- Email: [email protected]
Increased transfer of45Ca into brain and cerebrospinal fluid from plasma during chronic hypocalcemia in rats
Brain Research, 454 (1988) 315-320 Elsevier
315
BRE 13735
Increased transfer of 45Cainto brain and cerebrospinal fluid from plasma during chronic hypocalcemia in rats Vincent A. Murphy and Stanley I. Rapoport Laboratory of Neurosciences, National lnstitute on Aging, National Institutes of Health, Bethesda, MD 20892 (U.S.A.)
(Accepted 28 January 1988) Key words: Calcium; Hypocalcemia; Brain; Cerebrospinal fluid; Choroid plexus; Transport; Blood-brain barrier
Recent studies have shown regulation of central nervous system [Ca] after chronic hypo- and hypercalcemia. To investigate the mechanism of this regulation, 3-week-old rats were fed diets for 8 weeks that contained low or normal levels of Ca. Plasma [Ca] was 40% less in rats fed the low Ca diet than in animals fed normal diet. Unidirectional transfer coefficients for Ca (Kc~) and C1 (Kcl) into cerebrospinal fluid (CSF) and brain were determined from the 10 min uptake of intravenously injected 4SCaand 36C1in awake animals. Kca for CSF was 68% greater in low-Ca rats than in normal rats. Likewise, the values of/(Ca for brain regions with areas adjacent to the ventricles like the hippocampus and pons-medulla were 50% higher than in normal animals. On the other hand, Kcas for parietal cortex, a brain region distant from the choroid plexus and not expected to be influenced by Ca entry into CSF, were similar between the groups. Comparison of the regional ratios of Kca/Kcl revealed that a selective increase of Ca transport occurred into CSF and all brain regions except the parietal cortex in Ca-deficient rats. The results suggest that Ca homeostasis of CSF and brain [Ca] during chronic hypocalcemia is due to increased transfer of Ca from blood to brain, and that the regulation occurs via the CSF, possibly at the choroid plexus, but not via the cerebral capillaries. INTRODUCTION Multiple studies in different m a m m a l i a n species have shown that both cerebrospinal fluid (CSF) and brain [Ca] remain stable after acute and chronic alterations in plasma [Ca]. Large acute changes ( - 5 0 to 300%) in either total or ionized p l a s m a [Ca] in rat, dog, and man p r o d u c e little if any changes in C S F or brain [Ca] 1'4,s,12,15,25,26. Chronic changes in plasma [Ca] p r o d u c e greater changes in central nervous system (CNS) [Ca], but only 2 0 - 3 0 % of the changes in plasma n'18. This stability of CNS [Ca] suggests that mechanisms exist to regulate [Ca] in the brain and CSF. One possible factor is the low p e r m e a b i l i t y of the b l o o d - b r a i n barrier to Ca 23. The b l o o d - b r a i n barrier is c o m p o s e d of the cerebral capillaries, choroid plexus, and arachnoid m e m b r a n e , and restricts the exchange of ions and p o l a r solutes b e t w e e n b l o o d and CNS 3,20. The half-time for Ca entry into brain from blood is 6 - 8 h in rats 13. Slow entry would pre-
vent acute changes in plasma [Ca] from being reflected in the brain 23. Chronic alterations in p l a s m a [Ca], however, would require m o r e than just low permeability at the b l o o d - b r a i n barrier, as eventual equilibration between b l o o d and CNS would occur 3'18. Stability of CNS [Ca] after long-term changes in plasma [Ca] requires the presence of a calcium transport system 3. Preliminary evidence for c a r r i e r - m e d i a t e d transport of Ca into CSF and brain has been d e m o n s t r a t e d with ventriculocisternal perfusion in rats and cats 2'9'10. H o w e v e r , B a r k a i and Meltzer 2 concluded that Ca uptake from b l o o d is d e p e n d e n t on CSF [Ca], whereas Graziani et al. 93° describe saturable Ca transfer from p l a s m a into CSF, but u p t a k e was indep e n d e n t of CSF [Ca]. In addition, Tai et al. 23 failed to demonstrate saturable Ca t r a n s p o r t b e t w e e n b l o o d and CNS after intravenous injection of 45Ca in rats. The conflicting nature of prior findings indicates that further studies on the mechanism of Ca regulation at the b l o o d - b r a i n barrier are necessary.
Correspondence: V.A. Murphy, Laboratory of Neurosciences, National Institute on Aging, Building 10, Room 6C103, National Institutes of Health, Bethesda, MD 20892, U.S.A.
316 Because our laboratory has developed a model of chronic hypocalcemia in rats Is, we decided to investigate the mechanism of CNS [Ca] regulation after chronic hypocalcemia. We measured 45Ca uptake into the CSF and 7 brain regions after intravenous injection in rats fed diet with deficient Ca. The results demonstrate that unidirectional 45Ca transport from blood to CNS is elevated under such conditions, possibly accounting for Ca homeostasis of the CNS during hypocalcemia. Some of these results have been published in abstract form 17. MATERIALS AND METHODS Male Fischer-344 rats, weighing 30-40 g, were obtained from Charles River (Wilmington, MA) at 3 weeks of age. Animals were housed in a temperature- and humidity-controlled room with a 12-h light cycle. Specially formulated diets (Teklad, Madison, WI) containing either 0.01% (wt/wt) Ca (low calcium diet, LOCA) or 0.67% Ca (control diet, CONT), were fed to the animals for 8 weeks. Some rats were fed a reduced quantity of CONT diet (RCONT) to match the reduced food consumption of L O C A diet animals. Diet composition, animal growth, and food consumption have been described 18. Briefly, diets contained 20% (wt/wt) protein as casein, 15% fat as corn oil, 65% carbohydrate as corn starch and sucrose, and essential amounts of vitamins and minerals, except Ca. Before an experiment, a rat was anesthetized with ketamine HC1, 100 mg/kg i.p. (Warner-Lambert, Morris Plains, N J). Polyethylene catheters (PE-50) filled with heparinized isotonic NaCl solution (0.9% wt/v) were tied into the left femoral vein and artery, and a local anesthetic was applied to the incision, bupivacaine 0.25% (Breon, New York, NY). The hindquarters of the animal were wrapped in a plaster bandage that was taped to a wooden block. The rats were allowed to recover from anesthesia for 4-5 h. This procedure is a modification of a previous method 19. After recovery, an arterial blood sample was taken for analysis of pH and ionized Ca at 37 °C, using an ICA-1 ionized-calcium analyzer (Radiometer, Copenhagen, Denmark). Some arterial blood was centrifuged to obtain plasma for analysis of total plasma calcium. A 0.05 ml sample of plasma was diluted with 4 ml of 20 mM La(C104) 3 solution and analysed by
