Feature Article

Pharmacologic manipulation of lysosomal enzyme transport across the blood–brain barrier

Journal of Cerebral Blood Flow & Metabolism 2016, Vol. 36(3) 476–486 ! Author(s) 2015 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0271678X15614589 jcbfm.sagepub.com

Akihiko Urayama1, Jeffrey H Grubb2,3, William S Sly3 and William A Banks4

Abstract The adult blood–brain barrier, unlike the neonatal blood–brain barrier, does not transport lysosomal enzymes into brain, making enzyme replacement therapy ineffective in treating the central nervous system symptoms of lysosomal storage diseases. However, enzyme transport can be re-induced with alpha-adrenergics. Here, we examined agents that are known to alter the blood–brain barrier transport of large molecules or to induce lysosomal enzyme transport across the blood–brain barrier (()epinephrine, insulin, retinoic acid, and lipopolysaccharide) in 2-week-old and adult mice. In 2-week-old adolescent mice, all these pharmacologic agents increased brain and heart uptake of phosphorylated human b-glucuronidase. In 8-week-old adult mice, manipulations with ()epinephrine, insulin, and retinoic acid were significantly effective on uptake by brain and heart. The increased uptake of phosphorylated human b-glucuronidase was inhibited by mannose 6-phosphate for the agents ()epinephrine and retinoic acid and by L-NG-nitroarginine methyl ester for the agent lipopolysaccharide in neonatal and adult mice. An in situ brain perfusion study revealed that retinoic acid directly modulated the transport of phosphorylated human b-glucuronidase across the blood–brain barrier. The present study indicates that there are multiple opportunities to at least transiently induce phosphorylated human b-glucuronidase transport at the adult blood–brain barrier.

Keywords Blood–brain barrier, b-glucuronidase, enzyme replacement therapy, brain uptake, mannose 6-phosphate receptor Received 14 July 2015; Revised 23 September 2015; Accepted 24 September 2015

Introduction Lysosomal storage diseases (LSDs) are hereditary disorders that are characterized by metabolic abnormalities caused by enzymatic dysfunction in lysosomes.1,2 Most of the LSDs are caused by the lack of one of the lysosomal hydrolases, a group of enzymes involved in the metabolism of glycoproteins, sphingolipids, and glycosaminoglycans and whose absence results in diverse clinical manifestations. The symptoms of LSDs are commonly progressive in the central nervous system (CNS) and peripheral organs, leading to early death.1,2 Enzyme replacement therapy (ERT) by the intravenous route is effective in improving symptoms in peripheral organs,2,3 but has limited effects on the CNS because the blood–brain barrier (BBB) blocks access of the lysosomal enzyme to brain. In our prior studies,4,5 we discovered that luminally expressed cation-independent mannose 6-phosphate

(CI-M6P) receptor is a transporter at the neonatal BBB for lysosomal enzymes that contain M6P moieties. The presence of this transporter allows parenterally administered enzyme to access the neonatal brain in therapeutic amounts.6,7 However, the M6P 1 Department of Neurology, University of Texas Medical School at Houston, Houston, TX, USA 2 Lysosomal Research, Ultragenyx Pharmaceutical Inc., Novato, CA, USA 3 Edward A. Doisy Department of Biochemistry and Molecular Biology, Saint Louis University School of Medicine, St. Louis, MO, USA 4 Geriatric Research, Education, and Clinical Center, Veterans Affairs Puget Sound Health Care System, Seattle, WA, USA; Division of Gerontology and Geriatric Medicine, Department of Medicine, University of Washington School of Medicine, Seattle, WA, USA

Corresponding author: Akihiko Urayama, Department of Neurology, University of Texas Medical School at Houston, Houston, TX, USA. Email: [email protected]

Urayama et al. receptor-dependent transport becomes inactive with development and so is not available to deliver lysosomal enzymes across the adult BBB.4,5 Conceptually, if one can re-induce CI-M6P receptor-mediated transport of lysosomal enzymes in adult BBB, this strategy could provide a novel brain targeting approach for treating abnormal storage in the CNS regardless of the age of subjects. We previously showed that systemic epinephrine restored CI-M6P receptor-mediated transport of recombinant phosphorylated human b-glucuronidase (P-GUS) across the adult BBB.8 Recently, adrenergic stimulation was found to improve the therapeutic CNS outcome of lysosomal ERT in an animal model of Pompe disease, another disease in which enzyme delivery is dependent on the CI-M6P receptor. Interestingly, there is accumulating evidence indicating that some lysosomal enzymes may employ multiple mechanisms for their transport across the BBB.10 For example, arylsulfatase A employs both CI-M6P receptor transport and adsorptive transcytosis in an in vitro model of the BBB.11 The mechanism of adsorptive transcytosis is most likely to occur at high doses of ERTs, including mucopolysaccharidoses,6,12,13 aspartylglucosaminuria,14 metachromatic leukodystrophy,15 and a-mannosidosis.16,17 This suggests that if circulating levels of lysosomal enzyme can be maintained at high levels for a long time, the availability of the enzyme to be delivered into the CNS increases in some LSDs. In the present study, we have further investigated systemic responses to pharmacologic stimuli of known BBB transport mediators on the induction of M6P receptor-mediated and/or adsorptive-mediated transport processes in the BBB. Pharmacological agents for macromolecule transport across the BBB include () epinephrine, insulin, retinoic acid, lipopolysaccharide (LPS). The mechanisms of induced transport were investigated by inhibition studies with M6P and L-NG-nitroarginine methyl ester (L-NAME). We used P-GUS as a model lysosomal enzyme to be delivered into the brain. We here report evidence that multiple mechanisms can enhance the delivery of PGUS into the brain by manipulating endogenous characteristics of the BBB.

