Main

The mucopolysaccharidoses (MPS) are a group of inherited lysosomal storage disorders that have in common an inability to degrade glycosaminoglycan chains (1, 2). Mucopolysaccharidosis VI (MPS VI, Maroteaux-Lamy syndrome) results from the deficiency of the lysosomal hydrolase N-acetylgalactosamine-4-sulphatase (4S; EC 3.1.6.1). This enzyme is required for the lysosomal degradation of dermatan sulfate and chondroitin sulfate glycosaminoglycan chains. A reduction in the level of 4S results in the accumulation of undegraded glycosaminoglycans within the lysosomes of many tissues (3). Children with MPS VI present with a range of clinical symptoms, most notably skeletal abnormalities. For severely affected children this involves short stature and bone radiographic changes (dysostosis multiplex). Joint movement is also progressively limited (1, 2).

Enzyme replacement therapy (ERT) is currently in clinical use for at least two genetically transmitted disorders: Gaucher's disease and severe combined immunodeficiency/adenosine deaminase deficiency (4–6). Clinical trials of ERT in several lysosomal storage disorders are also currently underway (abstracts of the Fifth International Symposium on MPS and Related Diseases, unpublished proceedings). The potential for ERT to be an effective treatment modality for MPS disorders is high. All are examples of gene mutations in which pathology results from the deficiency of a single lysosomal hydrolase. The deficient enzymes have been purified and extensively characterized (7–17) and the gene for many of these enzymes has been cloned (13, 18–26). In addition, most of the enzymes have been expressed at high levels in recombinant systems, providing a source of human enzyme for therapy (13, 25, 27–34). MPS VI is a strong candidate for ERT. Unlike other MPS disorders, CNS involvement is minimal. However, to be effective in promoting normal skeletal growth, administered enzyme must be able to affect the storage of undegraded dermatan sulfate chains within cells in cartilage and bone.

ERT trials in various animal models of MPS have been conducted. The tissue distribution of the enzyme deficient in MPS I (α-L-iduronidase), MPS VI (4S), and MPS VII (β-glucuronidase) and the efficacy of treatment have been assessed (35–38). In all of these studies a reduction or resolution of lysosomal storage was observed in a variety of connective tissue cells, including bone cells. The radiographic appearance of bone also improved in the MPS VI cat and MPS VII mouse (36, 38). In the MPS VI cat, the improvement in bone pathology was clearly dependent upon dose, route of administration, and the age at which therapy was initiated (37–39). Overall, the animal studies suggest that ERT will prove an effective therapy for most of the pathology associated with MPS VI and other MPS with minimal CNS involvement. However, no improvement in chondrocyte vacuolation was observed in any animal study and lysosomes remained full of undegraded glycosaminoglycan, suggesting that minimal amounts of enzyme had been taken up by these cells. We suggest that this continuation of cartilage pathology in the presence of ERT is responsible for the progression of degenerative joint changes in MPS VI cats undergoing ERT, which are indistinguishable from those observed in untreated MPS VI animals (38). Clearly, cartilage must be targeted in any treatment program that seeks to improve the overall skeletal health of children with MPS VI.

In this study we have modified the structure of 4S to promote its diffusion through cartilage and have altered the dosing regimen of 4S in an attempt to redefine the distribution pattern of 4S after i.v. administration. The uptake of normal and modified forms of recombinant human 4S into cartilage chondrocytes and skin fibroblasts in vitro is examined. The ability of the various forms of 4S to remove accumulated substrate, their partitioning coefficient in cartilage, their distribution in tissues of the feline MPS VI model, and their efficacy in correcting skeletal pathology in vivo is reported.

