- Email: [email protected]
Studies of the slow bidirectional transport of iron and transferrin across the blood-brain barrier
Brain Research
Bulletin,
Vol. 21, pp. 881-885. Pergamon Press
1988. Printed
plc,
0361~9230/88 $3.00 + .OO
in the U.S.A.
Studies of the Slow Bidirectional Transport of Iron and Transferrin Across the Blood-Brain Barrier WILLIAM A. BANKS,’ ABBA J. KASTIN, MELITA B. FASOLD, CARLOS M. BARRERA AND GENEVIEVE AUGEREAU
Veterans Administration Medical Center and Tulane University School of Medicine 1601 Perdido St., New Orleans, LA 70146 Received
BANKS,
W. A., A. J. KASTIN,
bidirectional
transport
M. B. FASOLD,
of iron and transferrin
across
7 June
1988
C. M. BARRERA AND G. AUGEREAU. Studies of the slow barrier. BRAIN RES BULL 21(6) 881-885, 1988.-
the blood-brain
Although iron is involved in brain function, very little is known about the regulation of its concentrations in the central nervous system. We quantitatively measured the entry and exit rates of iron, transferrin (its major transport protein), and albumin in mice. The blood to brain transport of iron > transferrin > albumin and the brain to blood transport of transferrin > albumin > iron. The results suggest that iron and transferrin have slow, bidirectional, probably saturable, and to some degree independent transport systems, although iron introduced directly into the brain is not readily available for brain to blood transport. Iron
Blood-brain barrier
Transferrin
Carrier-mediated
transport
Mice
injections of radioactive 50Fe+3 (1.1 Ci/mMol; New England Nuclear, Boston, MA), transferrin labeled with lZ51(9.2 Ci/mMol; Amersham, Arlington Heights, IL), or albumin labeled with ggmTc (8.5 Ci/mMol; Medi+Physics, Paramus, NJ) into the jugular vein in a volume of 0.2 ml. Ferric iron, the form typically used in the literature, was studied here instead of ferrous iron because it is the form transported by transferrin. Arterial blood was collected every 10 min in a separate group of mice from 10 to 90 min after injection from a severed, isolated carotid artery, centrifuged at 2500 g for 10 min at 4”C, and the serum removed. Log 1,, CPM/ml was plotted against time and, for those substances following first order kinetics, the half-time disappearance and volume of distribution determined from (a), the slope, and (b), the intercept of the line. The mice were decapitated immediately after collection of blood and the whole brain minus the pineal and pituitary removed. Arterial serum and whole brain samples were counted in a gamma counter for 3 min. Some mice were treated with desferoxamine (DFO) 100 mgikg IV 1 hr before the administration of iron. Blood to brain transport rates measured in ml/g-min (Ki) were determined by the graphical method as described by Patlak et al. (17) and applied by Blasberg et al. (3). Briefly, the slope of the linear portion of the curve relating brain to blood
IRON participates in the function of the central nervous system (CNS). The distribution of iron suggests interactions with GABAergic (10) and peptide systems (11) in addition to its well known role in energy metabolism. Derangements in the availability of iron are associated with adverse effects on activity and behavior in man and rat (15). Iron deficiency results in a selective decrease in Q-dopaminergic receptors, an increased sensitivity to the antinociceptive properties of morphine and opiate peptides, alterations in the transport of insulin, glucose, and valine across the blood-brain barrier (BBB), and learning disabilities not due to its effects on hemoglobin (2, 21, 22). Despite these observations, very little is known about the movement of iron into or out of the CNS. We, therefore, studied the quantitative rate of entry into and the exit out of the brain of iron and its transport protein, transferrin, in mice. METHOD
Blood to Brain Transport
The right carotid artery and the left jugular vein were exposed in male, 17-25 g ICR mice (Blue Spruce Farms, Altamont, NY) anesthetized IP with urethane. Mice received
‘Requests for reprints should be addressed to William Banks, Veterans Administration 70146.
881
Medical Center, 1601 Perdido St., New Orleans, LA
BAN KS Ls‘7‘ A f,.
XX?
Transferrin
F?+++
FIG. 1. Halftime disappearance from the bIood of radioactive iron, Fe++ * (DFO), transferrin, and albumin. All compounds appeared to follow first order kinetics over the time period plotted with half times of 139 min, 101 min, 136 min, and 70 min respectively.
ratios (y axis) to their respective exposure times (x axis) gives the Ki, expressed in ml/g (of brain)-min. Exposure time is the value in blood integrated from 0 min to a given time divided by the actual blood value at that time. For those substances with a half-time disappearance from the blood that follows first order kinetics, a simplified equation can be used to find exposure time. For such substances, the integrated blood value between time 0 min (to) to any given time (t) is given by the equation:
(1) which can be simplified algebraically to:
=
(t)#atiz+,‘t~
(2)
Dividing the integrated value by the actual value in blood at that time gives exposure time: Exposure time = [(t)10’at’2+“)]iI~a’+h’,
(3)
which simplifies algebraically to:
studies, mice were studied 60 min after injection and the results expressed as a brain/blood ratio in microllg of brain.
