Main

The mucopolysaccharidoses (MPS) are inherited lysosomal storage diseases (LSD), each caused by the deficiency of one of the lysosomal enzymes that degrade glycosaminoglycans. Patients with MPS have progressive glycosaminoglycan accumulation in lysosomes and widespread organ dysfunction. A murine model of MPS VII lacks β-glucuronidase (EC3.2.1.3), shares clinical, biochemical, and pathological features with patients with MPS VII or Sly disease(1–5) and allows controlled tests of therapies in meaningful numbers of genetically identical mice. Data from such therapeutic studies may be generalizable to LSD in humans.

Gaucher disease, an LSD that affects mainly the fixed tissue macrophage system, can be treated with infusion of mannose-targeted glucocerebrosidase(6–9). The availability of recombinant β-glucuronidase with both mannose and mannose 6-phosphate recognition moieties for receptor-mediated endocytosis(10) allowed us to target enzyme to other tissues as well as to fixed tissue macrophages in murine MPS VII. Affected mice have storage in the brain and bone as well as the fixed tissue macrophage system and thus can be used to test therapeutic responses in multiple tissues and cell types. β-glucuronidase injected i.v. in newborn MPS VII mice reached sites of clinically important disease and the enzyme's tissue distribution was similar to that of the mannose 6-phosphate receptor (M 6-PR)(11). A course of six weekly injections of 28,000 U β-glucuronidase initiated in newborn pups reduced lysosomal storage in many cell types, including neurons in the central nervous system (CNS)(12). This enzyme therapy regimen also diminished behavioral and auditory dysfunction characteristic of murine MPS VII(13) and improved survival and growth, even in mice that received no further β-glucuronidase therapy after 5 wk of age(14). However, after only a single injection of 28,000 U β-glucuronidase at 5 wk of age, there was no morphological evidence of delivery of enzyme to neurons or glia in the CNS(12).

We define the age at which storage in the CNS in murine MPS VII becomes resistant to β-glucuronidase therapy by comparing the quantity of lysosomal storage in the brain in mice started on therapy at various times after birth. We also evaluate the effect of the time of initiation of therapy and the dose of enzyme on LSD in other viscera. In contrast to earlier studies, where a fixed-dose of enzyme was administered throughout the study, each β-glucuronidase infusion was now administered as a weight-adjusted dose, providing the older mice with a larger amount of enzyme than in previous experiments.

METHODS

Homozygous MPS VII mice from the B6.C-H-2bm 1/ByBir-gusmps/+ mutant strain were treated at The Jackson Laboratory (Bar Harbor, ME). MPS VII mice have less than 1% of normal tissue β-glucuronidase activity and were identified at birth by a fluorometric assay for β-glucuronidase activity using tissue obtained from a toe clip. Recombinant β-glucuronidase, prepared as previously described(10), was injected in buffer [10 mM Tris (pH 7.5), 150 mM NaCl, 1 mM β-glycerol-phosphate] into 17 MPS VII mice at weekly intervals according to the schedule in Table 1. The enzyme was assayed by adding 0.025 mL of an enzyme dilution to 0.1 mL of 10 mM 4-methylumbelliferyl β-D-glucuronide in 0.1 M sodium acetate buffer at pH 4.8. One unit of activity is defined as the amount that hydrolyzes 1 nmol of 4-methylumbelliferyl β-D-glucuronide/h at 37°C. The 4-methylumbelliferyl released was measured fluorometrically. The specific activity of the purified enzyme used in these experiments was 5000 U/µg. Mice were weighed immediately before each injection to calculate the enzyme dose at 16,000 U/g. The injection on the first day of life was given via a superficial temporal vein. Because skin pigmentation obscured scalp veins and tail veins are not large enough to inject at 7 d of age, the second injection in groups 1-6 was given i.p. Injections in mice 14 d of age and older were given via a tail vein.

Table 1 Injection schedule for groups of MPS VII mice receiving β-glucuronidase replacement
Full size table

Mice were observed for approximately 30 min after each injection for adverse reactions. Because reactions were observed with the second injection in mice in group 7, these mice were pretreated 1 h before injection with 1 mg/kg diphenhydramine and 0.5 mg/kg ephedrine, beginning with the third injection. All experiments were performed with the highest standards of humane animal care. This study was approved by The Jackson Laboratory Animal Care and Use Committee. The Jackson Laboratory is fully accredited by the American Association for Accreditation of Laboratory Animal Care.

Animals were killed 1 wk after their last injection, tissues were collected, fixed, and prepared for histology and biochemistry as previously described(12). Samples of liver, spleen, brain, eye, and rib were fixed in 2% glutaraldehyde and 4% paraformaldehyde in phosphate-buffered saline.

