SEARCH

SEARCH BY CITATION

Keywords:

  • alkaline phosphatase;
  • blood–brain barrier;
  • blood–CSF barrier;
  • capillary;
  • iron deficiency;
  • receptor

Abstract

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. References

Anti-transferrin receptor IgG2a (OX26) transport into the brain was studied in rats. Uptake of OX26 in brain capillary endothelial cells (BCECs) was > 10-fold higher than isotypic, non-immune IgG2a (Ni-IgG2a) when expressed as % ID/g. Accumulation of OX26 in the brain was higher in 15 postnatal (P)-day-old rats than in P0 and adult (P70) rats. Iron-deficiency did not increase OX26 uptake in P15 rats. Three attempts were made to investigate transport from BCECs further into the brain. (i) Using a brain capillary depletion technique, 6–9% of OX26 was identified in the post-capillary compartment consisting of brain parenchyma minus BCECs. (ii) In cisternal CSF, the volume of distribution of OX26 was higher than for Ni-IgG2a when corrected for plasma concentration. (iii) Immunohistochemical mapping revealed the presence of OX26 almost exclusively in BCECs; extravascular staining was observed only in neurons situated periventricularly. The data support the hypothesis of facilitated uptake of OX26 due to the presence of transferrin receptors at the blood–brain barrier (BBB). However, OX26 accumulation in the post-capillary compartment was too small to justify a conclusion of receptor-mediated transcytosis of OX26 occurring in BCECs. Accumulation of OX26 in the post-capillary component may result from a diphasic transport that involves high-affinity accumulation of OX26 by the BCECs, clearly exceeding that of Ni-IgG2a, followed by a second transport mechanism that releases OX26 non-specifically further into the brain. The periventricular localization suggests that OX26 probably also derives from transport across the blood–CSF barrier.

Abbreviations used
AUC

area under the plasma radioactivity curve

BBB

blood–brain barrier

BCEC

brain capillary endothelial cell

Ni-IgG2a

non-immune IgG2a

OX26

antitransferrin receptor IgG2a

P

postnatal.

Brain capillary endothelial cells (BCECs) form the major component of the blood–brain barrier (BBB), which constitutes a prominent barrier for macromolecules and hydrophilic drugs, leaving it virtually impossible for these molecules to enter the CNS by simple diffusion. Studies of cerebrovascular permeability indicate that the BBB may be circumvented by conjugating hydrophilic molecules to monoclonal antibodies raised against peptides expressed by BCECs, e.g. the OX26 antibody, which is a monoclonal antibody raised against the transferrin receptor expressed by BCECs (Jefferies et al. 1984). Hence, targeting of drugs conjugated to OX26 antibodies to the BBB should enable transport of the drugs through the BBB and into the CNS provided the OX26 undergoes transport through the BBB and further into the brain (Friden et al. 1991; Pardridge et al. 1991; Lee et al. 2000). By this means, the use of OX26 as a vector for transport through the BBB should, in principle, allow the treatment of neurodegenerative disorders such as amyotrophic lateral sclerosis and Parkinson's disease with neurotrophic factors, which are otherwise incapable of passage through the BBB (Pardridge 1998).

The mechanism by which OX26 gains access to the CNS remains unresolved. In liver cells, internalization of OX26 is thought to follow the generally accepted mechanism for the uptake of iron-containing transferrin, which involves receptor-mediated endocytosis and transport to the endosomal compartment (Trinder et al. 1988; Wu and Pardridge 1998). In terms of uptake of iron–transferrin by BCECs, quantitative studies performed on the transport of iron–transferrin strongly indicate that this is the principle mechanism for the uptake of iron–transferrin at the BBB. However, the fractional appearance of iron transported into the brain is much higher than that of transferrin, which implies that only the iron is transported into the brain parenchyma, whereas most of the transferrin recycles to the plasma (Taylor et al. 1991; Morris et al. 1992; Ueda et al. 1993). Nonetheless, it is claimed that the immunoglobulin OX26 undergoes receptor-mediated transcytosis through the BBB resulting in a higher transport than is achieved with non-immune immunoglobulins (Friden et al. 1991; Pardridge et al. 1991; Lee et al. 2000). There is also emerging evidence for iron transport across the blood–CSF barriers (Moos and Morgan 2000), which makes it possible that transport of OX26 through the choroid plexuses may account for OX26 in the brain.

In this study, we investigated the question of OX26 transport across the BBB. We examined both developing and adult rats, the former having a much higher rate of transport of transferrin and iron into the brain (Taylor and Morgan 1990), and studied the transport of OX26 and isotypic-specific, non-immune IgG2a (Ni-IgG2a) as a marker for non-specific transport. The appearance of these compounds in the brain was related to their presence in capillary-depleted fractions of brain homogenates and in CSF, and to their distribution within the brain parenchyma.

Materials and methods

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. References

Materials

Isotopes 125I and 131I were obtained from Amersham International (Amersham Pharmacia Biotech, Buckinghamshire, UK) as Na[125I] and Na[131I]. Purified monoclonal mouse-anti-rat transferrin receptor IgG2a (clone OX26) and non-immune mouse IgG2a (Ni-IgG2a) were purchased from Serotec (Oxford, UK).

Experimental procedures

OX26 and Ni-IgG2a were iodinated with 125I and 131I, respectively, using iodogen as described previously (Fraker and Speck 1978). The specific activities of the two proteins were: [125I]OX26: 2 × 105 and [131I]Ni-IgG2a: 1 × 105 c.p.m./µg. The proteins were injected simultaneously into male rats aged P0, P15 or P70 (adult rats) i.p. (any aged rats) or i.v. (P15 and P70 rats) injection into a lateral tail-vein using a 30 G needle. Iron-deficient rats aged P15 were also employed, as iron-deficiency at this age leads to a significantly higher transport of iron and transferrin into the brain (Taylor et al. 1991); they were made iron-deficient by feeding pregnant female rats a low iron diet with an iron content of 5 mg/kg (American Institute of Nutrition, diet AIN 76A lacking iron) from day 12 after observing spermatozoa in the vagina. The iron status of normal and iron-deficient rats was investigated by examination of the concentration of iron in samples of brain, liver and plasma.