[3H]-leucine (144 Ci/mmol) and [35S]-sulfate (550 mCi/mmol) were purchased from NEN-Dupont (North Ryde, N.S.W., Australia). Ethylene diamine (ED), poly-L-lysine (PL), and 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (EDC) were purchased from Sigma Chemical Co. (St. Louis, MO). N-hydroxysulfosuccinimide (NHS) was purchased from Pierce Immunotechnology (Rockford, IL). All other chemicals were of reagent grade. MPS VI cats were bred and maintained at the Institute of Medical and Veterinary Science, Adelaide, S.A., Australia (38). All animal studies were reviewed and approved by the Women's and Children's Hospital Animal Research Ethics Committee.

METHODS

Isolation of chondrocytes and fibroblasts.

Chondrocytes were isolated from normal and MPS VI cat costal cartilage with a sequential trypsin (0.1%) and collagenase (0.2%) digestion. Liberated cells were passed through a 100-μm nylon mesh followed by a 54-μm mesh to ensure a single-cell suspension was obtained. Chondrocytes were plated out in Dulbecco's modified Eagles medium (DMEM) supplemented with 10% FCS, 50 units/mL penicillin, and 50 μg/mL streptomycin.

Fibroblasts were isolated by explant outgrowth from normal and MPS VI skin biopsies as previously described (40). Cells were maintained in basal medium Eagle medium supplemented with 10% FCS, 50 units/mL penicillin, and 50 μg/mL streptomycin.

Preparation and purification of 4S and 3H-4S.

Recombinant human 4S (rh4S) was purified from the medium of an overexpressing CHO cell line by monocloncal antibody affinity chromatography using a modification of Anson et al. (27), employing antibody 58.3 in place of antibody 4.1. Radiolabeled rh4S was prepared by incubating the same CHO cell line in leucine-free minimum essential medium alpha modification containing 10% dialyzed FCS, 10 mM NH4Cl, and 20 μCi/mL [3H]-leucine for 48 h. After labeling, the media was collected and concentrated on an Amicon YM10 Diaflo membrane (Amicon Corporation, Danvers, MA). Radiolabeled rh4S was purified from the concentrated medium by MAb affinity chromatography as described above.

Protein concentration was determined by the method of Bradford (41) and 4-sulphatase activity was determined using the fluorogenic substrate 4-methylumbelliferyl sulfate as previously described (7). The specific activity of both the purified rh4S and 3H-rh4S preparations was 55 000 nmol/min/mg protein and the 3H-rh4S preparation contained 2.09 ± 0.46 × 107 cpm per mg protein (n = 9 preparations) resulting in 363 cpm/nmol/min.

Derivatization of 4-sulphatase.

Recombinant human 4S (unlabeled and radiolabeled) was coupled to either ED or to PL to increase the pI of the molecule using the water-soluble EDC crosslinker according to previously published methods (42, 43). The reaction was enhanced by the addition of NHS (44). Briefly, rh4S was mixed with ED or PL (average Mr 3000) and NHS and the pH adjusted to 5.0 with HCl. EDC was added to the reaction mix and coupling was allowed to proceed at room temperature for 6 h. Enzyme was dialyzed against PBS overnight at 4°C to remove unreacted components and activity was subsequently determined as previously described using the fluorogenic substrate (7). Unmodified rh4S exhibited a range of pI between 6.0 and 6.5. Coupling to ED (4SED) resulted in a slight increase in pI to between 6.5 and 6.8. Coupling to PL (4SPL) resulted in a pI of >9, yielding a molecule with a net positive charge at neutral pH. Modification also resulted in a decrease in enzyme activity, with 4SED retaining 65% of unmodified 4S activity and 4SPL retaining 48% of original activity, as measured using the fluorogenic substrate 4-methylumbelliferyl sulfate (7).

Uptake of 3H-4S into cartilage chondrocytes and skinfibroblasts in vitro.