Toluidine blue-stained 0.5-µm thick sections of tissue embedded in Spurr's resin were evaluated for lysosomal storage. For the brain, two sections were made, one a coronal section through the cerebral cortex including the hippocampus and the overlying parietal neocortex (Fig. 1) and the other a sagittal section taken 1-2 mm laterally from the midline through the cerebellum and the brain stem. Tissues were compared with those from age-matched untreated MPS VII mice. Histological sections were evaluated morphologically, with conventional light microscopy, without knowledge of the treatment group, using a semiquantitative scoring system similar to that previously described(14).

Figure 1
figure 1

Brain dissection method. The brain was dissected in the same fashion in each mouse and a coronal section of parietal cortex and hippocampus and including leptomeninges, as shown in the large cut-out was studied by light microscopy to determine the morphological score for neuronal storage as well as storage in other cell types. A section as in the inset was used for ultrastructural examination. Parietal neocortical neurons from this section were photographed and the amount of neuronal cytoplasmic storage was quantitated morphometrically.

Full size image

To validate our semiquantitative scoring system, we also evaluated the neocortical neuronal storage morphometrically using a method similar to that previously used to evaluate hepatocyte storage in feline and canine MPS models(15). Ultrathin uranyl acetate-lead citrate- stained sections of the parietal neocortex overlying the hippocampus from animals in groups 1-6 and age-matched untreated MPS VII controls were examined with a Joel 100CS transmission electron microscope. Because the amount of neuronal storage in MPS VII mice varies in different regions of the brain(5), sections of the same anatomic area from each mouse were used for the morphometric evaluation (Fig. 1, inset). Beginning at one corner of the superficial cortex, the section was scanned transversely to the deep cortex, photographing every intact neuron encountered. The field of view was then shifted laterally and scanning continued back to the superficial cortex. This process was repeated until 84-101 (mean 96) neocortical neurons were photographed. Prints of the electron photomicrographs at 3 900-15 000× final magnification were analyzed morphometrically. The neuronal cell membrane and nuclear border were traced using BioScan OPTIMAS software (Edmonds, WA) and cytoplasmic area calculated. The storage vacuoles in each cell were traced and the percentage of neuronal cytoplasm occupied by lysosomal storage was calculated. Ultrastructural examination and digitizing were performed without knowledge of the animal's morphological score or treatment group. The data were analyzed as a hierarchical design(16), where the neurons were treated as a random factor nested within mice and the group variable was treated as a fixed effect. Analyses were done using SAS PROC MIXED(17).

β-Glucuronidase activity was assayed(18) in liver, spleen, kidney, and brain. Plasma collected at the time of death of mice in groups 1, 3, 4, and 7 was assayed for antibodies to murine β-glucuronidase using immunoprecipitation and enzyme linked immunosorbent assay (ELISA) methods, as previously described(19). Although both methods detected antibody to β-glucuronidase in one mouse, the type of antibody was not characterized.

RESULTS

Neocortical neuronal storage is resistant to enzyme therapy by age 2 wk. Six doses of β-glucuronidase given from birth (group 1) markedly reduced the amount of lysosomal storage in both neocortical and hippocampal pyramidal neurons (Fig. 2, A-F and Tables 2 and 3). Two of the three mice treated beginning on d 7 (group 2) also had less neuronal storage based on the light microscopic morphological evaluation of the neocortex. None of the MPS VII mice whose treatment began after d 7 had a reduction in neuronal storage with β-glucuronidase therapy, including those in group 7, treated with 16,000 U/g β-glucuronidase for 6 wk beginning at d 35. The morphometric quantitation of neocortical neuronal storage correlated well with our "blinded" light microscopic morphological evaluation (Table 2). There was a significant reduction in neuronal storage as determined morphometrically in group 1, the mice treated from birth (p = 0.008). However, the groups of MPS VII mice whose treatment began at or after wk 1 had no significant difference in the measured quantity of neocortical neuronal storage compared with the group of untreated MPS VII mice (p > 0.05) (Table 2).

Figure 2
figure 2

Neuronal lysosomal storage is reduced with enzyme replacement therapy. Neurons from the parietal neocortex (A and B) and h2 Sommer sector of the hippocampus (C) from a 42-d-old untreated MPS VII mouse have prominent lysosomal distention in neurons (arrowheads). Parietal neocortex neurons (D and E) and hippocampal neurons (F) from the same area of a MPS VII mouse in group 1, treated with six doses of 16,000 U/g β-gluc from birth have inconspicuous lysosomes. (A, C, D, and F, toluidine blue, bar equals 24 μ; B and E, uranyl acetate-lead citrate; bar equals 3.5 μ).