For i.v. injections, rats were anaesthetized with halothane (5%) in a mixture of N2 : O2 (60 : 40%) before the injection. Except where noted, rats were injected i.p. or i.v. with a dose of 2 µg of each protein in a mixture dissolved in 0.15 m NaCl. P0 and P15 rats were injected with a volume of 20 µL and P70 rats with 50 µL. Blood was sampled from the ventral tail vein into microcapillary tubes from i.v.-injected rats 2 min after the injection to estimate the plasma volume of distribution, and successively at various time intervals ranging from 5 to 240 min to estimate a time-plasma concentration curve for [125I]OX26 and [131I]Ni-IgG2a.

At various time intervals after injection ranging from 5 to 240 min, the rats were deeply anaesthetized by an i.p. injection of sodium phenobarbiturate, and blood was sampled by cardiac puncture into a heparinized syringe. Subsequently, the right atrium was incised and the rats were slowly, manually perfused through the left ventricle with 60 mL (P70) or 20 mL (P15) of heparinized physiological saline (0.9% in distilled water), which led to a water-clear perfusate flowing from the right atrium, and tissues (entire brain devoid of meninges, liver, two femurs, spleen and hearts) were collected and weighed. To obtain information about the distribution of [125I]OX26 and [131I]Ni-IgG2a in the ventricular system of i.v.-injected rats, CSF was sampled from the cerebellomedullary cistern (cisterna magna) 120 min after injection before collecting blood from the heart (Moos and Morgan 1998). To exclude contamination of the CSF with blood, samples were centrifuged and carefully examined for the presence of blood using a microscope. CSF samples which contained blood were discarded.

Capillary depletion

The presence of [125I]OX26 in a brain parenchymal fraction devoid of BCECs was evaluated in P15 and adult rats using a protocol to separate brain capillaries from the remaining brain tissue (Triguero et al. 1990). In brief, brains were homogenized in 3.5 mL ice-cold buffer (10 mm HEPES, 141 mm NaCl, 4 mm KCl, 2.8 mm CaCl2, 1 mm MgSO4, 1 mm NaH2PO4 and 10 mm glucose; pH 7,4), and 2 mL of the homogenate was mixed with 2 mL of ice-cold 26% dextran (Mr 60 000, Sigma). The dextran-containing solution was centrifuged at 5400 g for 30 min in a Beckman Model J2-21M centrifuge at 4°C. The resulting supernatant and pellet were separated manually and counted for radioactivity. A small sample was taken from each fraction, fixed in 4% paraformaldehyde in 0.1 phosphate-buffered saline (PBS) and embedded in epon. One-micrometre sections were cut from the epon blocks and stained with toluidine blue to verify the separation of the majority of brain capillaries from the remaining brain parenchyma.

Alkaline phosphatase (EC 3.1.3.1) activity

In the brain, the enzyme alkaline phosphatase is selectively expressed by BCECs. Accordingly, the activity of this enzyme can be used to determine the contamination of the supernatant with BCECs that fails to pellet in the dextran density centrifugation used for the capillary depletion technique (Triguero et al. 1990). Alkaline phosphatase activity was determined using the protocol of Williams et al. (1980). In brief, samples of pellet and supernatants isolated by the capillary depletion technique were sonicated for 15 s × 3 at 4 °C, and a 0.1-mL suspension of each sample was added to a buffer consisting of 50 mm-MgCl2, 5 mm CaCl2, 100 mm KCl, 5 mmp-nitrophenyl phosphate (pNPP), 100 mm Tris, pH 9.0 in a 1-mL volume for 20 min at 37 °C. The reaction was stopped by adding 2.0 mL 1 m NaOH. Insoluble material was removed by spinning for 7 min at 2000 g. Finally, absorbance was determined at 420 nm and activity converted to nmol/min/mg protein using purified alkaline phosphatase (Sigma P-7640) as standards.

Chemical iron measurements

Plasma iron and liver non-heme iron were estimated as described previously (Taylor and Morgan 1990). Brain total iron was measured by atomic absorption spectrophotometry using a Varian Spectr AA 300/400 spectrometer after digestion with 35% (w/v) HNO3 and 15% (w/v) H2O2.

Morphological studies

The distribution of OX26 and Ni-IgG2a in the brain was further addressed using immunohistochemistry. Normal P15 rats were injected i.v. with 0.1 mg OX26 or 0.1 mg Ni-IgG2a and killed after 2 h by perfusion fixation with 4% paraformaldehyde in 0.1 m PBS, followed by processing of the brains for cryostat sectioning. OX26 and Ni-IgG2a were then detected by biotinylated goat-anti-mouse IgG absorbed with rat immunoglobulins (Sigma) diluted 1 : 100 followed by stepwise incubation with horseradish peroxidase–streptavidin–biotin complex (Vector) and diaminobenzidine (Moos and Høyer 1996).

Calculations

All the radioactive samples were precipitated with 10% trichloroacetic acid and counted in a two-channel γ-scintillation counter (LKB-Wallac 1281 Compu-gamma) for their protein-bound 125I and 131I content, with appropriate corrections for spill-over between the two channels (Crowe and Morgan 1992). The amount of non-precipitable radioactivity was not > 5% in any sample examined (not shown). The amounts of trichloroacetic acid-precipitable radioactivity present in plasma, CSF and tissue samples were calculated as % of the injected dose (% ID)/mL of plasma or CSF, or as % ID/organ.