atomic absorption spectroscopy (Instrumentation, Lab., Lexington, MA) to obtain total plasma calcium. Each rat received 45Ca and 36C1, 0.05 mCi/kg b. wt., as an intravenous bolus into the femoral vein after the arterial blood was taken. Arterial blood samples (0.05 ml) then were collected at 0.25, 0.5, 1, 2, 5 and 10 rain after injection, in 0.3 ml heparinized microcentrifuge tubes. A tracer to measure residual blood volume, [3H]raffinose, 0.25 mCi/kg b. wt., was injected at 7 min. Immediately after collecting the 10 min blood sample, 150 mg/kg b. wt. of Na pentobarbital was injected intravenously to kill the animal, and the heart was excised to reduce residual blood in the brain. Blood samples were centrifuged for 1 min at 10,000 g in a Microfuge B (Beckman, Fullerton, CA) to obtain plasma. CSF was removed from the cisterna magna into tapered capillary tubes 16. CSF samples from L O C A rats were between 0.005 and 0.015 ml; samples with similar volumes were taken from R C O N T and CONT rats. The brain was removed from the skull and dissected into specific regions on chilled filter paper, soaked in 5.1% mannitol solution 5. Aliquots of plasma at each time point, and blood at 10 min, as well as samples of CSF and brain, were placed in tared scintillation vials and reweighed. The contents of each vial were digested at 50 °C overnight after adding 1 ml of 1 M piperidine solution I6. The following day, 10 ml of Ready-Solv MP (Beckman) was added and the vials were mixed. Radioactivity was counted on a liquid scintillation counter (Beckman LS6800), using a triple label program. [45Ca]CaCI2 (spec. act. 24.7 mCi/mg) and [3H]raffinose (spec. act. 7.6 Ci/mmol, purity 99%) were purchased from New England Nuclear (Boston, MA). The [3H]raffinose was evaporated to dryness with a stream of dry N 2 gas to remove any [3H]H20.36C1 as NaC1 (18.1 mCi/g) came from ICN Biomedicals, (Irvine, CA). Plasma to brain transfer coefficients, Kca and Kcl, were calculated from the following equation, as described in previous reports 21'23 Cbr*
g c a or C1 - -
f~l~i~inCp*dt
where Cur* is the concentration of radioactivity in
317 brain parenchyma (dpm/g) at 10 min, Cp* is the plasma radioactivity (dpm/ml), andfis the fraction of isotope ionized in plasma (ionized/total, 0.5 for Ca and 1.0 for C1). CSF transfer coefficients were obtained by substituting CSF radioactivity at 10 min (dpm/g) for brain parenchyma radioactivity. Intravascular radioactivity was subtracted from total brain radioactivity to obtain brain parenchymal radioactivity (dpm/g). Intravascular radioactivity was calculated as the product of regional blood volume and the blood concentration of isotope at 10 min. Regional blood volume was determined from (dpm/g brain)/(dpm/g blood) of [3H]raffinose after 3 min. Raffinose had a K value of i x 10 -6 ml.g-l.s -1, less than that of 45Ca o r 36C121'23. If K values are calculated when the plasma concentration of isotope is much greater than the concentration in brain extracellular fluid or in CSF back diffusion can be ignored. We chose to do this by restricting circulation time to 10 min. Changes in cerebral blood flow can alter K, but only for rapidly penetrating compounds 19, with Kvalues near 1 x 10-3 ml.g-l.s -1. Neither Kca nor Kcl approximate this value 21,23. Regional brain K may reflect both cerebral capillary and choroid plexus transfer of isotope, with the distance from the ventricles determining the contribution of each route 22. Restriction of isotope circulation time to 10 min kept cortical gray regions relatively free of isotope contamination from C S F 22. One-way analysis of variance and Bonferroni ttests were used to compare multiple means TM. Levels of significance were set at P < 0.05. Integrated plasma radioactivity was calculated with a computer program that utilized the trapezoidal method 24. In order to perform this calculation, the plasma concentration at time 0 was estimated by visual approximation from a plot of plasma concentration vs time. The approximation could be altered by 100% and the integrated area would change by only 5%.
TABLE I
Body weight, blood pH, and plasma [Ca] Values are means _+ S.E.M. for 5 animals each. Diets consumed for 8 weeks were either deficient in Ca (LOCA) or sufficient in Ca at normal (CONT) or reduced (RCONT) intake.
Measurement
CONT
RCONT
LOCA
Body weight (g) Arterial blood pH Plasma ionized [Ca] (mmol/l)
227+8 7.40+0.01
114_+6a 7.38_+0.03
105+10 a 7.40+0.03
1.31+0.01
1.32+0.01
0.80+0.05 a'b
Plasma total [Ca] (mmol/l)
a Differs from value of CONT or Udiffers from value of RCONT by A N O V A (one-way) and Bonferroni t-test (twotailed), P < 0.05.
groups (Table I). Information on other ions in plasma, CSF, and brain after L O C A and CONT dietary intake can be found elsewhere TM. Mean blood pressure was significantly less in L O C A than in C O N T rats (108 + 2 torr, compared with 123 + 3). Fig. 1 plots m e a n 45Ca and 36C1 concentrations in plasma of CONT animals from 0.25 to 10 min after intravenous injection. Concentrations of these isotopes in plasma, blood, CSF, and 7 brain regions at
2000
O
Isotooe C o n c e n l r a t i o n at 1D min
X
(dz>rn/olx 10"3
45Ca
36CI
37.0 2 2.7 PARIETALCORTEX 1.2 i 0.1 CAUOATENUCLEUS 5.8 t 0.5 HIPPOCAMPUS 5.6 ¢ 0,5 THAI.AMUS-HYPOTHALJ~AUS 4,6 t; 0.3 MIDBR~QJN-COLLICULI 3.0 * 0+2 CEREBELLUM S.0 ± 0.6 PONS-MEDULLA 4.8 ± 0.5 BLOOD 244.0 2 ILl ~ P L A 1S .963 2M 11.4 CSF
E
1500
C
1000 ,
Q. "O v
~
A
118.0 ~: 7.0 10,0 ¢ 0.2 17.0 ± 1.0 15,$ Jt 0.4 15.0 ± 0.7 10.5 ± 0+5 16.1 1t 0.5 13.1 ± 0.6 288,6 ± $.0 397.3 i $.4
O C
g, X O
500 •
i m I
RESULTS
2.67+0.03 2.66_+0.05 1.48+_0.10a'u
o o. o
2.15
5 .Io
7.15
1 ol.o
Time (rain)
Rats fed L O C A diet for 8 weeks were about 50% lighter than rats fed C O N T diet (Table I), whereas R C O N T rats had similar weights to L O C A rats. Plasma [Ca]s, both total and ionized, were the same for CONT and R C O N T animals, but were 40% lower in L O C A animals. Blood pH was the same for all 3
Fig. 1. Arterial plasma concentrations of ~SCa and 36C1 after i.v. injection of 0.05 mCi/kg b. wt. into awake rats. Each point is the mean of 5 animals fed CONT diet. The points are connected by straight lines depicting the top of each trapezoid used to calculate the area under the curve, excluding the estimated points at 0 min. Mean concentrations + S.E.M. of 45Ca and 36C1 (n = 5) in brain parenchyma, CSF, blood, and plasma are listed for rats killed 10 min after isotope injection.
318 10 min are included in Fig. 1 as well. T h e brain values
TABLE III
are parenchymal concentrations (total - - intravascular). In general, the time courses of 45Ca and 36C1 in
Regional transfer coefficients for 36Cl
plasma of L O C A and R C O N T rats were similar to those in C O N T rats. T h e amounts of 45Ca and 36C1 in brain at 10 min were lowest in the parietal cortex and highest in caudate nucleus, whereas C S F concentra-
Values are means _+ S.E.M. for 5 animals each. Radioactive C1 was injected i.v. as a bolus into the femoral vein. Rats were killed at 10 min after injection.
