Overcoming the blood-brain barrier with high-dose enzyme replacement therapy in murine mucopolysaccharidosis VII

Abstract

Enzyme replacement therapy (ERT) effectively reverses storage in several lysosomal storage diseases. However, improvement in brain is limited by the blood-brain barrier except in the newborn period. In this study, we asked whether this barrier could be overcome by higher doses of enzyme than are used in conventional trials. We measured the distribution of recombinant human β-glucuronidase (hGUS) and reduction in storage by weekly doses of 0.3-40 mg/kg administered i.v. over 1-13 weeks to mucopolysaccharidosis type VII mice immunotolerant to recombinant hGUS. Mice given up to 5 mg/kg enzyme weekly over 3 weeks had moderate reduction in meningeal storage but no change in neo-cortical neurons. Mice given 20-40 mg/kg three times over 1 week showed no reduction in storage in any area of the CNS except the meninges. In contrast, mice receiving 4 mg/kg per week for 13 weeks showed clearance not only in meninges but also in parietal neocortical and hippocampal neurons and glia. Mice given 20 mg/kg once weekly for 4 weeks also had decreased neuronal, glial, and meningeal storage and averaged 2.5% of wild-type hGUS activity in brain. These results indicate that therapeutic enzyme can be delivered across the blood-brain barrier in the adult mucopolysaccharidosis type VII mouse if administered at higher doses than are used in conventional ERT trials and if the larger dose of enzyme is administered over a sufficient period. These results may have important implications for ERT for lysosomal storage diseases with CNS involvement.

The mucopolysaccharidoses (MPSs) are a group of lysosomal storage diseases (LSD) caused by the deficiency of enzymes needed for the stepwise degradation of glycosaminoglycans (GAGs). The widespread lysosomal accumulation of undegraded GAGs leads to progressive cellular and organ dysfunction (1). Current treatments for patients with MPSs include hematopoietic stem cell transplantation and enzyme replacement therapy (ERT) (2, 3). MPS type VII (also known as Sly disease) results from deficiency of β-d-glucuronoside glucuronosohydrolase (GUS; EC 3.2.1.31) and is inherited as an autosomal recessive trait. Affected patients share many clinical features with patients with other MPSs, including shortened life span, mental retardation, organomegaly, and bone and joint abnormalities, that are collectively referred to as dysostosis multiplex (4). The murine model of MPS VII has proven valuable for the evaluation of novel therapies for LSDs, including bone marrow transplantation, neural progenitor cell transplantation, somatic cell gene replacement therapy, and ERT (5).

Previous studies have shown that i.v. injection of a fixed-dose of recombinant murine β-glucuronidase (mGUS) initiated at birth reduced pathological evidence of disease and prevented some of the learning, memory, and hearing deficits in the MPS VII mouse (6). However, recombinant mGUS reduced lysosomal storage in the neurons of the brain only if treatment was begun before or during the second week of life. After that age, no therapeutic effect of ERT was seen in the neocortical and hippocampal pyramidal neurons in the CNS of MPS VII mice (7). We recently reported that delivery of infused human enzyme (hGUS) to mouse brain in the newborn period depends on transcytosis of phosphorylated enzyme across the blood-brain barrier by the mannose 6-phosphate/insulin-like growth factor 2 receptor (M6P/IGF2R) (8). Down-regulation of this receptor-mediated delivery system by age 2 weeks appears to explain the lack of response in the adult MPS VII mouse. Because most patients with LSD present long after the newborn period with well established disease, identifying treatment strategies that reduce brain storage and restore CNS function is essential.

In this study, we used adult MPS VII/E540A^Tg mice tolerant to the human enzyme (9) to study the morphological response in brain to varying doses of recombinant human enzyme (rhGUS) over short and long durations of treatment. Our aim was to determine whether the barrier to enzyme delivery to the brain in the adult MPS VII mouse could be overcome by administering higher doses of enzyme than are used in conventional trials.

