Main

Gaucher disease type 1 is the most common lysosomal disease and has high frequency among Ashkenazi Jews(1). The disease results from the defective activity of acid β-glucosidase(N-acylsphingosyl-β-D-glucopyranoside:glucohydrolase, EC 3.2.1.45)(1). Numerous point mutations and gene rearrangements at the acid β-glucosidase locus lead to the variants of Gaucher disease(1). The pathophysiology of Gaucher disease type 1, the nonneuronopathic variant, derives from the accumulation ofβ-glucosyl-terminated sphingolipids in cells of monocyte/macrophage origin(1, 2). Because of this single cell lineage involvement, bone marrow transplantation from normal related donors ameliorated several disease manifestations(3–6). This provided the rationale for gene therapy via gene transfer into bone marrow stem cells(7–10). The involvement of the monocyte/macrophage cells and the presence of an α-mannosyl receptor on the surface of macrophages(11, 12) is the basis for enzyme therapy of this disease using acid β-glucosidase withα-mannosyl-terminated oligosaccharides(13, 14).

To date, about 900 patients with Gaucher disease type 1 are receiving infusions of α-mannosyl-terminated human acid β-glucosidase from placenta or recombinant sources(15–20). A variety of doses and dosage schemes are efficacious for reversing hepatic, splenic, and bone marrow macrophage involvement. The great expense of the enzyme therapy products has sparked vigorous discussions about the dose level needed for effective therapy. These discussions are based primarily on clinical outcome measures of visceral and hematologic resolution with little data about the pharmacology(13, 14, 21), tissue distribution or cellular localization in the target organs. Histologic and enzyme data are available for only two treated patients; significant amounts of enzyme persisted in hepatic and/or splenic tissues for several days after enzyme injection but very little was in the lungs, brain, and bone marrow mononuclear cells(18, 22).

Because of the serendipitous nature of such information in humans, we undertook a systematic study in mice of the tissue distribution, time courses, and appearance and disappearance of enzyme activity in tissues. These studies show a nonuniform enzyme distribution with primary localization to liver and splenic monocyte/macrophage and endothelial cells. These studies also show complex uptake and disappearance of exogenously administered enzyme that include the inability to deliver fully active enzyme to the interior of cells.

METHODS

Materials. The following were from commercial sources: Ceredase(alglucerase for injection; 80 IU/mL) and 4MU-Glc (Genzyme Corp., Boston, MA); 4-methylumbelliferone, TX, human serum albumin, and Lowry reagents (Sigma Chemical Co., St. Louis, MO); sodium TC (Calbiochem, La Jolla, CA); gelatin(J. T. Baker Co., Phillipsburg, NJ); RPMI 1640 and Dulbecco's modified Eagle's medium (Life Technologies, Inc., Bethesda, MD); methoxyflurane (Metofane, Pitman-Moore, Terre Haute, IN); Ficoll-Paque and Phast System reagents(Pharmacia Biotech Inc., Picataway, NJ); immobilon polyvinyldine fluoride transfer membrane (Millipore, Bedford, MA); Vectastain Elite ABC kit, alkaline phosphatase-vector red system and peroxidase-DAB system (Vector Labs, Burlingame CA); FITC-conjugated goat affinity-purified antibody to rabbit IgG F(ab′)2 (Cappel, Durham, NC); rat anti-mouse macrophage 160-kD membrane protein (F4/80) MAb and rat antimouse CD31 (PECAM-1) MAb (PECAM-1 is expressed on endothelial cell surfaces) (Research Diagnostics Inc., Flander, NJ); and goat anti-rabbit IgG conjugated to alkaline phosphatase and alkaline phosphatase conjugate substrate kit (Bio-Rad, Richmond, CA). Other reagents were analytical grade or better. All statistical analyses were done with Systat (Systat, Inc., Evanston, IL).