477 then desalted over Bio Gel P6 sizing resin (Bio-Rad, Hercules, CA). Recombinant non-phosphorylated human GUS (NP-GUS) was produced in Sf21 insect cells by using the baculovirus system as described for mouse GUS.19 The concentration of the P-GUS was adjusted to 2.5  105 units per mL (1 unit ¼ 1 nmol of substrate cleaved per hour), and both purified enzymes were stored at 70 C. M6P-specific uptake of the P-GUS purified enzymes by human fibroblasts was 185 units per mg/h, respectively (data not shown).

Radioactive labeling P-GUS and NP-GUS were radioactively labeled using the iodobead method (Pierce, Rockford, IL) with [131I]Na (Perkin Elmer, Waltham, MA), as we reported earlier in the context of labeling GUS. The use of a single iodobead allows controlled iodination, so that both enzymatic activity and susceptibility to endocytosis by the M6P receptor are preserved.4,5,8 Labeled, active enzyme was separated from free iodine on a Sephadex G-10 column. Albumin was labeled with [125I]Na (Perkin Elmer) using the chloramine-T method and purified on a column of Sephadex G-10. Each reagent was freshly prepared on the day of the experiment.

Animals Male CD-1 mice from our in-house colony were studied at two days (0.8–1.3 g), two weeks (4.2–5.5 g), and 7–8 weeks (19.0–26.2 g) of age. The mice had free access to food and water and were maintained on a 12-h dark/light cycle in a room with controlled temperature. The studies were carried out in accordance with the recommendations in the guide for the Care and Use of Laboratory Animals of the National Institutes of Health: Reporting of In Vivo Experiments (ARRIVE) Guidelines. The protocols were approved by the St. Louis VA Animal Care and Use Committee and carried out in a facility approved by the Association for Assessment and Accreditation of Laboratory Animal Care.

Drug administration and experimental procedures Materials and methods Production of recombinant human lysosomal enzymes P-GUS was produced in overexpressing, CI-M6P receptor-deficient mouse L cells as described in Grubb et al.18 The enzyme was purified from conditioned media by anti-human GUS mAb affinity column chromatography. P-GUS was eluted with 3.5 M MgCl2,

Mice were anesthetized with 40% urethane before receiving an intravenous (i.v.) injection of [131I]P-GUS with [125I]albumin (5.5  105 cpm of each) into the superficial temporal vein (neonates) or jugular vein (adults). At designated times from 1 to 10 min after the injection, mice were sacrificed under anesthesia, and the blood, brain, and heart were collected immediately. Serum from mouse blood was isolated by centrifugation. To modulate brain uptake of

478

Journal of Cerebral Blood Flow & Metabolism 36(3)

[131I]P-GUS following pharmacologic agents were i.v. co-injected; racemic ()epinephrine (1.6 mmol/kg), M6P (2 mmol/injection), insulin (20 mmol/kg), and retinoic acid (0.01–10 nmol/injection). Due to poor water solubility of retinoic acid, the i.v. dose was not strictly proportional to the body weight of mice. The concentration of retinoic acid in injection solution was at 5.2 nmol/100 mL. However, graded-dosing of retinoic acid was given in neonate to adult mice with the limitation of the volume of injection up to 10 mL in neonates and 200 mL in adults not to complicate vascular volume in mice. LPS (3 mg/kg) was injected three times with 10 min interval 30 min prior to the injection of [131I]P-GUS, and L-NAME (10 mg/kg) was injected i.v. 30 min prior to the treatment with LPS. All reagents were from commercial sources. The doses of these agents were determined on the basis of data from our prior studies or established doses from representative pharmacologic studies.20–24

In situ transcardiac brain perfusion Mice at the age of two weeks were anesthetized with 40% urethane, the heart exposed, both jugulars severed, and the descending thoracic aorta ligated. A 26-gauge butterfly needle was inserted into the left ventricle of the heart, and the perfusion fluid, containing [131I]P-GUS and [125I]albumin with or without retinoic acid (3 M), was infused at a rate of 2 mL/min for 1, 3, or 5 min. After perfusion, the whole brain was removed and weighed. The level of radioactivity was determined in a g-counter. [131I]P-GUS (200,000 cpm/ mL) with or without retinoic acid was diluted in prewarmed physiological buffer (7.19 g/L NaCl, 0.3 g/L KCl, 0.28 g/L CaCl2, 2.1 g/L NaHCO3, 0.16 g/L KH2PO4, 0.17 g/L anhydrous MgCl2, 0.99 g/L d-glucose, and 1% wt/vol BSA). The perfusion fluid was freshly prepared each day.

