Abstract
Enzyme replacement therapy of lysosomal storage disorders is complicated by the lack of enzyme transport across the blood-brain barrier (BBB). The present studies evaluate the delivery of a model enzyme across the BBB following enzyme conjugation to a BBB receptor-specific monoclonal antibody (mAb). Bacterial β-galactosidase (116 kDa) was conjugated to the rat 8D3 mAb to the rat transferrin receptor (TfR) via a streptavidin-biotin linkage. The unconjugated β-galactosidase or the β-galactosidase-8D3 conjugate was injected intravenously in adult mice, and enzyme activity was measured at 1 and 4 h in brain and peripheral organs (liver, spleen, kidney, and heart). Unconjugated β-galactosidase was rapidly removed from the blood compartment owing to avid uptake by liver and spleen. There was minimal uptake of the unconjugated β-galactosidase by brain. Following conjugation of the enzyme to the 8D3 TfRmAb, there was a 10-fold increase in brain uptake of the enzyme based on measurement of enzyme activity. Histochemistry of brain showed localization of the enzyme in the intraendothelial compartment of brain following intravenous injection of the enzyme-mAb conjugate. The capillary depletion technique showed that more than 90% of the enzyme-8D3 conjugate that entered into the endothelial compartment of brain passed through the BBB to enter brain parenchyma. In conclusion, high molecular weight enzymes, such as bacterial β-galactosidase, can be conjugated to BBB targeting antibodies for effective delivery across the BBB in vivo. Fusion proteins comprised of BBB targeting antibodies and recombinant enzymes could be therapeutic in the treatment of the brain in human lysosomal storage disorders.
Lysosomal storage disorders are treated with recombinant enzyme replacement therapy. The majority of lysosomal storage disorders affect the brain (Cheng and Smith, 2003). A major limitation in the enzyme replacement therapy of lysosomal storage disorders is the lack of transport of the therapeutic enzyme across the brain capillary wall, which forms the blood-brain barrier (BBB). The involvement of the central nervous system is generally severe in lysosomal storage disorders (Cheng and Smith, 2003), and it is important to develop BBB drug delivery strategies for therapeutic enzymes. Recombinant proteins as large as 40,000 Da have been delivered across the BBB in vivo with molecular Trojan horses that access endogenous BBB receptor-mediated transport systems (Pardridge, 2001). A peptidomimetic monoclonal antibody (mAb) to the BBB transferrin receptor (TfR) mediated the delivery of several peptides and recombinant proteins across the BBB with in vivo central nervous system pharmacological effects following intravenous administration (Pardridge, 2002). The recombinant protein is attached to the TfRmAb via avidin-biotin technology. In this approach, the nontransportable protein drug is monobiotinylated in parallel with the production of a TfRmAb-streptavidin (SA) conjugate. Owing to the very high affinity of SA binding of biotin, there is instantaneous formation of the protein-TfRmAb conjugate following mixing of the monobiotinylated drug and the TfRmAb-SA (Pardridge, 2001).
Lysosomal enzymes have a molecular weight of 50 to 100 kDa, and it is not clear if BBB molecular Trojan horses such as a receptor-specific mAb can deliver a protein of this size across the BBB in vivo. Molecular Trojan horses have delivered liposomes carrying genes across the BBB (Pardridge, 2002). However, to date, the largest protein delivered across the BBB with a molecular Trojan horse is pegylated brain derived neurotrophic factor, which has a molecular weight of about 40,000 Da (Pardridge et al., 1998). The present studies use bacterial β-galactosidase, which has a molecular weight of 116,000 Da, as a model enzyme for delivery to mouse brain. The rat 8D3 mAb to the mouse TfR, which enters brain via the BBB TfR (Lee et al., 2000), is used in the present studies to deliver enzyme to the brain.
