchange of substances between blood and the brain because of the presence of tight junctions in the cerebral capillary endothelium.

During the past several decades, accumulating information has challenged the anatomic concept of an impermeable blood-brain barrier by showing that what is true for some vital dyes and HRP does not hold for many other biologically important molecules ^43,⁴⁴ or for cells of the immune system. In addition, several classes of metabolic substrates, regulatory peptides, transport plasma proteins, steroid hormones, ions, and various groups of centrally active pharmacotherapeutics are able to use specialized shuttle services at the blood-brain barrier. In fact, the concept of a blood-brain barrier should be replaced with that of a blood-brain interface because the endothelial layer in reality regulates homeostasis of the neural milieu by numerous highly specific transport, enzymatic, receptor, and cell-mediated mechanisms rather than simply impeding exchange of solutes between blood and the brain.

Formation of Cerebrospinal Fluid

Formation Sites

It is generally agreed that most CSF is formed within the ventricular system. Possible sites of origin include the choroid plexus, the ependyma, and the parenchyma. A method has not been developed to separate the function of the ependyma from the remainder of the parenchyma, so the role of the ependyma in bulk CSF formation is not known, although from a morphologic standpoint its contribution is most likely to be insignificant. However, the choroidal epithelium has histologic features characteristic of epithelia specialized for transcellular transport of solutes and solvents.^45,⁴⁶ The discussion that follows is limited solely to the bulk secretion of CSF.

Results from isolated choroid plexus preparations would indicate that 80% or more of CSF production is derived from this source alone.^26,^47,⁴⁸ However, perfusion of a portion of the ventricular system devoid of choroid plexus has demonstrated that 30% to 60% of CSF is produced from nonchoroidal sources,^49,⁵⁰ which may explain the failure of choroid plexectomy in the clinical setting to control progressive hydrocephalus.⁴⁹ It may be added that this operative procedure removes the choroid plexus only from the lateral ventricles and not from the third and fourth ventricles. The contributions of the remaining intact choroid plexus to the formation of CSF is not clear, and whether it can compensate for the portion of the choroid plexus removed is not known.

The various lines of evidence showing the ECS to be approximately 15% of the brain’s volume has been summarized by Welch.²⁶ The established presence of substantial ECS, the lack of ependymal resistance to free exchange between fluid in the ECS and CSF, and the similar composition of ECS fluid and CSF have a direct bearing on the possibility that the parenchyma may be the main source of nonchoroidal CSF formation.^26,^41,^51,⁵² In summary, it appears that normally roughly 80% of CSF secretion is derived from the choroid plexus, with the remaining portion probably originating from the parenchyma. The obvious candidate for the parenchymal source is the capillary endothelium because its high content of mitochondria could provide the metabolic energy required for such a function.⁵³