Methods

Mice and Treatment Protocols. MPS VII/E540A^Tg mice homozygous for a 1-bp deletion in exon 10 of the murine gus structural gene [identical to that in the B6.C-H-2^bm1/ByBir-gus^mps/mps MPS VII mice (10, 11)] and for a transgene that expresses hGUS cDNA with an amino acid substitution at the active site nucleophile (E540A) were used (9). These mice are totally deficient in mGUS activity, produce a low level of inactive hGUS, and were reported to be tolerant to immune challenge by rhGUS (9). Mice were screened by PCR for the gus^mps/mps mutation and for the hGUS transgene E540A in a sample of tissue obtained by tail clipping. All experiments were conducted with the highest standards of humane animal care. The rhGUS was purified from conditioned media from a Chinese hamster ovary cell line overexpressing rhGUS using a series of conventional chromatography steps.

Four treatment protocols (Groups I-IV) were used. Mice in Group I were 4-5 weeks old and were injected weekly for 3 weeks with 0.3-5 mg/kg rhGUS. Group II mice were 8-12 weeks old and were injected weekly for 13 weeks with 1-4 mg/kg rhGUS. Group III mice were 15-16 weeks old and were injected every other day with 3 doses of 20 or 40 mg/kg rhGUS. Mice in Group IV were 15-16 weeks old and were injected weekly for 4 weeks with 20 mg/kg rhGUS. Mice in Groups I, II, and IV were killed 1 week after the last injection, and those in Group III were killed 3 days after the last enzyme infusion. Eight untreated age-matched MPS VII mice served as controls. The weight-based enzyme dose was determined immediately before each injection and was delivered in 125 μl of buffer (10 mM Tris, pH 7.5/150 mM NaCl/1 mM β-glycerophosphate) by bolus infusion via the tail vein. Separate experiments showed that rhGUS at these doses was completely cleared from the circulation in <24 h (data not shown).

Biochemical Analysis. Tissues obtained at necropsy were snap-frozen in liquid nitrogen and stored at -70° for measurement of rhGUS activity. Quantitation of rhGUS was carried out on tissue homogenates of brain from mice in Groups I, II, and IV and six wild-type B6 mice (12). Plasma obtained at necropsy was snap-frozen in liquid nitrogen and stored at -70° for measurement of anti-GUS by ELISA (9). MPS VII (gus^mps/mps) mice injected i.p. with rhGUS as described in ref. 9 were used as positive controls for the ELISA.

Histological Analysis. For morphological evaluation of lysosomal storage, half of the brain was immersion fixed in cold 2% paraformaldehyde/4% glutaraldehyde in PBS, sectioned as described in ref. 13, postfixed in 1% osmium tetroxide, and embedded in Spurr resin (Polysciences). Sections of brain 0.5 μm thick, including parietal neocortex, hippocampus, cerebellum, meninges, and perivascular cells, were stained with toluidine blue and evaluated for lysosomal storage as indicated by cytoplasmic vacuolization by light microscopy. All sections were evaluated by two reviewers (C.V. and B.L.) without knowledge of the treatment group using a semiquantitative scale that identified reduction in lysosomal storage in cell types.

Results

GUS Levels in Brain After ERT with rhGUS. We first determined the tissue half-life (T_1/2) of the infused rhGUS enzyme in this model by administering 2 mg/kg to a group of animals and killing mice at intervals over the next 10 days. The data are presented in Table 1. The T_1/2 in most organs was ≈100 h. Based on these data, we chose weekly injections for Groups I, III, and IV and analyzed tissue levels 7 days after the last injection.

Table 1. Half-life of GUS in mouse tissues after infusion at a dose of 2 mg/kg.

Tissue	T1/2, hr
Brain	—
Liver	84
Spleen	80
Heart	124
Kidney	129
Lung	115
Muscle	86
Bone marrow	85
Bone	121
Eye	97

—, number too low to be accurate.