The method used was one adapted from Noble, et al. (16) and shown to be quantitatively reliable (1). The craniums were exposed in mice anesthetized with urethane. After incision of the scalp, a hole [email protected] mm deep was made 1.Omm lateral and 1.0 mm posterior to the bregma with a 26 gauge guarded syringe. An injection of 1.0 microiiter of lactated Ringer’s solution containing 25,000 CPM of radioactive material (iron, transfer& or albumin) was made through this hole into the lateral ventricle (ICV). Delivery with this method has a coefficient of variation of about 5% (1). Mice were decapitated between 2 min and 2 hr later and the brains were removed for counting in a gamma counter. Half time disappearances were determined from the relationship between log&PM rem~~ng in the brain and time. The amount of material available for transport (Al was taken to be the antilog of the intercept of this line. The percent of available material transported (ST) at any time t can then be determined: %T = lOO(A-MYA
Some mice received injections containing combinations of “HmTc-albumin, radioactive iron, and transferrin to directly compare the relative permeability of these compounds. Other mice were tested for a saturable component of the blood to brain transport of iron by injections of both radioactive iron and 9smTc-albumin with or without nonradioactive iron (200 pmol/kg or 2 micromo~kg). In these
(5)
where M is the CPM of material measured or computed to be in the brain at time t. Some mice were treated 1 hr before the ICV injection with IV DFO or with IP aluminum chloride (100 mg&g of elemental aluminum), which has been shown to interfere with the binding of iron to tr~sfe~n (6,20). Other mice received 10 nmol of nonradioactive iron in their ICV injections. The above method for determining %T, the percent of
IRON T~NSPORT
883
ACROSS THE BBB
Fe
Fe +++
l
+*
1
Transferrin
Albumin
( OF0 ) 0050
I
* 070
.
0 .
l
oo:* ~ 0
Exposure
Time
JO
60
90
!ZO
( min i
FIG. 2. Blood to brain transport. Rates (Ki’s), calculated as described in the text, were (1.17~0.15)10-4 ml/g-min (iron), (1.69~0.20)10-* ml/g-min (iron in DFO treated mice), (0.97kO.28)10-’ ml/g-min (transferrin), and (0.65~?0.32)10-4 ml/gmin (albumin).
material transported from the brain to blood, is based on residual CPM in the brain and can be checked for accuracy by comparing its results to those obtained by a method based on the appearance of CPM in the blood after ICV injection of material. Mice were injected ICV with about 2~,~ CPM of radioactive material and an arterial serum sample obtained 60 min later. The infusion rate Q (which in this case represents the brain to blood transport rate) needed to achieve the observed levels in serum can be calculated for a material with a disappearance from the blood that follows first order kinetics by the equation: Q = PKVd/(l-eWXt),
(6)
where P is the CPM/ml of arterial serum, K is the inverse of the half-time disappearance in blood multiplied by 0.693, Vd is the volume of distribution after IV injection, and t is the time from injection to decapitation (60 min). The percent of available material transported as determined from the ap pearance of material in the blood (%Q) is taken to be: %Q = (l~)Q(t)2S~/Ai
(7)
and can be compared with %T, the percent of available material transported as determined from the disappearance of material from the brain. Statistics Regression lines were computed by the least squares method and compared statistically with the BMDPlR program (BMDP Statistical Software, Los Angeles, CA). Means were initially compared by analysis of variance (ANOVA)
and then, if there were more than two means, by Duncan’s multiple range test (DMRT).
RESULTS
Blood to Brain Transport The disappearance from the blood of radioactive iron, iron in DFO treated mice, transferrin, and albumin followed first order kinetics (Fig. 1) with hares of 139 min, 101 min, 136 min, and 70 min and Vd’s of 1.44 ml, 1.44 ml, 1.42 ml, and 1.51 ml respectively. The plots between the brain to blood ratios and exposure times showed a linear relationship for the first 60 min after IV injection for iron, transferrin, and albumin and up to 90 min for iron in DFO treated mice (Fig. 2). The Ki’s (expressed in ml/g-min) were (1.17*0.15)10-4, (1.69+-0.20)10-4, (0.97-tO.28)10-*, and (0.65+0.32)10W4 with Vi’s (a measurement of the rapidly exchangeable space between the brain and blood expressed in ml/g) of 0.011, 0.011, 0.012, and 0.013 for iron, iron in DFO treated mice, transfertin, and albumin, respectively, ANOVA showed significant differences among these lines, F(6,19)=5.#, p
BANKS E’l AL.
ransferrtn 25 -
\
. ;
3b
-
Bo Time
9b
Go
(min)
FIG. 3. Half time disappearance from the brain of radioactive iron, transferrin, and albumin in otherwise untreated mice and of iron in DFO treated mice. Half times were 669 min. 30.1 min. 45.7 min, and 107 min, respectively.
