Mechanisms of Sodium Transport at the Blood-Brain Barrier Studied with In Situ Perfusion of Rat Brain

Abstract: The mechanism of unidirectional transport of sodium from blood to brain in pentobarbital-anesthetized rats was examined using in situ perfusion. Sodium transport followed Michaelis-Menten saturation kinetics with a V _max of 50.1 nmol/g/min and a K _m of 17.7 m M in the left frontal cortex. The kinetic analysis indicated that, at a physiologic sodium concentration, ∼26% of sodium transport at the blood-brain barrier (BBB) was carrier mediated. Dimethylamiloride (25 µ M ), an inhibitor of Na⁺/H⁺ exchange, reduced sodium transport by 28%, whereas phenamil (25 µ M ), a sodium channel inhibitor, reduced the transfer constant for sodium by 22%. Bumetanide (250 µ M ) and hydrochlorothiazide (1.5 m M ), inhibitors of Na⁺-K⁺-2Cl⁻/NaCl symport, were ineffective in reducing blood to brain sodium transport. Acetazolamide (0.25 m M ), an inhibitor of carbonic anhydrase, did not change sodium transport at the BBB. Finally, a perfusate pH of 7.0 or 7.8 or a perfusate P co ₂ of 86 mm Hg failed to change sodium transport. These results indicate that 50% of transcellular transport of sodium from blood to brain occurs through Na⁺/H⁺ exchange and a sodium channel in the luminal membrane of the BBB. We propose that the sodium transport systems at the luminal membrane of the BBB, in conjunction with Cl⁻/HCO₃⁻ exchange, lead to net NaCl secretion and obligate water transport into the brain.     

Kinetics of Neutral Amino Acid Transport Across the Blood-Brain Barrier

Neutral amino acid (NAA) transport across the blood-brain barrier was examined in pentobarbital-anesthetized rats with an in situ brain perfusion technique. Fourteen of 16 plasma NAAs showed measurable affinity for the cerebrovascular NAA transport system. Values of the transport constants (Vmax, Km, KD) were determined for seven large NAAs from saturation studies, whereas Km values for five small NAAs were estimated from inhibition studies. These data, together with our previous work, provide a complete set of constants for prediction of NAA influx from plasma. Among the NAAs, Vmax varied at least fivefold and Km varied approximately 700 fold. The apparent affinity (1/Km) of each NAA was related linearly (r = 0.910) to the octanol/water partition coefficient, a measure of NAA side-chain hydrophobicity. Predicted influx values from transport constants and average plasma concentrations agree well with values measured using plasma perfusate. These results provide accurate new estimates of the kinetic constants that determine NAA transport across the blood-brain barrier. Furthermore, they suggest that affinity of a L-alpha-amino acid for the transport system is determined primarily by side-chain hydrophobicity.     

Blood-Brain Barrier Transport of 1-Aminocyclohexanecarboxylic Acid, a Nonmetabolizable Amino Acid for In Vivo Studies of Brain Transport

Regional transport of 1-aminocyclohexanecarboxylic acid (ACHC), a nonmetabolizable amino acid, across the blood-brain barrier was studied in pentobarbital-anesthetized rats using an in situ brain perfusion technique. The concentration dependence of influx was best described by a model with a saturable and a nonsaturable component. Best-fit values for the kinetic constants of the frontal cortex equaled 9.7 X 10(-4) mumol/s/g for Vmax, 0.054 mumol/ml for Km, and 1.0 X 10(-4) ml/s/g for KD in the absence of competing amino acids. Saturable influx could be reduced by greater than 85% by either L-phenylalanine or 2-aminobicyclo[2.2.1]heptane-2-carboxylic acid, consistent with transport by the cerebrovascular neutral amino acid transport system. The transport Km for ACHC was one-fifth that for the more commonly used homologue, 1-aminocyclopentanecarboxylic acid, and was similar to values for several natural amino acids, such as L-methionine, L-isoleucine, and L-tyrosine. The results indicate that ACHC may be a useful probe for in vivo studies of amino acid transport into brain.     

