Chemically modified β-glucuronidase crosses blood–brain barrier and clears neuronal storage in murine mucopolysaccharidosis VII

Characterization of Purified Recombinant GUS.

The purified recombinant GUS used in these experiments was similar to that described (11, 19). The apparent molecular mass of the enzyme monomer was 75 kDa on reducing SDS/PAGE. The tetrameric enzyme had a molecular mass of ≈300 kDa when analyzed by sizing gel filtration chromatography (data not shown). The specific activity of the purified enzyme was 4.9 × 10⁶ units/mg. The K_uptake, where K is the concentration of enzyme giving half-maximal uptake, was 1.25- 2.50 nM, calculated from uptake saturation curves by using human MPS VII fibroblasts in which the uptake is almost entirely M6PR-dependent.

Treatment of Purified GUS with Periodate and Borohydride.

To inactivate the mannose and M6P recognition markers that normally mediate clearance of infused GUS, we treated the enzyme sequentially with sodium metaperiodate followed by sodium borohydride (18). Sodium metaperiodate treatment inactivates the exposed carbohydrate recognition markers on glycoproteins by oxidative degradation (18). Subsequent reduction by sodium borohydride prevents aggregation that results from cross-linking of the reactive aldehydes on periodate-treated enzyme with free amino groups on other enzyme molecules (20). This sequential treatment of GUS resulted in only 14% loss in specific activity (4.9 × 10⁶ units/mg to 4.22 × 10⁶ units/mg).

To determine the impact of this treatment on receptor-mediated uptake, we measured the M6PR-mediated uptake in human fibroblasts (Table 1) and the MR-mediated uptake in mouse J774E macrophages (Table 2). The data presented in Table 1 show that the PerT-GUS had essentially no uptake by fibroblasts in the presence or absence of M6P, confirming that the M6P recognition marker had been completely inactivated. Similarly, Table 2 shows that PerT-GUS had lost its mannan-inhibitable uptake by macrophages. This result indicates that all of the exposed mannose residues on the carbohydrate chains of GUS had been inactivated.

Comparative Stability of GUS and PerT-GUS.

The carbohydrates on glycoproteins often confer enhanced thermal stability, and removal of oligosaccharide chains often destabilizes glycoproteins (21). Human GUS has been shown to be relatively stable to thermal inactivation at 65°C (22–26). As shown in Fig. 1, recombinant GUS retained 90% of initial activity after 3 h at 65°C, whereas PerT-GUS retained 40% of its activity under these conditions (Fig. 1).

To compare the stability of GUS and PerT-GUS in lysosomes of living cells at 37°C, we studied their half-life after uptake by MPS VII fibroblasts. The low rate of endocytosis of PerT-GUS by fibroblasts required exposure to 100,000 units/ml PerT-GUS per plate for 48 h to accumulate sufficient enzyme by fluid phase pinocytosis (28 units per plate) to allow measurement of its half-life. By contrast, fibroblasts exposed to 500 units/ml M6P containing native GUS for 48 h contained 228 units per plate. Fig. 2 shows the half-life for the two enzymes in fibroblasts upon subsequent incubation at 37°C. The t½ of GUS was 18.9 days. The t½ of PerT-GUS was shorter (12.9 days), but nearly one-third of the initial activity was still present at 21 days.

Clearance of GUS vs. PerT-GUS from the Circulation After Infusion.

Plasma clearance was measured in MPS VII mice after infusion of 4 mg/kg of each enzyme (Fig. 3). GUS was cleared with a t½ of 11.7 min. This rapid clearance has been shown to depend on the MR and the M6PR clearance systems (15). In contrast, the clearance of PerT-GUS was dramatically prolonged to a t½ of 18.5 h.

Tissue Distribution of GUS vs. PerT-GUS 48 h After Infusion.

Previously, we observed that treating MPS VII mice with high-dose GUS slowed the plasma clearance of the enzyme and facilitated enzyme delivery to the brain (11). In these experiments, it was not clear whether it was the higher dose of enzyme itself or the delayed plasma clearance of the enzyme that accounted for improved delivery to brain. To address this question, we compared the distribution of GUS and PerT-GUS in brain and other tissues 48 h after infusion into MPS VII mice (Table 3). Mice were perfused with Tris-buffered saline before collection of tissues to ensure that tissue was not contaminated with residual plasma enzyme.