Animal treatment and tissue dissection. Enzyme activity, CRIM, and CRIM-specific activity determinations represent the means from duplicate samples from each of three mice at each time point for each enzyme source and experiment. The data points in the graphs represent mean values and standard errors. Balb/c mice (Harlan Sprague-Dawley, Indianopolis, IN; 8-9 mo old, 28-32 g) were anesthetized (Metofane) before injection, dissection, and sampling. Alglucerase in saline (100 μL) was immediately injected as a bolus via tail vein. In all animals, blood (100 μL) was obtained for enzyme activity from the retro-orbital plexus 5 min after injection. The enzyme activity was stable in serum from clotted blood (1 h, 4°C) separated for 15 min at 12,000 rpm. Serum was then rapidly frozen in small aliquots. Preliminary studies showed minimal (<10%) losses of added enzyme activity during this procedure. Anesthetized animals were exsanguinated via subaxillary vessels at 10, 20, 30 min and 1, 2, 4, 16, and 24 h. Blood samples were processed as above. The brain, kidney, liver, lung, spleen, and thymus were immediately recovered, rinsed with cold saline, blotted dry, snap frozen, and stored at -80°C until assay. On the day of the assays, the total tissue weight was determined and aliquots (50-100 mg) were minced, diluted 1:10(wt/vol) in 1% TC/1% TX and homogenized (15 strokes). These suspensions were sonicated (4°C, Sonifier 450, Branson Ultrasonics Corp., Danbury, CT) at 80% output for 30-, 20-, and 10-s pulses at 1-min intervals. The supernatants(15 min, 12,000 rpm) were collected for enzyme activity and protein and immunoblot analyses. Dilution experiments showed the lack of soluble inhibitors and activators in the tissue sample.

Bone marrow cells were flushed from femurs and tibias with RPMI 1640 medium. The mononuclear-enriched fraction was isolated in a Ficoll-Paque gradient developed by centrifugation (1200 rpm, 20 min). The light density mononuclear cells at the interface were collected, treated with 0.15 M NH4Cl to lyse red blood cells and washed twice with saline. The mononuclear cell pellets were stored (-20°C). For analysis, these pellets were sonicated (100 μL, 1% TC/1% TX) and processed as above. All tissue samples were processed at 4°C. The uninjected animals were zero time point controls. Several animals with failed injections (s.c. injections) were used as additional sham-injected controls.

Enzyme assays. Acid β-glucosidase activity was determined fluorometrically with 4MU-Glc (4 mM) as substrate(23). Briefly, samples (5 μL of tissue lysate supernatant or 50 μL of serum) were assayed (30 min, 37°C) in 20 mM citrate/phosphate, pH 5.5, 0.25%TC/0.25% TX 100, 4 mM 4MU-Glc, as a final volume of 200 μL. Acidβ-glucosidase-specific activity is expressed as nanomoles/h/mg of tissue protein(24) or per mL of serum. To allow direct comparison to the therapeutic results in humans, the amount of enzyme injected was expressed in units taken directly from the label (Ceredase) as micromoles/min p-nitrophenyl-β-D-glucoside cleaved. In comparison, the enzyme activities in all animal and/or human sources were determined and expressed as nanomoles/h 4MU-Glc substrate cleaved. The 0 point activities in the graphs represent the mean background activity in the uninjected and failed injected animals. These values varied by <10% from the mean.

Stability of acid β-glucosidase activity. Buffered and unbuffered serum or saline samples were incubated at 37°C. At time points 0, 10, 20, 30, 60, and 120 min, aliquots (4 μL) were mixed with 196 μL of 1% TC/1% TX and assayed for acid β-glucosidase activity. The pH of saline was 5.6. Serum samples or saline were buffered as needed to pH 7.4 or 5.6 with 1 MN-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid buffer, pH 5.6 or pH 7.6, respectively.

Immunoblots. Equal amounts of supernatant protein from tissues at each time point were subjected to SDS-PAGE (12.5% gels) and then electroblotted onto membranes (Phast system). Blocking (1 h) was with 3% gelatin in TBS. Human acid β-glucosidase CRIM was detected after 1 h of incubation at room temperature with rabbit antisera raised against native human placental acid β-glucosidase(25) in T-TBS containing 0.1% gelatin. This was followed by reaction with alkaline phosphatase-conjugated secondary antibody (1/3,000) in T-TBS with 1% gelatin. The membranes were washed three times in T-TBS for 10 min after each antibody incubation. The alkaline phosphatase conjugate substrate kit was used for visualization according to the manufacturer's instructions.

Densitometry and quantitation of CRIM. The amount of acidβ-glucosidase was quantified by transmission densitometry of membranes. The area under the human acid β-glucosidase peak was determined by integration (Pharmacia). Standards were of known amounts of acidβ-glucosidase protein and activity. The mass of CRIM was referenced to these standards on each gel. CRIM-specific activities were calculated as nanomoles/h/mg of acid β-glucosidase based on the measured enzyme activity due to administered enzyme and the mass of CRIM. Using 4 mM 4MU-Glc as substrate, the specific activity of homogenous acid β-glucosidase purified from alglucerase was 1.507 ± 0.268 (SD) × 106 nmol/h/mg of acid β-glucosidase protein. The determined CRIM-specific activities of the acid β-glucosidase in alglucerase preparations were±15% of this mean value.