Some Characteristics of Threonine Transport Across the Blood-Brain Barrier of the Rat

Threonine entry into brain is altered by diet-induced changes in concentrations of plasma amino acids, especially the small neutrals. To study this finding further, we compared effects of various amino acids (large and small neutrals, analogues, and transport models) on transport of threonine and phenylalanine across the blood-brain barrier. Threonine transport was saturable and was usually depressed more by natural large than small neutrals. Norvaline and 2-amino-n-butyrate (AABA) were stronger competitors than norleucine. 2-Aminobicyclo[2.2.1]heptane-2-carboxylate (BCH), a model in other preparations for the large neutral (L) system, and cysteine, a proposed model for the ASC system only in certain preparations, reduced threonine transport; 2-(methylamino)isobutyrate (MeAIB; a model for the A system for small neutrals) did not. Phenylalanine transport was most depressed by cold phenylalanine and other large neutrals; threonine and other small neutrals had little effect. Norleucine, but not AABA, was a strong competitor; BCH was more competitive than cysteine or MeAIB. Absence of sodium did not affect phenylalanine transport, but decreased threonine uptake by 25% (p less than 0.001). Our results with natural, analogue, and model amino acids, and especially with sodium, suggest that threonine, but not phenylalanine, may enter the brain partly by the sodium-dependent ASC system.     

Isolated Brain Microvessels as In Vitro Equivalents of the Blood-Brain Barrier: Selective Removal by Collagenase of the A-System of Neutral Amino Acid Transport

On treatment with collagenase, brain microvessels, together with several protein components, lose some enzymatic activities such as alkaline phosphatase and gamma-glutamyltranspeptidase, whereas no change occurs in the activities of 5'-nucleotidase and glutamine synthetase. The energy-requiring "A-system" of polar neutral amino acid transport is also severely inactivated, whereas the L-system for the facilitated exchange of branched chain and aromatic amino acids is preserved. In the collagenase-digested microvessels, this leads to loss of the transtimulation effect of glutamine on the transport of large neutral amino acids, because such transtimulation is due to a cooperation between the A- and L-systems. By contrast, NH4+ maintains (and even enhances) its ability to stimulate the L-system of amino acid transport, presumably through glutamine synthesis within the endothelial cells.     

Rapid Brain Uptake of Manganese(II) Across the Blood-Brain Barrier

Abstract: ⁵⁴Mn²⁺ uptake into brain and choroid plexus from the circulation was studied using the in situ rat brain perfusion technique. Initial uptake from blood was linear with time (30 s to 6 min) and extrapolated to zero with an average transfer coefficient of ∼6 × 10^-5 ml/s/g for brain and ∼7 × 10^-3 ml/s/g for choroid plexus. Influx from physiologic saline was three- to fourfold more rapid and exceeded that predicted for passive diffusion by more than one order of magnitude. The lower uptake rate from blood could be explained by plasma protein binding as the free fraction of ⁵⁴Mn²⁺ in rat plasma was ≤30%. Purified albumin, transferrin, and α₂-macroglobulin were each found to bind ⁵⁴Mn²⁺ significantly and to restrict brain ⁵⁴Mn²⁺ influx. The results demonstrate that ⁵⁴Mn²⁺ is readily taken up into the CNS, most likely as the free ion, and that transport is critically affected by plasma protein binding. The results support the hypothesis that Mn²⁺ transport across the blood-brain barrier is facilitated by either an active or a passive mechanism.     

Transport of Leucine-Enkephalin Across the Blood-Brain Barrier in the Perfused Guinea Pig Brain

Transport of [tyrosyl-3,5-3H]enkephalin-(5-L-leucine) [( 3H]Leu-enkephalin) across the blood-brain barrier was studied in the adult guinea pig, by means of vascular perfusion of the head in vivo. The unidirectional transfer constant (Kin) estimated from the multiple-time uptake data for [3H]Leu-enkephalin ranged from 3.62 X 10(-3) to 3.63 X 10(-3) ml min-1 g-1 in the parietal cortex, caudate nucleus, and hippocampus. Transport of [3H]Leu-enkephalin was not inhibited by unlabelled L-tyrosine (the N-terminal amino acid) at a concentration as high as 5 mM, or by the inhibitor of aminopeptidase activity bacitracin (2 mM), suggesting that there was no enzymatic degradation of peptide at the blood-brain barrier. By contrast, 2 mM unlabelled Leu-enkephalin strongly inhibited the unidirectional blood-to-brain transport of [3H]Leu-enkephalin by 74-78% in the parietal cortex, caudate nucleus, and hippocampus. The tetrapeptide tyrosyl-glycyl-glycyl-phenylalanine (without the C-terminal leucine of Leu-enkephalin), at a concentration of 5 mM, caused a moderate inhibition ranging from 15 to 29% in the brain regions studied, whereas the tetrapeptide glycyl-glycyl-phenylalanyl-leucine (without the N-terminal tyrosine) at 5 mM was without effect on Leu-enkephalin transport. Unidirectional brain uptake of Leu-enkephalin was not altered in the presence of naloxone at a concentration as high as 3 mM (1 mg/ml), suggesting that there is no binding of Leu-enkephalin to opioid receptors at the blood-brain barrier. It is concluded that there is a specific transport mechanism for Leu-enkephalin at the blood-brain barrier in the guinea pig.(ABSTRACT TRUNCATED AT 250 WORDS)     