TABLE
1
TRANSPORT OF MATERIAL FROM THE BRAIN TO THE BLOOD (EXPRESSED AS A PERCENT OF THE AMOUNT AVAILABLE) AS DETERMINED BY THE DISAPPEARANCE RATE FROM THE BRAIN (%T) OR THE APPEARANCE IN BLOOD (%Q)
89, I r 1.3% precipitated with 3oLT trifmoroacetic acid, indicating that it was the intact molecule that had been transported. DISCUSSION
Compound Transferrin Albumin Iron
the two compounds, was not significant.
74.9 59.9 6.1
74.7 61.0 11.7
F( 1,20)= 16.5,p
Brain to Blood Transport
The rates of disappearance from the brain after ICV injection (Fig. 2) had slopes of -0.00045~0.00012 (half life of 669 min) for iron, -0.00282~0.00045 (half life of 107 min) for iron in mice treated with DFO, -O.OlOO+-0.0016 (half life of 30.1 min) for transferrin, and -0.0066+0.0013 (half life of 45.7 min) for albumin (Fig. 3). The DMRT done on these 4 groups showed that each comparison differed significantly (~~0.05) except that for iron vs. DFO. Pretreatment of mice with aluminum or inclusion of non~d~oactive iron in the ICV injection did not appear to alter brain to blood transport. The percent of available material transported from the brain to the blood (%T) after 1 hr as determined from the half time disappearance rate from the brain was compared to %Q for iron (6.1 vs. 11.7), transferrin (74.9 vs. 74.7), and albumin (59.9 vs. 61.0), Table 1. The relationship was: %T= l.O9(%Q) - 6.68, r=.999, p
These studies quantitatively determined the blood to brain and brain to blood transport rates of radioactive iron, transferrin, and albumin. They showed an order of rates that differed with the direction of transport: iron > transfer& > albumin for blood to brain transport, but t~nsfer~n > albumin > iron for brain to blood transport. For blood to brain transport, this order was found for both the Ki’s and for the brain to blood ratios after simultaneous injection of the compounds, which allowed direct comparison of two compounds within the same animal. For brain to blood transport, this order was found for both the half time disappearance rate from the brain and for %Q which is based on the rate of appearance in the circulation of ICV-injected material. The low transport rate of iron into the brain was not unexpected since almost all of a peripherally administered dose of iron can be found in the blood, liver, and a few other tissues (5). The transport rate did exceed that of transferrin, however, as has been found for some peripheral tissues (5). Since Ki measures only the unidirectional influx (blood to brain transport), this difference between the transport rates of iron and transferrin cannot be explained by different rates of efflux (brain to blood transport), and suggests that at least some iron is taken up at the BBB independently of transferrin uptake. By contrast, the brain to blood transport of iron was much slower than that of albumin, a substance commonly used as an indicator of the reabsorption of cerebrospinal fluid (bulk flow). This indicates that iron introduced directly into the brain is not readily available for brain to blood transport. This retention cannot be explained by a rapid influx masking the effhrx, since the Ki for iron was low. Paradoxically, the
IRON TRANSPORT
885
ACROSS THE BBB
rate of influx of iron was noted to decrease about 60 min after IV injection, consistent with a significant efflux by that time. An intermediate step may occur between the initial uptake of iron and the subsequent avid retention, This step probably would be saturable, since pretreatment with DFO, which decreases free iron (19), extended the time during which the influx rate of iron was linear, and could be the main determinant in the exchange rates of iron between brain and blood. The finding that unlabeled iron did not significantly affect transport in either direction suggests that both processes may be near saturation under normal conditions. The results are also consistent with a saturable transport of transferrin across the BBB. Both the entry and exit of transferrin slightly exceeded that of albumin, even though transferrin is a much larger molecule and, therefore, might be expected to move across the BBB with more difficulty. However, the similarities between the brain/blood ratios at 60 min and the Vi’s as well as the low Ki’s indicate a very low blood to brain penetration rate for these two compounds. Jefferies et al. (7) showed that transferrin receptors exist on brain capillaries; transcytosis of transferrin was found by Fishman et al. (7,8) and confirmed by Pardridge et al. (18). Our results suggest that transport of transferrin across the BBB is bidirectional. The Ki that we obtained for transferrin (0.975 x 10m4)is similar to that which can be computed from the data of Fishman et al. (8). They found 22,000 CPM in the brain after a 60 min infusion, which gives a PA=0.611 x 10e4. At low entry rates, PA and Ki are equivalent (3). Transferrin is synthesized in the CNS (4,14)
and has a different pattern of distribution within the brain from that of iron (11,12). Our findings that the blood to brain transport of iron exceeded that of transferrin while brain to blood transport of iron was much slower than that of transferrin or even albumin is consistent with a role for transferrin in the CNS that may include actions in addition to that of transporting iron. Finally, the excellent agreement between %Q and %T shows that the method used here to determine %T is quantitatively accurate. This method was developed to measure the brain to blood transport of materials. Its quantitative accuracy has only been determined previously for one other compound (1). The method does not assume the location or number of sites within the CNS involved in transport, nor does it make assumptions about the individual components of transport (CSF to blood, blood to CSF, brain to blood, blood to brain, brain to CSF, or CSF to brain). It measures the net movement of material from the CNS to the peripheral circulation. The similar results obtained for %T and %Q indicates that the accuracy of the method was not impaired for these compounds even though some areas of the CNS, such as the spinal cord, were not included in the sample.
