H. Reiber
Serum Proteins in CSF
Factors Affecting the Normal Blood CSF–Barrier Function
The concentration of a serum protein in CSF depends on its concentration in the serum: the higher the serum concentration, the higher the CSF concentration. The molecular size of a protein determines the relationship between its concentrations in CSF and serum: the larger the protein molecule, the more slowly it passes through the barriers and the steeper the concentration gradient between blood and CSF. This is demonstrated in Table 3.1 for normal conditions with an albumin quotient of 5 × 10−3. When the serum/CSF gradient for albumin is 200:1, the ratio observed for the much larger IgM molecule is 3300:1. The diagram in Fig. 2.5 illustrates how the ratio between the CSF/serum quotients of serum proteins and the albumin quotient (e. g., QIgG: QAlb) follows a theoretically based hyperbolic function. Age-related variations in the CSF flow rate and length of flow path make it necessary to determine an age-related reference range of the normal albumin quotient (normal barrier function).
Blood–CSF Barrier Dysfunction: Leakage vs. Reduced CSF Flow Rate
Many neurological diseases are accompanied by an increased protein concentration in CSF. In the past, this increased protein concentration was mistakenly thought to be caused by a change in morphological structure (leakage model). Today, this increasing concentration of serumderived proteins in CSF is explained quantitatively by a reduced CSF flow rate (Reiber, 1994 a and b; Reiber, 2003).
The leakage model was based on the assumption that the dysfunction of the blood–CSF barrier is due to a change in morphological structure—in other words, to holes (leaks) in the barrier. The model was also supposed to explain a seemingly reduced selectivity dependent on molecular size during protein transfer between blood and CSF. However, careful analysis of the empirical CSF protein data already contradicts this assumption of a morphological change. The following arguments, in particular, are immediately apparent:
• Unchanged selectivity: Even in the most severe blood–CSF barrier dysfunction, selectivity for molecules of the various sizes remains unchanged (Fig. 3.1 a; Table 3.2).
• Dynamics of proteins in blood–CSF barrier dysfunction: The data obtained from CSF of a patient with bacterial meningitis (who underwent lumbar puncture on days 1 and 2 after the first clinical symptoms) contradict the leakage model (Table 3.3). At the time of the first puncture, the albumin concentration was 47% of the value on day 2, as compared to 21% for IgG, 12% for IgA, and only 5% for IgM. According to the leakage model (with concomitant loss of selectivity), the relative increase between normal values and those obtained on day 1 (Table 3.3) should as a matter of course be larger for a larger molecule (IgM), with the steeper blood/CSF gradient, than for a smaller molecule (albumin). Clearly, however, this is not the case: the empirical changes in the quotient diagrams (Fig. 3.1 a) do not agree with the hypothesis of a leakage model, the consequences of which are presented in Fig. 3.1 b. The dynamics of the process in Fig. 3.1 a are explained quantitatively by the biophysical model (Reiber, 1994a and b; Reiber, 2003).
• Blood–brain barrier in newborns: The high total protein concentration in CSF of newborns is caused not by an immature barrier (as animal studies have shown, the barrier is already formed during the early fetal phase), but by the fact that the CSF flow starts only around time of birth with the maturation of the arachnoid villi (see Chap. 1). In CSF of the newborn, the ratio of molecules of different sizes (QIgG: QAlb) already follows the same hyperbolic function as in adults (see Fig. 1.2).
Fig. 3.1 a, b Changes in the CSF concentration of immunoglobulins (QIg) as a function of increasing albumin quotients (QAlb) (blood–CSF barrier dysfunction).
a Mean values of the empirical data for the immunoglobulin CSF/serum quotients with increasing albumin quotients, which may be described as hyperbolic functions (Reiber, 1994 a).
b Experimental simulation of a leakage model in which serum proteins pass in bulk flow from the tissue into the CSF. The immunoglobulin CSF/serum quotients follow a linear function of the increasing albumin quotient. In this in-vitro experiment a patient's serum was added stepwise to the patient's CSF sample, and the resulting concentrations of IgG, IgA, IgM, and albumin were measured. Calculation of the CSF/serum quotients for each protein reveals a linear increase, in each case parallel to the 45° line. The empirical data from a large group (a), which show that the molecular-size-dependent discrimination (selectivity) is maintained even in case of the most severe barrier dysfunction, are confirmed by the extreme data from individual patients with different diseases given in Table 3.2. The leakage model (b) cannot explain the reality seen in patients, which is explained quantitatively by the diffusion/flow model.
• Early detection of serum proteins in lumbar CSF: Serum proteins can be detected earlier in lumbar CSF than in ventricular CSF (for references see Reiber, 1994 a). This is explained in the new biophysical model by the steeper local concentration gradient in the spinal subarachnoid space (Reiber, 2003).