Kinetics of Neutral Amino Acid Transport Through the Blood-Brain Barrier of the Newborn Rabbit

Abstract: Since protein synthesis in the developing brain may, under certain conditions, be limited by amino acid availability, the present studies were undertaken to characterize the kinetics of large neutral amino acid transport through the blood-brain barrier (BBB) of the newborn rabbit. The K_m, V_max, and K_D of the transport of eight amino acids were determined by a nonlinear regression analysis of data obtained with the carotid injection technique. Compared with kinetic parameters observed for the adult rat, the K_m, V_max, and K_D of amino acid transport were all two- to threefold higher in the newborn. Albumin was found to bind tryptophan actively in vitro , but had no inhibitory effect on tryptophan transport through the newborn BBB. Glutamine was transported through the BBB of the newborn at rates severalfold higher than are seen in the adult rat. However, glutamine transport was not inhibited by high concentrations of N -methylaminoisobutyric acid (NMAIB), a model amino acid that is specific for the alanine-preferring or A-system present in peripheral tissues. In conclusion, these studies show that the BBB neutral amino acid transport system of the newborn rabbit has a lower affinity and higher capacity than does the BBB of the adult rat. Under conditions of high plasma amino acids, the increased capacity of the newborn transport system allows for a higher rate of amino acid transport into brain than would occur via the lower capacity system present in the adult rat brain.     

Facilitated Transport of the Neurotoxin, β-N-Methylamino-l-Alanine, Across the Blood-Brain Barrier

beta-N-Methylamino-L-alanine (BMAA) is a neurotoxic plant amino acid that has been implicated in the pathogenesis of the high incidence amyotrophic lateral sclerosis and related parkinsonism dementia of the western Pacific. Previous studies have demonstrated that BMAA is taken up into brain following intravenous or oral administration. To examine the kinetics and mechanism of brain transfer, BMAA influx across the blood-brain barrier was measured in rats using an in situ brain perfusion technique. BMAA influx was found to be saturable with a maximal transfer rate (Vmax) of 1.6 +/- 0.3 x 10(-3) mumol/s/g and a half-saturation constant (Km) of 2.9 +/- 0.7 mM based on total perfusate BMAA concentration. Uptake was sodium independent and inhibitable by excess L-leucine, but not by L-lysine, L-glutamate, or methylaminoisobutyric acid, indicative of transfer by the cerebrovascular large neutral amino acid carrier. L-BMAA competitively reduced brain influx of L-[14C]leucine, as expected for cross-inhibition. The results demonstrate that BMAA is taken up into brain by the large neutral amino acid carrier of the blood-brain barrier and suggest that uptake may be sensitive to the same factors that affect neutral amino acid transport, such as diet, metabolism, disease, and age.     

Slow Penetration of Thyrotropin-Releasing Hormone Across the Blood-Brain Barrier of an In Situ Perfused Guinea Pig Brain