Monolayer cultures of normal feline chondrocytes were preincubated in DMEM (10% FCS) ± 5 mM mannose-6-phosphate (M-6-P) for 1 h at 37°C (n = 3). After preincubation, 3H-rh4S (50 nmol/min/mL) was added to the cell cultures and incubated for a further 48 h at 37°C. Monolayer cultures of normal human skin fibroblasts were preincubated in BME (10% FCS) ± 5 mM M-6-P for 1 h at 37°C. After preincubation, 3H-rh4S (n = 3), 3H-4SED (n = 2), or 3H-4SPL (n = 2) at a concentration of 50 nmol/min/mL was added to the cultures and incubation continued for a further 44 h at 37°C. Both cell types were harvested with trypsin-versene. The cell pellets were resuspended in 500 μL 20 mM Tris, 0.5 M NaCl, pH 7.0 and freeze/thawed seven times. The amount of 3H-enzyme transported into the cells in the presence and absence of M-6-P was determined by liquid scintillation counting of an aliqout of the cell lysate.

Correction of storage in MPS VI cells.

Cells (feline costal chondrocytes or human skin fibroblasts) were incubated in Ham's F-12 containing 20 μCi/mL [35S]-sulfate for 24 h to label and accumulate glycosaminoglycan chains within lysosomes. At the end of the labeling period, cells were harvested using trypsin-versene and replated in the absence of [35S]-sulfate in DMEM medium (chondrocytes) or BME medium (fibroblasts). Recombinant human 4S, both unmodified and modified forms, was added at a concentration of 50 nmol/min/mL and the incubation continued for a further 48 h. Cells were harvested with trypsin-versene and the pellets resuspended in 500 μL 20 mM Tris, 0.5 M NaCl, pH 7.0 and freeze/thawed seven times. The amount of [35S]-labeled glycosaminoglycan chains remaining within the cells was determined by liquid scintillation counting of an aliquot of the cell lysate.

Partitioning coefficient of modified 4S.

The partitioning coefficient of rh4S between the extracellular matrix of articular cartilage and tissue culture medium was determined as described by Maroudas (45). Articular cartilage was dissected from the leg bones of fetal sheep obtained from the abattoir. Fetal sheep were used as a convenient source of cartilage, as only small amounts of tissue can be removed from cats. The tissue was minced and freeze/thawed to kill the chondrocytes and thus prevent cell metabolism. Tissue was then incubated in DMEM containing 0.125 mg/mL 3H-4S, 3H-4SED, or 3H-4SPL for 24 h. The tissue was then washed in PBS and incubated in DMEM containing 0.125 mg/mL of unlabeled 4S, 4SED, or 4SPL for a further 24 h. The radioactivity appearing in the desorbate solution was determined by liquid scintillation counting and the dry weight of the tissue measured. The partitioning coefficient (k) for rh4S in articular cartilage was calculated according to Maroudas (45).

Tissue distribution of 4-sulphatase.

Unmodified and modified forms of 3H-4S were injected into the cephalic vein of normal, MPS VI, and ERT MPS VI cats at a dose of 27 500 nmol/min/kg, containing up to 10 × 106 cpm/kg. After 1–4 h the animals were killed and tissues removed. Radiolabeled 4S was extracted from the tissues of interest into 1 M NaOH (using the ratio of 5 mL NaOH per g of tissue), overnight at room temperature. An aliquot of the supernatant was taken for liquid scintillation counting. The amount of 4S recovered from each tissue as a percentage of that injected was calculated from the amount of 3H radioactivity in the sample and the tissue weight. The amount of 4S in serum was calculated assuming total blood volume is 7% of body weight and that serum volume represents 60% of total blood volume (46).

Enzyme replacement therapy in MPS VI cats with unmodified andmodified forms of 4S.