ACKNOWLEDGEMENTS
We would like to thank Dr. 0. A. Correa, Dr. 0. M. Garcia, and C. L. Gaspard for their kind help in acquisition of materials and Karl Fasold for graphics. Supported by the VA.
REFERENCES 1. Banks, W. A.; Kastin, A. J.; Horvath, A.; Michals, E. A. Carrier-mediated transport of vasopressin across the bloodbrain barrier of the mouse. J. Neurosci. Res. 18:326332; 1987. 2. Ben-Shachar, D.; Yehuda, S.; Finberg, J. P. M.; Spanier, 1.; Youdim, M. B. H. Selective alteration in blood-brain barrier and insulin transport in iron-deficient rats. J. Neurochem. 50: 1434-1437; 1988. 3. Blasberg, R. G.; Fenstermacher, J. D.; Patlak, C. S. Transport of a-aminoisobutyric acid across brain capillary and cellular membranes. J. Cereb. Blood Flow Metab. 38-32; 1983. 4. Bloch, B.; Popovici, T.; Chouham, S.; Levin, M. J.; Tuil, D.; Kahn, A. Transferrin gene expression in choroid plexus of the adult rat brain. Brain Res. Bull. 18:573-576; 1987. 5. Cheney, B. A.; Lothe, K.; Morgan, E. H.; Sood, S. K.; Finch, C. A. Internal iron exchange in the rat. Am. J. Physiol. 212:376-380; 1967. 6. Co&ran, M.; Coates, J.; Neoh, S. The competitive equilibrium between aluminum and ferric ions for the binding sites of transferrin. FEBS Lett. 176:12%132; 1984. 7. Fishman, J. B.; Handrahan, J. V.; Connor, J.; Dickey, B. F.; Fine, R. E. Receptor-mediated transcytosis of transferrin across the blood-brain barrier. J. Cell Biol. 101:423a: 1985. 8. Fishman, J. B.; Rubin, J. B.; Hand&an, J. V.; Connor, J. R.; Fine, R. E. Receptor-mediated transcytosis of transferrin across the blood-brain barrier. J. Neurosci. Res. 18:299-304; 1987. 9. Fleming, J.; Joshi, J. G. Ferritin: isolation of aluminum-ferritin complex from brain. Proc. Natl. Acad. Sci. USA 84:78667870; 1987. 10. Francois, C.; Nguyen-Legros, J.; Percheron, G. Topographical and cytological localization of iron in rat and monkey brains. Brain Res. 215:317-322; 1981. 11. HiIl, J. M.; Switzer, R. C. The regional distribution and cellular localization of iron in the rat brain. Neuroscience 11:595-603; 1984.
12. Hill, J. M.; Ruff, M. R.; Weber, R. J.; Pert, C. B. Transferrin receptors in rat brain: neuropeptide-like pattern and relationship to iron distribution. Proc. Natl. Acad. Sci. USA 82:4553-4557; 1985. 13. Jefferies, W. A.; Brandon, M. R.; Hunt, S. V.; Williams, A. F.; receptor on Gatter, K. C.; Mason, D. Y. Transfetrin endothelium of brain capillaries. Nature 312:162-163; 1984. 14. Levin, M. J.; Tuil, D.; Uzan, G.; Dreyfus, J. C.; Kahn, A. Expression of the transferrin gene during development of nonhepatic tissues: high level of transferrin mRNA in fetal muscle and adult brain. Biochem. Biophys. Res. Commun. 122:212217; 1984. 15. Lozoff, B.; Brittenham, G. M. Behavioral aspects of iron deticiency. In: Brown, E. B., ed. Progress in hematology. vol. XIV. Orlando: Harcourt Brace Jovanovich; 1986:23-53. 16. Noble, E. P.; Wurtman, A. J.; Axelrod, J. A simple and rapid method for injecting [H3]norepinephrine into the lateral ventricle of the rat brain. Life Sci. 6:281-291; 1967. 17. Patlak, C. S.; Blasberg, R. G.; Fenstermacher, J. D. Graphical evaluation of blood-to-brain transfer constants from multipletime uptake data. J. Cereb. Blood Flow Metab. 3:1-7; 1983. 18. Pardridge, W. M.; Eisenberg, J.; Yang, J. Human blood-brain barrier transfenin receptor. Metabolism 36:892-895; 1987. 19. Peters, G.; Keberle, H.; Schmid, K.; Brunner, H. Distribution and renal excretion of desferrioxamine and ferrioxamine in the dog and in the rat. Biochem. Pharmacol. 15:93-109; 1%6. 20. Trapp, G. A. Plasma aluminum is bound to transfenin. Life Sci. 33:311-316; 1983. 21. Yehuda, S.; Youdim, M. B. H. The increased opiate action of p-endorphin in iron-deficit rats: the possible involvement of dopamine. Eur. J. Pharmacol. 104:245-251; 1984. 22. Yehuda, S.; Youdim, M. B. H.; Mostofsky, D. I. Brain irondeficiency causes reduced learning capacity in rats. Pharmacol. Biochem. Behav. 25:141-144; 1986.