Transport of 3H-labelled thyrotropin-releasing hormone (TRH) across the blood-brain barrier was studied in the ipsilateral perfused in situ guinea pig forebrain. The unidirectional transfer constant (Kin) calculated from the multiple time brain uptake analysis ranged from 1.14 X 10(-3) to 1.22 X 10(-3) ml min-1 g-1, in the parietal cortex, caudate nucleus, and hippocampus. Regional Kin values for [3H]TRH were significantly reduced by 43-48% in the presence of an aminopeptidase and amidase inhibitor, 2 mM bacitracin, suggesting an enzymatic degradation of tripeptide during interaction with the blood-brain barrier. In the presence of unlabelled 1 mM TRH and 2 mM bacitracin together, a reduction of [3H]TRH regional Kin values similar to that obtained with 2 mM bacitracin alone was obtained . L-Prolinamide, the N-terminal residue of tripeptide, at a 10 mM level had no effect on the kinetics of entry of [3H]TRH into the brain. The data indicate an absence of a specific saturable transport mechanism for TRH presented to the luminal side of the blood-brain barrier. It is concluded that intact TRH molecule may slowly penetrate the blood-brain barrier, the rate of transfer being some three times higher than that of D-mannitol.     

Carrier-Mediated Transport of Chloride Across the Blood-Brain Barrier

36Cl concentrations in each of eight brain regions and in cisternal cerebrospinal fluid (CSF) were determined 30 min after the intravenous injection of 36Cl in dialyzed-nephrectomized rats with plasma Cl concentrations between 14 and 120 mumol X ml-1. CSF 36Cl exceeded 36Cl concentrations in brain extracellular fluid. The calculated blood-to-brain transfer constants for Cl, kCl, ranged from 1.8 X 10(-5) S-1 at the parietal cortex to 3.8 X 10(-5) S-1 at the thalamus-hypothalamus. kCl fell by 42-62% when mean plasma [Cl] was elevated from 16 to 114 mumol X ml-1. Brain uptake of [14C]mannitol or of 22Na was independent of plasma [Cl], but 22Na influx into CSF fell when plasma [Cl] was reduced. Cl flux into brain and CSF could be represented by Michaelis-Menten saturation kinetics, where, for the parietal cortex, Km = 43 mumol X ml-1 and Vmax = 2.5 X 10(-3) mumol X S-1 X g-1, and for CSF Km = 68 mumol X ml-1. At least 80% of 36Cl influx into the parietal cortex was calculated to occur at the cerebrovascular endothelium, whereas the remainder was derived from tracer that first entered CSF. The CSF contribution was greater at brain regions adjacent to cerebral ventricles. The results show that Cl transport at the cerebrovascular endothelium as well as at the choroid plexus epithelium is a saturable concentration-dependent process, and that the CSF is a significant intermediate pathway for Cl passage from blood to brain.     

Regional Blood-Brain Barrier Permeability to Amino Acids After Portacaval Anastomosis

Abstract: The influx of phenylalanine, tryptophan, leucine, and lysine across the blood-brain barrier of individual brain structures was studied in rats 7–8 weeks after a portacaval shunt or sham operation. The method involved a brief infusion of labeled amino acid in tracer quantity and quantitative autoradiography. The clearance rates of phenylalanine, tryptophan, and leucine were increased in proportion to each other in every region examined, but not by the same factor. Tryptophan clearance increased the most (about 200%) and leucine the least (about 30%), compared with phenylalanine (about 80%). This was unexpected, as all three amino acids are believed to be transported by the same mechanism. The changes were most marked in several limbic structures and the reticular formation, whereas the hypothalamus was least affected. Plasma clearance of lysine was decreased in all areas by about 70%. Since the circulating lysine concentration was decreased by 13%, the actual rate of lysine influx was even more reduced. The results demonstrate specific alterations in two different amino acid transport systems. The resulting excess brain neutral amino acids, some of which are neurotransmitter precursors, as well as reduced basic amino acid availability, may be of etiological significance in hepatic encephalopathy.     

Transport of Lead-203 at the Blood-Brain Barrier During Short Cerebrovascular Perfusion with Saline in the Rat

Lead transport at the blood-brain barrier has been studied by short (less than 1.5 min) vascular perfusion of one cerebral hemisphere of the rat with a buffered physiological salt solution at pH 7.4 without calcium, magnesium, or bicarbonate and containing 203 Pb-labelled lead chloride. In the absence of complexing agents, 203Pb uptake was rapid, giving a space of 9.7 ml/100 g of wet frontal cortex at 1 min. Lead-203 influx was linear with lead concentration up to 4 microM. Five percent albumin, 200 microM cysteine, or 1 mM EDTA almost abolished 203Pb uptake. Lead-203 entry into brain was uninfluenced by varying the calcium concentration or by magnesium or the calcium blocker methoxyverapamil. Similarly, 1 mM bicarbonate or 50 microM 4,4'-diisothiocyanostilbene-2,2'-disulphonic acid was without effect. Increasing the potassium concentration reduced 203Pb uptake. Vanadate at 2 mM, 2 microM carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (a metabolic uncoupler), or 2 microM stannic chloride all markedly enhanced lead entry into brain, as did a more alkaline pH (7.80). In conclusion, there is a mechanism allowing rapid passive transport of 203Pb at the brain endothelium, perhaps as PbOH+. Lead uptake into brain via this system is probably made less important by active transport of lead back into the capillary lumen by the calcium-ATP-dependent pump.     