As evident from the data in Table 3, delivery of native GUS to brain at 48 h was minimal. However, native GUS was delivered to other tissues at levels similar those reported. PerT-GUS was delivered to heart, kidney, muscle, lung, and eye at levels higher than those seen with native GUS. The levels in liver and spleen were nearly 4-fold lower after PerT-GUS infusion than after GUS infusion. This result undoubtedly reflects the curtailment of receptor-mediated uptake by the MPR and M6PR that are highly expressed in these two tissues. By contrast, brain levels were greatly increased (7.8% of wild type) in PerT-GUS-infused animals. This result suggests that the long circulating PerT-GUS has an advantage in crossing the BBB. Thus, it was of great interest to study its effectiveness in clearing storage from cells in the CNS.

Comparison of the Effectiveness of GUS and PerT-GUS in Clearing Neuronal Storage.

Fig. 4 compares the morphology of neocortical (Fig. 4 A–C) and hippocampal (Fig. 4 D–F) neurons and meninges (Fig. 4 G–I) in untreated (Fig. 4 A, D, and G) and GUS-treated (Fig. 4 B, E, and H) and PerT-GUS-treated (Fig. 4 C, F, and I) MPS VII mice. There was a marked reduction in storage in the neocortical (Fig. 4C) and hippocampal (Fig. 4F) neurons in the PerT-GUS-treated mice and a slight reduction in storage in neocortical neurons (Fig. 4B) with native GUS treatment. Meningeal storage was moderately reduced by both GUS and PerT-GUS treatment (Fig. 4 H and I).

To obtain a more objective measure of the reduction in neuronal storage in the CNS in response to the two types of GUS, we devised a morphometric approach described in Materials and Methods. Table 4 summarizes the results of assessment of storage in neocortical and hippocampal neurons of untreated, GUS-treated, and PerT-GUS-treated MPSVII mice. ERT with GUS over 12 weeks with both 2 mg/kg and 4 mg/kg GUS reduced storage in neocortical neurons compared with untreated MPS VII mice (P = 0.002 and P = 0.003, respectively), although hippocampal neurons failed to show a morphological response to this therapy. PerT-GUS at 2 mg/kg also reduced neocortical neuronal storage (P = 0.001). At 4 mg/kg, the therapeutic effect of PerT-GUS was even more striking (P = 0.003 for 2 vs. 4 mg of PerT-GUS and P < 0.001 compared with untreated). In addition, there was virtually no storage in the hippocampal neurons in the three PerT-GUS-treated mice available for quantitation (the CA2 region was not present in the section and was therefore not available for quantitation in two of the five PerT-GUS-treated mice). These results indicate that ERT with PerT-GUS is remarkably more effective than traditional GUS at clearing storage in the neocortical and especially hippocampal neurons in the MPS VII mouse. As a group, the PerT-GUS-treated mice also had slightly less storage in glial and perivascular cells than the GUS-treated mice. However, the dose-dependent reduction in storage in meninges, which was moderate at 4 mg/kg, was equivalent in the PerT-GUS- and the GUS-treated animals.

Clearance of Storage in Nonneuronal Tissues.

Clearance of storage was nearly complete in liver and spleen sinusoidal cells and kidney at the 2 mg/kg and 4 mg/kg doses with both GUS and PerT-GUS. In bone, the chondrocytes showed no improvement with either enzyme. Storage was more consistently cleared by PerT-GUS in osteocytes at both doses than by GUS, although 4 mg/kg GUS also produced moderate to marked decrease in storage in most osteocytes.

Generation of Stable Cell Lines Secreting GUS.

Using DNA-cloning techniques, we subcloned the cDNA sequence encoding the full-length cDNA for human GUS (GenBank accession no. NM_000181) into the mammalian expression vector pCXN (29). This expression vector contains an expression cassette consisting of the chicken β-actin promoter coupled to the CMV intermediate–early enhancer. pCXN also contains a selectable marker for G418 allowing selection of stably expressing mammalian cells.

This plasmid was introduced into the Chinese hamster ovary cell line CHO-K1 by electroporation (30). After selection in growth medium consisting of minimal essential medium + 35 μg/ml proline + 15% FBS + 400 μg of G418, colonies were picked and grown to confluence in 48-well plates. High-level expressing clones were identified by measuring GUS activity secreted into the conditioned medium from these clones (19). The highest-producing clone was scaled up, and secreted enzyme was collected in protein-free collection medium, PF-CHO. Conditioned medium collected in this way was pooled and centrifuged at 5,000 × g for 20 min, and the supernatant was collected and frozen at −20°C until sufficient quantities were accumulated for purification.

Measurement of GUS Activity and Protein.