Bulletin,
Vol. 21, pp. 881-885. Pergamon Press
1988. Printed
plc,
0361~9230/88 $3.00 + .OO
in the U.S.A.
Studies of the Slow Bidirectional Transport of Iron and Transferrin Across the Blood-Brain Barrier WILLIAM A. BANKS,’ ABBA J. KASTIN, MELITA B. FASOLD, CARLOS M. BARRERA AND GENEVIEVE AUGEREAU
Veterans Administration Medical Center and Tulane University School of Medicine 1601 Perdido St., New Orleans, LA 70146 Received
BANKS,
W. A., A. J. KASTIN,
bidirectional
transport
M. B. FASOLD,
of iron and transferrin
across
7 June
1988
C. M. BARRERA AND G. AUGEREAU. Studies of the slow barrier. BRAIN RES BULL 21(6) 881-885, 1988.-
the blood-brain
Although iron is involved in brain function, very little is known about the regulation of its concentrations in the central nervous system. We quantitatively measured the entry and exit rates of iron, transferrin (its major transport protein), and albumin in mice. The blood to brain transport of iron > transferrin > albumin and the brain to blood transport of transferrin > albumin > iron. The results suggest that iron and transferrin have slow, bidirectional, probably saturable, and to some degree independent transport systems, although iron introduced directly into the brain is not readily available for brain to blood transport. Iron
Blood-brain barrier
Transferrin
Carrier-mediated
transport
Mice
injections of radioactive 50Fe+3 (1.1 Ci/mMol; New England Nuclear, Boston, MA), transferrin labeled with lZ51(9.2 Ci/mMol; Amersham, Arlington Heights, IL), or albumin labeled with ggmTc (8.5 Ci/mMol; Medi+Physics, Paramus, NJ) into the jugular vein in a volume of 0.2 ml. Ferric iron, the form typically used in the literature, was studied here instead of ferrous iron because it is the form transported by transferrin. Arterial blood was collected every 10 min in a separate group of mice from 10 to 90 min after injection from a severed, isolated carotid artery, centrifuged at 2500 g for 10 min at 4”C, and the serum removed. Log 1,, CPM/ml was plotted against time and, for those substances following first order kinetics, the half-time disappearance and volume of distribution determined from (a), the slope, and (b), the intercept of the line. The mice were decapitated immediately after collection of blood and the whole brain minus the pineal and pituitary removed. Arterial serum and whole brain samples were counted in a gamma counter for 3 min. Some mice were treated with desferoxamine (DFO) 100 mgikg IV 1 hr before the administration of iron. Blood to brain transport rates measured in ml/g-min (Ki) were determined by the graphical method as described by Patlak et al. (17) and applied by Blasberg et al. (3). Briefly, the slope of the linear portion of the curve relating brain to blood
IRON participates in the function of the central nervous system (CNS). The distribution of iron suggests interactions with GABAergic (10) and peptide systems (11) in addition to its well known role in energy metabolism. Derangements in the availability of iron are associated with adverse effects on activity and behavior in man and rat (15). Iron deficiency results in a selective decrease in Q-dopaminergic receptors, an increased sensitivity to the antinociceptive properties of morphine and opiate peptides, alterations in the transport of insulin, glucose, and valine across the blood-brain barrier (BBB), and learning disabilities not due to its effects on hemoglobin (2, 21, 22). Despite these observations, very little is known about the movement of iron into or out of the CNS. We, therefore, studied the quantitative rate of entry into and the exit out of the brain of iron and its transport protein, transferrin, in mice. METHOD
Blood to Brain Transport
The right carotid artery and the left jugular vein were exposed in male, 17-25 g ICR mice (Blue Spruce Farms, Altamont, NY) anesthetized IP with urethane. Mice received
‘Requests for reprints should be addressed to William Banks, Veterans Administration 70146.
881
Medical Center, 1601 Perdido St., New Orleans, LA
BAN KS Ls‘7‘ A f,.
XX?
Transferrin
F?+++
FIG. 1. Halftime disappearance from the bIood of radioactive iron, Fe++ * (DFO), transferrin, and albumin. All compounds appeared to follow first order kinetics over the time period plotted with half times of 139 min, 101 min, 136 min, and 70 min respectively.
ratios (y axis) to their respective exposure times (x axis) gives the Ki, expressed in ml/g (of brain)-min. Exposure time is the value in blood integrated from 0 min to a given time divided by the actual blood value at that time. For those substances with a half-time disappearance from the blood that follows first order kinetics, a simplified equation can be used to find exposure time. For such substances, the integrated blood value between time 0 min (to) to any given time (t) is given by the equation:
(1) which can be simplified algebraically to:
=
(t)#atiz+,‘t~
(2)
Dividing the integrated value by the actual value in blood at that time gives exposure time: Exposure time = [(t)10’at’2+“)]iI~a’+h’,
(3)
which simplifies algebraically to:
studies, mice were studied 60 min after injection and the results expressed as a brain/blood ratio in microllg of brain.