Materials and Methods
Materials. Recombinant Escherichia coli β-galactosidase (β-d-galactoside galactohydrolase, EC 3.2.1.23) with a specific activity of 700 to 1200 units/mg protein was purchased from Roche Diagnostics (Indianapolis, IN). Sulfo-N-hydroxysuccinimide (NHS)-long-chain (LC)-LC-biotin, EZ biotin quantitation kit, Traut's reagent, sulfosuccinimidyl-4-(p-malimidophenyl)butyrate (S-SMPB) and the bicinchoninic acid (BCA) protein assay reagents were from Pierce (Rockford, IL). [3H]Biotin (60 Ci/mmol) was purchased from PerkinElmer Life and Analytical Sciences (Boston, MA). Protein G-Sepharose and Sephacryl S-300HR were obtained from Amersham Biosciences (Piscataway, NJ). Sephadex G-25, recombinant streptavidin, and dextran (60,000–90,000 Da) were obtained from Sigma-Aldrich (St. Louis, MO). The β-galactosidase histochemistry kit was from Invitrogen (Carlsbad, CA). The Beta-Glo β-galactosidase luminescence assay system, lysis buffer, and the β-galactosidase enzyme assay system were purchased from Promega (Madison, WI).
Synthesis of 8D3-SA Conjugate. The rat hybridoma line secreting the 8D3 mAb to the mouse TfR was cultured on a feeder layer of mouse thymocytes and peritoneal cells in Dulbecco's modified Eagle's medium with 10% fetal bovine serum (Lee et al., 2000). The hybridoma cells were propagated as ascites in nude mice, and the 8D3 mAb was purified by protein G affinity chromatography. A 1:1 conjugate of the 8D3 mAb and SA was prepared by stable thiol-ether linkage using 8D3 thiolated with Traut's reagent at a 40:1 molar ratio of Traut's reagent. The SA was activated with S-SMPB at a 24:1 molar ratio, and the 8D3-SA conjugate was purified with a 2.5 × 95-cm column of Sephacryl S-300HR in PBS + 0.05% Tween 20 (0.01 M Na2HPO4, 0.15 M NaCl, pH = 7.4, 0.05% Tween 20). The elution of the 8D3-SA conjugate and unconjugated SA were monitored by adding a trace amount of [3H]biotin to the mixture prior to addition to the column. The fractions containing the 8D3-SA conjugate (Fig. 1A, reaction I) were pooled and stored at –20°C.
Monobiotinylation of β-Galactosidase and Biotin Quantitation. Bacterial β-galactosidase was homogeneous on SDS-polyacrylamide gel electrophoresis and migrated with a molecular weight of 116,00 Da (Fig. 1B). The β-galactosidase was dissolved in 0.05 M NaHCO3/8.5, and the protein concentration was determined with the BCA assay. The sulfo-NHS-LC-LC-biotin (45 nmol/μl) was prepared in 0.05 M NaHCO3/8.5, and 19 μl of sulfo-NHS-LC-LC-biotin solution (855 nmol) was added to 5 mg (43 nmol) of β-galactosidase, which was a 20:1 molar ratio of biotin/β-galactosidase. The mixture was capped and rocked end over end for 60 min at room temperature. The sample was applied to a 0.7 × 15-cm Sephadex G-25 column, eluted with 10 ml of 0.01 M PBS/7.4 at 0.5 ml/min, and 0.5-ml fractions were collected. The three fractions comprising the first A280 peak were pooled, the protein concentration was determined, and the biotin-LC-LC-β-galactosidase (Fig. 1A, reaction II) was stored at –20C. The enzymatic activity of β-galactosidase or biotinylated β-gal (biotin-LC-LC-β-galactosidase) was measured with either the spectrophotometric method or the Beta-Glo luminescence assay system.
The molar ratio of sulfo-NHS-LC-LC-biotin to β-galactosidase was determined to yield 1 to 1.5 biotin moieties per enzyme molecule. The degree of biotinylation was determined with Pierce EZ biotin quantitation kit per the manufacturer's instructions, which monitors the binding of 2-(4′-hydroxyazobenzene)benzoic acid (HABA) to avidin by absorbance at 500 nm with an extinction coefficient of 34 mM–1. The displacement of HABA from avidin is proportional to the biotin content in the biotin-LC-LC-β-galactosidase.
The β-galactosidase-8D3 conjugate, also designated β-gal-8D3 (Fig. 1, reaction III), was formed by mixing a 1:1 molar ratio of biotin-LC-LC-β-galactosidase and the 8D3-SA conjugate at 15 min at room temperature. There was no loss in β-galactosidase enzyme activity following monobiotinylation and attachment to the 8D3-SA conjugate (Fig. 1C).