CNS region
CONT
tions were substantially above those in the brain. Mean regional blood volumes as m e a s u r e d by [3H]raffinose ranged b e tw e e n 0.011 and 0.021 for various brain regions, and did not differ significantly between L O C A , C O N T and R C O N T . T h e values were similar to those obtained with [3H]inulin21. CSF concentrations of [3H]raffinose at 3 min were 1.38 + 0.2% of plasma for L O C A rats, 0.52 + 0.08 for R C O N T rats, and 0.33 _+ 0.04 for C O N T rats. Values of Kca at CSF and various brain regions are
Kcl (ml.g-l.s -1) ×lOs
CSF Parietalcortex Caudate nucleus Hippocampus Thalamushypothalamus Midbrain-colliculi Cerebellum Pons-medulla
RCONT
LOCA
42.30_+2.00 47.40_+1.70 52.80+2.00 a 3.60-+0.05 3.71-+0.14 3.24+0.28 6.10+0.32 5.19-+0.61 5.11+0.42 5.92+0.11 5.35-+0.12 5.28+0.29 5.38-+0.19 6.02-+0.24 3.78_+0.14 4.37_+0.13 5.77+0.16 6.54+0.39 4.71-+0.16 5.24_+0.26
5.35-+0.35 3.77-+0.37 5.46-+0.30 5.41_+0.29
a Differs from transfer coefficient of CONT by ANOVA (oneway) and Bonferroni t-test (two-tailed), P < 0.05.
listed in Table II for C O N T , R C O N T , or L O C A rats. Kca at CSF was substantially larger than at any brain
eral brain regions. Like values for Kca, KCI w a s great-
region in all diet groups. Transfer of 45Ca from blood
est in CSF and least in parietal cortex. No significant
was greatest into the caudate nucleus, hippocampus,
differences were found b e t w e e n the values of Kcl
and pons-medulla, and least into the parietal cortex
into the brain among the 3 groups. H o w e v e r , Kcl at
and midbrain-colliculi a m o n g the various brain re-
CSF was significantly higher in L O C A rats c o m p a r e d
gions. CSF and brain regional/(Ca values for L O C A
to C O N T rats, but not c o m p a r e d with R C O N T rats.
were significantly increased 4 0 - 6 0 % above corre-
If we assume some c o m m o n factors influence Kca
sponding values for C O N T and R C O N T , except at the parietal cortex, midbrain-colliculi, and cerebel-
and Kcl, such as surface area, CSF formation rate, and CSF or brain extracellular volumes, the ratio of
lum.
g c a over Kcl can be used to d et er m i n e if changes in
Table III displays the values of K a at C S F and sev-
Kca are specific for calcium. Calculated values for
Kca/KcI are illustrated in Table IV. Th e CSF and TABLE II
Regional transfer coefficients for 45Ca
TABLE IV
Values are means _+ S.E.M. for 5 animals each. Radioactive Ca was injected i.v. as a bolus into the femoral vein. Rats were killed at 10 min after injection.
Ratio of regional transfer coefficients"
CNS region
CNS region
Kca (ml.g-l.s -1) x lO5 CONT
CSF Parietal cortex Caudatenucleus Hippocampus Thalamushypothalamus Midbrain-colliculi Cerebellum Pons-medulla
RCONT
4.32_+0.10a'b 2.64-+0.13 4.48_+0.61 4.91-+0.21a'b
a Differs from transfer coefficient of CONT or b differs from transfer coefficient of RCONT by ANOVA (one-way) and Bonferroni t-test (two-tailed), P < 0.05.
KcjKcl CONT
LOCA
22.70+1.20 27.40_+1.40 41.90_+2.90a'b 0.77-+0.07 0.90-+0.15 0.83_+0.17 3.69+0.30 3.41-+0.45 5.18_+0.31a'b 3.53+0.29 2.91_+0.20 4.80_+0.40~'b 2.90-+0.20 32.7_+0.15 1.91+0.15 2.24-+0.28 3.17+0.35 3.96-+0.52 2.86_+0.25 3.55_+0.32
Values are means _+S.E.M. for 5 animals each.
CSF Parietal cortex Caudatenucleus Hippocampus Thalamushypothalamus Midbrain-colliculi Cerebellum Pons-medulla
RCONT
LOCA
0.54-+0.01 0.58-+0.01 0.21 _+0.02 0.24+0.03 0.60_+0.02 0.66+0.02 0.60-+0.05 0.54+0.03
0.79+0.04 a'b 0.25_+0.05 1.03-I-0.05a'b 0.91+0.07 a'b
0.54-+0.02 0.54-+0.02 0.50+0.02 0.51_+0.05 0.55+0.05 0.59-+0.05 0.60-+0.04 0.68_+0.05
0.82+0.05 a'b 0.71_+0.03a'b 0.81+0.07 a'b 0.91_+0.03a'b
a Differs from ratio of CONT or bdiffers from ratio of RCONT by ANOVA (one-way) and Bonferroni t-test (two-tailed), P < 0.05.
319 brain regional ratios of Kca/Kcl are significantly higher in L O C A animals compared to both CONT and RCONT rats with the exception of parietal cortex. DISCUSSION This paper demonstrates that in 11 week old rats, fed a low Ca diet for the prior 8 weeks and made hypocalcemic, the transfer coefficients (Kca) of calcium from blood into CSF and brain regions adjacent to ventricular CSF, are elevated. These increases in/(Ca are specific to Ca, as Kcl values into affected brain regions were unchanged and the Kca/Kcl ratio was elevated for both CSF and brain. Stability of CSF and brain [Ca] after chronic hypocalcemia 11,18, therefore, is partly due to a relatively constant rate of Ca entry from blood into the CNS, and Ca specificity suggests that active transport of Ca is altered. Finding that the Kca for the parietal cortex, a brain region distant from the ventricles where Ca is elaborated into CSF, remains unchanged during chronic hypocalcemia suggests that enhanced Ca transfer occurs at the choroid plexus and not at the cerebral capillaries. One way to maintain stable [Ca] in the CNS during chronic hypocalcemia is to keep the unidirectional flux of Ca into the brain and CSF relatively constant. Unidirectional flux, Jca, is the product of Kca and plasma ionized [Ca]. Jca into CSF for L O C A is 34 x 10-5 ymol.g-l.s -1, whereas for CONT and RCONT, it equals 30 x 10-5 and 36 x 10 -5/~mol.g-l.s -1, respectively. Flux of Ca into CSF is not different among the 3 groups, which implies Ca entry into CSF is regulated. On the other hand, parietal cortex uptake is linearly related to plasma [Ca], as was found by Tai et al. 3. Values of K for CSF depend on other factors, in addition to uptake. Smith and Rapoport 22 have described these in detail in relation to K values for 22Na and 36C1. One of these factors is CSF volume. L O C A animals had an apparent reduction in CSF volume, as sampling of CSF was more difficult in L O C A rats than in RCONT rats of similar size. A reduction in CSF volume might explain the increase in Kcl and Kca in these animals. Comparison of the ratios of Kca to Kcl suggests that factors in addition to a decreased CSF volume contributed to the elevated gca value in the CSF.