MPS VI animals were treated from birth (38) with weekly i.v. infusions of 4SED (n = 2) or 4SPL (n = 2) at a dose of 55 000 nmol/min/kg body weight or unmodified 4S at a dose of 27 500 nmol/min/kg body weight twice weekly (n = 3). To enable measurement of bone formation rate, all cats received s.c. injections of the fluorochrome dyes, calcein, and oxytetracycline, on days 10/11 and 2/3, respectively, before death. Cats were killed at 6 mo of age and tissues collected and processed for electron microscopy as previously described (37, 38). The L5 vertebra was also taken and parameters describing structural trabecular elements were measured on von Kossa stained sections (39). Bone formation rate was measured by fluorescent microscopy (λ = 420 nm) on unstained undecalcified tissue using a routine manual point counting technique. Standardized histomorphometric nomenclature and formulas were used (47). Statistical analysis was performed using a one-way ANOVA and Tukey's-HSD (honestly significant difference) test (p < 0.05). Clinical and histomorphometric information obtained for the ERT MPS VI cats in this study was compared with data obtained for the L5 vertebra of 6-mo-old normal (n = 9), untreated MPS VI (n = 15), and ERT MPS VI at 55 000 nmol/min/kg (1 mg/kg) rh4S weekly (n = 4) cats analyzed in a previous study (38, 39).

RESULTS

Recombinant 4S uptake into normal cells and the correction ofstorage in MPS VI cells.

The uptake of 3H-rh4S into normal cat chondrocytes was markedly reduced in the presence of 5 mM M-6-P, representing 14% of control levels (Fig. 1). Similarly, the modified forms of 4S were also taken up by cells, in this instance skin fibroblasts, in a M-6-P-dependent manner. The uptake of 3H-rh4SED and 3H-rh4SPL was inhibited to 17% and 23%, respectively, of control levels by 5 mM M-6-P.

Figure 1
figure 1

3H-4S, 3H-4SED, and 3H-4SPL uptake into normal cells. Unmodified 3H-4S (open bars), 3H-4SED (filled bars), and 3H-4SPL (hatched bars) was added to the culture medium of either normal chondrocytes or normal skin fibroblasts at a concentration of 50 nmol/min/mL in the presence and absence of 5 mM M-6-P. The amount of 4S transported into the cell after 48 h (chondrocytes) or 44 h (fibroblasts) was determined by scintillation counting of the cell lysate. Results are presented as the mean ± SD.

Full size image

Chondrocytes isolated from MPS VI cat cartilage were incubated with 35SO4 for 24 h to label and accumulate glycosaminoglycan chains within their lysosomes and then incubated in the presence or absence of rh4S for a further 48 h. In the absence of added rh4S, a significant amount of [35S]-labeled material remained within the chondrocytes (Table 1). The addition of 50 nmol/min/mL of rh4S resulted in the clearance of [35S]-labeled glycosaminoglycan from the chondrocytes, with 7% of the control level remaining. Similarly, the modified forms of rh4S were able to efficiently clear stored [35S]-labeled glycosaminoglycans from the lysosomes of MPS VI fibroblasts (Table 1). No significant difference was observed between the ability of unmodified or modified forms of 4-sulphatase to correct storage (t test, p < 0.05).

Table 1 Clearance of 35S-labeled storage material from MPS VI cells by modified and unmodified forms of 4S MPS VI cells were incubated with 35SO4 for 24 h, as described in “Methods,” to label and accumulate intracellular glycosaminoglycans. Cells were then incubated in the absence or presence of 4S, 4SED, or 4SPL at a final concentration of 50 nmol/min/mL for a further 48 h. Cells were harvested by trypsin-versene digestion and the amount of 35S-labeled material remaining within the cells was determined by scintillation counting of an aliquot of the cell lysate. Results are presented as the mean ± SD of n = 2* or n = 3† observations.
Full size table

Partitioning coefficient of rh4S and modified rh4S into articularcartilage.

The ratio of enzyme within cartilage tissue to surrounding culture medium was expressed as the partitioning coefficient (k). For unmodified recombinant 4S the value of k for articular cartilage was 0.06. Coupling to ED resulted in an increase in partitioning coefficient to 0.106, whereas coupling to PL resulted in a partitioning coefficient of 0.09, increases of 77% and 50%, respectively.