Evaluation of tissue uptake The brain/serum, tissue/serum, or brain/perfusion ratios for [131I]P-GUS and [125I]albumin were calculated by dividing the radioactivity in a gram of brain by the radioactivity in a milliliter of serum, having a dimension of the volumes of distribution in brain. To estimate the blood-to-brain unidirectional influx rate (Kin) of radioactively labeled agents in the in situ perfusion study, the brain/perfusion ratios were plotted against perfusion time T in min using the following equation. Am=Cp ¼ Kin T þ Vi

ð1Þ

where Am and Cp are the cpm/g of brain and the cpm/mL of perfusion fluid at time T, respectively. Kin was measured as the slope for the linear portion of the relation between the brain/perfusion ratios and respective perfusion time T. The y-intercept of the line represents Vi, the distribution volume in the brain at T ¼ 0.

Brain clearance study Brain clearance was measured as previously described in Banks et al.25 Mice at the age of 7–8 weeks old received an intracerebroventricular (i.c.v.) injection of [131I]P-GUS or [131I]NP-GUS into lateral ventricle. [125I]albumin was simultaneously injected with the radioactively labeled enzyme. Radioactivity of each agent injected was 2  104 cpm per animal, and the brain was dissected at 0, 2, 5, 10, 20, and 30 min after i.c.v. injection. Radioactivity remaining in the brain was normalized with the percent of injected dose (%ID) per brain and plotted against time (t) to measure the efflux rate from brain with the equation. Log ½%ID=brain ¼ mt þ i

ð2Þ

where m is the slope and i is the y-intercept of the line.

Statistical analysis The mean values are presented with their SEs and compared by one-way analysis of variance followed by Dunnet’s or Newman-Keuls multiple comparison tests, or by two-tailed unpaired t test for linear regression results using the Prism 5.0 program (GraphPad Inc, San Diego, CA).

Results Manipulation of brain uptake of -glucuronidase by endogenous characteristics of the BBB We studied the acute effects of pharmacological manipulations on [131I]P-GUS delivery into brain using known modulators of macromolecule transport across the BBB, including ()epinephrine, insulin, retinoic acid, and LPS. Doses of these agents were based on previous studies.20–24 [131I]P-GUS transport across the BBB was measured in mice at the ages from 2 to 8 weeks, the time period when mice gradually lose the functional transport of lysosomal enzymes mediated through the CI-M6P receptor at the BBB.4,5 Figure 1 shows that brain and heart uptakes of [131I]P-GUS 10 min after simultaneous i.v. injection of insulin or racemic ()epinephrine in 2- and 8-week-old mice. Tissue/serum ratios for [131I]P-GUS and [125I]albumin

Urayama et al.

479

Figure 1. [131I]P-GUS uptake by brain and heart induced by insulin and ()epinephrine treatment. Tissue-to-serum ratios for [131I]P-GUS and [125I]albumin were simultaneously measured 10 min after i.v. co-injection with insulin (20 mmol/kg), ()epinephrine (1.6 mmol/kg), and ()epinephrine (1.6 mmol/kg) with M6P (2 mmol/injection). Tissue uptakes of [131I]P-GUS and [125I]albumin were shown. Delta values represent the uptake of [131I]P-GUS normalized with tissue vascular space estimated by the vascular marker [125I]albumin. Panels (a) and (c) show brain- and heart-to-serum ratios in 2-week-old mice, respectively. Panels (b) and (d) show that brain- and heart-to-serum ratios in 8-week-old mice, respectively. Values are expressed as mean  SE of 8–10 determinations. Oneway ANOVA followed by Dunnet’s post hoc test was employed for statistical evaluation. Asterisks show that statistical difference from control values, *P < 0.05, **P < 0.01, and ***P < 0.001.

(vascular space marker) are shown. Delta values, calculated by subtracting the [125I]albumin values from respective [131I]P-GUS values, represent the net uptake of [131I]P-GUS, indicating extravascular localization of [131I]P-GUS in tissues. The [131I]P-GUS uptake values in the brain and heart were significantly increased by both agents, while [125I]albumin levels remained unchanged, confirming that vascular spaces in tissues were not affected by either insulin or ()epinephrine treatment. A lack of effect on the [125I]albumin levels also suggests that the integrity of the BBB was not altered by these pharmacological manipulations. The increases in net uptake of [131I]PGUS in brain in 2- and 8-week-old mice were 1.4–2.0 fold for insulin and 2.4–2.9 fold for ()epinephrine, compared to those in untreated age-matched control mice. Increased uptake was also observed in the heart where insulin increased the net uptake 1.5–1.6 fold and the increase for ()epinephrine was 3.9–4.5 fold in neonatal and adult mice. Epinephrine-induced uptake of [131I]P-GUS in the brain and heart in adult mice was inhibited by M6P (Figure 1) which was consistent with