The brain from mice in Group II that received 4 mg/kg rhGUS contained, on average, 1.38% of wild-type GUS activity 7 days after the last of 13 weekly doses (Table 2). One mouse in this group had 1.8% of wild-type activity in brain. The mice in Group IV had even higher enzyme levels in brain, with an average of 2.5% of wild-type activity (Table 2). The brains of mice in Group I had much lower levels of enzyme. Thus, the enzyme levels in brain correlated with both dose and duration of treatment.

Table 2. Reduction in storage in the CNS with enzyme replacement therapy in MPS VII mice.

Group	GUS dose, mg/kg	Doses, n	Total cumulative dose, mg/kg	Interval between doses, days	Mice, n	Wild-type GUS activity, %	Meninges and perivascular	Glia	Parietal neocortical neurons	Hippocampal neurons
I	0.3	3	0.9	7	3	0.3	↓	NC	NC	NC
	1	3	3.0	7	2	0.4	↓	NC	NC	NC
	2.5	3	7.5	7	2	0.3	+/− to ↓	NC	NC	NC
	5	3	15	7	3	0.7	↓ to ↓ ↓	NC to +/−	NC	NC
II	1	13	13	7	4	0.92	↓ to ↓ ↓	NC	NC	NC
	2	13	26	7	5	1.21	↓ ↓ to ↓ ↓ ↓	NC to ↓	NC to ↓	NC
	4	13	52	7	5	1.38	↓ ↓ ↓	↓ ↓	↓ to ↓ ↓	NC to ↓
III	20	3	60	2	2	NA	↓ to ↓ ↓	NC	NC	NC
	40	3	120	2	2	NA	↓ to ↓ ↓ ↓	NC	NC	NC
IV	20	4	80	7	2	2.5	↓ ↓ to ↓ ↓ ↓	↓	↓ ↓	NC

Wild-type brain GUS levels were 16.7 ± 2 units expressed per mg of protein. Untreated MPS VII TgE540A hGUS mice have 0.37% of wild-type enzyme activity. Enzyme levels were obtained 1 week after the last injection. +/−, Very slight decrease in storage/may be focal; ↓, slight decrease in cytoplasmic vacuolization; ↓ ↓, moderate decrease in cytoplasmic vacuolization; ↓ ↓ ↓, marked decrease in cytoplasmic vacuolization/essentially normal; NC, no change in storage; NA, not applicable.

Data on levels of rhGUS in other organs (summarized in Table 3) show a dose-dependent increase in hGUS content in each group. Because the values represent residual enzyme 7 days after the last injection, it is likely that levels two to four times higher were achieved shortly after injection.

Table 3. Tissue GUS levels after GUS infusion into MPS VII/E540A^Tg mice at different doses and regimens.