Distribution of unmodified and modified forms of rh4S invivo.

The tissue distribution of recombinant human 3H-4S, 3H-4SED, and 3H-4SPL into various tissues of normal cats, MPS VI cats, and MPS VI cats on ERT, 1 or 4 h after administration, is shown in Table 2. In all cases, enzyme was taken up primarily by the liver with much lower amounts appearing in other tissues. This is in agreement with previous reports of lysosomal enzyme tissue distribution studies, which show the liver to be the primary site of enzyme uptake (35, 37, 48). No major difference was noted in the distribution of 3H-4S into tissues of normal, MPS VI, or MPS VI-ERT cats. The distribution of 3H-4SED and 3H-4SPL into treated animals did display some differences compared with the distribution of 3H-4S. The amount remaining in the serum was increased as was the level of enzyme distributing to the kidney for both forms of modification. The amount of 3H-4SPL in the lung was very high compared with other distribution groups and was thought to be a consequence of the anaphylactic response observed in this animal at the time of final injection. A small increase in the amount of 3H-4SED recovered from the brain was also observed. In all cases, the amount of enzyme recovered from cartilage was negligible and at the limits of detection of the assay.

Table 2 Distribution of 3H-4S, 3H-4SED, and 3H-4SPL in cat tissues: percentage of enzyme injected recovered per tissue The tissue distribution of unmodified 3H-4S and 3H-4SED was determined 1 h after intravenous enzyme administration. The tissue distribution of 3H-4SPL was determined 4 h after intravenous enzyme administration. Tissues were extracted with 1 M NaOH, and the amount of enzyme present was determined by scintillation counting. Results are presented as the mean ± SD, where n = 2. * Percentage of radioactivity recovered per gram of tissue not total tissue.
Full size table

Response to ERT with modified forms of rh4S.

MPS VI cats were treated with 55 000 nmol/min/kg 4SED or 4SPL, administered i.v. on a weekly basis for 6 mo, or with 27 500 nmol/min/kg unmodified rh4S, administered i.v. twice weekly for 6 mo (Table 3). The response of animals to these treatment regimens was compared with the previous treatment protocol of 55 000 nmol/min/kg unmodified enzyme once a week for 6 mo and to untreated normal and MPS VI animals (38, 39). In support of these previous observations, lysosomal vacuolation was lessened in various soft connective tissue cells such as Kupffer cells in the liver and perivascular cells of the brain in animals undergoing ERT with modified enzyme or the altered dosing regimen of unmodified 4S. In addition, 4SPL appeared to improve lysosomal clearance in dense connective tissues such as aorta and heart valve when compared with unmodified enzyme. In contrast, treatment with 4SED and the altered dosing schedule of unmodified 4S appeared to result in a worse outcome than unmodified enzyme in brain, skin, dura, and heart valve. No resolution of or reduction in cartilage chondrocyte lysosomal vacuolation was observed in any treated animal.

Table 3 Lysosomal storage in tissues from ERT-treated MPS VI cats compared with age-matched, untreated MPS VI animals NA, not available; 0, no lysosomal storage; ±, very mild vacuolation; +, mild; ++, moderate; +++, severe lysosomal vacuolation. * Data from these animals previously reported in Crawley et al (38).
Full size table

The radiographic appearance of animals on i.v. ERT with rh4SED or rh4SPL demonstrated an improvement compared with untreated animals but was indistinguishable from animals treated with unmodified rh4S (data not shown). Quantitative histomorphometric analysis of bone structure and formation revealed that individual animals treated with 4SED or 4SPL displayed variation in the final bone volume reached (Table 4). This is most likely due to variability in skeletal disease in untreated MPS VI cats (37–39). All treated animals had higher final bone volumes than untreated MPS VI animals, however, modification of enzyme structure did not yield an appreciable advantage over unmodified rh4S as a therapy protocol. Varying the dosing schedule to 27 500 nmol/min/kg twice weekly did not demonstrate any advantage over a 55 000 nmol/min/kg once weekly regimen.