our prior study.8 Next, we tested the effects of LPS and L-NAME on brain uptake of [131I]P-GUS (Figure 2). LPS treatments significantly increased brain and heart uptakes of [131I]P-GUS 10 min after injection of [131I]P-GUS in 2-week-old mice. LPS-induced tissue uptake of [131I]P-GUS was significantly inhibited by L-NAME injected 30 min prior to LPS dosing, suggesting that the increased tissue uptake of [131I]P-GUS by LPS was mediated through nitric oxide activity. However, these effects on [131I]P-GUS uptake were not observed in adult mice. Neither of these two agents affected vascular spaces ([125I]albumin) in the brain and heart, regardless of the age of mice. Similar results although less robust were obtained with sulfamidase (data not shown), suggesting these agents work on other lysosomal enzymes.

Retinoic acid enhanced brain uptake of -glucuronidase Brain endothelial cells highly express retinoic acidinduced transcriptomes,26 and retinoic acid has diverse

480

Journal of Cerebral Blood Flow & Metabolism 36(3)

Figure 2. [131I]P-GUS uptake by brain and heart induced by lipopolysaccharide treatment. Tissue-to-serum ratios for [131I]P-GUS and [125I]albumin were simultaneously measured 10 min after i.v. co-injection with lipopolysaccharide (3 mg/kg) and L-NAME (10 mg/kg). Tissue uptakes of [131I]P-GUS and [125I]albumin were shown. Delta values represent the uptake of [131I]P-GUS normalized with tissue vascular space estimated by the vascular marker albumin. Panels (a) and (c) show that brain- and heart-to-serum ratios in 2week-old mice, respectively. Panels (b) and (d) show that brain- and heart-to-serum ratios in 8-week-old mice, respectively. Values are expressed as mean  SE of 5–8 determinations. One-way ANOVA followed by Dunnet’s post hoc test was employed for statistical evaluation. Asterisks show that statistical difference from control values, *P < 0.05 and **P < 0.01.

biologic effects on cellular functions. For example, retinoic acid promotes intracellular redistribution of CI-M6P receptors.27,28 Thus, we tested the effect of retinoic acid on brain uptake of P-GUS in neonate and adult mice. Figure 3 shows the distribution of [131I]P-GUS in the brain, heart, and serum 10 min after i.v. co-injection with retinoic acid in 2-day-, 2-week-, and 8-week-old mice. The dashed line represents the vascular space as measured by [125I]albumin in each tissue at different ages. Brain uptake of [131I]PGUS induced by retinoic acid increased in a dosedependent manner in younger mice at the ages of two days or two weeks, and there was a significant increase in [131I]P-GUS uptake with the dose of >1 nmol of retinoic acid. Such induction of [131I]P-GUS uptake was inhibited by the presence of M6P and reduced to the levels of vascular space as measured by [125I]albumin in two-day-old mice. The serum [131I]P-GUS levels were significantly increased by M6P (2 mmol per mouse), implying that retinoic acid-induced uptake of [131I]PGUS by brain and heart is mediated through CI-M6PR. Dose-dependency in [131I]P-GUS uptake was not observed in 8-week-old mice, but retinoic acid at the highest dose (10 nmol) significantly

increased [131I]P-GUS uptake both in the brain and heart, and such increases in tissue uptake were higher than respective vascular spaces in these tissues.

Retinoic acid directly interacted with the BBB to increase brain uptake of -glucuronidase from vasculature Next, we performed in situ transcardiac brain perfusion of [131I]P-GUS and [125I]albumin with retinoic acid in 2-week-old mice (Figure 4). This approach enables one to measure direct effects of retinoic acid on [131I]P-GUS uptake at the luminal surface of vascular endothelial cells. In this method, perfusion fluid containing [131I]P-GUS and [125I]albumin with or without retinoic acid goes directly to the cerebral vasculature, negating first pass effects through lung or liver as well as peripheral factors such as binding proteins which may interact with the substance of interest and so alter its BBB transport. We used serum-free artificial perfusion fluid both to negate the interaction of study substances with blood components and also to limit the site of action for P-GUS uptake by retinoic acid at the cerebral vasculature. In a prior study, the transport of P-GUS was

Urayama et al.

481

Figure 3. Changes in [131I]P-GUS levels in the brain, heart, and serum by retinoic acid treatment. Tissue-to-serum ratios of [131I]PGUS in the brain (a), heart (b), and serum (c) after i.v. co-injection with retinoic acid in 2-day-old mice. Brain (panels (d) and (f)) and heart (panels (e) and (g)) uptakes of [131I]P-GUS in 2- and 8-week-old mice are shown, respectively. Each column represents mean  SE of 4–13 determinations. The dashed line represents tissue vascular space simultaneously measured by [125I]albumin co-injection. One-way ANOVA followed by Dunnet’s post hoc test was employed for statistical evaluation among control and retinoic acid treated groups. A two-tailed unpaired t test was used to compare the highest concentration of retinoic acid and M6P treated groups. Asterisks show a statistical difference from control values, *P < 0.05 and **P < 0.01. Daggers show a statistical difference between the groups receiving retinoic acid (1 nmol/injection) and retinoic acid with M6P (2 mmol/injection), yyyP < 0.001.