The concentration of [3H]raffinose in CSF, 3 min after injection, was much greater in L O C A than in either RCONT or CONT. The difference between the raffinose concentrations suggests blood contamination of CSF (1.38-0.33 = 1.05%). The product of 0.0105 and the concentration of 45Ca or 36C1in plasma represents radioactivity due to contamination. Subtraction of this estimate of contamination from total radioactivity in CSF of L O C A animals and recalculation yields values of 39 x 10-5 ml.g-l.s -1 for gca and 51.9 x 10-5 for Kcl, respectively. As CSF K values corrected for possible contamination are not different from the uncorrected values, the elevations of gca and K a at the CSF in L O C A rats are not due to blood contamination. Reduced CSF volume could at least partly explain the elevated [3H]raffinose concentration in L O C A rats. Our values for/(Ca and Kcl in awake rats are similar to prior measurements in anesthetized animals 21'23. Values for gel at CSF and various brain regions are the same with the exception of the parietal cortex which is 50-100% higher in our study. Dissec' tion or strain differences or a decrease in surface area by anesthesia 7 could explain this difference in Kcl values. Brain regional Kca was the same in our study and that of Tai et al. 23, but CSF/(Ca was 100% greater in our study (23 x 10-5 ml.g-l.s -1 vs 12 x 10-5). This difference could be due to a shorter circulation time, 10 min vs 15, or possibly anesthesia. Smith and Rapoport 22 suggested that if the times of isotope circulation were kept under 10 min, cortical regions distant from the choroid plexuses can be used to estimate cerebral capillary exchange, free of contamination by isotope from CSF. In our study, /(Ca for the parietal cortex did not significantly increase after chronic hypocalcemia. Autoradiographic data of both 45Ca and 36C1indicate that the major route of entry into CSF is across the choroid plexus, rather than the arachnoid or pial membranes 22'23. Diffusion of Ca from subarachnoid CSF into adjacent brain areas appears to be insignificant at 15 min 23. The elevations of Kca in other brain regions probably reflect isotope entering from ventricular CSF. Thus, a Caspecific transport system at the choroid plexus appears to regulate Ca entry into the CNS, whereas the cerebral capillaries restrict exchange. Indeed, some reports suggest that a carrier for Ca exists at the choroid plexus. In cats, Ames and Nes-
320 bitt found that Ca secreted by the choroid plexus is not an ultrafiltrate of plasma (quoted in Cserr6), whereas in rats, Barkai and Meltzer 2 demonstrated
tering Ca extrusion from the choroidal epithelium. Assuming this model is correct, previous results can be reconciled by noting that CSF [Ca] must change
regulated Ca uptake into CSF which could be in-
for the carrier to alter the rate of Ca transport. In all but one study 23, CSF [Ca] was changed 2A2. With ex-
hibited by R u t h e n i u m red, a Ca-ATPase inhibitor. The frog choroid plexus has a large paracellular pathway for Ca m o v e m e n t , but transcellular transport also exists 27. Barkai and Meltzer 2 suggest that a C a - A T P a s e is
tended hypocalcemia, CSF [Ca] begins to decrease. In response, the carrier may slow Ca expulsion from the choroid plexus, allowing more Ca to enter the CSF from blood. This process would be noted as an
located on the basolateral m e m b r a n e of the choroid plexus and can respond to changes in CSF [Ca] by al-
increase in Kca for CSF as in our study.
REFERENCES
14 Miller, R.G., Simultaneous Statistical Inference, McGrawHill, New York, 1966. 15 Morgulus, S. and Perley, A.M., Studies on cerebrospinal fluid and serum calcium with special reference to the parathyroid hormone, J. Biol. Chem., 88 (1930) 169-188. 16 Murphy, V.A. and Johanson, C.E., Adrenergic-induced enhancement of brain barrier system permeability to small nonelectrolytes: choroid plexus versus cerebral capillaries, J. Cereb. Blood Flow Metab., 5 (1985) 401-412. 17 Murphy, V.A. and Rapoport, S.I., Central nervous system uptake of Ca-45 in rats fed calcium deficient diet, Soc. Neurosci. Abstr., 13 (1987) 1249. 18 Murphy, V.A., Smith, Q.R. and Rapoport, S.I., Homeostasis of brain and cerebrospinal fluid calcium concentrations during chronic hypo- and hypercalcemia, J. Neurochem., 47 (1986) 1735-1741. 19 Ohno, K., Pettigrew, K.D. and Rapoport, S.I., Lower limits to cerebrovascular permeability to nonelectrolytes in the conscious rat, Am. J. Physiol., 235 (1978) H299-H307. 20 Rapoport, S.I., Blood-Brain Barrier in Physiology and Medicine, Raven, New York, 1976. 21 Smith, Q.R. and Rapoport, S.I., Carrier-mediated transport of chloride across the blood-brain barrier, J. Neurochem., 42 (198)I) 754-763. 22 Smith, Q.R. and Rapoport, S.I., Cerebrovascular permeability coefficients to sodium, potassium, and chloride, J. Neurochem., 46 (1986) 1732-1742. 23 Tai, C.-Y., Smith, Q.R. and Rapoport, S.I., Calcium influxes into brain and cerebrospinal fluid are linearly related to plasma ionized calcium concentration, Brain Research, 385 (1986) 227-236. 24 Wagner, J., Fundamentals of Clinical Pharmacokinetics, Hamilton, Hamilton, IL, 1975. 25 Wallach, S., Bellavia, J.V., Schorr, J. and Reizenstein, D.L., Tissue distribution of electrolytes, C a 47 and Mg28 in acute hypercalcemia, Am. J. Physiol., 207 (1964) 553-560. 26 Wong, R.P.K. and Bradbury, M.W.B., Permeability of the blood-brain barrier to calcium in adrenal insufficiency, Brain Research, 84 (1974) 361-364. 27 Wright, E., Transport processes in the formation of cerebrospinal fluid, Rev. Physiol. Biochem. Pharmacol., 83 (1978) 1-34.
1 Arieff, A.I. and Massry, S.G., Calcium metabolism of brain in acute renal failure: effects of uremia, hemodialysis, and parathyroid hormone, J. Clin. Invest., 53 (1974) 387-392. 2 Barkai, A.I. and Meltzer, H.L., Regulation of calcium entry into the extracellular environment of the rat brain, J. Neurosci., 2 (1982) 1322-1328. 3 Bradbury, M., The Concept of a Blood-Brain Barrier, Wiley, New York, 1979. 4 Cameron, A.T. and Moorhouse, V.H.K., The relation between plasma and cerebrospinal fluid calcium, J. Physiol. (Lond.), 91 (1937) 90-100. 5 Chiueh, C.C., Sun, C.L., Kopin, I.J., Fredericks, W.R. and Rapoport, S.I., Entry of [3H]norepinephrine, [125I]albumin and Evans blue from blood to brain following unilateral osmotic opening of the blood-brain barrier, Brain Research, 145 (1978) 291-301. 6 Cserr, H.F., Physiology of the choroid plexus, Physiol. Rev., 51 (1971) 273-311. 7 Gjedde, A. and Rasmussen, M., Pentobarbital anesthesia reduces blood-brain glucose transfer in the rat, J. Neurochem., 35 (1980) 1382-1387. 8 Goldstein, D.A., Romoff, M., Bogin, E. and Massry, S.G., Relationship between the concentration of calcium and phosphorous in blood and cerebrospinal fluid, J. Clin. Endocrinol. Metab., 49 (1979) 58-62. 9 Graziani, L., Escriva, A. and Katzman, R.A., Exchange of calcium between blood, brain and cerebrospinal fluid, Am. J. Physiol., 208 (1965) 1058-1064. 10 Graziani, L.J., Kaplan, R.K., Escriva, A. and Katzman, R., Calcium flux into CSF during ventricular and ventriculocisternal perfusion, Am. J. Physiol., 213 (1967) 629-636. 11 Harris, R.A., Carnes, D.L. and Forte, L.R., Reduction of brain calcium after consumption of diets deficient in calcium or vitamin D, J. Neurochem., 36 (1981) 460-466. 12 Herbert, F.K., The total and diffusible calcium of serum and the calcium of cerebrospinal fluid in human cases of hypocalcemia and hypercalcemia, J. Biochem., 27 (1934) 1978-1991. 13 Katzman, R. and Pappius, H.M., Brain Electrolytes and Fluid Metabolism, Williams and Wilkins, Baltimore, 1973.