Group	Dose mg/kg	Doses, n	Liver	Spleen	Heart	Kidney	Bone	Bone marrow	Muscle	Eye
I	0	3	0.35 ± 0.08	0.41 ± 0.36	0.053 ± 0.030	0.04 ± 0.02	0.39 ± 0.05	—	—	0.019 ± 0.004
			0.2	0.1	0.25	0.03	0.39	—	—	0.39
	0.3	3	13.29 ± 4.15	1.63 ± 0.61	0.36 ± 0.12	0.39 ± 0.15	1.27 ± 0.03	—	—	0.074 ± 0.027
			7.2	0.5	1.72	0.36	1.27	—	—	1.5
	1.0	2	41.88 ± 14.76	4.49 ± 1.63	0.83 ± 0.05	0.64 ± 0.1	3.57 ± 0.56	—	—	0.126 ± 0.025
			22.6	1.5	4.0	0.6	3.6	—	—	2.6
	2.5	2	96.02 ± 5.44	44.38 ± 24.5	2.16 ± 0.09	3.18 ± 1.18	11.78 ± 0.03	—	—	0.166 ± 0.049
			51.9	14.7	10.4	2.9	11.8	—	—	3.4
	5.0	3	211.8 ± 121	63.83 ± 22.3	3.36 ± 1.29	5.47 ± 2.98	8.66 ± 4.75	—	—	0.231 ± 0.024
			114	21.2	16.2	5.1	8.7	—	—	4.7
II	0		0.31 ± 0.02	0.096 ± .016	0.023 ± 0.023	0.06 ± 0.01	—	0.044 ± 0.024	0.024 ± .010	0.056 ± .023
			0.17	0.03	0.1	0.06	—	0.03	0.48	1.1
	1	4	65.1 ± 28.4	19.0 ± 9.7	1.10 ± 0.35	1.71 ± 1.07	—	6.30 ± 2.84	0.64 ± 0.38	0.19 ± .006
			35.2	6.3	5.3	1.58	—	3.91	12.9	3.9
	2	5	108 ± 25.9	132 ± 105	1.78 ± 0.70	3.69 ± 1.14	—	14.34 ± 9.62	0.81 ± 0.43	0.28 ± .08
			58.4	43.9	8.6	3.42	—	8.9	16.4	5.7
	4	5	230 ± 31.3	113 ± 52	4.18 ± 2.09	7.83 ± 1.13	—	46.44 ± 20.51	2.04 ± 1.56	0.60 ± .23
			124	37.5	20.1	7.25	—	28.8	41.2	12.3
IV	20	2	975 ± 442	374 ± 38	11.5 ± 1	20 ± 1.9	—	173 ± 25	2.95 ± 1.16	1.26 ± 0.39
			527	124	55.3	18.5	—	107	60.0	25.8
WtB6	—	4	185 ± 11.9	301 ± 26.6	20.8 ± 12.5	108 ± 7.5	100 ± 18	161 ± 35	4.95 ± 1.80	4.88 ± 0.68

Data are given as means ± SD, with the percentage of wild-type (Wt) values shown in italic type. —, not applicable.

Reduction in GAG levels was also determined in liver, kidney, and urine in mice in Group II (Table 4). Again, a dose-dependent response is evident. In liver, the maximum reduction was achieved with only 1 mg/kg. In kidney and urine, 2 mg/kg was required.

Table 4. Tissue and urine GAG levels (±SD) of rhGUS at 1, 2, and 4 mg/kg doses.

	Liver		Kidney		Urine
Dose	Protein, μg/mg	Decrease, %	Protein, μg/mg	Decrease, %	Creatinine, μg/mg	Decrease, %
0	13.12 ± 1.50	—	17.04 ± 4.41	—	1,178 ± 118	—
1	5.69 ± 0.79	57	12.11 ± 3.64	29	885 ± 356	29
2	5.21 ± 1.18	60	9.24 ± 3.85	46	641 ± 178	46
4	4.33 ± 1.25	67	8.76 ± 3.08	49	816 ± 31	49

—, not applicable.

Immune Response to rhGUS ERT. Although antibody response to rhGUS was dramatically lower in the transgenic, tolerant MPS VII/E540A^Tg mice than in the nontransgenic, nontolerant controls that received a maximal immune challenge i.p (9). Low levels of antibody to rhGUS were detected in the mice in Group II that received 4 mg/kg rhGUS for 13 weeks (Fig. 1 A). Four of the five mice had detectable ELISA reactions at 10^-2 dilution and one of the mice had a slightly higher reaction. A very faint ELISA reaction at 10^-2 dilution was also seen in the two mice that received 20 mg/kg rhGUS for 4 weeks (Fig. 1B). No clinically apparent hypersensitivity reactions to the GUS infusions were seen in any of the mice.

Fig. 1.