Brain Delivery of β-Galactosidase and β-Galactosidase-8D3 in Adult Mice. Adult female BALB/c mice weighing 20 to 25 g (Charles River Laboratories, Wilmington, MA) were anesthetized with 100 mg/kg ketamine and 10 mg/kg xylazine intraperitoneal. The mice were injected via the jugular vein with either unconjugated β-galactosidase or the β-gal-8D3 conjugate. In the high dose treatment, mice were administered either 150 μg/mouse of unconjugated β-galactosidase or 150 μg/mouse of biotinylated β-galactosidase conjugated to 300 μg/mouse of 8D3-SA. In the low dose treatment, mice were administered either 15 μg/mouse of unconjugated β-galactosidase or 15 μg/mouse of biotinylated β-galactosidase conjugated to 30 μg/mouse of 8D3-SA. The mice were sacrificed at either 1 or 4 h after intravenous (i.v.) injection. The brain, liver, spleen, heart, and kidney were removed, weighed, and frozen on dry ice. The blood from each mouse was collected, heparinized, and stored at –20°C. Organs and blood were also removed from uninjected mice to determine the activity of endogenous β-galactosidase at neutral pH.
β-Galactosidase Enzyme Activity Measurements. The spectrophotometric assay for β-galactosidase enzyme activity (β-galactosidase enzyme assay system; Promega) was not used owing to interference in the absorbance readings by endogenous tissue pigments. Enzyme activity was measured with Promega Beta-Glo luminescence assay system. The tissue was extracted with Promega lysis buffer at a ratio of 2 ml of buffer to 0.5 g of tissue, followed by homogenization with a Brinkmann Polytron PT3000. The homogenate was centrifuged for 10 min at 12,000g, and the supernatant was used to measure β-galactosidase activity with the Promega Beta-Glo assay solution at pH = 7.6. The mixture was incubated in the dark at room temperature for 1 h. The relative light units were measured with a luminometer (PerkinElmer Life and Analytical Sciences, Boston, MA), and the relative light units were converted to milliunits (mU) of enzyme activity based on a β-galactosidase standard curve. The protein content in the organ extract was measured with the BCA reagent. Organ enzyme activity was measured as milliunits per milligram of protein, milliunits per gram of organ weight, or percent of injected dose (ID) per gram of organ weight. The ID was computed from the known specific activity (milliunits of microgram) of the unconjugated β-galactosidase or the β-gal-8D3 conjugate. The endogenous β-galactosidase enzyme activity in uninjected mice was also measured in each organ.
Brain β-Galactosidase Histochemistry. Mice were anesthetized and injected with maximal doses of either unconjugated β-galactosidase (300 μg/mouse) or the β-gal-8D3 conjugate (300 μg/mouse of biotin-LC-LC-β-galactosidase mixed with 600 μg/mouse of 8D3-SA conjugate) via the jugular vein. At 60 min after i.v. injection, the brain plasma volume was cleared with a 4-min infusion of 4 ml of ice-cold PBS into the ascending aorta at a rate of 1 ml/min, followed by a 20-min perfusion of 20 ml of fixative (2% paraformaldehyde in 0.01 M PBS/7.4 with 0.5% glutaraldehyde and 2 mM MgCl2) at a rate of 1 ml/min. The brain was removed and divided into four coronal slabs, and the slabs were immersion-fixed in the same fixative at 4°C for 4 h. The tissue was washed briefly in 0.1 M phosphate-buffered water/7.4 and then placed in 30% sucrose/0.1 M PBS/7.4 for 24 h at 4°C. The brain slab was frozen in Tissue-Tek O.C.T. compound and stored at –70°C until sectioning. Frozen section of 40 μm was prepared on an HM505 microtome (Mikron, San Diego, CA), and β-galactosidase histochemistry was performed with the Invitrogen β-galactosidase staining kit. The frozen section was fixed with 2% formaldehyde and 0.2% glutaraldehyde in 0.01 M PBS/7.4 for 5 min. After washing in PBS, the section was incubated in X-gal staining solution (4 mM potassium ferricyanide, 4 mM potassium ferrocyanide, 2 mM MgCl2, 0.02% IGEPAL CA-630, 0.01% sodium deoxycholate, and 1 mg/ml X-gal, pH 7.4) at 37°C overnight, where X-gal = 5-bromo-4-chloro-3-indoyl-β-d-galactoside. The pH of the incubation was maintained at 7.4 throughout the incubation. After staining with X-gal, the section was briefly washed in distilled water, mounted without counter-staining, and photographed.