315
BRE 13735
Increased transfer of 45Cainto brain and cerebrospinal fluid from plasma during chronic hypocalcemia in rats Vincent A. Murphy and Stanley I. Rapoport Laboratory of Neurosciences, National lnstitute on Aging, National Institutes of Health, Bethesda, MD 20892 (U.S.A.)
(Accepted 28 January 1988) Key words: Calcium; Hypocalcemia; Brain; Cerebrospinal fluid; Choroid plexus; Transport; Blood-brain barrier
Recent studies have shown regulation of central nervous system [Ca] after chronic hypo- and hypercalcemia. To investigate the mechanism of this regulation, 3-week-old rats were fed diets for 8 weeks that contained low or normal levels of Ca. Plasma [Ca] was 40% less in rats fed the low Ca diet than in animals fed normal diet. Unidirectional transfer coefficients for Ca (Kc~) and C1 (Kcl) into cerebrospinal fluid (CSF) and brain were determined from the 10 min uptake of intravenously injected 4SCaand 36C1in awake animals. Kca for CSF was 68% greater in low-Ca rats than in normal rats. Likewise, the values of/(Ca for brain regions with areas adjacent to the ventricles like the hippocampus and pons-medulla were 50% higher than in normal animals. On the other hand, Kcas for parietal cortex, a brain region distant from the choroid plexus and not expected to be influenced by Ca entry into CSF, were similar between the groups. Comparison of the regional ratios of Kca/Kcl revealed that a selective increase of Ca transport occurred into CSF and all brain regions except the parietal cortex in Ca-deficient rats. The results suggest that Ca homeostasis of CSF and brain [Ca] during chronic hypocalcemia is due to increased transfer of Ca from blood to brain, and that the regulation occurs via the CSF, possibly at the choroid plexus, but not via the cerebral capillaries. INTRODUCTION Multiple studies in different m a m m a l i a n species have shown that both cerebrospinal fluid (CSF) and brain [Ca] remain stable after acute and chronic alterations in plasma [Ca]. Large acute changes ( - 5 0 to 300%) in either total or ionized p l a s m a [Ca] in rat, dog, and man p r o d u c e little if any changes in C S F or brain [Ca] 1'4,s,12,15,25,26. Chronic changes in plasma [Ca] p r o d u c e greater changes in central nervous system (CNS) [Ca], but only 2 0 - 3 0 % of the changes in plasma n'18. This stability of CNS [Ca] suggests that mechanisms exist to regulate [Ca] in the brain and CSF. One possible factor is the low p e r m e a b i l i t y of the b l o o d - b r a i n barrier to Ca 23. The b l o o d - b r a i n barrier is c o m p o s e d of the cerebral capillaries, choroid plexus, and arachnoid m e m b r a n e , and restricts the exchange of ions and p o l a r solutes b e t w e e n b l o o d and CNS 3,20. The half-time for Ca entry into brain from blood is 6 - 8 h in rats 13. Slow entry would pre-
vent acute changes in plasma [Ca] from being reflected in the brain 23. Chronic alterations in p l a s m a [Ca], however, would require m o r e than just low permeability at the b l o o d - b r a i n barrier, as eventual equilibration between b l o o d and CNS would occur 3'18. Stability of CNS [Ca] after long-term changes in plasma [Ca] requires the presence of a calcium transport system 3. Preliminary evidence for c a r r i e r - m e d i a t e d transport of Ca into CSF and brain has been d e m o n s t r a t e d with ventriculocisternal perfusion in rats and cats 2'9'10. H o w e v e r , B a r k a i and Meltzer 2 concluded that Ca uptake from b l o o d is d e p e n d e n t on CSF [Ca], whereas Graziani et al. 93° describe saturable Ca transfer from p l a s m a into CSF, but u p t a k e was indep e n d e n t of CSF [Ca]. In addition, Tai et al. 23 failed to demonstrate saturable Ca t r a n s p o r t b e t w e e n b l o o d and CNS after intravenous injection of 45Ca in rats. The conflicting nature of prior findings indicates that further studies on the mechanism of Ca regulation at the b l o o d - b r a i n barrier are necessary.
Correspondence: V.A. Murphy, Laboratory of Neurosciences, National Institute on Aging, Building 10, Room 6C103, National Institutes of Health, Bethesda, MD 20892, U.S.A.
316 Because our laboratory has developed a model of chronic hypocalcemia in rats Is, we decided to investigate the mechanism of CNS [Ca] regulation after chronic hypocalcemia. We measured 45Ca uptake into the CSF and 7 brain regions after intravenous injection in rats fed diet with deficient Ca. The results demonstrate that unidirectional 45Ca transport from blood to CNS is elevated under such conditions, possibly accounting for Ca homeostasis of the CNS during hypocalcemia. Some of these results have been published in abstract form 17. MATERIALS AND METHODS Male Fischer-344 rats, weighing 30-40 g, were obtained from Charles River (Wilmington, MA) at 3 weeks of age. Animals were housed in a temperature- and humidity-controlled room with a 12-h light cycle. Specially formulated diets (Teklad, Madison, WI) containing either 0.01% (wt/wt) Ca (low calcium diet, LOCA) or 0.67% Ca (control diet, CONT), were fed to the animals for 8 weeks. Some rats were fed a reduced quantity of CONT diet (RCONT) to match the reduced food consumption of L O C A diet animals. Diet composition, animal growth, and food consumption have been described 18. Briefly, diets contained 20% (wt/wt) protein as casein, 15% fat as corn oil, 65% carbohydrate as corn starch and sucrose, and essential amounts of vitamins and minerals, except Ca. Before an experiment, a rat was anesthetized with ketamine HC1, 100 mg/kg i.p. (Warner-Lambert, Morris Plains, N J). Polyethylene catheters (PE-50) filled with heparinized isotonic NaCl solution (0.9% wt/v) were tied into the left femoral vein and artery, and a local anesthetic was applied to the incision, bupivacaine 0.25% (Breon, New York, NY). The hindquarters of the animal were wrapped in a plaster bandage that was taped to a wooden block. The rats were allowed to recover from anesthesia for 4-5 h. This procedure is a modification of a previous method 19. After recovery, an arterial blood sample was taken for analysis of pH and ionized Ca at 37 °C, using an ICA-1 ionized-calcium analyzer (Radiometer, Copenhagen, Denmark). Some arterial blood was centrifuged to obtain plasma for analysis of total plasma calcium. A 0.05 ml sample of plasma was diluted with 4 ml of 20 mM La(C104) 3 solution and analysed by