The method used was one adapted from Noble, et al. (16) and shown to be quantitatively reliable (1). The craniums were exposed in mice anesthetized with urethane. After incision of the scalp, a hole [email protected] mm deep was made 1.Omm lateral and 1.0 mm posterior to the bregma with a 26 gauge guarded syringe. An injection of 1.0 microiiter of lactated Ringer’s solution containing 25,000 CPM of radioactive material (iron, transfer& or albumin) was made through this hole into the lateral ventricle (ICV). Delivery with this method has a coefficient of variation of about 5% (1). Mice were decapitated between 2 min and 2 hr later and the brains were removed for counting in a gamma counter. Half time disappearances were determined from the relationship between log&PM rem~~ng in the brain and time. The amount of material available for transport (Al was taken to be the antilog of the intercept of this line. The percent of available material transported (ST) at any time t can then be determined: %T = lOO(A-MYA
Some mice received injections containing combinations of “HmTc-albumin, radioactive iron, and transferrin to directly compare the relative permeability of these compounds. Other mice were tested for a saturable component of the blood to brain transport of iron by injections of both radioactive iron and 9smTc-albumin with or without nonradioactive iron (200 pmol/kg or 2 micromo~kg). In these
(5)
where M is the CPM of material measured or computed to be in the brain at time t. Some mice were treated 1 hr before the ICV injection with IV DFO or with IP aluminum chloride (100 mg&g of elemental aluminum), which has been shown to interfere with the binding of iron to tr~sfe~n (6,20). Other mice received 10 nmol of nonradioactive iron in their ICV injections. The above method for determining %T, the percent of
IRON T~NSPORT
883
ACROSS THE BBB
Fe
Fe +++
l
+*
1
Transferrin
Albumin
( OF0 ) 0050
I
* 070
.
0 .
l
oo:* ~ 0
Exposure
Time
JO
60
90
!ZO
( min i
FIG. 2. Blood to brain transport. Rates (Ki’s), calculated as described in the text, were (1.17~0.15)10-4 ml/g-min (iron), (1.69~0.20)10-* ml/g-min (iron in DFO treated mice), (0.97kO.28)10-’ ml/g-min (transferrin), and (0.65~?0.32)10-4 ml/gmin (albumin).
material transported from the brain to blood, is based on residual CPM in the brain and can be checked for accuracy by comparing its results to those obtained by a method based on the appearance of CPM in the blood after ICV injection of material. Mice were injected ICV with about 2~,~ CPM of radioactive material and an arterial serum sample obtained 60 min later. The infusion rate Q (which in this case represents the brain to blood transport rate) needed to achieve the observed levels in serum can be calculated for a material with a disappearance from the blood that follows first order kinetics by the equation: Q = PKVd/(l-eWXt),
(6)
where P is the CPM/ml of arterial serum, K is the inverse of the half-time disappearance in blood multiplied by 0.693, Vd is the volume of distribution after IV injection, and t is the time from injection to decapitation (60 min). The percent of available material transported as determined from the ap pearance of material in the blood (%Q) is taken to be: %Q = (l~)Q(t)2S~/Ai
(7)
and can be compared with %T, the percent of available material transported as determined from the disappearance of material from the brain. Statistics Regression lines were computed by the least squares method and compared statistically with the BMDPlR program (BMDP Statistical Software, Los Angeles, CA). Means were initially compared by analysis of variance (ANOVA)
and then, if there were more than two means, by Duncan’s multiple range test (DMRT).
RESULTS
Blood to Brain Transport The disappearance from the blood of radioactive iron, iron in DFO treated mice, transferrin, and albumin followed first order kinetics (Fig. 1) with hares of 139 min, 101 min, 136 min, and 70 min and Vd’s of 1.44 ml, 1.44 ml, 1.42 ml, and 1.51 ml respectively. The plots between the brain to blood ratios and exposure times showed a linear relationship for the first 60 min after IV injection for iron, transferrin, and albumin and up to 90 min for iron in DFO treated mice (Fig. 2). The Ki’s (expressed in ml/g-min) were (1.17*0.15)10-4, (1.69+-0.20)10-4, (0.97-tO.28)10-*, and (0.65+0.32)10W4 with Vi’s (a measurement of the rapidly exchangeable space between the brain and blood expressed in ml/g) of 0.011, 0.011, 0.012, and 0.013 for iron, iron in DFO treated mice, transfertin, and albumin, respectively, ANOVA showed significant differences among these lines, F(6,19)=5.#, p
BANKS E’l AL.
ransferrtn 25 -
\
. ;
3b
-
Bo Time
9b
Go
(min)
FIG. 3. Half time disappearance from the brain of radioactive iron, transferrin, and albumin in otherwise untreated mice and of iron in DFO treated mice. Half times were 669 min. 30.1 min. 45.7 min, and 107 min, respectively.
TABLE
1
TRANSPORT OF MATERIAL FROM THE BRAIN TO THE BLOOD (EXPRESSED AS A PERCENT OF THE AMOUNT AVAILABLE) AS DETERMINED BY THE DISAPPEARANCE RATE FROM THE BRAIN (%T) OR THE APPEARANCE IN BLOOD (%Q)
89, I r 1.3% precipitated with 3oLT trifmoroacetic acid, indicating that it was the intact molecule that had been transported. DISCUSSION
Compound Transferrin Albumin Iron
the two compounds, was not significant.