A dot blot assay was developed to determine the minimal β-galactosidase enzyme activity that could be detected with a colorimetric histochemical assay. Enzyme (100 μl) was spotted with a Bio-Rad dot blot apparatus in a 3-mm circle to nitrocellulose filter paper in the following amounts: 68, 6.8, 0.68, 0.068, and 0.0068 mU with or without fixation of the blotted filter paper in 0.2% glutaraldehyde in 0.1 M Na2HPO4/7.4/2 mM MgCl2 for 2 min. Enzyme activity in the filter paper was measured with the colorimetric technique (β-galactosidase enzyme assay system; Promega). The amount of enzyme that was barely detected by eye was >2 mU with fixation and >1mU without fixation. A 40-μm section of mouse brain weighs approximately 1 mg. Therefore, it would be necessary to achieve a β-galactosidase enzyme activity >2000 mU/g brain to visualize the enzyme in brain parenchyma with a colorimetric technique such as histochemistry.
Brain Capillary Depletion Method. Mice were anesthetized and injected with the β-gal-8D3 conjugate (150 μg/mouse of biotin-LC-LC-β-galactosidase mixed with 300 μg/mouse of 8D3-SA conjugate) via the jugular vein. At 60 min after i.v. injection, the brain plasma volume was cleared with a 4-min infusion of 4 ml of ice-cold PBS into the ascending aorta at a rate of 1 ml/min. The brain was removed, weighed, and homogenized in an ice-cold physiological buffer (10 mM HEPES, 141 mM NaCl, 4 mM KCl, 2.8 mM CaCl2, 1 mM MgSO4, 1 mM NaH2O4, and 10 mM d-glucose, pH 7.4) with a glass tissue grinder, followed by the addition of ice-cold dextran to a final concentration of 40%. After removal of an aliquot of the homogenate, the remainder was centrifuged at 3200g for 10 min at 4°C, and the supernatant was carefully separated from the capillary pellet with the capillary depletion technique described previously (Triguero et al., 1990). The homogenate, postvascular supernatant, and the capillary pellet were solubilized in Promega lysis buffer. The β-galactosidase enzymatic activity was measured with the Promega Beta-Glo assay system as described above and reported as milliunits per gram of brain for each of the three fractions. The enzyme activity of a brain microvascular-enriched enzyme, γ-glutamyl transpeptidase (GTP), was measured with a spectrophotometric assay using l-γ-glutamyl-p-nitroanilide as a substrate (Diagnostic Chemicals Ltd., Charlottetown, PE, Canada). GTP is highly enriched at the brain microvasculature and is also expressed in brain parenchymal cells. The volume of the microvascular endothelium in brain is ∼0.1% of the total intracellular volume in brain. Therefore, GTP enzyme activity in the postvascular supernatant approximates the GTP enzyme activity in the starting brain homogenate; however, the ratio of GTP enzyme activity in the vascular pellet, relative to the brain homogenate, is a quantitative measure of the extent to which the postvascular supernatant has been depleted of the microvascular component of brain (Triguero et al., 1990).
Results
Following the i.v. administration of the low dose (15 μg/mouse) of the unconjugated β-galactosidase, the enzyme was rapidly cleared from blood by liver, spleen, and kidney (Fig. 2). The enzyme was cleared by liver and spleen after the i.v. administration of the high dose (150 μg/mouse) of the unconjugated β-galactosidase (Fig. 3). The high dose caused minimal saturation of the uptake of the unconjugated enzyme by liver and spleen. The 60-min enzyme activity in liver was 1144 ± 190 mU/g (7.7 ± 0.9 mU/mg protein) and 41,086 ± 8497 mU/g (386 ± 84 mU/mg protein) after the i.v. injection of the low dose and high dose, respectively, of unconjugated enzyme. The 60-min enzyme activity in spleen was 3038 ± 384 mU/g (21 ± 3 mU/mg protein) and 32,686 ± 5777 mU/g (367 ± 68 mU/mg protein) after the i.v. injection of the low dose and high dose, respectively, of unconjugated enzyme. The brain uptake of the unconjugated enzyme was minimal at both the low dose (Fig. 2, inset) and the high dose of enzyme (Fig. 3, inset). The 60-min enzyme activity in brain was 121 ± 3 mU/g (2.0 ± 0.1 mU/mg protein) and 116 ± 26 mU/g (2.8 ± 0.6 mU/mg protein) after the i.v. injection of the low dose and high dose, respectively, and both values approximated the endogenous enzyme activity in the uninjected mouse brain, 85 ± 3 mU/g (1.5 ± 0.1 mU/mg protein).