Fatty Acid Transport Through the Blood-Brain Barrier

Across the cerebral capillaries, the anatomical locus of the blood-brain barrier, the unidirectional influxes of the saturated fatty acids, octanoic and myristic acids, and the unsaturated essential fatty acid, linoleic acid, were measured. Employing an in situ rat brain perfusion technique that allows control of perfusate composition and accurate measurement of perfusate-to-brain fatty acid transport, we found that both [¹⁴C]octanoic and [¹⁴C]myristic acids were transported through the blood-brain barrier in vivo, in large part, by a specific, probenecid-sensitive transport system. However, the transport of [¹⁴C]linoleic acid was not probenecid sensitive. With 0.5 μM fatty acid but no plasma proteins in the perfusate, the permeability-surface area constant was higher for myristic acid (4.8 × 10^--²× s^-1) than for octanoic and linoleic acids (1.5 and 1.2 × 10^--²× s^-1, respectively). Approximately 70, 30, and 25% of the [¹⁴C]myristic, [¹⁴C]octanoic, or [¹⁴C]linoleic acids, respectively, were extracted from the perfusate.     

Selective Alteration in Blood-Brain Barrier and Insulin Transport in Iron-Deficient Rats

Nutritional iron deficiency induced in rats causes a significant reduction in level of brain nonheme iron and is accompanied by selective reduction of dopamine D2 receptor Bmax. Our previous studies have clearly demonstrated that these alterations can be restored to normal by supplementation with ferrous sulfate; however, neither brain nonheme iron level nor dopamine D2 receptor Bmax can be increased beyond control values even after long-term iron therapy. The possibility that iron deficiency can induce the breakdown of the blood-brain barrier (BBB) was examined. A 70 and 100% increase in brain uptake index (BUI) for L-glucose and insulin, respectively, were noted in iron-deficient rats. However, the BUI for valine was decreased by 40%, and those for L-norepinephrine and glycine were unchanged. In addition, it was demonstrated that in normal rats insulin is transported into the brain. The data show that iron deficiency selectively affects the integrity of the BBB for insulin, glucose, and valine transport. Whether the effect of iron deficiency on the BBB is at the level of the capillary endothelial cell tight junction is not yet known. However, this study has shown that an important nutritional disorder (iron-deficiency anemia) has a profound effect on the BBB and brain function.     

Effect of Hyperammonemia and Methionine Sulfoximine on the Kinetic Parameters of Blood-Brain Transport of Leucine and Phenylalanine

The activity of the blood-brain neutral amino acid transport system is increased in rats infused with ammonium salts or rendered hyperammonemic by a portacaval anastomosis. This effect may be due to a direct action of ammonia or to some metabolic consequence of high ammonia levels, such as increased brain glutamine synthesis. To test these possibilities we evaluated the kinetic parameters of blood-brain transport of leucine and phenylalanine in control rats, in rats after continuous 24 h infusion of ammonium salts (NH4+ = 2.5 mmol X kg-1 X h-1), and in rats treated with methionine sulfoximine, an inhibitor of glutamine synthetase, before infusion of ammonium salts. In ammonia-infused rats without methionine sulfoximine treatment, the KD and Vmax of phenylalanine transport were increased, respectively, about 170% and 80% compared to controls, whereas the Km and Vmax of leucine transport were increased, respectively, about 100% and 200%. Electron microscopy demonstrated marked swelling of astrocytic processes around brain capillaries of ammonia-infused rats; however, capillary permeability to horseradish peroxidase apparently was not increased by ammonia infusion. Administration of methionine sulfoximine before ammonia infusion inhibited glutamine synthesis and prevented the changes in transport of leucine and phenylalanine, but apparently did not reverse the perivascular swelling. These results suggest that the ammonia-induced increase in the activity of transport of large neutral amino acids across the blood-brain barrier requires glutamine synthesis in brain, and is not a direct effect of ammonia.     