atomic absorption spectroscopy (Instrumentation, Lab., Lexington, MA) to obtain total plasma calcium. Each rat received 45Ca and 36C1, 0.05 mCi/kg b. wt., as an intravenous bolus into the femoral vein after the arterial blood was taken. Arterial blood samples (0.05 ml) then were collected at 0.25, 0.5, 1, 2, 5 and 10 rain after injection, in 0.3 ml heparinized microcentrifuge tubes. A tracer to measure residual blood volume, [3H]raffinose, 0.25 mCi/kg b. wt., was injected at 7 min. Immediately after collecting the 10 min blood sample, 150 mg/kg b. wt. of Na pentobarbital was injected intravenously to kill the animal, and the heart was excised to reduce residual blood in the brain. Blood samples were centrifuged for 1 min at 10,000 g in a Microfuge B (Beckman, Fullerton, CA) to obtain plasma. CSF was removed from the cisterna magna into tapered capillary tubes 16. CSF samples from L O C A rats were between 0.005 and 0.015 ml; samples with similar volumes were taken from R C O N T and CONT rats. The brain was removed from the skull and dissected into specific regions on chilled filter paper, soaked in 5.1% mannitol solution 5. Aliquots of plasma at each time point, and blood at 10 min, as well as samples of CSF and brain, were placed in tared scintillation vials and reweighed. The contents of each vial were digested at 50 °C overnight after adding 1 ml of 1 M piperidine solution I6. The following day, 10 ml of Ready-Solv MP (Beckman) was added and the vials were mixed. Radioactivity was counted on a liquid scintillation counter (Beckman LS6800), using a triple label program. [45Ca]CaCI2 (spec. act. 24.7 mCi/mg) and [3H]raffinose (spec. act. 7.6 Ci/mmol, purity 99%) were purchased from New England Nuclear (Boston, MA). The [3H]raffinose was evaporated to dryness with a stream of dry N 2 gas to remove any [3H]H20.36C1 as NaC1 (18.1 mCi/g) came from ICN Biomedicals, (Irvine, CA). Plasma to brain transfer coefficients, Kca and Kcl, were calculated from the following equation, as described in previous reports 21'23 Cbr*
g c a or C1 - -
f~l~i~inCp*dt
where Cur* is the concentration of radioactivity in
317 brain parenchyma (dpm/g) at 10 min, Cp* is the plasma radioactivity (dpm/ml), andfis the fraction of isotope ionized in plasma (ionized/total, 0.5 for Ca and 1.0 for C1). CSF transfer coefficients were obtained by substituting CSF radioactivity at 10 min (dpm/g) for brain parenchyma radioactivity. Intravascular radioactivity was subtracted from total brain radioactivity to obtain brain parenchymal radioactivity (dpm/g). Intravascular radioactivity was calculated as the product of regional blood volume and the blood concentration of isotope at 10 min. Regional blood volume was determined from (dpm/g brain)/(dpm/g blood) of [3H]raffinose after 3 min. Raffinose had a K value of i x 10 -6 ml.g-l.s -1, less than that of 45Ca o r 36C121'23. If K values are calculated when the plasma concentration of isotope is much greater than the concentration in brain extracellular fluid or in CSF back diffusion can be ignored. We chose to do this by restricting circulation time to 10 min. Changes in cerebral blood flow can alter K, but only for rapidly penetrating compounds 19, with Kvalues near 1 x 10-3 ml.g-l.s -1. Neither Kca nor Kcl approximate this value 21,23. Regional brain K may reflect both cerebral capillary and choroid plexus transfer of isotope, with the distance from the ventricles determining the contribution of each route 22. Restriction of isotope circulation time to 10 min kept cortical gray regions relatively free of isotope contamination from C S F 22. One-way analysis of variance and Bonferroni ttests were used to compare multiple means TM. Levels of significance were set at P < 0.05. Integrated plasma radioactivity was calculated with a computer program that utilized the trapezoidal method 24. In order to perform this calculation, the plasma concentration at time 0 was estimated by visual approximation from a plot of plasma concentration vs time. The approximation could be altered by 100% and the integrated area would change by only 5%.
TABLE I
Body weight, blood pH, and plasma [Ca] Values are means _+ S.E.M. for 5 animals each. Diets consumed for 8 weeks were either deficient in Ca (LOCA) or sufficient in Ca at normal (CONT) or reduced (RCONT) intake.
Measurement
CONT
RCONT
LOCA
Body weight (g) Arterial blood pH Plasma ionized [Ca] (mmol/l)
227+8 7.40+0.01
114_+6a 7.38_+0.03
105+10 a 7.40+0.03
1.31+0.01
1.32+0.01
0.80+0.05 a'b
Plasma total [Ca] (mmol/l)
a Differs from value of CONT or Udiffers from value of RCONT by A N O V A (one-way) and Bonferroni t-test (twotailed), P < 0.05.
groups (Table I). Information on other ions in plasma, CSF, and brain after L O C A and CONT dietary intake can be found elsewhere TM. Mean blood pressure was significantly less in L O C A than in C O N T rats (108 + 2 torr, compared with 123 + 3). Fig. 1 plots m e a n 45Ca and 36C1 concentrations in plasma of CONT animals from 0.25 to 10 min after intravenous injection. Concentrations of these isotopes in plasma, blood, CSF, and 7 brain regions at
2000
O
Isotooe C o n c e n l r a t i o n at 1D min
X
(dz>rn/olx 10"3
45Ca
36CI
37.0 2 2.7 PARIETALCORTEX 1.2 i 0.1 CAUOATENUCLEUS 5.8 t 0.5 HIPPOCAMPUS 5.6 ¢ 0,5 THAI.AMUS-HYPOTHALJ~AUS 4,6 t; 0.3 MIDBR~QJN-COLLICULI 3.0 * 0+2 CEREBELLUM S.0 ± 0.6 PONS-MEDULLA 4.8 ± 0.5 BLOOD 244.0 2 ILl ~ P L A 1S .963 2M 11.4 CSF
E
1500
C
1000 ,
Q. "O v
~
A
118.0 ~: 7.0 10,0 ¢ 0.2 17.0 ± 1.0 15,$ Jt 0.4 15.0 ± 0.7 10.5 ± 0+5 16.1 1t 0.5 13.1 ± 0.6 288,6 ± $.0 397.3 i $.4
O C
g, X O
500 •
i m I
RESULTS
2.67+0.03 2.66_+0.05 1.48+_0.10a'u
o o. o
2.15
5 .Io
7.15
1 ol.o
Time (rain)
Rats fed L O C A diet for 8 weeks were about 50% lighter than rats fed C O N T diet (Table I), whereas R C O N T rats had similar weights to L O C A rats. Plasma [Ca]s, both total and ionized, were the same for CONT and R C O N T animals, but were 40% lower in L O C A animals. Blood pH was the same for all 3
Fig. 1. Arterial plasma concentrations of ~SCa and 36C1 after i.v. injection of 0.05 mCi/kg b. wt. into awake rats. Each point is the mean of 5 animals fed CONT diet. The points are connected by straight lines depicting the top of each trapezoid used to calculate the area under the curve, excluding the estimated points at 0 min. Mean concentrations + S.E.M. of 45Ca and 36C1 (n = 5) in brain parenchyma, CSF, blood, and plasma are listed for rats killed 10 min after isotope injection.
318 10 min are included in Fig. 1 as well. T h e brain values
TABLE III
are parenchymal concentrations (total - - intravascular). In general, the time courses of 45Ca and 36C1 in
Regional transfer coefficients for 36Cl
plasma of L O C A and R C O N T rats were similar to those in C O N T rats. T h e amounts of 45Ca and 36C1 in brain at 10 min were lowest in the parietal cortex and highest in caudate nucleus, whereas C S F concentra-
Values are means _+ S.E.M. for 5 animals each. Radioactive C1 was injected i.v. as a bolus into the femoral vein. Rats were killed at 10 min after injection.