74.9 59.9 6.1
74.7 61.0 11.7
F( 1,20)= 16.5,p
Brain to Blood Transport
The rates of disappearance from the brain after ICV injection (Fig. 2) had slopes of -0.00045~0.00012 (half life of 669 min) for iron, -0.00282~0.00045 (half life of 107 min) for iron in mice treated with DFO, -O.OlOO+-0.0016 (half life of 30.1 min) for transferrin, and -0.0066+0.0013 (half life of 45.7 min) for albumin (Fig. 3). The DMRT done on these 4 groups showed that each comparison differed significantly (~~0.05) except that for iron vs. DFO. Pretreatment of mice with aluminum or inclusion of non~d~oactive iron in the ICV injection did not appear to alter brain to blood transport. The percent of available material transported from the brain to the blood (%T) after 1 hr as determined from the half time disappearance rate from the brain was compared to %Q for iron (6.1 vs. 11.7), transferrin (74.9 vs. 74.7), and albumin (59.9 vs. 61.0), Table 1. The relationship was: %T= l.O9(%Q) - 6.68, r=.999, p
These studies quantitatively determined the blood to brain and brain to blood transport rates of radioactive iron, transferrin, and albumin. They showed an order of rates that differed with the direction of transport: iron > transfer& > albumin for blood to brain transport, but t~nsfer~n > albumin > iron for brain to blood transport. For blood to brain transport, this order was found for both the Ki’s and for the brain to blood ratios after simultaneous injection of the compounds, which allowed direct comparison of two compounds within the same animal. For brain to blood transport, this order was found for both the half time disappearance rate from the brain and for %Q which is based on the rate of appearance in the circulation of ICV-injected material. The low transport rate of iron into the brain was not unexpected since almost all of a peripherally administered dose of iron can be found in the blood, liver, and a few other tissues (5). The transport rate did exceed that of transferrin, however, as has been found for some peripheral tissues (5). Since Ki measures only the unidirectional influx (blood to brain transport), this difference between the transport rates of iron and transferrin cannot be explained by different rates of efflux (brain to blood transport), and suggests that at least some iron is taken up at the BBB independently of transferrin uptake. By contrast, the brain to blood transport of iron was much slower than that of albumin, a substance commonly used as an indicator of the reabsorption of cerebrospinal fluid (bulk flow). This indicates that iron introduced directly into the brain is not readily available for brain to blood transport. This retention cannot be explained by a rapid influx masking the effhrx, since the Ki for iron was low. Paradoxically, the
IRON TRANSPORT
885
ACROSS THE BBB
rate of influx of iron was noted to decrease about 60 min after IV injection, consistent with a significant efflux by that time. An intermediate step may occur between the initial uptake of iron and the subsequent avid retention, This step probably would be saturable, since pretreatment with DFO, which decreases free iron (19), extended the time during which the influx rate of iron was linear, and could be the main determinant in the exchange rates of iron between brain and blood. The finding that unlabeled iron did not significantly affect transport in either direction suggests that both processes may be near saturation under normal conditions. The results are also consistent with a saturable transport of transferrin across the BBB. Both the entry and exit of transferrin slightly exceeded that of albumin, even though transferrin is a much larger molecule and, therefore, might be expected to move across the BBB with more difficulty. However, the similarities between the brain/blood ratios at 60 min and the Vi’s as well as the low Ki’s indicate a very low blood to brain penetration rate for these two compounds. Jefferies et al. (7) showed that transferrin receptors exist on brain capillaries; transcytosis of transferrin was found by Fishman et al. (7,8) and confirmed by Pardridge et al. (18). Our results suggest that transport of transferrin across the BBB is bidirectional. The Ki that we obtained for transferrin (0.975 x 10m4)is similar to that which can be computed from the data of Fishman et al. (8). They found 22,000 CPM in the brain after a 60 min infusion, which gives a PA=0.611 x 10e4. At low entry rates, PA and Ki are equivalent (3). Transferrin is synthesized in the CNS (4,14)
and has a different pattern of distribution within the brain from that of iron (11,12). Our findings that the blood to brain transport of iron exceeded that of transferrin while brain to blood transport of iron was much slower than that of transferrin or even albumin is consistent with a role for transferrin in the CNS that may include actions in addition to that of transporting iron. Finally, the excellent agreement between %Q and %T shows that the method used here to determine %T is quantitatively accurate. This method was developed to measure the brain to blood transport of materials. Its quantitative accuracy has only been determined previously for one other compound (1). The method does not assume the location or number of sites within the CNS involved in transport, nor does it make assumptions about the individual components of transport (CSF to blood, blood to CSF, brain to blood, blood to brain, brain to CSF, or CSF to brain). It measures the net movement of material from the CNS to the peripheral circulation. The similar results obtained for %T and %Q indicates that the accuracy of the method was not impaired for these compounds even though some areas of the CNS, such as the spinal cord, were not included in the sample.