not detected in adult mice by the in situ brain perfusion method.4 However, we found that 2-week-old mice retained active transport of [131I]P-GUS at the BBB in the present study (Figure 4). While [131I]P-GUS was transported across the BBB at the influx rate of 3.76  0.71 mL/g/min in 2-week-old mice, retinoic acid at 3 M significantly increased the influx rate of

[131I]P-GUS about 1.7 times (6.49  0.33 mL/g/min) without changing cerebral vascular space simultaneously measured by [125I]albumin. These results suggest that retinoic acid acts directly on the cerebral vasculature to affect CI-M6P-dependent luminal-to-abluminal transport of [131I]P-GUS across the BBB (Figures 3 and 4).

482

Journal of Cerebral Blood Flow & Metabolism 36(3)

Figure 4. In situ transcardiac brain perfusion of [131I]P-GUS and [125I]albumin with retinoic acid in 2-week-old mice. Time-courses of the brain/perfusate ratio of [131I]P-GUS and [125I]albumin with and without retinoic acid at the dose of 3 M (a), the unidirectional brain influx rate of [131I]P-GUS (b), and the initial volumes of distribution of [131I]P-GUS (c) are shown, respectively. Each point and column represents the mean  SE of six determinations. A Neuman-Keuls test was employed to evaluate statistical significance. Asterisks show a statistical difference from control values, **P < 0.01.

Brain clearance of -glucuronidase was non-specific and independent from its phosphorylation state To investigate subarchinoidal clearance of GUS by the bulk flow of CSF,29 we further investigated the clearance of P-GUS and NP-GUS after i.c.v. injection into the lateral ventricle in adult mice at the age of 7–8 weeks. Either [131I]P-GUS with [125I]albumin or [131I]NP-GUS with [125I]albumin were injected so as to compare directly with albumin clearance (Figure 5). The time courses of brain clearance of [131I]P-GUS, [131I]NP-GUS, and [125I]albumin were identical, suggesting that there was no phosphorylation state dependency in CSF clearance and that no measurable active elimination mechanism for GUS from CSF compartment, such as albumin, is cleared from brain by the mechanism of bulk flow.

Discussion The difficulty in delivering lysosome enzymes across the BBB has been a pivotal issue in the treatment of the

CNS component of LSDs. This is due to the lack of universal methodologies for CNS delivery of therapeutic molecules and incomplete understanding of CNS barriers. In our previous studies,4,5 we have characterized the mechanisms of BBB transport for lysosomal enzymes using the enzyme P-GUS which was resistant to metabolism and retained enzymatic activities during the course of brain uptake. Although the half-life of P-GUS in the systemic circulation is about 5 min, this short half-life is mainly mediated by peripheral tissue sequestration, not by degradation in the blood stream.30,31 This metabolically stable feature of P-GUS aids in studies of its pharmacological manipulations. In earlier studies, we also measured the half-life of GUS in brain and found it to be 4.5 days, suggesting that GUS activity lasts a long time once it gets in brain.32 In the present study, we further investigated pharmacological manipulations for brain delivery of intravenously administered P-GUS into adult brain. Pharmacologic agents for the manipulation of BBB transport include ()epinephrine, retinoic acid, insulin,

Urayama et al.

Figure 5. Brain clearance of [131I]P-GUS, [131I]NP-GUS, and [125I]albumin examined by i.c.v. injection in 7-week-old mice. Mice received an i.c.v. injection of [131I]P-GUS or [131I]NP-GUS into the lateral ventricle. [125I]albumin was co-injected with each enzyme. Time-courses of radioactivity in the brain were plotted against clock time after the injection. There were not statistically significant differences among the groups. Each point represents mean  SE of 4–9 determinations.

and LPS. The doses of pharmacologic agents were determined based on our prior studies testing their effect on BBB permeation.8,20–24 Although it is known that the injection of epinephrine transiently increases systemic blood pressure, the dose (1.6 mmol/kg) of epinephrine used in this study did not compromise the integrity of the BBB in vivo as seen in our previous study.8 LPS (3 mg/kg) potentially increases a vascular permeability in the brain and periphery and would produce sickness behavior. Insulin (20 mmol/kg) could induce hypoglycemia. We used the endogenous characteristics of the BBB to induce brain uptake of a lysosomal enzyme into the adult mouse and addressed potential mechanisms of the induction of transport across the BBB. First, we confirmed that radioactively labeled P-GUS retained saturable transport properties which are re-inducible by the pharmacologic agent ()epinephrine as seen in prior study.8 Thus, we could use epinephrine-induced uptake of P-GUS as a reference for restored transport activity both in the CNS and peripheral tissue. We showed the successful re-induction of CI-M6P receptor-mediated transport for [131I]P-GUS in the brain and heart in 2- and 8-week-old mice. In related experiments, Koeberl et al.9 recently showed enhanced uptake of activity of acid alphaglucosidase by chronic treatment with adrenergic agents in the brain and muscle, and also found that the effect of b2-adrenergics in enhancing the efficacy