The area under the plasma radioactivity curve (AUC) was determined for both [125I]OX26 and [131I]Ni-IgG2a in curves obtained by plotting plasma concentration vs. time in P15 and P70 rats injected i.v., using a software program (figp, Biosoft, UK). The average trichloroacetic acid-precipitable radioactivity in plasma within a time interval of 120 min was calculated by multiplying the AUC value with the activity in plasma at 0 min. The activity in plasma at 0 min was calculated by using the plasma volume obtained by extrapolation of the Ni-IgG2a values of the plasma-concentration vs. time curve to 0 min and dividing by the time interval.

Values of the distribution volume, Vd, for [125I]OX26 and [131I]Ni-IgG2a in total brain, in the fractions of the brain capillary depletion experiment and CSF were estimated by dividing the radioactivity per g brain or mL CSF in the samples with the average plasma radioactivity. To compare the uptake of OX26 by the differently aged rats, the rate of uptake expressed as the transfer constant Kin (mL/kg/min) was calculated for the brain and liver by dividing the % ID/g with the plasma AUC.

Statistical evaluation of the data was performed by analysis of variance (anova). When differences were detected (p < 0.05), means were tested with Student's–Newman-Keuls test for differences between individual means. Three rats were used in all experimental groups except where noted.

Results

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. References

The rapid growth of the Wistar rats was reflected in their body weight, which increased as they aged from P0 to P70 (Table 1). The increase in weight was accompanied by an increase in the weight of the brain and liver, and an increase in plasma volume (Table 1).

Table I.  Body, brain, and liver weights of different aged rats together with plasma volumes and hematocrits
 Body weight (g)Brain weight (g)Liver weight (g)Hematocrit (%)Plasma volume (ml)
  1. Each value is the mean ± SD of measurements from eight to 10 rats. 1 indicates statistically significant difference (p < 0.05) between normal and iron-deficient P15 rat. n.d., not determined.

P06.21 ± 0.620.27 ± 0.020.19 ± 0.02n.d.n.d.
p15 normal18.64 ± 2.801.01 ± 0.110.33 ± 0.0529.67 ± 2.001.69 ± 0.22
P15 iron-deficient18.29 ± 1.830.99 ± 0.050.31 ± 0.0617.20 ± 2.3911.75 ± 0.19
P70233.11 ± 9.301.82 ± 0.0710.65 ± 1.5439.56 ± 2.0711.76 ± 0.72

To study the effects of variations in dosage of OX26 on the uptake of OX26 in the brain and other organs, adult rats were injected with 0.7, 2, 6 and 18 µg 125I-labelled OX26 and killed after 2 h. Varying the dose from the 2 µg used for the remaining experiments did not significantly alter the % ID of OX26 taken up by the brain and heart (Fig. 1) or liver, spleen and femur (not shown). In the heart, where capillaries contain fenestrations but are devoid of transferrin receptors, the uptake was extremely low, ≈ 0.003% ID, compared with ≈ 0.4% ID for the total brain (Fig. 1). The lack of a decline in terms of uptake of % ID in the brain when the dose of OX26 was increased indicates that the OX26 antibody did not saturate all the transferrin receptors.

image

Figure 1. Effect of injected amount (dose) of OX26 on the uptake of [125I]OX26 in brain and heart expressed as % ID/organ ± SD. The different values for the individual organs are not statistically different. Each value is the mean ± SD of measurements from three rats. □, brain; ▪, heart.

Download figure to PowerPoint

Kinetics of OX26 and Ni-IgG2a in plasma

Following i.v. injection with the mixture of OX26 and Ni-IgG2a into adult rats, the fractional appearance of the compounds in plasma was found to differ within 2 min (Fig. 2a). OX26 decreased to 4.20% ID/mL plasma 2 min after the injection (Fig. 2a). This amount of OX26 in plasma equalled 49% of the ID. Hence, although attempts were not made to estimate the T2 of OX26 in this initial phase, it is evident that virtually half of the ID had disappeared from plasma around 2 min post-injection. In contrast, 8.70% ID/mL plasma of Ni-IgG2a, which equals almost 100% of the ID, was still present in plasma after 2 min. Following the rapid initial decrease, the OX26 concentration decreased exponentially at a more rapid rate than Ni-IgG2a until 60–120 min after injection. Thereafter, the two proteins decreased more slowly. After 240 min, ≈ 4.5% ID/mL plasma of Ni-IgG2a and 1.9% ID/mL plasma of OX26 were present. The AUC values (0 [RIGHTWARDS ARROW] 120 min) for the time-plasma concentration were measured as: OX26(P70) = 326% ID * min/mL; Ni-IgG2a(P70) = 806% ID * min/mL. When expressed as % ID * min in plasma, the values were OX26(P70) = 3850% ID * min and Ni-IgG2a(P70) = 9519% ID * min.

image

Figure 2. Changes in plasma concentration of the OX26 and Ni-IgG2a with time after injection expressed as % ID/mL plasma ± SD in adult (a) and P15 (b) rats. Initially, the concentration of OX26 decreases markedly to a significantly lower concentration than Ni-IgG2a, and from 2 min and onwards, the plasma concentration of OX26 and Ni-IgG2a are statistically different at any time point. The diamonds marked by an arrow represent the corresponding values of OX26 and Ni-IgG2a (identical colour) of iron-deficient P15 rats; these values are not statistically different to the corresponding values of P15 rats having a normal iron status. Each value is the mean ± SD of measurements from 3 to 5 rats. □, OX26; ▪, Ni-Ig2a.

Download figure to PowerPoint

As in the case of adult rats, similar patterns of decrease in the plasma concentration of OX26 and Ni-IgG2a were observed in the P15 rats (Fig. 2b). Within 2 min, OX26 decreased to 28% ID/mL plasma 2 min after the injection, which equalled 47% of the ID. In contrast, 62% ID/mL plasma of Ni-IgG2a equalling almost 100% of the ID, was still present in plasma after 2 min. Following this rapid decrease, the OX26 concentration of the P15 rats decreased exponentially at a slightly more rapid rate than Ni-IgG2a until 60–120 min after injection. Thereafter, the two proteins decreased more slowly. After 240 min, ≈ 4.5% ID/mL plasma of Ni-IgG2a and 1.9% ID/mL plasma of OX26 were present.