GUS activity was measured by using 10 mM 4-methyl-umbelliferyl β-d-glucuronide as substrate in 0.1 M sodium acetate buffer (pH 4.8), 1 mg/ml crystalline BSA as described in ref. 31. Protein levels were assayed by the method of Lowry (32).

Purification of GUS.

GUS was purified by a multistep procedure with conventional column chromatography. Medium collected from stable cell lines was thawed and filtered through a 0.2-μm microculture capsule filter (Pall Corporation). Filtrate was concentrated over an Amicon YM-100 membrane to minimize volume, then diafiltrated with 2 × 2.25-liter 20 mM NaPO₄ + 150 mM NaCl + 0.025% NaN₃ (pH 5.5). Diafiltrated concentrate was then passed over a Blue Sepharose FF (GE Healthcare) column preequilibrated with 20 mM NaPO₄ (pH 5.5). After washing with 10 column volumes of 20 mM NaPO₄ + 150 mM NaCl (pH 5.5), the column was eluted with 10 mM NaPO₄ + 800 mM NaCl (pH 7.5). The fractions containing GUS activity were pooled and applied to a phenyl-Sepharose high-Sub FF column preequilibrated with 10 mM NaPO₄ + 1,000 mM NaCl (pH 8.0). After washing with 10 column volumes of 10 mM NaPO₄ + 1,000 mM NaCl (pH 8.0), the column was eluted with 10 mM Tris + 1 mM sodium-β-glycerophosphate (pH 8.0). Fractions containing GUS activity were pooled and dialyzed with three changes of 10 mM Tris + 1 mM Na-β-glycerophosphate (pH 8.0). The dialyzed phenyl Sepharose elute was loaded onto a DEAE-Sephacel (Sigma–Aldrich) column preequilibrated with 10 mM Tris + 1 mM Na-β-glycerophosphate (pH 8.0). After washing with 10 column volumes of 10 mM Tris + 1 mM Na-β-glycerophosphate (pH 8.0), the column was eluted with a 0–0.4 M NaCl gradient in the same buffer. The GUS peak was pooled and then dialyzed against three changes of 25 mM sodium acetate + 1 mM Na-β-glycerophosphate + 0.025% NaN₃ (pH 5.5). The dialyzed DEAE eluate was applied to a CM-Sepharose (Sigma–Aldrich) column preequilibrated with 25 mM sodium acetate + 1 mM Na-β-glycerophosphate + 0.025% NaN₃ (pH 5.5). After washing with 10 column volumes of the same buffer, the column was eluted with a 0–0.3 M gradient of NaCl in the equilibration buffer. Fractions containing the peak GUS activity were pooled and dialyzed against three changes of 25 mM Tris (pH 7.5), 1 mM β-glycerophosphate, 150 mM NaCl, 0.25% sodium azide, and then frozen at −80°C where it was stable indefinitely.

Treatment of GUS with Periodate and Borohydride.

The mannose and M6P recognition sites on GUS are located in the carbohydrate portion of the enzyme. To inactivate the exposed carbohydrates, the enzyme was treated with sodium metaperiodate followed by sodium borohydride (17, 18). Approximately 10 mg of purified GUS was treated with a final concentration of 20 mM sodium metaperiodate in 20 mM sodium phosphate, 100 mM NaCl (pH 6.0) for 6.5 h on ice, in the dark. The reaction was quenched by the addition of 200 mM final concentration ethylene glycol and incubated for an additional 15 min on ice, in the dark. Afterward, this mixture was dialyzed against two changes of 20 mM sodium phosphate, 100 mM NaCl (pH 6.0) at 4°C. Sodium borohydride was added to the dialyzed enzyme to a final concentration of 100 mM, and the enzyme was held on ice overnight in the dark to reduce reactive aldehyde groups. The enzyme was then dialyzed against two changes of 20 mM sodium phosphate, 100 mM NaCl (pH 7.5) at 4°C and was stable to storage in this buffer at 4°C.

GUS Uptake by Human Primary Fibroblasts and Mouse J774E Macrophages.

M6PR-mediated uptake was determined by adding 4,000 units of GUS or PerT-GUS ± 2 mM M6P in 1 ml of growth medium to 35-mm dishes of confluent GM-2784 GUS-deficient fibroblasts. After incubation at 37°C and 5% CO₂ for 2 h, the cells were cooled on ice, washed five times with cold PBS, then solubilized in 0.5 ml of 1% sodium desoxycholate. Extracts were assayed for GUS activity and protein, and values were expressed as units of enzyme taken up per mg of cell protein per hour of uptake.