483 of the ERT in Pompe disease model mice was, at least in part, mediated through adrenergic-induced, increased expression of CI-M6P receptors in the brain. There was about a 1.4-fold increase in the CI-M6P receptor level by clenbuterol in the cerebellum. Adrenergic receptors are G protein-coupled receptors which participate in numerous signal transduction pathways as well as change the functionality of the surface membrane in endothelial cells. Brain microvessel endothelial cells forming the BBB express both alphaand beta- adrenergic receptors at the lumen. In an early study, adrenergic receptors participated in regulation of barrier function as well as promoted pinocytotic activity at endothelial cell surfaces.33,34 Internalization of adrenergic receptors through the turnover of endocytic vesicles was initially studied as a mechanism of desensitization, which leads to either short-term recycling of receptors or long-term down-regulation by the decrease of the receptor population. The increase in the rate of cell surface receptor disappearance by endocytic vesicles shows the changes in vesicular translocation and activity which is induced by the ligands. In our prior study, we demonstrated that the epinephrine-induced brain uptake of P-GUS in adults was CI-M6P receptor-mediated transport.8 The stimulated uptake of GUS showed that a dependence on phosphorylation, thus indicating M6P moieties on the enzyme GUS regulates the transport, and that epinephrine-induced P-GUS uptake in brain was completely inhibited by co-injection of M6P in a dose-dependent manner.8 That study also suggested that loss of functional transport across the adult BBB of lysosomal enzyme by CI-M6P receptor could be partially reversed by change in intracellular distribution of the receptor, as it was too rapid (10 min) to reflect new synthesis of receptor at the plasma membrane surface in brain microvessel endothelial cells. Retinoic acid plays important biological roles that include regulation of cellular functions. For example, brain tissue contains cellular retinol-binding proteins and a nuclear receptor protein for retinoic acid. Cellular retinol-binding protein was localized to brain microvessel endothelial cells and to the choroid plexus epithelial cell layers. A transcriptome study of the BBB showed that the RXR alpha signaling cascade is specifically enriched in the brain microvessel endothelial cells.26 Retinoic acid promotes intracellular redistribution of CI-M6P receptors which are detected in endosomes, in the trans Golgi element, and at the plasma membrane in varying proportions in different cell types.27,28 In this study, we measured retinoic acidinduced brain uptake of [131I]P-GUS in neonates and adults. The magnitude of increase in brain [131I]P-GUS uptake by retinoic acid decreased with maturation regardless of a substantial increase in doses of retinoic

484

Journal of Cerebral Blood Flow & Metabolism 36(3)

Figure 6. Schematic presentation of pharmacologic inductions of P-GUS transport across adult BBB. Epinephrine-induced brain uptake of P-GUS was mediated through coordinated stimuli of adrenoceptors. Insulin and LPS also contribute P-GUS uptake by brain. Retinoic acid directly interacts with brain microvasculature and retinoic acid-induced brain uptake of P-GUS was M6P sensitive, suggesting the restoration of M6P receptor-mediated uptake process across the BBB.

acid. However, an increase in [131I]P-GUS uptake by brain and heart did occur and was M6P sensitive. Considering the fact that retinoic acid is known to induces relocation of intracellular CI-M6P, we propose that the retinoic acid-induced [131I]P-GUS uptake was mediated through a CI-M6P receptor-mediated process that resulted from redistribution of CI-MPR to the luminal surface of brain microvessel endothelial cells (Figure 4). Furthermore, retinoic acid did not change brain vascular space, suggesting that increased brain uptake of [131I]P-GUS was not due to a compromise in the barrier integrity. In addition, cation-dependent M6P receptor does not bind M6P at neutral pH, suggesting physiologic pH7.4 only allows CI-M6P receptors to be responsible for P-GUS uptake.35,36 Currently, the mechanisms of action for brain uptake of [131I]P-GUS by insulin and LPS remain speculative. The insulin receptor belongs to the receptor tyrosine kinase family which has endothelial nitric oxide synthase (NOS) as its downstream signal transduction. Insulin in the CNS is neuroprotective and retains many of its ancient growth hormone-like properties.37 It is well known that ligand binding to the receptor tyrosine kinases promotes internalization of the activated receptors through clathrin-dependent or caveolae-mediated pathways which modulate the intracellular vesicle activity and re-localization.38 Also, LPS is a potent inducer of NOS activity and L-NAME is a specific inhibitor of NOS. Nitric oxide (NO) can influence vesicle activity and re-location processes, and NO