ACKNOWLEDGEMENTS
We would like to thank Dr. 0. A. Correa, Dr. 0. M. Garcia, and C. L. Gaspard for their kind help in acquisition of materials and Karl Fasold for graphics. Supported by the VA.
REFERENCES 1. Banks, W. A.; Kastin, A. J.; Horvath, A.; Michals, E. A. Carrier-mediated transport of vasopressin across the bloodbrain barrier of the mouse. J. Neurosci. Res. 18:326332; 1987. 2. Ben-Shachar, D.; Yehuda, S.; Finberg, J. P. M.; Spanier, 1.; Youdim, M. B. H. Selective alteration in blood-brain barrier and insulin transport in iron-deficient rats. J. Neurochem. 50: 1434-1437; 1988. 3. Blasberg, R. G.; Fenstermacher, J. D.; Patlak, C. S. Transport of a-aminoisobutyric acid across brain capillary and cellular membranes. J. Cereb. Blood Flow Metab. 38-32; 1983. 4. Bloch, B.; Popovici, T.; Chouham, S.; Levin, M. J.; Tuil, D.; Kahn, A. Transferrin gene expression in choroid plexus of the adult rat brain. Brain Res. Bull. 18:573-576; 1987. 5. Cheney, B. A.; Lothe, K.; Morgan, E. H.; Sood, S. K.; Finch, C. A. Internal iron exchange in the rat. Am. J. Physiol. 212:376-380; 1967. 6. Co&ran, M.; Coates, J.; Neoh, S. The competitive equilibrium between aluminum and ferric ions for the binding sites of transferrin. FEBS Lett. 176:12%132; 1984. 7. Fishman, J. B.; Handrahan, J. V.; Connor, J.; Dickey, B. F.; Fine, R. E. Receptor-mediated transcytosis of transferrin across the blood-brain barrier. J. Cell Biol. 101:423a: 1985. 8. Fishman, J. B.; Rubin, J. B.; Hand&an, J. V.; Connor, J. R.; Fine, R. E. Receptor-mediated transcytosis of transferrin across the blood-brain barrier. J. Neurosci. Res. 18:299-304; 1987. 9. Fleming, J.; Joshi, J. G. Ferritin: isolation of aluminum-ferritin complex from brain. Proc. Natl. Acad. Sci. USA 84:78667870; 1987. 10. Francois, C.; Nguyen-Legros, J.; Percheron, G. Topographical and cytological localization of iron in rat and monkey brains. Brain Res. 215:317-322; 1981. 11. HiIl, J. M.; Switzer, R. C. The regional distribution and cellular localization of iron in the rat brain. Neuroscience 11:595-603; 1984.
12. Hill, J. M.; Ruff, M. R.; Weber, R. J.; Pert, C. B. Transferrin receptors in rat brain: neuropeptide-like pattern and relationship to iron distribution. Proc. Natl. Acad. Sci. USA 82:4553-4557; 1985. 13. Jefferies, W. A.; Brandon, M. R.; Hunt, S. V.; Williams, A. F.; receptor on Gatter, K. C.; Mason, D. Y. Transfetrin endothelium of brain capillaries. Nature 312:162-163; 1984. 14. Levin, M. J.; Tuil, D.; Uzan, G.; Dreyfus, J. C.; Kahn, A. Expression of the transferrin gene during development of nonhepatic tissues: high level of transferrin mRNA in fetal muscle and adult brain. Biochem. Biophys. Res. Commun. 122:212217; 1984. 15. Lozoff, B.; Brittenham, G. M. Behavioral aspects of iron deticiency. In: Brown, E. B., ed. Progress in hematology. vol. XIV. Orlando: Harcourt Brace Jovanovich; 1986:23-53. 16. Noble, E. P.; Wurtman, A. J.; Axelrod, J. A simple and rapid method for injecting [H3]norepinephrine into the lateral ventricle of the rat brain. Life Sci. 6:281-291; 1967. 17. Patlak, C. S.; Blasberg, R. G.; Fenstermacher, J. D. Graphical evaluation of blood-to-brain transfer constants from multipletime uptake data. J. Cereb. Blood Flow Metab. 3:1-7; 1983. 18. Pardridge, W. M.; Eisenberg, J.; Yang, J. Human blood-brain barrier transfenin receptor. Metabolism 36:892-895; 1987. 19. Peters, G.; Keberle, H.; Schmid, K.; Brunner, H. Distribution and renal excretion of desferrioxamine and ferrioxamine in the dog and in the rat. Biochem. Pharmacol. 15:93-109; 1%6. 20. Trapp, G. A. Plasma aluminum is bound to transfenin. Life Sci. 33:311-316; 1983. 21. Yehuda, S.; Youdim, M. B. H. The increased opiate action of p-endorphin in iron-deficit rats: the possible involvement of dopamine. Eur. J. Pharmacol. 104:245-251; 1984. 22. Yehuda, S.; Youdim, M. B. H.; Mostofsky, D. I. Brain irondeficiency causes reduced learning capacity in rats. Pharmacol. Biochem. Behav. 25:141-144; 1986.