Compared with adult rats, plasma concentrations were clearly higher in the P15 rats because of the smaller plasma volume at P15 (compare Figs 2a and b). The AUC values (0 [RIGHTWARDS ARROW] 120 min) estimated from plasma concentration vs. time in the P15 rats were: OX26(P15) = 2158% ID * min/mL; Ni-IgG2a(P15) = 5069% ID * min/mL. When expressed as % ID * min in plasma, the values were OX26(P15) = 3647% ID * min and Ni-IgG2a(P15) = 8567% ID * min.

Uptake of OX26 in different aged rats

Following i.v. injection of OX26 into the adult rat, the uptake in brain increased from 0.25% ID by 30 min to reach a significantly higher plateau of ≈ 0.4% ID after 60–240 min (Fig. 3a). Ni-IgG2a occurred in the brain only at 0.03% of the injected dose. In the adult liver, the concentration of OX26 was much higher than in the brain, e.g. after 60–120 min ≈ 10–12% of the injected dose was present in the liver (Fig. 3b). As in the brain, the amount of OX26 present in the liver was significantly higher than that of Ni-IgG2a at any time point (Fig. 3b).

image

Figure 3. Accumulation of OX26 and Ni-IgG2a expressed as % ID/organ ± SD in adult brain (a) and adult liver (b). The accumulation of both OX26 and Ni-IgG2a is much higher in the liver than in the brain (*). At 60 min, the accumulation of OX26 in the brain is significantly higher than at 30 min post-injection. □, OX26; ▪, Ni-Ig2a.

Download figure to PowerPoint

When examined in differently aged rats, OX26 uptake clearly exhibited an age-dependent variation. Both Kin and Vd for OX26 in the brain and liver of the i.v.-injected P15 rats were higher than in the adult brain (Tables 2 and 3).

Table II.  The rate of uptake expressed as the transfer constant Kin (ml/kg/min) of OX26 and IgG2a in brain and liver of adult and P15 rats
 P70 P15 
  1. Data are from rats at 2 h post-injection. Each value is the mean ± SD of measurements from 3-4 rats. 1Significant difference between OX26 and Ni-IgG2a on either age (p < 0.05). 2Significant difference between P70 and P15 of OX26 or Ni-IgG2a (p < 0.05).

 OX26Ni-IgG2aOX26Ni-IgG2a
Brain0.79 ± 0.100.02 ± 0.0111.78 ± 0.2120.05 ± 0.011,2
Liver3.17 ± 0.490.15 ± 0.0515.94 ± 1.2620.24 ± 0.061,2
Table III.  Uptake of OX26 in total brain and brain fractions separated by the capillary depletion technique in adult (P70) and young (P15) rats at 2 hr post-injection
 Amount (% ID) P70P15
Total brain0.47 ± 0.063.89 ± 0.511
Capillary depletion
Pellet0.44 ± 0.063.54 ± 0.461
Supernatant0.03 ± 0.000.35 ± 0.05
Supernatant/total brain ratio6.4 %9.0 %
 Vd (μL/g) P70P15
  1. The results are expressed as the amount taken up (% dose) and as the volume of distribution, Vd (μL/g). Each value is the mean ± SD. 1Significant difference between P70 and P15 (p < 0.05).

Total brain173 ± 22.1216 ± 28.31
Capillary depletion
Pellet160 ± 20.5197 ± 25.8
Supernatant12.3 ± 1.5719.4 ± 2.55
Supernatant/total brain ratio7.1 %9.0 %

The use of i.p. injection allowed examination of the OX26 distribution even in P0 animals (Fig. 4). Accumulation tended to be highest at 240 min post-injection, especially in the P15 rats, which contrasted that of i.v.-injected rats in which the amount of OX26 and Ni-IgG2a were higher at 60 and 120 min (not shown). When expressed as % ID/g brain, the accumulation in i.p.-injected rats was significantly higher in P15 rats than in P0 or adult rats (Fig. 4). The uptake by the P0 rat brain was approximately half that of the P15 rat, but nonetheless much higher than in the adult rat. When expressed as % ID in total brain, mean OX26 values at 240 min were P70 = 0.07% ID, normal P15 = 1.2% ID, P0 = 0.12% ID. These values were statistically different (p < 0.001) when compared among different ages. Hence, uptake of OX26 in the brain clearly exhibited age-dependent variation.

image

Figure 4. OX26 in brains of differently aged rats injected i.p. expressed as % ID/g weight ± SD. The uptake of OX26 is age-dependent being highest in P15 brains. Iron-deficiency did not affect the uptake of OX26 in i.p.-injected P15 rats. □, adult rat; ▪, P15 normal rat; ○, P15 iron-deficient rat; ●, P0 rat.

Download figure to PowerPoint

Uptake of OX26 in iron-deficient rats

The haematocrit levels of P15 rats subjected to an iron-deficient diet were significantly lower than those in the normal P15 animals (Table 1). Moreover, measurements of the iron-content in brain, liver and plasma revealed significantly lower values in the group of iron-deficient rats, thus confirming the iron-deplete conditions of these rats (Table 4). Nonetheless, in i.v.-injected rats an absence of an influence of iron-deficiency on the uptake of OX26 in the brain was seen as the accumulation at 120 min post-injection was 3.23% ID/g in normal P15 rats and 3.53% ID/g in iron-deficient P15 rats; these values were not statistically different (p < 0.01). Plasma AUC was not measured in iron-deficient P15 rats, but the plasma concentration of OX26 at 120 min was not different from that of the normal P15 rat (Fig. 2b), which strongly suggests that OX26 distributes from plasma to different organs in normal and iron-deficient rats to same extent. In studies of i.p.-injected rats, the mean OX26 values at 240 min were: normal P15 = 1.2% ID and iron-deficient P15 = 1.3% ID (Fig. 4); this difference between normal and iron-deficient P15 was not significant (p < 0.001).