directly stimulates insulin transport in vascular endothelial cells.39 One of our prior studies showed that LPS affects insulin transport across the BBB by modulating NOS isozyme activity.23 NO released by endothelial NOS and inducible NOS acts to stimulate insulin transport. Considering the increased brain uptake of [131I]PGUS by insulin and LPS (Figures 1 and 2), endocytic activity stimulated by these agents may enable either receptor- or adsorptive-mediated cellular uptake of [131I]P-GUS at luminal surface of brain microvessel endothelial cells which may be mediated through the changes in intracellular vesicle activity and re-localization of CI-M6P receptors.40 It is of interest that epinephrine and insulin affected both adolescents and adults, while LPS had a detectable effect only in adolescents in this single dose study. Figure 6 illustrates some of the complexities of the induction of vesicular transport at the BBB. Epinephrine-induced brain uptake of [131I]P-GUS was mediated through coordinated stimuli of adrenergic receptors as a nature of ()epinephrine binding to receptor subtypes. Retinoic acid directly interacts with the brain microvasculature and retinoic acid-induced brain uptake of [131I]P-GUS was M6P sensitive, suggesting the restoration of CI-M6P receptor-mediated uptake processes across the BBB. Our results suggest that insulin and LPS also contribute to [131I]P-GUS uptake by brain. While we conducted in vivo determinations of transport mechanisms for [131I]P-GUS, one clear finding in the present study is that there are

Urayama et al. multiple opportunities to induce the transport at the adult BBB by modulating endogenous mechanisms. Further studies confirming the effectiveness of these pharmacological manipulations in higher vertebrates are needed. Also, chronic multiple dosing studies such as those of Koeberl et al.9 are necessary to determine which of these pharmacological manipulations might potentially have clinical application to enhancing enzyme delivery to brain. Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by The Sanfilippo Syndrome Medical Research Foundation (WAB and WSS), Veterans Affairs Merit Review (WAB), National Institutes of Health Grants NS050547 and AG029839 (to WAB), and National Institutes of Health Grant GM34182 (to WSS).

Declaration of conflicting interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Authors’ contributions AU and WAB designed the research; AU and JHG performed research; AU and WAB analyzed data; and AU, WSS, and WAB wrote the paper.

References 1. Neufeld EF. Lysosomal storage diseases. Annu Rev Biochem 1991; 60: 257–280. 2. Desnick RJ and Schuchman EH. Enzyme replacement and enhancement therapies: lessons from lysosomal disorders. Nat Rev Genet 2002; 3: 954–966. 3. Desnick RJ and Schuchman EH. Enzyme replacement therapy for lysosomal diseases: lessons from 20 years of experience and remaining challenges. Annu Rev Genomics Hum Genet 2012; 13: 307–335. 4. Urayama A, Grubb JH, Sly WS, et al. Developmentally regulated mannose 6-phosphate receptor-mediated transport of a lysosomal enzyme across the blood–brain barrier. Proc Natl Acad Sci USA 2004; 101: 12658–12663. 5. Urayama A, Grubb JH, Sly WS, et al. Mannose 6-phosphate receptor-mediated transport of sulfamidase across the blood-brain barrier in the newborn mouse. Mol Ther 2008; 16: 1261–1266. 6. Vogler C, Levy B, Galvin NJ, et al. Enzyme replacement in murine mucopolysaccharidosis type VII: neuronal and glial response to beta glucuronidase requires early initiation of enzyme replacement therapy. Pediatr Res 1999; 45: 838–844. 7. Gliddon BL and Hopwood JJ. Enzyme-replacement therapy from birth delays the development of behavior and learning problems in mucopolysaccharidosis type IIIA mice. Pediatr Res 2004; 56: 1–8.

485 8. Urayama A, Grubb JH, Banks WA, et al. Epinephrine enhances lysosomal enzyme delivery across the bloodbrain barrier by up-regulation of the mannose 6-phosphate receptor. Proc Natl Acad Sci USA 2007; 104: 12873–12878. 9. Koeberl DD, Li S, Dai J, et al. b2 agonists enhance the efficacy of simultaneous enzyme replacement therapy in murine Pompe disease. Mol Genet Metab 2012; 105: 221–227. 10. Urayama A. Toward the successful delivery of lysosomal enzymes across the blood-brain barrier. Clin Exp Neuroimmunol 2013; 4: 228–238. 11. Matthes F, Wolte P, Bockenhoff A, et al. Transport of arylsulfatase A across the blood-brain barrier in vitro. J Biol Chem 2011; 286: 17487–17494. 12. Vogler C, Levy B, Grubb JH, et al. Overcoming the blood-brain barrier with high dose enzyme replacement therapy in murine mucopolysaccharidosis VII. Proc Natl Acad Sci USA 2005; 102: 14777–14782. 13. Polito VA, Abbondante S, Polishchuk RS, et al. Correction of CNS defects in the MPSII mouse model via systemic enzyme replacement therapy. Hum Mol Genet 2010; 19: 4871–4885. 14. Dunder U, Kaartinen V, Valtonen P, et al. Enzyme replacement therapy in a mouse model of aspartylglycosaminuria. FASEB J 2000; 14: 361–367. 15. Matzner U, Herbst E, Hedayati KK, et al. Enzyme replacement improves nervous system pathology and function in a mouse model for metachromatic leukodystrophy. Hum Mol Genet 2005; 14: 1139–1152. 16. Roces DP, Lullmann-Rauch R, Peng J, et al. Efficacy of enzyme replacement therapy in alpha-mannosidosis mice: a preclinical animal study. Hum Mol Genet 2004; 13: 1979–1988. 17. Blanz J, Stroobants S, Lullmann-Rauch R, et al. Reversal of peripheral and central neural storage and ataxia after recombinant enzyme replacement therapy in alpha-mannosidosis mice. Hum Mol Genet 2008; 17: 3437–3445. 18. Grubb JH, Kyle JW, Cody LB, et al. Large-scale purification of phosphorylated recombinant human beta-glucuronidase from overexpressing mouse L-cells. FASEB J 1993; 71225a (Abstr). 19. Sands MS, Vogler CA, Ohlemiller KK, et al. Biodistribution, kinetics, and efficacy of highly phosphorylated and non-phosphorylated beta-glucuronidase in the murine model of mucopolysaccharidosis VII. J Biol Chem 2001; 276: 43160–43165. 20. Kaya M, Chang L, Truong A, et al. Chemical induction of fenestrae in vessels of the blood-brain barrier. Exp Neurol 1996; 142: 6–13. 21. Banks WA. Enhanced leptin transport across the bloodbrain barrier by alpha1-adrenergic agents. Brain Res 2001; 899: 209–217. 22. Kastin AJ and Akerstrom V. Glucose and insulin increase the transport of leptin through the blood-brain barrier in normal mice but not in streptozotocin-diabetic mice. Neuroendocrinology 2001; 73: 237–242. 23. Xaio H, Banks WA, Niehoff ML, et al. Effect of LPS on the permeability of the blood-brain barrier to insulin. Brain Res 2001; 896: 36–42.