Conjugation of the enzyme to the TfRmAb accelerated uptake in peripheral tissues and the largest increases in uptake were observed for liver and spleen at the low dose of enzyme (Fig. 2). At the low dose, the brain uptake of β-galactosidase was increased 10-fold following conjugation to the TfRmAb (Fig. 2, inset). At the high dose, the uptake of the enzyme-TfRmAb conjugate by liver and spleen showed saturation (Fig. 3), whereas the brain uptake was still increased 10-fold following conjugation to the TfRmAb (Fig. 3, inset).
The β-galactosidase enzyme activity was rapidly eliminated from mouse organs in vivo, and organ enzyme activity at 4 h after an i.v. injection was no greater than the endogenous organ enzyme activity at the high dose injection (Table 1). The rapid clearance of β-galactosidase from blood in the adult mouse is shown in Table 2, which indicates the blood concentration of the unconjugated enzyme is <1% ID/ml blood at 60 min after i.v. injection.
The brain uptake of the unconjugated β-galactosidase or the β-galactosidase-TfRmAb conjugate was measured with histochemistry after treatment with maximal doses (see Materials and Methods). At 60 min after an i.v. injection of the unconjugated enzyme, there is no measurable enzyme activity in brain in either the parenchymal or capillary compartment (Fig. 4C). At 60 min after an i.v. injection of the high dose of the β-galactosidase-TfRmAb conjugate, the enzyme product is detected by histochemistry in the capillary compartment throughout the entire brain, including cerebellum (data not shown) and a representative low magnification view is shown in Fig. 4B. High magnification microscopy (Fig. 4A) shows the enzyme within the microvascular endothelium; this enzyme activity is localized to the intraendothelial compartment, and not the plasma compartment, because the brain was saline cleared prior to perfusion fixation for histochemistry (see Materials and Methods). The brain vasculature was effectively cleared of enzyme as shown by the absence of vascular enzyme product following injection of the unconjugated enzyme (Fig. 4C). Histochemical product in brain parenchyma was not clearly detectable because the brain β-galactosidase enzyme activity was less than the threshold for colorimetric detection, 2000 mU/g (see Materials and Methods). However, brain parenchymal enzyme activity was demonstrated with the capillary depletion technique and a luminescence-based assay of brain β-galactosidase enzyme activity (Table 3). More than 90% of the brain β-galactosidase enzyme activity was localized to the postvascular supernatant compartment at 60 min following intravenous administration of the high dose of the β-galactosidase-TfRmAb conjugate (Table 3). Measurement of the activity of a brain microvascular-enriched enzyme, GTP, showed that the vascular pellet GTP activity was 16-fold greater than GTP activity in the brain homogenate, indicating the postvascular supernatant was 94% depleted of the capillary compartment (Table 3).
Discussion
The results of this study are consistent with the following conclusions. First, unconjugated β-galactosidase is rapidly cleared from blood in vivo (Table 2), owing to rapid uptake of the unconjugated enzyme by liver and spleen (Figs. 2 and 3). Second, once inside cells, β-galactosidase is rapidly degraded in vivo such that 99% of the organ enzyme activity is lost at 4 h after an intravenous injection (Table 1). Third, the 116-kDa β-galactosidase (Fig. 1B) can be conjugated to the 8D3 TfRmAb without loss of enzyme activity (Fig. 1C). Fourth, there is minimal brain uptake of the unconjugated β-galactosidase, but there is a 10-fold increase in brain uptake of enzyme following conjugation to the 8D3 TfRmAb (Table 1; Figs. 2 and 3).