CNS region
CONT
tions were substantially above those in the brain. Mean regional blood volumes as m e a s u r e d by [3H]raffinose ranged b e tw e e n 0.011 and 0.021 for various brain regions, and did not differ significantly between L O C A , C O N T and R C O N T . T h e values were similar to those obtained with [3H]inulin21. CSF concentrations of [3H]raffinose at 3 min were 1.38 + 0.2% of plasma for L O C A rats, 0.52 + 0.08 for R C O N T rats, and 0.33 _+ 0.04 for C O N T rats. Values of Kca at CSF and various brain regions are
Kcl (ml.g-l.s -1) ×lOs
CSF Parietalcortex Caudate nucleus Hippocampus Thalamushypothalamus Midbrain-colliculi Cerebellum Pons-medulla
RCONT
LOCA
42.30_+2.00 47.40_+1.70 52.80+2.00 a 3.60-+0.05 3.71-+0.14 3.24+0.28 6.10+0.32 5.19-+0.61 5.11+0.42 5.92+0.11 5.35-+0.12 5.28+0.29 5.38-+0.19 6.02-+0.24 3.78_+0.14 4.37_+0.13 5.77+0.16 6.54+0.39 4.71-+0.16 5.24_+0.26
5.35-+0.35 3.77-+0.37 5.46-+0.30 5.41_+0.29
a Differs from transfer coefficient of CONT by ANOVA (oneway) and Bonferroni t-test (two-tailed), P < 0.05.
listed in Table II for C O N T , R C O N T , or L O C A rats. Kca at CSF was substantially larger than at any brain
eral brain regions. Like values for Kca, KCI w a s great-
region in all diet groups. Transfer of 45Ca from blood
est in CSF and least in parietal cortex. No significant
was greatest into the caudate nucleus, hippocampus,
differences were found b e t w e e n the values of Kcl
and pons-medulla, and least into the parietal cortex
into the brain among the 3 groups. H o w e v e r , Kcl at
and midbrain-colliculi a m o n g the various brain re-
CSF was significantly higher in L O C A rats c o m p a r e d
gions. CSF and brain regional/(Ca values for L O C A
to C O N T rats, but not c o m p a r e d with R C O N T rats.
were significantly increased 4 0 - 6 0 % above corre-
If we assume some c o m m o n factors influence Kca
sponding values for C O N T and R C O N T , except at the parietal cortex, midbrain-colliculi, and cerebel-
and Kcl, such as surface area, CSF formation rate, and CSF or brain extracellular volumes, the ratio of
lum.
g c a over Kcl can be used to d et er m i n e if changes in
Table III displays the values of K a at C S F and sev-
Kca are specific for calcium. Calculated values for
Kca/KcI are illustrated in Table IV. Th e CSF and TABLE II
Regional transfer coefficients for 45Ca
TABLE IV
Values are means _+ S.E.M. for 5 animals each. Radioactive Ca was injected i.v. as a bolus into the femoral vein. Rats were killed at 10 min after injection.
Ratio of regional transfer coefficients"
CNS region
CNS region
Kca (ml.g-l.s -1) x lO5 CONT
CSF Parietal cortex Caudatenucleus Hippocampus Thalamushypothalamus Midbrain-colliculi Cerebellum Pons-medulla
RCONT
4.32_+0.10a'b 2.64-+0.13 4.48_+0.61 4.91-+0.21a'b
a Differs from transfer coefficient of CONT or b differs from transfer coefficient of RCONT by ANOVA (one-way) and Bonferroni t-test (two-tailed), P < 0.05.
KcjKcl CONT
LOCA
22.70+1.20 27.40_+1.40 41.90_+2.90a'b 0.77-+0.07 0.90-+0.15 0.83_+0.17 3.69+0.30 3.41-+0.45 5.18_+0.31a'b 3.53+0.29 2.91_+0.20 4.80_+0.40~'b 2.90-+0.20 32.7_+0.15 1.91+0.15 2.24-+0.28 3.17+0.35 3.96-+0.52 2.86_+0.25 3.55_+0.32
Values are means _+S.E.M. for 5 animals each.
CSF Parietal cortex Caudatenucleus Hippocampus Thalamushypothalamus Midbrain-colliculi Cerebellum Pons-medulla
RCONT
LOCA
0.54-+0.01 0.58-+0.01 0.21 _+0.02 0.24+0.03 0.60_+0.02 0.66+0.02 0.60-+0.05 0.54+0.03
0.79+0.04 a'b 0.25_+0.05 1.03-I-0.05a'b 0.91+0.07 a'b
0.54-+0.02 0.54-+0.02 0.50+0.02 0.51_+0.05 0.55+0.05 0.59-+0.05 0.60-+0.04 0.68_+0.05
0.82+0.05 a'b 0.71_+0.03a'b 0.81+0.07 a'b 0.91_+0.03a'b
a Differs from ratio of CONT or bdiffers from ratio of RCONT by ANOVA (one-way) and Bonferroni t-test (two-tailed), P < 0.05.
319 brain regional ratios of Kca/Kcl are significantly higher in L O C A animals compared to both CONT and RCONT rats with the exception of parietal cortex. DISCUSSION This paper demonstrates that in 11 week old rats, fed a low Ca diet for the prior 8 weeks and made hypocalcemic, the transfer coefficients (Kca) of calcium from blood into CSF and brain regions adjacent to ventricular CSF, are elevated. These increases in/(Ca are specific to Ca, as Kcl values into affected brain regions were unchanged and the Kca/Kcl ratio was elevated for both CSF and brain. Stability of CSF and brain [Ca] after chronic hypocalcemia 11,18, therefore, is partly due to a relatively constant rate of Ca entry from blood into the CNS, and Ca specificity suggests that active transport of Ca is altered. Finding that the Kca for the parietal cortex, a brain region distant from the ventricles where Ca is elaborated into CSF, remains unchanged during chronic hypocalcemia suggests that enhanced Ca transfer occurs at the choroid plexus and not at the cerebral capillaries. One way to maintain stable [Ca] in the CNS during chronic hypocalcemia is to keep the unidirectional flux of Ca into the brain and CSF relatively constant. Unidirectional flux, Jca, is the product of Kca and plasma ionized [Ca]. Jca into CSF for L O C A is 34 x 10-5 ymol.g-l.s -1, whereas for CONT and RCONT, it equals 30 x 10-5 and 36 x 10 -5/~mol.g-l.s -1, respectively. Flux of Ca into CSF is not different among the 3 groups, which implies Ca entry into CSF is regulated. On the other hand, parietal cortex uptake is linearly related to plasma [Ca], as was found by Tai et al. 3. Values of K for CSF depend on other factors, in addition to uptake. Smith and Rapoport 22 have described these in detail in relation to K values for 22Na and 36C1. One of these factors is CSF volume. L O C A animals had an apparent reduction in CSF volume, as sampling of CSF was more difficult in L O C A rats than in RCONT rats of similar size. A reduction in CSF volume might explain the increase in Kcl and Kca in these animals. Comparison of the ratios of Kca to Kcl suggests that factors in addition to a decreased CSF volume contributed to the elevated gca value in the CSF.