Human and mouse tissue preservation. Tissues from two Gaucher disease type 2 siblings were obtained at autopsy. Case 1 had received injections of alglucerase (60 IU/kg of body weight) over a period of 2 mo preceding death. The final regularly scheduled dose was given about 48 h before death and about 60 h before autopsy. Case 2 had been treated from d 4 after birth until 15 mo of age(22). A regularly scheduled dose (30 IU/kg of body weight) of alglucerase was given 48 h preceding her demise. The tissues for enzyme, histologic, and/or immunohistochemistry studies were obtained by snap freezing or fixed in formalin and embedded in paraffin. Mouse tissues for enzyme assay were as described above. For immunofluorescence studies, fresh mouse tissues were fixed at 4°C in 4% paraformaldehyde in PBS overnight, cryoprotected in 30% sucrose/M-1 embedding matrix (1:1 vol/vol) for 2 d, embedded in this matrix for 2 d, and quickly snap frozen in liquid nitrogen. The embedded tissue blocks were stored at -80°C until used.

Immunohistochemistry. Mouse anti-human acid β-glucosidase MAb 41(26) was used in the studies of Gaucher disease tissues. The formalin-fixed human tissues were deparaffinized with xylene and ethanol washes. Precipitated pigments were removed by soaks (2 h, room temperature) with absolute ethanol saturated with picric acid and rinsed with distilled water. The Vectastain Elite ABC kit (Vector Labs) was used for immunochemical staining. When using the peroxidase-diaminobenzidine system and before applying the blocking serum, the sections were incubated for 15 min at room temperature with 0.5% H2O2 in methanol to remove endogenous peroxidase activity and then treated with 0.1% trypsin/0.1% CaCl2 for 15 min at 37°C. When using the alkaline phosphatase vector red system, 1 mM levamisole was added during the substrate incubation to inhibit endogenous alkaline phosphatase activity while staining frozen tissue sections.

Immunofluorescence and immunohistochemical staining of mouse tissues. The fixed mouse tissues sections (7-10 μm) were postfixed (1 h) in 4% paraformaldehyde at room temperature and then incubated with 0.2% TX in PBS (10 min). The sections were blocked (3% BSA and 10% goat serum in PBS) at 37°C (30 min) and incubated with rabbit anti-human acid β-glucosidase antiserum diluted (1/250) in PBS containing 3% BSA and 3% goat serum at 37°C (1 h) or anti-mouse macrophage or anti-mouse PECAM-1 antibodies. With the latter two antibodies, visualization was as above with alkaline phosphatase. For the anti-acid β-glucosidase antibodies, the tissue sections were washed in washing buffer (0.05% Tween 20 in PBS) at room temperature, thrice for 15 min each. FITC-conjugated goat anti-rabbit antibody diluted to 1/20 in 3% BSA/3% goat serum was applied on each slide and incubated at 37°C (1 h). The sections were washed at room temperature, thrice for 15 min each. Propidium iodide/Antifade (Oncor, Gaithersburg, MD) was added on the section to obtain a nuclear contrast stain and to preserve fluorescence. The signals were visualized by fluorescence microscopy.

RESULTS

After the bolus injection of alglucerase (60 IU/kg of body weight), enzyme activity disappeared from plasma [t½ ≈ 3-5 min(data not shown)]. The results conformed to an exponential decay curve consistent with a single primary or several kinetically similar distribution compartments. Enzyme reentry into the circulation was not detected by enzyme assay or immunoblotting. A t½ ≈ 20 min was obtained with 300 IU/kg bolus doses in two mice, indicating saturation of the primary distribution compartments. At the peak level, over 98% of the recovered enzyme(60 IU/kg body weight) was in serum, i.e. not in red or white blood cells. Using serum peak acid β-glucosidase activities (60 IU/kg), recoveries of injected enzyme were >80% in most animals. These studies were done at 3-5 min and before enzyme protein or activity was detectable in tissues. Concurrent CRIM determinations in serum showed a similar disappearance profile (t½ ≈ 5 min) with enzyme protein being undetectable by ≈20 min.

Figure 1 shows the pH stability (37°C) of acidβ-glucosidase in serum or saline. In saline (pH 5.6) with or without added albumin, the enzyme was stable for over 2 h. In serum (pH 5.6) thet½ was ≈90 min. At pH 7.4 in either serum or saline, acid β-glucosidase activity had a t½ ≈ 10 min. In unbuffered serum the pH rapidly rose to >8.2 and the enzyme activity t½ was <2 min. After exposure to buffered serum, pH 7.4 (37°C) for 120 min, the pH was readjusted to 5.6 without recovery of enzyme activity. The loss of enzyme activity in all of these experiments was irreversible, because the activity could not be recovered under numerous pH and detergent conditions. The enzyme protein withMr ≈ 60 000 was unaltered on Western blots for >24 h.