Analysis of plasma for hGUS antibodies. (A) ELISA for anti-hGUS antibody in plasma of untreated MPS VII mice (lanes 1-5) and MPS VII mice in Group II treated over 13 weeks with rhGUS at 1 mg/kg (lanes 6-10), 2 mg/kg (lanes 11-15) and 4 mg/kg (lanes 16-20). Although mice treated with 4 mg/kg for 13 weeks had detectable ELISA reactions at 10^-2 (lanes 17-20) and 10^-3 (lane 16) dilutions, none of the mice treated with lower doses of enzyme had detectable ELISA reactions, nor did mice in Group I treated for only 3 weeks at comparable doses (data not shown). (B) ELISA reactions for anti-hGUS antibody in plasma for mice in Group III (lanes 1, 3, and 4) and Group IV (lanes 5 and 6).

Reduction in Lysosomal Storage in the Brain Is a Function of rhGUS Dose and Duration of ERT. The analysis of the histological data are summarized in Table 1 and Figs. 2 and 3. In the parietal neocortex, neuronal storage was reduced in six mice in Group II, including all those treated with 4 mg/kg per week GUS over 13 weeks (and one treated with 2 mg/kg GUS). However, none of the mice in Group I (0.3-5 mg/kg weekly for 3 weeks) had a reduction in neuronal storage (Table 1 and Fig. 2 A-C). One of the mice in Group II also had less storage in hippocampal pyramidal neurons (data not shown); this mouse had 1.8% of wild-type GUS in the brain. Mice in Group III (treated three times for 1 week with 20 and 40 mg/kg; thus, treated with very high doses of enzyme but for only 1 week) had no morphologic response in neurons in the CNS (Fig. 2D). However, mice in Group IV (treated over 4 weeks with weekly injections of 20 mg/kg) had reduction in neocortical neuronal storage (Fig. 2E). There was a similar enzyme dose and treatment duration response in storage in the glia in the CNS (Fig. 2). No change in cerebellar Purkinje cell storage was seen in any of the ERT-treated mice (data not shown).

Fig. 2.

Lysosomal storage reduction in the parietal neocortex in MPS VII/E540A^Tg mice treated with ERT. Shown are brain sections stained with toluidine blue. (A) In the parietal neocortex of an untreated MPS VII mouse, small vacuoles are apparent in the neocortical neurons (arrow) and the cytoplasm of the oligodendroglial cells is distended with abundant lysosomal storage (arrowhead). (B) Mice in Group I treated for only 3 weeks had no reduction in storage in the neurons (arrow) or glia (arrowhead). (C) Mice in Group II treated with 4 mg/kg for 13 weeks had consistent reduction in storage in glia (arrowhead) and neocortical neurons (arrow). (D) Mice in Group III receiving three infusions of 20-40 mg/kg GUS given over 1 week had no reduction in storage in glia (arrowhead) or neurons (arrow). (E) Mice in Group IV treated with 20 mg/kg for 4 weeks had less storage in glia (arrowhead) and neocortical neurons (arrow) than untreated MPS VII mice. (Scale, 1 cm = 25 μm.)

Fig. 3.

Lysosomal storage reduction in the brain meninges and perivascular cells in MPS VII/E540A^Tg mice treated with ERT. Shown are brain sections stained with toluidine blue. (A) An untreated MPS VII/E540A^Tg mouse had extensive vacuolization of the cells in the meninges (arrow). (B) The meninges of a mouse in Group I had slight improvement in lysosomal storage (arrow) after only 3 weeks of treatment with 5 mg/kg GUS. (C) In Group II, there was marked reduction in the amount of meningeal storage (arrow) after 4 mg/kg GUS for 13 weeks. (D) The mice in Group III had slightly less meningeal storage (arrow) after only 1 week of treatment with three large doses of GUS. (E) Group IV mice had marked reduction in storage in the meninges after being treated for 4 weeks with 20 mg/kg GUS (arrow). (Scale, 1 cm = 25 μm.)