Table IV.  Iron measurements in P15 rats with different iron-status
 Brain total iron (μg/g wet weight)Liver non-heme iron (μg/g wet weight)Plasma iron (μg/mL)
  1. Each value is the mean ± SD of measurements from nine to 10 rats. 1 Indicates statistically significant difference (p < 0.05).

P15 normal6.82 ± 0.6746.60 ± 13.541.32 ± 0.19
P15 iron-deficient5.25 ± 1.01111.41 ± 0.9710.42 ± 0.091

Capillary depletion experiments

The capillary depletion technique provided two different fractions from each brain sample which were embedded in epon and examined in routinely stained sections. In the pellet that is supposed to contain the brain capillaries (Triguero et al. 1990), an abundance of vessel-like structures and cellular nuclei was observed. Many of these structures clearly shared the morphological characteristics of brain capillaries (Fig. 5a). In the supernatant fraction, extremely few structures resembling brain capillaries were observed (< 1 capillary-like structure per section). Instead, a plethora of cross-sectioned neuronal extensions from both white and grey matter areas were observed (Fig. 5b). In the assay for the capillary-specific enzyme, alkaline phosphatase, the vast majority of enzymatic activity was found in the pellet of both adult and P15 rat brain samples (Table 5). From these observations, it was concluded that sufficient separation of the brain capillaries was obtained to allow for the use of the brain capillary technique to distinguish between the brain capillaries and the brain parenchyma or post-capillary compartment.

image

Figure 5. Nissl-stained sections from material obtained from brain capillary depletion experiment showing the pellet (a) and supernatant meant to contain neuronal and glial debris (b). (a) Cellular elements corresponding to capillaries (arrows) and major vessels (double arrows) are seen. The pellet also section also contain nuclei of brain parenchymal cells (curved arrow). (b) The neuronal debris is mainly seen as transected myelinated fibres. Scale bar = 10 µm.

Download figure to PowerPoint

Table V.  Activity of the BCEC specific enzyme, alkaline phosphatase in brain fractions separated as pellet and supernatant using the capillary depletion technique.
 Enzyme activity (nmol/min/mg)
 P70P15
  1. Each value is the mean ± SD of measurements from four to six rats.

Capillary depletion
Pellet8.99 ± 1.4112.59 ± 1.52
Supernatant0.65 ± 0.110.88 ± 0.29
Supernatant/pellet6.7%6.5%

When the appearance of OX26 was studied in brain homogenates in which brain capillaries had been separated from the remaining brain tissue, the fraction containing the capillaries contained 90–95% of the amount of radioactivity in both adult and P15 rats (Table 3). Comparing the different aged rats in terms of spaces, the values for OX26 in both the supernatant and pellet were significantly higher in young rats than adult rats (Table 3).

Cerebrospinal fluid

In the CSF samples, the % ID/mL of OX26 and IgG2a were significantly different in adult and P15 rats, and the amounts of OX26 and IgG2a were significantly lower in adult rat CSF than P15 rat CSF (Fig. 6a). When uptake was calculated in terms of volumes, the Vd of OX26 in the adult rat was approximately equal to that of the P15 rat (Fig. 6b). The Vd values of Ni-IgG2a were of similar magnitude in the adult and the P15 rats and not significantly different, but they were significantly lower than the values for OX26 (Fig. 6b).

image

Figure 6. Accumulation of OX26 and Ni-IgG2a in CSF of adult and P15 rats expressed as % ID/mL CSF + SD (a) or Vd expressed as µL plasma/mL CSF + SD (b). (a) The uptake of both proteins are significantly higher in the P15 rat than in the adult (*). OX26 accumulates to a significantly higher degree than Ni-IgG2a in both the adult and P15 rat (o). (b) When expressed in terms of Vd, OX26 accumulation is greater than Ni-IgG2a at both ages (o). The Vd values of OX26 and Ni-IgG2a are not statistically different.

Download figure to PowerPoint

Immunohistochemical detection of OX26 and Ni-IgG2a in brain

OX26 was consistently detected in brain capillaries throughout the brain parenchyma (Fig. 7a). Ni-IgG2a could not be detected in BCECs at 2 h after injection (not shown). Neurons with high levels of transferrin receptors situated at a distance from the ventricular system, e.g. neurons of the neocortex and the red nucleus (Moos et al. 1998), did not contain OX26 (Fig. 7b). In contrast, neuronal OX26 immunoreactivity was observed in cells with proximity to the ventricular system, e.g. neurons of the hippocampal cortex (Figs 7c and d). Staining was also seen outside of neuronal perikarya in the near ventricular surface of the hippocampal cortex (Fig. 7d). Given the fact that neurons and also their extensions labelled with OX26 (Fig. 7d), it is likely that extraperikaryal staining can be attributed to uptake and axonal transport of OX26 from the ventricular system (Moos and Morgan 1998). Neuronal staining was never observed in animals injected with Ni-IgG2a. OX26 labelled epithelial cells of the plexus epithelial cells. The staining was faint but greater than that of IgG. Cells with the morphology of astrocytes, oligodendrocytes, or microglia were not labelled by either of the antibodies. When the anti-mouse antibody that had been absorbed with rat immunoglobulins was omitted in the immunohistochemical reaction, it was not possible to detect labelling of neuronal perikarya or extraperikaryal elements anywhere in the brain.