The β-galactosidase is rapidly removed from the blood due to the avid uptake of the enzyme by liver and spleen (Figs. 2 and 3), which confirms the earlier observation of Onodera et al. (1983). The blood concentration of the β-galactosidase-TfRmAb is 5- to 10-fold higher than the corresponding blood concentration of unconjugated β-galactosidase at 1 to 4 h after injection (Table 2). The delayed clearance of the enzyme-TfRmAb conjugate may be related to the larger size of the enzyme-TfRmAb conjugate, compared with the unconjugated enzyme. Following transport of the unconjugated β-galactosidase into cells, the enzyme is rapidly degraded in vivo, and the organ enzyme activity at 4 h after intravenous injection is no greater than the endogenous enzyme activity (Table 1). In liver, the organ enzyme activity following intravenous injection of unconjugated β-galactosidase decreases 275-fold between 1 and 4 h after administration (Table 1). This observation confirms an earlier study on the rapid degradation of bacterial β-galactosidase by mouse organs in vivo (Scherpereel et al., 2001).
In contrast to the rapid uptake of the unconjugated enzyme by peripheral tissues, the brain uptake of unconjugated β-galactosidase is nil, as shown by the absence of any change in brain enzyme activity following injection of the low and high enzyme doses (see Results). The brain delivery of β-galactosidase was accomplished in the present work by conjugating the enzyme to the TfRmAb. The bacterial β-galactosidase was conjugated to the 8D3 TfRmAb via a streptavidin-biotin linker. The 50-kDa recombinant SA was conjugated to the 8D3 TfRmAb via a stable thiol-ether linker (Fig. 1A, reaction I) in parallel with the monobiotinylation of β-galactosidase (Fig. 1A, reaction II). The enzyme was monobiotinylated with NHS-LC-LC-biotin (see Materials and Methods) which positions a 14-atom spacer between the biotin and the ϵ-amino group of a surface lysine residue on the enzyme. This long spacer arm eliminates steric hindrance of the enzyme caused by attachment to the SA. Consequently, there is no loss of enzyme activity following attachment to the TfRmAb (Fig. 1C). The monobiotinylation of the enzyme was confirmed with the HABA assay (see Materials and Methods). It is important that the enzyme is conjugated with only one biotin moiety. Higher degrees of biotinylation cause the formation of high molecular weight aggregates, owing to the multivalency of SA binding of biotin (Green, 1975). Such high molecular weight aggregates are rapidly removed in vivo. Conversely, attachment of the β-galactosidase to the TfRmAb via the SA-biotin linkage delayed the removal of the enzyme from blood, as the 60-min blood concentration of the β-galactosidase-TfRmAb was 5-fold greater than the corresponding blood concentration for the unconjugated enzyme (Table 2).
The molecular weight of the recombinant SA is 50 kDa and the molecular weight of the recombinant β-galactosidase is 116 kDa (Fig. 1B). The combined size of the SA and the β-galactosidase is 166 kDa or approximately 10% greater than the size of the 150 kDa 8D3 TfRmAb. Despite the large size of the β-galactosidase, the 8D3 TfRmAb targets the enzyme to mouse brain in vivo (Fig. 2, inset). There is a 10-fold increase in the brain uptake of the β-galactosidase following conjugation of the 8D3 TfRmAb, relative to the unconjugated enzyme (Fig. 2, inset). When expressed per gram of brain tissue, the peak β-galactosidase enzyme activity in brain was 484 ± 62 mU/g brain at 60 min following the i.v. injection of the high dose of the β-galactosidase-8D3 conjugate. This level of β-galactosidase enzyme activity in brain cannot be detected with histochemistry using colorimetric methods such as the standard X-gal technique, where a minimal enzyme activity level of 2000 mU/g is required (see Materials and Methods). It was not possible to inject even larger amounts of enzyme-mAb conjugate, because the dose used for the histochemical study in Fig. 3 is a saturating concentration of the TfRmAb. The dose of 300 μg of β-galactosidase conjugated to 600 μg of 8D3-SA per mouse is equivalent to 12 mg/kg β-galactosidase and 24 mg/kg 8D3-SA conjugate, and this dose of 8D3 TfRmAb completely saturates the BBB TfR. The BBB transport of the mAb is 50% saturated at a systemic dose of 2 to 4 mg/kg 8D3 mAb (Lee et al., 2000). Although the β-galactosidase enzyme activity could not be detected in brain parenchyma with the histochemical method, the presence of the enzyme in the intraendothelial compartment of brain could be detected following the intravenous administration of the high dose of the β-galactosidase-8D3 conjugate (Fig. 4, A and B). This histochemical assay demonstrates the targeting of the enzyme to the BBB compartment of brain, whereas no measurable enzyme activity was detected in the endothelial compartment following intravenous injection of the unconjugated enzyme (Fig. 4C). The histochemical product in the endothelial compartment of brain was not due to entrapment of the enzyme in the blood compartment because the brain was saline cleared prior to perfusion fixation for the histochemistry (see Materials and Methods). The adequacy of the saline clearance is demonstrated by the inability to detect histochemical product in the capillary compartment following injection of the unconjugated enzyme (Fig. 4C).