The meninges and perivascular cells in the brain had a dose-related response in all groups (Table 1 and Fig. 3). A slight to moderate reduction in meningeal storage occurred in mice treated weekly for only 3 weeks with doses of up to 5 mg/kg in Group I (Fig. 3B). The most marked and consistent decrease in storage in meningeal and perivascular cells occurred in the mice in Group II, which were treated weekly with 4 mg/kg for 13 weeks (Fig. 3C). Tissue GAG levels in this group are presented in Table 4.

Discussion

ERT is already an accepted treatment for several LSDs. Delivery of corrective enzyme to lysosomes depends on receptor-mediated endocytosis. The nearly ubiquitous M6P/IGF2R (14), the mannose receptor primarily on macrophages (15), and the asialoglycoprotein (galactose) receptor on hepatocytes (16) have been implicated in the uptake and/or intracellular targeting of lysosomal enzymes. Although ERT targeted at these receptors can reduce visceral storage, most CNS storage has been very refractory to this therapy to date (17). Most LSDs, including the MPSs, affect the CNS, and most patients are not diagnosed until they have established storage in the brain and functional CNS defects. Thus, arresting progressive damage due to lysosomal storage in brain and recovery of CNS function are key challenges. Delivering enzyme beyond the blood-brain barrier might achieve this goal.

In the murine MPS VII model, correction of clinical, biochemical, and morphological abnormalities in the CNS has been produced by several approaches. Systemic gene therapy in newborn MPS VII mice was effective in decreasing CNS storage and dysfunction (18, 19), but the success of such treatments in adult animals has been limited (20). However, MPS VII mice treated as adults with a recombinant adeno-associated virus (serotype 2) GUS viral vector that expressed supraphysiological levels of GUS in the liver and that survived beyond 50 days after injection of the viral vector into liver had 2% of wild-type levels of GUS in the brain and reduced CNS neuronal storage (21).

Adult MPS VII mice also showed reduced storage in neurons and glia after intraventicular transplantation of neural progenitor cells in the newborn period (22). Retrovirally transduced syngeneic murine fibroblasts that overexpress GUS implanted into the neocortex of adult MPS VII mice also resulted in a decrease in neuronal storage (23). Adult MPS VII mice treated with gene transfer using a feline immunodeficiency virus-based vector injected directly into brain showed not only reduced storage in neurons and glia but also a functional correction of established CNS deficits (24). Although the results are promising, these strategies require invasive injections into brain tissue or implantation of transduced cells. Nonetheless, these experiments provide hope that strategies that allow sufficient quantities of enzyme to cross the blood-brain barrier in the adult could reduce established storage and reverse established CNS deficits.

We have previously shown in MPS VII mice that a weekly, fixed dose of recombinant mGUS begun in the newborn period and continued for the first 6 weeks of life reduced storage in the CNS, an effect still evident 1 year after treatment was discontinued (25, 26). Treatment begun in the newborn period also improved cognitive function (6). However, treatment begun after the second week of life (weekly injections of 4 mg/kg for 6 weeks) had no effect on neuronal storage (7). Similar restriction of the therapeutic effect of ERT to the newborn period has been observed in a murine model of MPS IIIA (27).

Recently, we showed that this enzyme delivery to brain in the newborn period depends on transcytosis of phosphorylated enzyme across the blood-brain barrier by the M6P/IGF2R (8). Down-regulation of this receptor-mediated delivery system by age 2 weeks appears to explain the lack of response in the adult MPS VII mouse. In the current studies, we attempted to overcome the previously observed barrier to enzyme delivery of GUS to the CNS in the adult MPS VII mouse by administering higher doses of enzyme over a longer duration.