MR-mediated uptake was measured by adding 10,000 units of GUS or PerT-GUS ± 1.7 mg/ml yeast mannan (Sigma–Aldrich) in 1 ml of growth medium to 35-mm dishes of confluent J774E mouse macrophages (33). After incubation at 37°C and 5% CO₂ for 4 h, the cells were washed as above and then solubilized in 1 ml of 1% sodium desoxycholate and assayed for GUS activity.

Stability of GUS and PerT-GUS at 65°C.

Purified GUS or PerT-GUS were diluted in equal volumes of heat inactivation buffer [40 mM Tris·HCl (pH 7.5), 150 mM NaCl, 10 mg/ml BSA], and aliquots were incubated for 0, 0.5, 1, 2, or 3 h at 65°C. After treatment, aliquots were cooled on ice and then assayed for GUS activity. Results were expressed as the percentage of original units of GUS activity remaining at the indicated times.

Stability of GUS vs. PerT-GUS After Uptake by Human MPS VII Fibroblasts.

Tissue culture dishes (35 mm) of confluent GM-2784 GUS-deficient fibroblasts were incubated with 500 units of GUS or 100,000 units of PerT-GUS in 1 ml of growth medium at 37°C and 5% CO₂ for 48 h under sterile conditions. The plates were washed twice with sterile growth medium and then fed with 2 ml of the same. Duplicate plates were taken off at 0, 2, 5, 7, 14, and 21 days, washed five times with PBS and frozen at −20°C. Remaining plates were fed twice weekly with 2 ml of growth medium. After all plates had been collected, the cells were solubilized in 0.5 ml of 1% desoxycholate and assayed for GUS activity. Values were expressed as percentage of zero time cell-associated GUS activity remaining at the indicated time points.

Clearance of GUS vs. PerT-GUS from the Circulation After Infusion into Mice.

MPS VII mice were infused via tail vein with GUS or PerT-GUS at a dose of 4 mg/kg in a total volume of 125 μl of PBS. After infusion, blood samples were taken by supraorbital puncture at 2, 5, 10, 20, 60, 90, and 120 min for GUS and 4, 240, 1,440, and 2,880 min for PerT-GUS into heparinized capillary tubes. Plasma was collected after centrifugation and assayed for GUS activity. Values were expressed as a percentage of GUS activity remaining compared with the first time point.

Tissue Distribution of GUS vs. PerT-GUS 48 h After Infusion.

MPS VII mice were infused via tail vein with GUS or PerT-GUS at a dose of 4 mg/kg in a total volume of 125 μl of PBS. At 48 h after infusion, the mice were perfused with 30 ml of 25 mM Tris (pH 7.2), 140 mM NaCl. Perfused tissues were collected and flash frozen in liquid nitrogen until further processing. Tissues were thawed, weighed, and homogenized for 30 s with a Polytron homogenizer in 10–20 volumes of 25 mM Tris (pH 7.2), 140 mM NaCl, 1 mM phenylmethylsulfonyl fluoride. Total homogenates were frozen at −80°C, thawed, and then sonicated for 20 s to produce a homogeneous extract. Extracts were assayed for GUS activity and protein, and the results were expressed as units/milligrams of tissue protein.

Treatment of MPS VII Mice with GUS vs. PerT-GUS at Doses of 2 and 4 mg/kg.

MPS VII mice were treated with 12 weekly infusions of GUS or PerT-GUS at doses of 2 or 4 mg/kg body weight. One week after the last infusion, tissues from untreated MPS VII (n = 3), 2 mg/kg (n = 3) or 4 mg/kg GUS (n = 2), or 2 mg/kg (n = 2) or 4 mg/kg PerT-GUS (n = 3) were obtained at necropsy after normal saline perfusion, fixed in 2% paraformaldehyde and 4% glutaraldehyde, postfixed in osmium tetroxide, and embedded in Spurr resin. For evaluation of lysosomal storage by light microscopy, toluidine blue-stained 0.5-μm-thick sections of liver, spleen, kidney, brain, heart, rib, and bone marrow were assessed without knowledge of treatment group. To evaluate storage in cortical neurons, 500 contiguous parietal neocortical neurons were scored for the number of lucent cytoplasmic vacuoles, indicating lysosomal storage. A maximum of seven vacuoles were counted per cell, and results were evaluated by ANOVA or Student's t test. We also evaluated hippocampal neurons by counting the number of vacuoles in 100 neurons in CA2 sector. Other tissues were examined by using a semiquantitative scale, as described in ref. 11.