• CSF-flow-dependent dynamics of brain proteins: The connection between barrier dysfunction and protein concentration in the CSF is valid for the passage of serum proteins from serum into CSF, but not for the passage of brain proteins into CSF. Whereas the leakage model gives no explanation of the dynamics of brain proteins, the diffusion/flow model provides a sufficient explanation of the dynamics of brain proteins of various origins. Thus, the diffusion/flow model is a more general theory explaining the dynamics of all proteins, both blood- and brain-derived.
The following physiological observations in the context of actual neurological diseases confirm our understanding of the barrier function as changes in the CSF flow rate:
• Leukemia of the CNS: This disease is primarily associated with changes in the trabeculae of the arachnoidea mater. Histopathological studies suggest a reduced CSF flow rate.
• Purulent bacterial meningitis: This disease is associated with increased CSF viscosity and meningeal adhesions. Post-mortem studies reveal protein complexes and cellular depositions in the arachnoid villi. All these features impede CSF drainage.
• Guillain–Barré syndrome: The high protein concentrations associated with this disease, too, are related to reduced CSF turnover caused by diminished outflow into the veins accompanying the spinal nerve roots, due to swellings in the area around the spinal roots.
• Complete spinal block: In the case of spinal stenosis or complete spinal block, high protein levels are measured in the lumbar CSF caudal to the blockade, despite normal cisternal and ventricular CSF levels. In contrast to bloodderived proteins, proteins originating from the brain, such as transthyretin (formerly called prealbumin), decrease relative to albumin caudal to the blockade. Here, too, the molecular-size-dependent discrimination (selectivity) for protein transfer between blood and CSF is undisturbed.
Brain Proteins in CSF
Site of synthesis of brain proteins. About 20% of the proteins in the CSF originate from the CNS (Thompson, 2005). Of these, only a few are brain-specific (i. e., synthesized exclusively in the CNS). Diagnostically relevant brain proteins come from three main sources (Reiber, 2001; Reiber, 2003):
• Proteins synthesized in the plexus epithelium (Aldred et al., 1995):
– Transthyretin (formerly called prealbumin)
– Transferrin (asialo form)
– Cystatin C
• Proteins synthesized in brain cells:
– Neuron-specific enolase (NSE) (γ-homodimer of enolase, derived from neurons)
– S-100B (β-homodimer derived from glial cells)
– Tau protein (axonal microtubules of neurons)
• Leptomeningeal proteins:
– β-Trace protein (prostaglandin D synthetase activity)
– Cystatin C
Concentration of brain proteins. Brain-derived proteins in CSF shown in Table 3.4 are characterized by the following:
• A higher concentration in the CSF than in the serum or, at least, a brain-derived fraction in the CSF that contributes more than 90%.
• A concentration that decreases from the ventricular to the lumbar CSF for proteins that originate from brain cells and choroid plexus and diffuse into the ventricular and cisternal CSF, but increases for leptomeningeal proteins (Reiber, 2001; Reiber, 2003).
• No or linear changes in cases of “barrier dysfunction” due to reduced CSF flow rate; i. e., with increasing QAlb in the lumbar CSF (Fig. 3.2), the CSF concentrations of brain cell proteins remain constant, while those of leptomeningeal proteins show a linear increase (Reiber, 2001; Reiber, 2003).
Dynamics of brain proteins. The biophysical model of the blood–CSF barrier function for blood-derived proteins allows us for the first time to provide a theoretically sound explanation of the dynamics of brain-derived proteins in the CSF (Fig. 3.2). Here, too, the change in CSF flow is sufficient to explain quantitatively the dynamics of proteins derived from glial cells and neurons, and from leptomeninges (Reiber, 2001; Reiber, 2003). The direct influence of CSF flow rate is shown by an example in which the physiologically increasing CSF flow rate induces an increasing albumin concentration: during the first few months of life, the concentration of the leptomeningeal protein cystatin C decreases in parallel with the concentration of albumin (Fig. 1.2). The positive feedback that is typical of serum proteins passing from blood into CSF does not exist for brain proteins, released from brain cells into extracellular fluid and CSF; the dynamics of these proteins therefore remain linear with changing CSF flow rates (leptomeningeal proteins, such as β-trace protein) (Reiber, 2001). Proteins derived from brain cells (e. g., tau protein, NSE) show a rostrocaudal drop in concentration (Table 3.4) and are constant in their lumbar CSF concentrations, i. e., they are independent of the CSF flow rate (Fig. 3.2)—as the theory predicts.