In multiple previous studies of ERT in this model, recombinant mGUS was used (5, 7, 25, 26). Here, we used rhGUS, which would be more appropriate for clinical trials of human MPS VII. To avoid a strong immunological reaction to the heterologous hGUS in the mouse, we used the transgenic mouse model that we previously showed was immunotolerant to strong challenging doses of the human enzyme given i.p. in complete Freund's adjuvant (9). This strategy seemed important because a strong antibody response has been shown to abrogate response to ERT (28-30). Such antibodies may alter enzyme binding to and uptake by the M6P/IGF2R and can alter organ distribution and intracellular targeting of infused enzyme (31). Despite the fact that the transgenic MPS VII mice used in this study were designed to be tolerant to hGUS (9), the mice in Groups II and IV had weakly positive ELISA reactions. These mice did not manifest clinical hypersensitivity reactions to enzyme infusions, as had some of the MPS VII mice in earlier treatment trials with the murine enzyme (25). Although only partial, the immune tolerance of the MPS VII/E540A^Tg mice may well have contributed to the successful delivery of enzyme to brain at high-dose/long-duration combinations.

What is most encouraging in this study is that, if given in high enough doses over a sufficient duration of treatment, some enzyme did reach brain and produced clearance of CNS storage. The greatest clearance of storage in the CNS correlated with the highest levels of brain GUS, i.e., in Groups II and IV. Perhaps the saturation of the relatively low number of M6P/IGF2R in the adult mouse brain (32) over a sufficient time can drive enzyme uptake into the brain and achieve correction, at least in some areas of the brain.

The observed delivery of enzyme across the blood-brain barrier in the adult mouse may not be mediated by the M6P/IGF2R, as it is in the newborn. It is also possible that phagocytic cell uptake of some of the enzyme infused at high doses allows subsequent transfer to neurons and glia by another mechanism and contributes to the observed reduction in preexisting brain storage. A third possibility is that enzyme is taken up by the so-called extracellular pathways that allow small amounts of large molecules, such as serum albumin, to enter the CNS (33). Banks (33) has summarized the evidence that exclusion of albumin by the blood-brain barrier is not absolute, even though the cerebral spinal fluid:serum albumin ratio is 1:200. Some therapeutic agents thought to be delivered to the CNS by these extracellular pathways in small but effective amounts are erythropoietin, which has a neuroprotective effect in stroke, and antibodies to amyloid-β protein, which reverses cognitive impairment in the transgenic mouse model of Alzheimer's disease. The most important determinant for delivery via this nonsaturable pathway is the T_1/2 of the protein in the circulation. Enzyme doses that exceed saturation of the mannose 6-phosphate and mannose receptors that mediate enzyme clearance from the circulation and delivery to visceral tissues prolong the T_1/2 of enzyme in circulation and favor delivery to CNS by this nonsaturable mechanism (33). Identifying the mechanism of delivery of rhGUS to brain may suggest strategies to enhance the process and make ERT for CNS storage a more realistic goal.

Whatever the mechanism, this study provides clear evidence that i.v. ERT with rhGUS can result in a reduction in storage in the CNS in MPS VII mice. Longer treatment and higher doses of enzyme resulted in the most consistent therapeutic response. Somewhat similar responses have been reported in preclinical trials of i.v. ERT in mouse models of aspartylglycosaminuria and α-mannosidosis (34, 35). Biochemical evidence showed that CNS substrate accumulation was reduced in both of these studies, although morphological evidence of reduced storage was not documented.

Important questions for future studies are the following: (i) What is the impact of high-dose ERT on spatial learning and memory abnormalities, which are well characterized in the adult MPS VII mouse (6, 24, 25)? (ii) Can one enhance the observed enzyme uptake and clearance of storage in the adult brain by manipulating the receptor-mediated or cell-mediated process that effects this transport? (iii) Could strategies to block rapid enzyme clearance by fixed tissue macrophages allow longer exposure of circulating enzyme to brain and enhance enzyme delivery to brain? (iv) Does the immune tolerance in this model actually contribute to the enhanced delivery of corrective enzyme to the brain, or could the breakdown of tolerance and the appearance of antibody to GUS at the highest doses actually lead to enhanced delivery to brain? Answers to all of these questions could have an important impact on therapy for LSDs with CNS involvement.