Full size image
Table 2 Correlation of morphologic (light microscopic) evaluation and morphometric (ultrastructural) measurements of the impact of β-glucuronidase replacement therapy in individual mice on amount of cortical neuronal lysosomal storage.
Full size table
Table 3 Reduction in lysosomal storage is a function of age at initiation and duration of β-glucuronidase replacement therapy.
Full size table

Comparison of the neuronal storage in mice in groups 1 and 7, both treated with six weight-based doses of β-glucuronidase, but beginning therapy as newborns or adults, showed that the time of initiation of enzyme therapy was more important for neuronal accessibility or response than the total dose of enzyme. The mice in group 7 showed no decrease in neuronal storage even though they received a total β-glucuronidase dose averaging 2,200,000 U in six weekly injections of 16,000 U/g beginning at age 35 d. By contrast, the mice in group 1, which also received 16,000 U/g of β-glucuronidase for six injections beginning at birth (but a total dose averaging only 41% of that of mice in group 7) showed marked reduction in neuronal storage.

Influence of time of initiation and total dose on response to enzyme therapy on other cell types and tissues. Glial cells had less storage if treatment began at or before 21 d. Those mice whose treatment started later, i.e. groups 5-7 had no reduction in storage in glial cells (Table 3). In contrast, the leptomeningeal and perivascular cell lysosomal storage reduction by enzyme replacement was dose dependent. Three or more β-glucuronidase injections reduced storage in these sites regardless of the age therapy was initiated. Cerebellar Purkinje cell storage did not decrease regardless of the total enzyme dose or the time of initiation of therapy (data not shown).

Mice in all the treatment groups had a reduction in storage in hepatic Kupffer cells and spleen and bone marrow sinus lining cells (Table 3). However, the corneas were unimproved and were indistinguishable from corneas of untreated MPS VII mice regardless of the treatment schedule or the total enzyme dose (data not shown). Three or more doses of β-glucuronidase were required to reduce storage in the retinal pigment epithelium. Maximum response in osteoblasts lining the bone trabeculae in the rib required five to six enzyme doses. As with Purkinje cells and corneal fibroblasts, chondrocytes failed to show any morphological response to β-glucuronidase therapy regardless of the dose or treatment schedule (data not shown).

Biochemical responses to enzyme replacement. Tissue β-glucuronidase was still elevated in the MPS VII mice 1 wk after their last injection (Fig. 3A). The livers had 17-34% and the spleens 3-7% of normal levels of β-glucuronidase. The amount of β-glucuronidase in the brain of mice that received injections beginning on d 7 or later and ending at d 35 (and therefore fewer injections than mice in group 1) was generally less than that in the mice treated from birth, with the exception of one mouse in group 4 (Fig. 3B). This animal had 1.3% of normal β-glucuronidase level in brain (and 8.4% of normal β-glucuronidase level in kidney) for reasons that are unclear. Although these mice (groups 2-6) received progressively fewer injections and a lower total enzyme dose, the mice in group 7 received the same weight-based dose as the mice in group 1, treated from birth. The lack of improvement in glia and neurons in group 7 suggests that the enzyme detected in brain in this group (which was approximately the same amount as was present in group 1) may have been localized in meningeal or perivascular cells that were cleared of their storage, or in blood vessels.

Figure 3
figure 3

Tissue β-gluc levels after enzyme replacement therapy. A, β-gluc levels in the viscera in MPS VII mice in the seven treatment groups 1 wk after their last injections showed a general correlation with the amount of enzyme received. B, β-gluc levels in the brain were generally lower in the mice that received less than six β-gluc injections but levels were higher than those of untreated MPS VII mice (mps/mps: untreated MPS VII control mice).

Full size image

Immune response to enzyme therapy. Two of three mice that received weekly enzyme injections beginning at age 5 wk (group 7) had anaphylactoid-type reactions to enzyme infusions. The reactions, characterized by lethargy and respiratory distress, began at 5-15 min, lasted up to 1 h after the infusion and were less severe after pretreatment with diphenhydramine and epinephrine as described. However, one mouse in group 7 died 0.5 h after its fifth injection. Antibodies to β-glucuronidase were not detected in serum from this mouse or in sera from the other mice in group 7, suggesting that the reactions in this group were to another, unidentified antigen in the enzyme preparation.

None of the animals in groups 1-6 had a clinical reaction to the enzyme injections although one of the three mice in group 1 developed antibodies to β-glucuronidase demonstrable both with ELISA and immunoprecipitation assays.

DISCUSSION

Multiple therapies including bone marrow transplantation (BMT), somatic cell gene therapy, and enzyme replacement have been effective in treating some aspects of LSD in murine MPS VII(12,13,20–22). Studies using animal models of MPS I(23), MPV VII(21), α-mannosidosis(24), and fucosidosis(25) demonstrated reduced neuronal storage after BMT. Our previous studies showed that recombinant β-glucuronidase administered in a fixed-dose to MPS VII mice for the first 5 wk of life, when clinical and morphological evidence of LSD are minimal, resulted in a decrease in lysosomal storage in multiple tissues including the brain(12,14). This treatment also greatly reduced the cognitive and hearing defects characteristic of murine MPS VII(13). Mice that received only a single enzyme dose at age 5 wk had no evidence of neuronal storage reduction(12). These observations raised the questions of whether 1) the fixed-dose of β-glucuronidase was simply too small at age 5 wk, given the weight of the mice at that age, or 2) the neurons were not as accessible or responsive to enzyme therapy by 5 wk of age. The experiments presented address these questions.