Table 4 Bone mineral volume (BV/TV) and bone formation rate (BFR/BS) in treated cats Results are expressed as the mean ± SD. * Data from these animals previously reported in Byers et al. (39). † Denotes significant difference MPS VI versus normal (Tukey's-HSD, p < 0.05). ‡ Denotes significant difference MPS VI treated versus MPS VI untreated (Tukey's-HSD, p < 0.05). § Denotes no significant difference between MPS VI treatment groups (Tukey's-HSD, p < 0.05).
Full size table

DISCUSSION

MPS VI is similar to the majority of MPS disorders (with the exception of MPS III) in that the skeletal system is a major site of pathology. Short stature resulting from the impaired transition of cartilage to bone is observed in the growing MPS VI child. In severe cases this can lead to cessation of growth by age 7–8 and an ultimate height of 107–138 cm (2). Bone shape and density are also altered in MPS VI and cartilage pathology is evident in the MPS VI cat fetus (38). From our ERT studies in the feline model of MPS VI, it is clear that although conventional i.v. injection of enzyme ameliorates most pathology associated with MPS VI, including partial resolution of bone disease, it does not result in sufficient enzyme reaching cells within cartilage to alter the progression of pathology in this tissue (37–39).

In support of a previous study (49), we have shown that cartilage cells possess M-6-P cell surface receptors capable of mediating the transport of exogenous 4S into chondrocytes. Once inside the cell, 4S was targeted correctly to the lysosome where it was effective in removing stored glycosaminoglycans. In these respects, chondrocytes have mechanisms analogous to those demonstrated in skin fibroblasts for the uptake and processing of exogenous lysosomal enzymes (27–30). Thus we can expect that ERT would be effective in improving cartilage pathology given that the enzyme can reach chondrocytes within the cartilage tissue.

Cartilage is an avascular tissue within which its constituent cells are surrounded by an extensive extracellular matrix. The movement of nutrients and other molecules must therefore occur by diffusion through the matrix from the synovial fluid in the joint space. The rate of diffusion of proteins through cartilage extracellular matrix is dependent upon several factors including the hydrodynamic size (45) and charge (50) of the diffusing protein and the concentration and charge density of proteoglycans within the cartilage matrix (51). Recombinant human 4S has a molecular mass of 67 kD (27) and with a pI in the range 6.0–6.5. 4S will have an overall negative charge at neutral pH. The diffusion of 4S through cartilage matrix should therefore be retarded both on the basis of size and charge. Indeed, the partitioning coefficient for 4S in articular cartilage was found to be low (0.06). This value is comparable to that reported for BSA, a molecule of similar size (64 kD) and negative charge (pI = 4) (45).

Recombinant human 4S was therefore coupled to either ED or PL to yield a net positive surface charge and promote diffusion through cartilage extracellular matrix. The modified enzyme was found to retain activity and was taken up by MPS VI cells in a M-6-P-dependent manner and delivered to the lysosome, where it was capable of removing stored glycosaminoglycans. Modification of 4-sulphatase also improved its partitioning into articular cartilage tissue, with a 77% increase noted with ED coupling and a 50% increase noted with PL coupling. However, the in vivo tissue distribution of modified forms of 4-sulphatase displayed a similar pattern to unmodified enzyme, with small differences noted only in the level recovered from serum and kidney. The level recovered from cartilage was minimal regardless of the physical structure of 4S. This is reflected in the lack of difference in the overall clinical and skeletal response of animals to ERT with modified enzyme compared with unmodified 4S. In particular, cartilage remained unaffected. Small differences, however, were observed in the in vivo response to 4SPL, which appeared to have an improved ability to clear storage in dense connective tissues such as aorta and heart valve. As heart failure is a frequent cause of death in MPS VI patients, these results may suggest a role for enzyme modification in the treatment of specific tissue pathology.