Blood-Brain Barrier Transport of Basic Amino Acids Is Selectively Inhibited at Low pH

The transport of amino acids across the blood-brain barrier was measured with the single-pass carotid injection method. The pH of the injected bolus varied between 4.5 and 8.5. Arginine and lysine uptakes were inhibited 24% at pH 5.5 and 59% at pH 4.5. The uptakes of 2-aminobicyclo (2,2,1) heptane-2-carboxylic acid and phenylalanine were unaffected at this pH. There were also no changes observed in choline, glucose, or butanol transport. The Ki of arginine transport inhibition by H+ was 2.4 +/- 0.5 microM; i.e., pH 5.6 +/- 0.1. No change with pH occurred in the Km of arginine transport, while a significant decrease (p less than 0.01) was observed in the Vmax (10.2 +/- 2.3 nmol min-1 g-1 and 5.6 +/- 2.3 nmol min-1 g-1 at pH 7.5 and pH 5.5, respectively). This noncompetitive inhibition was found to be transient as arginine uptake at pH 7.5; it was measured by carotid injection 30 sec following a previous bolus which was buffered to pH 4.5, and was not significantly different from the control. This selective inhibition of the blood-brain barrier basic amino acid carrier demonstrates the advantage of the carotid injection approach in exposing the capillary exchange site to extreme alterations in chemical composition which could not be tolerated systemically.     

Transport of Zinc-65 at the Blood-Brain Barrier During Short Cerebrovascular Perfusion in the Rat: Its Enhancement by Histidine

Abstract: Zinc-65 transport into different regions of rat brain has been measured during short vascular perfusion of one cerebral hemisphere with an oxygenated HEPES-containing physiological saline at pH 7.40. The [Zn²⁺] was buffered with either bovine serum albumin or histidine. In each case uptake was linear with time up to 90 s. ⁶⁵Zn flux into brain in the presence of albumin followed Michaelis-Menten kinetics and for parietal cortex had a K _m of 16 n M and a V _max of 44 nmol/kg/min. Increasing concentrations of l -histidine enhanced ⁶⁵Zn flux into brain at [Zn²⁺] values between 1 and 1,000 n M . The combined effect of [histidine] and [Zn²⁺] was best accounted for by a function of [ZnHis⁺], i.e., flux = 64.4 · [ZnHis⁺]/(390 + [ZnHis⁺]) + 0.00378 · [ZnHis⁺], with concentrations being nanomolar. d -Histidine had an influence similar to that of l -histidine. ⁶⁵Zn flux in the presence of 100 µ M l -histidine was not affected by either 500 µ M l -arginine or 500 µ M l -phenylalanine. The results indicate specific transport of Zn²⁺ across the plasma membranes of brain endothelium. The enhancement due to histidine has been attributed to diffusion of ZnHis⁺ across unstirred layers "ferrying" zinc to and from transport sites.     

Saturable Transport of Manganese(II) Across the Rat Blood-Brain Barrier

Unanesthetized adult male rats were infused intravenously with solutions containing 54Mn (II) and one of six concentrations of stable Mn(II). The infusion was timed to produce a near constant [Mn] in plasma for up to 20 min. Plasma was collected serially and on termination of the experiment, samples of CSF, eight brain regions, and choroid plexus (CP) were obtained. Influx of Mn (JMn) was calculated from uptake of 54Mn into tissues and CSF at two different times. Plasma [Mn] was varied 1,000-fold (0.076-78 nmol/ml). Over this plasma concentration range, JMn increased 123 times into CP, 18-120 times into brain, and 706 times into CSF. CP and brain JMn values fit saturation kinetics with Km (nmol/ml) equal to 15 for CP and 0.7-2.1 for brain, and Vmax (10(-2) nmol.g-1.s-1) of 27 for CP and 0.025-0.054 for brain. Brain JMn except at cerebral cortex had a nonsaturable component. CSF JMn varied linearly with plasma [Mn]. These findings suggest that Mn transport into brain and CP is saturable, but transport into CSF is nonsaturable.