The concentration of [3H]raffinose in CSF, 3 min after injection, was much greater in L O C A than in either RCONT or CONT. The difference between the raffinose concentrations suggests blood contamination of CSF (1.38-0.33 = 1.05%). The product of 0.0105 and the concentration of 45Ca or 36C1in plasma represents radioactivity due to contamination. Subtraction of this estimate of contamination from total radioactivity in CSF of L O C A animals and recalculation yields values of 39 x 10-5 ml.g-l.s -1 for gca and 51.9 x 10-5 for Kcl, respectively. As CSF K values corrected for possible contamination are not different from the uncorrected values, the elevations of gca and K a at the CSF in L O C A rats are not due to blood contamination. Reduced CSF volume could at least partly explain the elevated [3H]raffinose concentration in L O C A rats. Our values for/(Ca and Kcl in awake rats are similar to prior measurements in anesthetized animals 21'23. Values for gel at CSF and various brain regions are the same with the exception of the parietal cortex which is 50-100% higher in our study. Dissec' tion or strain differences or a decrease in surface area by anesthesia 7 could explain this difference in Kcl values. Brain regional Kca was the same in our study and that of Tai et al. 23, but CSF/(Ca was 100% greater in our study (23 x 10-5 ml.g-l.s -1 vs 12 x 10-5). This difference could be due to a shorter circulation time, 10 min vs 15, or possibly anesthesia. Smith and Rapoport 22 suggested that if the times of isotope circulation were kept under 10 min, cortical regions distant from the choroid plexuses can be used to estimate cerebral capillary exchange, free of contamination by isotope from CSF. In our study, /(Ca for the parietal cortex did not significantly increase after chronic hypocalcemia. Autoradiographic data of both 45Ca and 36C1indicate that the major route of entry into CSF is across the choroid plexus, rather than the arachnoid or pial membranes 22'23. Diffusion of Ca from subarachnoid CSF into adjacent brain areas appears to be insignificant at 15 min 23. The elevations of Kca in other brain regions probably reflect isotope entering from ventricular CSF. Thus, a Caspecific transport system at the choroid plexus appears to regulate Ca entry into the CNS, whereas the cerebral capillaries restrict exchange. Indeed, some reports suggest that a carrier for Ca exists at the choroid plexus. In cats, Ames and Nes-
320 bitt found that Ca secreted by the choroid plexus is not an ultrafiltrate of plasma (quoted in Cserr6), whereas in rats, Barkai and Meltzer 2 demonstrated
tering Ca extrusion from the choroidal epithelium. Assuming this model is correct, previous results can be reconciled by noting that CSF [Ca] must change
regulated Ca uptake into CSF which could be in-
for the carrier to alter the rate of Ca transport. In all but one study 23, CSF [Ca] was changed 2A2. With ex-
hibited by R u t h e n i u m red, a Ca-ATPase inhibitor. The frog choroid plexus has a large paracellular pathway for Ca m o v e m e n t , but transcellular transport also exists 27. Barkai and Meltzer 2 suggest that a C a - A T P a s e is
tended hypocalcemia, CSF [Ca] begins to decrease. In response, the carrier may slow Ca expulsion from the choroid plexus, allowing more Ca to enter the CSF from blood. This process would be noted as an
located on the basolateral m e m b r a n e of the choroid plexus and can respond to changes in CSF [Ca] by al-
increase in Kca for CSF as in our study.
REFERENCES
14 Miller, R.G., Simultaneous Statistical Inference, McGrawHill, New York, 1966. 15 Morgulus, S. and Perley, A.M., Studies on cerebrospinal fluid and serum calcium with special reference to the parathyroid hormone, J. Biol. Chem., 88 (1930) 169-188. 16 Murphy, V.A. and Johanson, C.E., Adrenergic-induced enhancement of brain barrier system permeability to small nonelectrolytes: choroid plexus versus cerebral capillaries, J. Cereb. Blood Flow Metab., 5 (1985) 401-412. 17 Murphy, V.A. and Rapoport, S.I., Central nervous system uptake of Ca-45 in rats fed calcium deficient diet, Soc. Neurosci. Abstr., 13 (1987) 1249. 18 Murphy, V.A., Smith, Q.R. and Rapoport, S.I., Homeostasis of brain and cerebrospinal fluid calcium concentrations during chronic hypo- and hypercalcemia, J. Neurochem., 47 (1986) 1735-1741. 19 Ohno, K., Pettigrew, K.D. and Rapoport, S.I., Lower limits to cerebrovascular permeability to nonelectrolytes in the conscious rat, Am. J. Physiol., 235 (1978) H299-H307. 20 Rapoport, S.I., Blood-Brain Barrier in Physiology and Medicine, Raven, New York, 1976. 21 Smith, Q.R. and Rapoport, S.I., Carrier-mediated transport of chloride across the blood-brain barrier, J. Neurochem., 42 (198)I) 754-763. 22 Smith, Q.R. and Rapoport, S.I., Cerebrovascular permeability coefficients to sodium, potassium, and chloride, J. Neurochem., 46 (1986) 1732-1742. 23 Tai, C.-Y., Smith, Q.R. and Rapoport, S.I., Calcium influxes into brain and cerebrospinal fluid are linearly related to plasma ionized calcium concentration, Brain Research, 385 (1986) 227-236. 24 Wagner, J., Fundamentals of Clinical Pharmacokinetics, Hamilton, Hamilton, IL, 1975. 25 Wallach, S., Bellavia, J.V., Schorr, J. and Reizenstein, D.L., Tissue distribution of electrolytes, C a 47 and Mg28 in acute hypercalcemia, Am. J. Physiol., 207 (1964) 553-560. 26 Wong, R.P.K. and Bradbury, M.W.B., Permeability of the blood-brain barrier to calcium in adrenal insufficiency, Brain Research, 84 (1974) 361-364. 27 Wright, E., Transport processes in the formation of cerebrospinal fluid, Rev. Physiol. Biochem. Pharmacol., 83 (1978) 1-34.
1 Arieff, A.I. and Massry, S.G., Calcium metabolism of brain in acute renal failure: effects of uremia, hemodialysis, and parathyroid hormone, J. Clin. Invest., 53 (1974) 387-392. 2 Barkai, A.I. and Meltzer, H.L., Regulation of calcium entry into the extracellular environment of the rat brain, J. Neurosci., 2 (1982) 1322-1328. 3 Bradbury, M., The Concept of a Blood-Brain Barrier, Wiley, New York, 1979. 4 Cameron, A.T. and Moorhouse, V.H.K., The relation between plasma and cerebrospinal fluid calcium, J. Physiol. (Lond.), 91 (1937) 90-100. 5 Chiueh, C.C., Sun, C.L., Kopin, I.J., Fredericks, W.R. and Rapoport, S.I., Entry of [3H]norepinephrine, [125I]albumin and Evans blue from blood to brain following unilateral osmotic opening of the blood-brain barrier, Brain Research, 145 (1978) 291-301. 6 Cserr, H.F., Physiology of the choroid plexus, Physiol. Rev., 51 (1971) 273-311. 7 Gjedde, A. and Rasmussen, M., Pentobarbital anesthesia reduces blood-brain glucose transfer in the rat, J. Neurochem., 35 (1980) 1382-1387. 8 Goldstein, D.A., Romoff, M., Bogin, E. and Massry, S.G., Relationship between the concentration of calcium and phosphorous in blood and cerebrospinal fluid, J. Clin. Endocrinol. Metab., 49 (1979) 58-62. 9 Graziani, L., Escriva, A. and Katzman, R.A., Exchange of calcium between blood, brain and cerebrospinal fluid, Am. J. Physiol., 208 (1965) 1058-1064. 10 Graziani, L.J., Kaplan, R.K., Escriva, A. and Katzman, R., Calcium flux into CSF during ventricular and ventriculocisternal perfusion, Am. J. Physiol., 213 (1967) 629-636. 11 Harris, R.A., Carnes, D.L. and Forte, L.R., Reduction of brain calcium after consumption of diets deficient in calcium or vitamin D, J. Neurochem., 36 (1981) 460-466. 12 Herbert, F.K., The total and diffusible calcium of serum and the calcium of cerebrospinal fluid in human cases of hypocalcemia and hypercalcemia, J. Biochem., 27 (1934) 1978-1991. 13 Katzman, R. and Pappius, H.M., Brain Electrolytes and Fluid Metabolism, Williams and Wilkins, Baltimore, 1973.