Based on our previous observation that the tissue T1/2 of recombinant 4S was on the order of 2–4 d (37), enzyme was administered twice weekly at half the dose to achieve a more uniform tissue presentation of 4S. However, no obvious advantage in clinical response was observed between this dose regimen and previously published dosing regimens (38, 39).

Although not specifically addressed in the design of this study, modification of 4S to yield a net positive charge also has the potential to improve penetration of the enzyme through the blood-brain barrier (BBB) (52–54). This is thought to occur via an electrostatic mechanism between cationic proteins and anionic sites on the surface of brain endothelial cells (55). Movement through the BBB is an important aspect of ERT for MPS disorders with CNS involvement such as MPS I, II, and III (2). Although an increase in the level of 4SED recovered from the brain was observed in the distribution study, this value was not corrected for enzyme remaining in contaminating vasculature (56). Additional studies have shown that after correction for enzyme remaining within the blood vessels, modification of 4S with either ED or PL did not significantly increase transport across the BBB (data not shown). Thus it appears that modification with ED localizes 4S onto the surface of vascular endothelial cells. Interestingly, modification with PL did not have the same effect, confirming the results of Poduslo et al. (57) in which modifications to yield higher positive charge did not lead to a concomitant increase in BBB penetration. Presumably, interactions other than an electrostatic mechanism play a role in the transport of these modified proteins across the BBB.

The in vivo cartilage distribution results and response to ERT are surprising in light of the in vitro results demonstrating that modification improves cartilage penetration without significant loss of enzyme activity as measured by enzyme assay and clearance of stored glycosaminoglycan. Antibody titre measured in serum of animals treated with 4SED or 4SPL (data not shown) was observed to be within the range of that observed in animals treated with unmodified 4S (58). Thus, although the physical structure of 4S in this study had been modified, the immune response was no different from that previously observed in animals undergoing i.v. ERT from birth with unmodified enzyme. This supports our previous finding that cats treated from birth with i.v. enzyme had antibody titres that were not significantly different from either normal cats or untreated MPS VI cats (58). Thus, the lack of improvement in clinical response to modified enzyme or dose regimen compared with unmodified 4S cannot be explained by a changed immune response altering the clinical outcome. Rather, the observation that enzyme was similarly distributed throughout the body regardless of enzyme structure implies that modification had not significantly changed enzyme penetration into the tissues, cartilage in particular, under study. We suggest that the poor distribution of enzyme into cartilage in vivo can be attributed to several factors. First, the rapid clearance of enzyme from the circulation and the large amount of enzyme distributing into the liver (∼50% of that injected), irrespective of enzyme modification or dose, leaves little enzyme available for uptake by other tissues. Second, the structure of the synovial joint represents a barrier to the uptake of enzyme from the circulation, which must be crossed before penetration of the cartilage matrix. Proteins are excluded from the joint space by the synovial membrane on the basis of size (59). Although modification of 4S improved penetration of cartilage matrix in vitro, the in vivo results suggest that the modifications had minimal effect on penetration of the synovial membrane and therefore did not significantly increase the amount of enzyme reaching the synovial space from the circulation. Procedures that result in enzyme bypassing the liver and remaining in circulation for long time periods may be more effective in promoting enzyme uptake into tissues such as cartilage by making more enzyme available for uptake across the synovial membrane. To target therapy to this important site of pathology, specific strategies must also be developed to improve penetration of the synovial membrane and increase the localized levels of 4S within the joint space.

Acknowledgments.

The authors thank Professor M. Haskins, University of Pennsylvania, Philadelphia, Pennsylvania, for the original MPS VI cats and the staff at Gilles Plains, Institute of Medical and Veterinary Science, Adelaide, S.A., Australia, for care and maintenance of the cat colony. We also thank Krystyna Niedzielski and Chris Boulter for the purification of rh4S, Peter McNeil for preparing bone sections, and Richard Davey for processing of tissues for electron microscopy.