image

Figure 7. Distribution of OX26 in P15 rat brain examined at 2 h after injection. (a) OX26 distribution in the neocortex shown at low power magnification. The OX26 accumulates profoundly in BCECs identified by the arrow. (b) The identical region shown at high power. Even though the neurons of the neocortex are known to contain transferrin receptors in the P15 rat (Moos et al. 1998), neuronal labelling by OX26 is not observed to an extent that exceeds the background staining. Asterisks identify the nuclei of neocortical neurons. Labelling is seen in BCECs. (c) In contrast, OX26 is seen in neurons of the hippocampal cortex situated close to the ventricular system, which may reflect uptake from the CSF (Moos and Morgan 1998). The region identified with a curved arrow is shown in higher magnification in (b) where OX26-labelled neurons of the hippocampal cortex are seen (thin arrows), including neuronal extensions (small arrows). Labelling is also seen in capillaries (curved arrow) and in the neuropil (asterisk). Scale bars: a = 60 µm; b, d = 10 µm; c = 300 µm of measurements from 6 to 8 rats.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. References

The results support the concept of facilitated uptake of OX26 in various tissues, including the brain, due to the attachment of OX26 to transferrin receptors (Friden et al. 1991; Pardridge et al. 1991). The highest uptake of OX26 takes place in the liver, but there is also high uptake in the spleen and femurs (erythropoietic organs in the rat), which is consistent with the high expression of transferrin receptors in these organs (Trinder et al. 1988; Morgan 1996; Wu and Pardridge 1998). In or experiments, OX26 accumulated in the adult rat brain in amounts comparable with results obtained in previous experiments in this species (Friden et al. 1991; Pardridge et al. 1991; Bickel and Kang 1999).

The plasma disappearance curves of OX26 and Ni-IgG2a reflect the different physiological behaviour of the two proteins after intravenous injection. The three phases in the decrease of plasma OX26 concentration (Fig. 2) can be attributed to: (i) binding to transferrin receptors on cell membranes exposed to blood plasma (BCECs; and hepatocytes of the liver, plus erythropoietic cells of the bone marrow and spleen where the sinusoids are fenestrated and highly permeable to plasma); (ii) binding to receptors that recycle to the cell membrane during the first hours after the injections; and (iii) transcapillary passage of the protein into the interstitial space of the animals. In the case of Ni-IgG2a, it is probable that only transcapillary passage followed by some return to the plasma via the lymph, which accounts for a slowing of the disappearance rate after 1–2 h, are involved. The rate of transcapillary transfer appears to be greater in P15 rats than adult rats (Fig. 2). It is of interest that the pattern of changes in the plasma concentration of OX26 found in this study is very similar to that reported previously (Pardridge et al. 1991) and that, in both investigations, 20–25% of the injected amount of OX26 remained in the plasma when the second phase of disappearance ended (after 1 h).

The experiment in which differing amounts of OX26 were injected (Fig. 1) showed that the 2 µg dose of the antibody used in the time-course study did not saturate all the binding sites on the receptors. This, plus the retention of 20–25% of the OX26 in the plasma 2 h after injection, suggest that a proportion of the antibody preparation lacked the ability to bind to the transferrin receptor.

The transport of OX26 through the BBB is restricted

Our results show that accumulation of OX26 in the brain occurred mainly in BCECs, which contained > 90% of the total uptake 2 h after injecting the protein into both adult and P15 rats. Some OX26 was also found in the post-capillary supernatant of brain homogenates, accounting for 7 and 9% of the total uptake in adult and P15 rats, respectively. It should, however, be noted that these fractions were almost the same as the ratios for the endothelial-specific enzyme alkaline phosphatase, suggesting that most, if not all, of OX26 in the post-vascular compartment may due to an artefactual leakage of OX26 from the BCECs into the supernatant.

The fraction of the antibody taken up by the BBB into the brain was small and definitively lower than reported previously when 30–40% was found in the post-capillary supernatant 2 h after injection (Friden et al. 1991; Pardridge et al. 1991). The reason for this difference is uncertain as the same capillary depletion technique was used in all investigations. Two reasons that may contribute to this difference are that the vasculature was perfused with physiological saline before the brain was removed and only trichloroacetic acid precipitable radioactivity was used to estimate protein levels in this study but not in the earlier ones. Unperfused brains would contain more plasma with its content of OX26, leading to a higher level of OX26 in the supernatant fraction, and protein breakdown products would be found mainly in this fraction. However, the effect of these differences in experimental technique would be considerably smaller than that required to explain the difference in results obtained in the two laboratories. Hence, the reason for the difference is unresolved at the present time.

The reported transfer of OX26 into the post-capillary supernatant 2 h after injection, whether 7 or 40% of the amount taken up by the BCECs, is low compared with the rate of transport of iron derived from transferrin taken up by these vessels through the agency of the transferrin receptors. Two hours after injection of 59Fe- and 125I-labelled transferrin, the Vd values for iron uptake by the brain were 10 and 5 times as great as those for transferrin in P15 and adult rats, respectively (Morgan 1999). This indicates that the transferrin receptors and transferrin had recycled at least these numbers of times, with iron being released from the transferrin and transported across the BBB with each uptake. Hence, it is clear that the processes involved in the transport of OX26 into the brain interior are much slower than those of iron derived from transferrin, even though they both involve the transferrin receptor.

The kinetics of transferrin uptake by the brain are similar to those of OX26, both increasing rapidly during the first 30–60 min after injection with little further change during the next few hours (Taylor and Morgan 1990; Friden et al. 1991). After releasing its iron, the transferrin must recycle to the plasma, thus accounting for the much higher Vd for iron than transferrin in the brain (Taylor and Morgan 1990; Morgan 1999). In contrast, OX26 does not recycle to the plasma because, if it did, the quantity in the brain would decrease rapidly as it was taken up by tissues with much higher numbers of transferrin receptors such as erythropoietic tissue and the liver.