It is possible to detect the β-galactosidase enzyme activity in the endothelial cell of brain because this compartment has such a small volume. The intraendothelial compartment in brain, <1 μl/g, is about 1000-fold lower than the extravascular volume in brain (Pardridge, 2001). Therefore, when the enzyme-TfRmAb conjugate passes through the endothelial compartment, the enzyme activity is concentrated in the small endothelial volume, which allows for light microscopic histochemical detection. An identical intraendothelial vascular staining pattern was reported previously following systemic administration of a TfRmAb conjugated to 5 nm of gold (Bickel et al., 1994). The transport of the β-galactosidase-TfRmAb conjugate across the BBB and into brain parenchyma was demonstrated with the capillary depletion technique (see Materials and Methods) as shown in Table 3. The β-galactosidase enzyme activity in the postvascular supernatant is >90% of the corresponding enzyme activity in the homogenate following i.v. injection of the β-galactosidase-TfRmAb conjugate (Table 3). Therefore, more than 90% of the β-galactosidase-8D3 conjugate that enters into the endothelial compartment passes through the BBB to enter brain parenchyma. This observation is in accord with prior work, which showed that >80% of the TfRmAb undergoes transcytosis through the BBB and into brain parenchyma within a 10-min internal carotid artery perfusion of brain (Skarlatos et al., 1995).
The β-galactosidase enzyme activity in brain, and in other organs, returns to the endogenous level by 4 h after an intravenous injection of the enzyme-TfRmAb conjugate (Table 1), which confirms earlier work by Scherpereel et al. (2001). This rapid inactivation of either the unconjugated enzyme or of the β-galactosidase-TfRmAb conjugate contrasts with prior work showing peak β-galactosidase enzyme activity in brain is observed between 4 and 8 h following the i.v. injection in mice of a fusion protein of β-galactosidase and the cationic peptide of the TAT protein of the human immunodeficiency virus (Schwarze et al., 1999). Conjugation of the cationic TAT peptide to a protein causes a >20-fold increase in the organ clearance of the protein in vivo (Lee and Pardridge, 2001). Therefore, a conjugate of the TAT cationic peptide and β-galactosidase might be rapidly eliminated from blood and organs such as brain in vivo.
In conclusion, these studies demonstrate the feasibility of delivering enzymes across the BBB with receptor-specific monoclonal antibodies that cross the BBB via endogenous receptor-mediated transport systems. With respect to protein targeting to the human brain, a mAb to the human insulin receptor (HIR) is 9-fold more active as a BBB drug delivery vector than is the TfRmAb (Pardridge, 2001). A genetically engineered form of the HIRmAb has been produced and is equally effective as a targeting vector as the original murine HIRmAb (Coloma et al., 2000). A fusion protein comprised of a therapeutic enzyme and a genetically engineered HIRmAb could prove to be of therapeutic value for the treatment of lysosomal storage disorder of the brain in humans.
Footnotes
-
doi:10.1124/jpet.104.082974.
-
ABBREVIATIONS: BBB, blood-brain barrier; mAb, monoclonal antibody; TfR, transferrin receptor; SA, streptavidin; NHS, N-hydroxysuccinimide; LC, long-chain; S-SMPB, sulfosuccinimidyl-4-(p-malimidophenyl)butyrate; BCA, bicinchoninic acid; PBS, phosphate-buffered saline; HABA, 2-(4′-hydroxyazobenzene)benzoic acid; β-gal, β-galactosidase; ID, injected dose; GTP, γ-glutamyl transpeptidase; 8D3-SA, conjugate of 8D3 TfRmAb and SA; HIR, human insulin receptor.
- Received December 29, 2004.
- Accepted February 14, 2005.
- The American Society for Pharmacology and Experimental Therapeutics