We show that a weight-based dose of 16,000 U/g of β-glucuronidase results in a consistent reduction in neocortical and hippocampal pyramidal neuronal storage in MPS VII mice treated from birth. However, neurons show no response if treatment is begun at 14 d or later, even if the enzyme dose is increased to adjust for increasing weight and even if continued for six doses. The lack of neuronal response in those treated from the second week on suggests that the receptors responsible for the neuronal response are either less abundant or not accessible to infused enzyme after the second week of life. The blood brain barrier may not be completely developed in the mouse brain until 10 d after birth(26) and this immaturity may allow intravascular β-glucuronidase access to cell surface receptors in the very young animal that are not accessible in the older mouse. In addition, it is known that the M 6-PR decreases dramatically in heart in late fetal and early neonatal life and the concentration of this receptor is also less in the postnatal than the fetal brain(27,28). A reduction in receptors could contribute to the diminished central nervous system response to infused enzyme as mice age. In vitro studies of enzyme uptake and response in murine MPS VII brain cells, similar to those performed to study uptake of α-L-iduronidase by glial cells and neurons(29), could help determine impediments to β-glucuronidase deliver to neurons, e.g. paucity of cell surface receptors.

The variable response in the central nervous system of mice treated from 1 wk of age (group 2) may reflect variability in enzyme absorption after the i.p. injection at 7 d or in neuronal accessibility or responsiveness to infused enzyme. The failure of neurons to respond to enzyme therapy in mice treated with six weight-based doses of β-glucuronidase initiated at 5 wk of age (group 7) indicates that the time of initiation of therapy is more important than the cumulative dose of β-glucuronidase administered.

The light microscopic morphologic assessment of the amount of lysosomal storage correlated well with and was validated by the morphometric method we used for quantitation of neuronal storage. The fact that carefully controlled light microscopic observations provide meaningful information is reassuring because the morphometric quantitation method is costly and labor intensive, making its wide-spread application limited.

Fixed tissue macrophages in the liver and spleen remained quite sensitive to infused β-glucuronidase at all ages examined. Retinal pigment epithelium also responded but required multiple enzyme doses. The failure of Purkinje cells, corneal fibroblasts, and chondrocytes to respond to β-glucuronidase means that another strategy needs to be developed to target these sites for enzyme replacement. The avascular nature of cartilage and cornea may preclude access of circulating β-glucuronidase to cells in these tissues. Alternatively, or in addition, the paucity of M 6-PR in cartilage(27) may contribute to resistance to therapy. Because mouse Purkinje cells express M 6-PR transcript(30), it is surprising that these cells fail to respond. Others have observed (in a model of α-mannosidosis) that Purkinje cell lysosomal storage is more resistant to BMT than that in cortical neurons(24). Perhaps Purkinje cells are less accessible to circulating enzyme than cortical neurons or their lysosomal storage, which is morphologically distinct from that elsewhere in the central nervous system(5), is more refractory to enzymatic degradation.

The development of immunological reactions to enzyme infusion has previously been observed in this and other LSD models(19,31,32). Reactions were not seen in the MPS VII mice in this study treated from birth. Possibly early initiation of enzyme treatment in the newborn period induced tolerance to the antigens in the infused enzyme that provoked anaphylactoid responses in the older animals. The development of antibodies to β-glucuronidase (although it occurred in only one mouse) did not appear to be a detriment to enzyme therapy in mice treated from birth to 5 wk. However, enzyme replacement by gene therapy methods in another LSD model led to development of antibodies to the enzyme with abrogation of further response to the corrective enzyme(33). Circumventing the immunological barrier to continued enzyme replacement is a challenge for which the MPS VII model may be suitable.

The most important conclusion from these studies in murine MPS VII is that treatment must be initiated early in life to produce optimal neuronal response, which has been shown to correlate with functional improvement(13). In MPS I patients, BMT is more effective at preventing neurological deterioration if performed early in the course of the disease(34) and the most important determinant for intellectual outcome is the amount of neurological impairment before BMT(35). Optimal response to enzyme therapy in human patients with MPS or similar LSD will likely require early intervention. A strong argument can be made for early diagnosis, e.g. with newborn screening for LSD(36), and, for families at risk, in utero screening and possibly even initiation of therapy in utero, to allow the most effective treatment of these disorders.