486 24. Banks WA, Dohgu S, Nakaoke R, et al. Nitric oxide isoenzymes regulate LPS-enhanced insulin transport across the blood-brain barrier. Endocrinology 2008; 149: 1514–1523. 25. Banks WA, Kastin AJ, Fischman AJ, et al. Carriermediated transport of enkephalins and N-Tyr-MIF-1 across blood-brain barrier. Am J Physiol 1986; 251: E477–E482. 26. Daneman R, Zhou L, Agalliu D, et al. The mouse bloodbrain barrier transcriptome: a new resource for understanding the development and function of brain endothelial cells. PLoS One 2010; 5: e13741. 27. Rajawat Y, Hilioti Z and Bossis I. Autophagy: a target for retinoic acids. Autophagy 2010; 6: 1224–1226. 28. Rajawat Y, Hilioti Z and Bossis I. Retinoic acid induces autophagosome maturation through redistribution of the cation-independent mannose-6-phosphate receptor. Antioxid Redox Signal 2011; 14: 2165–2177. 29. Rudick RA, Zirretta DK and Herndon RM. Clearance of albumin from mouse subarachnoid space: a measure of CSF bulk flow. J Neurosci Methods 1982; 6: 253–259. 30. Stahl P, Six H, Rodman JS, et al. Evidence for specific recognition sites mediating clearance of lysosomal enzymes in vivo. Proc Natl Acad Sci USA 1976; 73: 4045–4049. 31. Sly WS, Vogler C, Grubb JH, et al. Enzyme therapy in mannose receptor-null mucopolysaccharidosis VII mice defines roles for the mannose 6-phosphate and mannose receptors. Proc Natl Acad Sci USA 2006; 103: 15172–15177. 32. Vogler C, Sands M, Higgins A, et al. Enzyme replacement with recombinant beta-glucuronidase in the

Journal of Cerebral Blood Flow & Metabolism 36(3)

33.

34.

35.

36.

37. 38. 39.

40.

newborn mucopolysaccharidosis type VII mouse. Pediatr Res 1993; 34: 837–840. Sarmento A, Borges N and Azevedo I. Adrenergic influences on the control of blood-brain barrier permeability. Naunyn-Schmiedebergs Arch Pharmacol 1991; 343: 633–637. Borges N, Shi F, Azevedo I, et al. Changes in brain microvessel endothelial cell monolayer permeability induced by adrenergic drugs. Eur J Pharmacol 1994; 269: 243–248. Tong PY and Kornfeld S. Ligand interactions of the cation-dependent mannose 6-phosphate receptor. Comparison with the cation-independent mannose 6-phosphate receptor. J Biol Chem 1989; 264(14): 7970–7975. Tong PY, Gregory W and Kornfeld S. Ligand interactions of the cation-independent mannose 6-phosphate receptor. The stoichiometry of mannose 6-phosphate binding. J Biol Chem 1989; 264(14): 7962–7969. Banks WA. Insulin in the brain: there and back again. Pharmacol Ther 2012; 136: 82–93. Goh LK and Sorkin A. Endocytosis of receptor tyrosine kinases. Cold Spring Harb Perspect Biol 2013; 5: a017459. Wang H, Wang AX, Aylor K, et al. Nitric oxide directly promotes vascular endothelial insulin transport. Diabetes 2013; 62: 4030–4042. Giudice J, Barcos LS, Guaimas FF, et al. Insulin and insulin like growth factor II endocytosis and signaling via insulin receptor B. Cell Commun Signal 2013; 11: 18.