The difference in behaviour of OX26 and transferrin taken up by BCECs is due to the difference in the chemical nature of the interaction between the transferrin receptor and either OX26 or transferrin. The former is a high-affinity antibody–antigen interaction that is not easily reversed, whereas the latter is readily reversed depending on pH and the iron content of transferrin (Morgan 1996). This raises the question of how OX26 is released from the receptor to enter the post-capillary fraction of the brain. One possibility is that an acidic pH within the endosomes of BCECs could weaken the interaction between OX26 and the receptor, a proportion of the OX26 being released and transported across the cells by transcytosis of the endosomes or a second system of vesicles. Alternatively, the interaction of OX26 with the transferrin receptor may lead to an altered intracellular distribution and increased rate of degradation of the receptor, as observed in lymphoma (Lesley and Schulte 1985) and K562 (Weissman et al. 1986) cells, leading to the relatively slow release of OX26. Whether the antibody is also degraded in the process in uncertain. This hypothesis is supported by the observation that the staining pattern of OX26 in BCECs changes at 4 h after injection, from a uniform to a more punctate distribution, whereas the quantity of radiolabelled OX26 in the capillary pellet falls progressively and that in the post-capillary brain parenchyma increases slowly after this time (Friden et al. 1991). Thus, binding of OX26 to transferrin receptors on BCECs could result in a relatively slow transfer of OX26 into the brain parenchyma but, as indicated below, transfer across the brain–CSF barrier may be equally or more important during the first 2 h after injection of the antibody.

The presence of OX26 in CSF

The presence of OX26 in the CSF confirms earlier conclusions that antibodies against the transferrin receptor can undergo transport into the brain (Friden et al. 1991; Pardridge et al. 1991; Lee et al. 2000), but they do not show whether this occurs by transport across the BBB or the blood–CSF barrier. The CSF in the ventricles represents a mixture of interstitial fluid, which drains from the brain to the ventricles, and the CSF proper generated by choroid plexus epithelial cells (Bradbury 1997). Judging from the findings on BBB transport described above, it is possible that more OX26 than Ni-IgG2a was transported to the interstitial fluid subsequent to uptake at the BBB. This could give rise to relatively higher concentration of OX26 than Ni-IgG2a in CSF as observed when the results were expressed as Vd values (Fig. 6b). OX26 and Ni-IgG2a may also escape into the ventricular system via transport across the blood–CSF barrier as both proteins were detected in the choroid plexus epithelium. Because of the presence of transferrin receptors on choroid plexus epithelial cells (Moos et al. 1998), OX26 might enter the CSF after facilitated uptake and this could also contribute to the higher Vd values for OX26 in CSF. Indeed, this is probably the major source of OX26 in CSF as the immunoglobulin was detected only on neurons close to the ventricular system not on neurons elsewhere in the brain with close proximity to BCECs. The absence of OX26 in neurons of the neocortex, a brain region that does not drain to the ventricular CSF (Zhang and Pardridge 2001), also suggest that OX26 in CSF is responsible for the presence of OX26 in neurons in periventricular regions, and OX26 in the post-capillary compartment in the capillary depletion experiments.

The volume of CSF in P15 and adult rats is ≈ 250 µL (Bass and Lundborg 1973; Johanson and Woodbury 1974). Hence, the quantity of OX26 found in the CSF 2 h after injection could account for more than the amount found in the post-capillary supernatant of the brains of P70 rats, and most of that found at P15. However, only the CSF within the ventricular system would be included in the brain homogenates, not that present in the subarachnoid space. This probably accounts for the apparent discrepancy in the P70 animals. Nevertheless, the CSF and immunohistochemical results support the view that a large proportion of the OX26 found in the brain, at least early after injection (up to 2 h), is the result of transfer of the protein across the blood–CSF barrier, probably mediated by transferrin receptors on choroid plexus epithelial cells. That the accumulation of both OX26 and Ni-IgG2a was higher in the CSF of the P15 rat can be explained by the fact that the rate of turnover of CSF is slower at this age than in adult animals (Johanson and Woodbury 1974).

Age affects brain uptake of OX26

The accumulation of OX26 in brain was at its highest in P15 rats, which is in accordance with previous studies showing that transferrin receptors are upregulated in 2–3-week-old postnatal rats (Taylor and Morgan 1990; Crowe and Morgan 1992). Iron-deficiency did not increase brain uptake of OX26, which may seem surprising given the fact that iron-deficiency clearly leads to a higher transport of iron and transferrin into the brain (Taylor et al. 1991). However, this effect with respect to transferrin uptake by the brain is not observed in P15 rats, probably because normal rats of this age have a physiological iron deficiency so that transferrin receptor expression is maximally upregulated and does not change further with added dietary iron depletion (Taylor et al. 1991).

Conclusions

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. References

The transport of OX26 from blood to brain across BCECs seems to involve a two-fold process, which at first results in transport into BCECs to an extent that clearly exceeds that of Ni-IgG2a. The process by which OX26 may move from BCECs and further into the brain is still uncertain although tentative mechanisms are suggested here. Future experiments targeting the use of OX26 as a transport vector for hydrophilic drugs for the treatment of patients suffering from brain disorders without an impaired BBB should focus on the processes by which OX26 is released from the abluminal surface of the BCECs into the brain interstitial fluid. Finally, it should not be overlooked that the periventricular localization of OX26 in brain and the higher presence of OX26 than of Ni-IgG2a in CSF may indicate that OX26 in brain also derives from transport across the blood–CSF barrier.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. References

The excellent technical assistance of Grazyna Hahn and Alan Light is greatly acknowledged. This work was supported by grants from the Novo Nordic Foundation and the Vera and Carl Johan Michaelsens Legat.