Fig. 3.2 Changes in the mean CSF concentrations of brain proteins (P) at increasing albumin quotients (QAlb) (“blood–CSF barrier dysfunction” due to decreasing CSF flow rate). Proteins getting into the CSF from the choroid plexus or from regions near the ventricles (tau protein, S-100, NSE) remain constant in their lumbar CSF concentrations, irrespective of CSF flow rate (compensation of molecular diffusion in and out of CSF). Proteins with pronounced leptomeningeal release (β-trace protein, cystatin C) show increasing CSF concentrations with decreasing CSF flow rate. In contrast to the hyperbolic function for serum-derived proteins (Fig. 2.5), this increase is linear because the primary increase of protein concentration in CSF in response to reduced CSF turnover does not exercise positive feedback on the rate of intracellular production or release of the proteins from the brain cells (Reiber, 2003).
Brain Proteins in the Blood
Because of their diagnostic relevance as marker proteins for cerebral dysfunctions that are measurable in blood (Schaarschmidt et al., 1994), some brain proteins are also of theoretical interest for understanding the dynamics of brain-derived proteins in blood (Reiber, 2003). Brain proteins enter the blood in two ways: by drainage of CSF into venous blood, and by diffusion through the blood–brain barrier (i. e., in the reverse direction, from brain to blood).
CSF-mediated passage. The β-trace protein is an example of CSF-mediated passage from brain to blood. The β-trace protein concentration (Table 3.4) is about 34 times higher in lumbar CSF than in blood. Based on the volume of CSF produced daily (500 mL) and the total volume of the blood (4–5 L), it has been calculated that the blood concentration is accounted for by CSF drainage (Reiber, 2003). This pathway of CSF-mediated passage into the blood is supported by an earlier report that the mean vitamin C concentration in blood decreases with decreasing CSF flow rate (barrier dysfunction), since the vitamin C concentration in CSF is 12 times higher than in blood (Reiber et al., 1993).
Diffusion from brain to blood. The second pathway by which brain-derived proteins can enter the blood is exemplified by the neuron-specific enolase (NSE). Even in the case of massive hypoxia with high NSE values in the CSF (Jacobi and Reiber, 1988), the NSE concentrations measured in the blood are too great to be explained by CSF-mediated passage. In fact, the increased release of NSE from brain cells leads to an increased concentration in the extracellular fluid and this creates a brain–blood concentration gradient, thus leading to diffusion of NSE from the brain directly into the blood vessels. For the rapid increase in NSE concentration observed in blood in, e. g., hypoxia (Schaarschmidt et al., 1994), no morphological blood–brain barrier dysfunction is required—again, diffusion is a sufficient explanation (Reiber, 2003).
For proteins moving only by diffusion from one compartment into another, the blood–brain barrier has by definition the same permeability in both directions (blood–brain and brain–blood), since—by definition—the direction of net diffusion depends only on the local concentration gradient of the protein.
References
Aldred AR, Brack CM, Schreiber G. The cerebral expression of the plasma protein genes in different species. Comp Biochem Physiol B Biochem Mol Biol 1995;111:1–15
Felgenhauer K. Protein size and cerebrospinal fluid composition. Klin Wochenschrift 1974;52:1158–1164
Jacobi C, Reiber H. Clinical relevance of increased neuron-specific enolase concentration in cerebrospinal fluid. Clin Chim Acta 1988;177:49–54
Reiber H. Flow rate of cerebrospinal fluid (CSF): a concept common to normal blood–CSF barrier function and to dysfunction in neurological diseases. J Neurol Sci 1994 a;122:189–203
Reiber H. The hyperbolic function: a mathematical solution of the protein flux/CSF flow model for blood–CSF barrier function. J Neurol Sci 1994 b;126:240–242
Reiber H. Dynamics of brain-derived proteins in cerebrospinal fluid. Clin Chim Acta 2001;310:173–186
Reiber H. Proteins in cerebrospinal fluid and blood: barriers, CSF flow rate and source-related dynamics. Restor Neurol Neurosci 2003;21:79–96
Reiber H, Felgenhauer K. Protein transfer at the blood–CSF barrier and the quantitation of the humoral immune response within the central nervous system. Clin Chim Acta 1987;163: 319–328
Reiber H, Peter JB. Cerebrospinal fluid analysis: disease-related data patterns and evaluation programs. J Neurol Sci 2001;184:101–122
Reiber H, Ruff M, Uhr M. Ascorbite concentration in human CSF and serum. Intrathecal accumulation and CSF flow rate. Clin Chim Acta 1993;217:163–173
Reiber H, Walther K, Althaus H. Beta-trace protein as sensitive marker for CSF rhinorhea and CSF otorhea. Acta Neurol Scand 2003;108:359–362
Schaarschmidt H, Prange H, Reiber H. Neuron-specific enolase concentrations in blood as a prognostic parameter in cerebrovascular diseases. Stroke 1994;24:558–565
Thomas L (ed). Labor und Diagnose. 5th ed. Frankfurt: TH-Books; 1998
Thompson, EJ. The CSF proteins: a biochemical approach. 2nd ed. Amsterdam: Elsevier; 2005