References

  1. Top of page
  2. Abstract
  3. Materials and methods
  4. Results
  5. Discussion
  6. Conclusions
  7. Acknowledgements
  8. References
  • Bass N. H. & Lundborg P. (1973) Postnatal development of bulk flow of the cerebrospinal fluid system of the albino rat: clearance of carboxyl-[14C]inulin after intrathecal infusion. Brain Res. 52, 323332.
  • Bickel U. & Kang Y.-S. (1999) Use of chimeric peptides in drug delivery to the brain, in Brain Barrier Systems. Alfred Benzon Symposium, Vol. 45 (Paulson, O. B., Knudsen, G. M., Moos, T., eds), pp. 478488. Munksgaard, Copenhagen.
  • Bradbury M. W. B. (1997) Transport of iron in the blood–brain–cerebrospinal fluid system. J. Neurochem. 69, 443454.
  • Crowe A. & Morgan E. H. (1992) Iron and transferrin uptake by brain and cerebrospinal fluid in the rat. Brain Res. 592, 816.
  • Fraker P. J. & Speck J. C. (1978) Protein and cell membrane iodination with a sparingly soluble chloroamide, 1,3,4,6-tetrachloro-3a,6a-diphenylglycoluril. Biochem. Biophys. Res. Commun. 80, 849857.
  • Friden P. M., Walus L. R., Musso G. F., Taylor M. A., Malfroy B., Starzyk R. M. (1991) Anti-transferrin receptor antibody and antibody–drug conjugates cross the blood–brain barrier. Proc. Natl Acad. Sci. USA 88, 47714775.
  • Jefferies W. A., Brandon M. R., Hunt S. V., Williams A. F., Gatter K. C., Mason D. Y. (1984) Transferrin receptor on endothelium of brain capillaries. Nature 312, 162163.
  • Johanson C. E. & Woodbury D. M. (1974) Changes in CSF flow and extracellular space in the developing rat, in Drugs and the Developing Brain (Vernadakis, A., Weiner, N., eds), pp. 281287. Plenum Press, New York.
  • Lee H. J., Engelhardt B., Lesley J., Bickel U., Pardridge W. M. (2000) Targeting rat anti-mouse transferrin receptor monoclonal antibodies through blood–brain barrier in mouse. J. Pharmacol. Exp. Ther. 292, 10481052.
  • Lesley J. F. & Schulte R. J. (1985) Inhibition of cell growth by monoclonal anti-transferrin receptor antibodies. Mol. Cell. Biol. 5, 18141821.
  • Moos T. & Høyer P. E. (1996) Detection of plasma proteins in CNS neurons: conspicuous influence of tissue-processing parameters and the utilization of serum for blocking nonspecific reactions. J. Histochem. Cytochem. 44, 591603.
  • Moos T. & Morgan E. H. (1998) The kinetics and distribution of [59Fe-125I]transferrin injected into the ventricular system of the rat. Brain Res. 790, 115128.
  • Moos T. & Morgan E. H. (2000) Transferrin and transferrin receptor function in brain barrier systems. Cell. Mol. Neurobiol. 20, 7795.
  • Moos T., Oates P., Morgan E. H. (1998) The expression of the intraneuronal transferrin receptor is age-dependent and susceptible to iron-deficiency. J. Comp. Neurol. 398, 420430.
  • Morgan E. H. (1996) Iron metabolism and transport, in Hepatology. A Textbook of Liver Disease, 3rd edn (Zakim, D., Bayer, T., eds), pp. 526554. Saunders, Philadelphia.
  • Morgan E. H. (1999) Iron and transition metal transport into the brain, in Brain Barrier Systems. Alfred Benzon Symposium, Vol. 45 (Paulson, O. B., Knudsen, G. M., Moos, T., eds), pp. 357366. Munksgaard, Copenhagen.
  • Morris C. M., Keith A. B., Edwardson J. A., Pullen R. G. (1992) Uptake and distribution of iron and transferrin in the adult rat brain. J. Neurochem. 59, 300306.
  • Pardridge W. M. (1998) Drug delivery to the brain. J. Cereb. Blood Flow Metab. 17, 713731.
  • Pardridge W. M., Buciak J. L., Friden P. M. (1991) Selective transport of an anti-transferrin receptor antibody through the blood–brain barrier in vivo. J. Pharmacol. Exp. Ther. 259, 6670.
  • Taylor E. M. & Morgan E. H. (1990) Developmental changes in transferrin and iron uptake by the brain in the rat. Dev. Brain Res. 55, 3542.
  • Taylor E. M., Crowe A., Morgan E. H. (1991) Transferrin and iron uptake by the brain: effects of altered iron status. J. Neurochem. 57, 15841592.
  • Triguero D., Buciak J., Pardridge W. M. (1990) Capillary depletion method for quantification of blood–brain barrier transport of circulating peptides and plasma proteins. J. Neurochem. 54, 18821888.
  • Trinder D., Morgan E. H., Baker E. (1988) The effects of an antibody to the rat transferrin receptor and of rat serum albumin on the uptake of diferric transferrin by rat hepatocytes. Biochim. Biophys. Acta 943, 440446.
  • Ueda F., Raja K. B., Simpson R. J., Trowbridge I. S., Bradbury M. W. (1993) Rate of 59Fe uptake into brain and cerebrospinal fluid and the influence thereon of antibodies against the transferrin receptor. J. Neurochem. 60, 106113.
  • Weissman A. M., Klausner R. D., Rao K., Harford J. B. (1986) Exposure of K562 cells to anti-receptor monoclonal antibody OKT9 results in rapid redistribution and enhanced degradation of the transferrin receptor. J. Cell Biol. 102, 951958.
  • Williams S. K., Gillis J. F., Matthews M. A., Wagner R. C., Bitensky M. W. (1980) Isolation and characterization of brain endothelial cells: morphology and enzyme activity. J. Neurochem. 35, 374381.
  • Wu D. & Pardridge W. M. (1998) Pharmacokinetics and blood–brain barrier transport of an anti-transferrin receptor monoclonal antibody (OX26) in rats after chronic treatment with the antibody. Drug Metab. Dispos. 26, 937939.
  • Zhang Y. & Pardridge W. M. (2001) Rapid transferrin efflux from brain to blood across the blood–brain barrier. J. Neurochem. 76, 15971600.