Figure 1
figure 1

The effect of pH on acid β-glucosidase stability. Alglucerase was added to serum or saline buffered at pH 7.4 or 5.6 and incubated at 37°C for the indicated time. Aliquots were removed and assayed at pH 5.6 for acid β-glucosidase activity. In saline at pH 5.6(•) or saline in the presence of albumin at pH 5.6 (♦;), serum at pH 5.6 (·), serum at pH 7.4 (★), or saline at pH 7.4 (▪). The same amounts of initial enzymatic activity were added to each sample and the results are shown as the percent of the initial acid β-glucosidase activity as determined from aliquots obtained within 1 min of dilution into the indicated solution.

Full size image

The uptake and turnover of administered acid β-glucosidase was assessed in various tissues. After a bolus injection, liver enzyme activity peaked at 10-20 min (Fig. 2A). Activity levels at 10 and 20 min were not statistically different (p = 0.563). The activity disappearance had rapid (t½ ≈ 45 min) and slow(t½ ≈ 12 h) phases with return to baseline levels of activity by ≈ 18-24 h (Fig. 2A). The rapid disappearance phase accounted for the loss of ≈80% of the peak of acidβ-glucosidase activity. In spleen, enzyme activity peaked at 10-30 min. The activities at 10, 20, and 30 min were not significantly different(p = 0.487). The activity disappearance curve was biphasic(Fig. 2B) but not as smooth as that in liver. The first phase had a t½ ≈ 70 min and the second phase had at½ ≈ 14 h (Fig. 2B). Baseline enzyme activity was reached by 24 h. In spleen, the rapid and slow phases each accounted for ≈50% losses of the recovered acid β-glucosidase activity. In liver and spleen, the peaks of activity were 1.5-2-fold above background. The points at 24 h were not statistically different (p> 0.1) from background, but all other points were (p < 0.05). Because of the relative sizes of the organs, the majority (≈80%) of the total recoverable enzyme (60 IU/kg) was localized to the liver (60%) and spleen (20%) at 20 min postinjection. Compared with 60 IU/kg doses, a dose of 300 IU/kg led to only an additional 50% increase in enzyme activity recovered from the liver and spleen at 30 min. These results indicate that a substantial portion of the enzyme was not distributed to these tissues by this time or that it had been inactivated.

Figure 2
figure 2

Time courses of acid β-glucosidase activities in liver (A) and spleen (B). The main figure of A shows the detail for the 24 h time course after a bolus of 60 IU/kg of alglucerase was injected. The inset shows the detail for the first 4 h. The corresponding data for the spleen are in B. Each data point represents the mean ± SE for duplicate assays for each of three mice.

Full size image

Uptake also was assessed in isolated bone marrow mononuclear cells(Fig. 3). Maximum enzyme activity in cells was at≈50-60 min with a disappearance t½ ≈ 40-60 min. The increase in enzymatic activity in this cell fraction was ≈10-30% over baseline. The compartment size could not be estimated but the total enzyme accounted for in bone marrow mononuclear cells was probably small (10%). In thymus and kidney, enzyme activity was present transiently(t½ < 30 min) with activity being detectable(≈20% increases) only at the earliest time points (Fig. 3). No increase in enzyme activity over baseline could be detected in brain or lung samples.

Figure 3
figure 3

Time courses for the appearance and disappearance of injected acid β-glucosidase activity in various tissues including bone marrow mononuclear cells (·), lung ([diamond]), thymus ([diamond]), kidney (★), spleen (▪), and liver (•). Each data point represents the mean± SE for duplicate assays in each of three different mice for the entire time course in every tissue. The complete results for liver and spleen are in Figure 2.

Full size image

The bolus method of administration was used to evaluate the appearance times for enzymatic activity in tissues. We had anticipated that enzyme activity would appear in tissues nearly instantaneously. However, ≈20 min was required to achieve peak activities in the liver and spleen and ≈50-60 min in bone marrow mononuclear cells. Our activity assay method provides an overall organ assessment of the enzyme and is not dependent on cellular organelle localization. Consequently, the assessment of enzyme activity uptake kinetics in an organ depends only on the properties of the enzyme and not on its distribution within tissue or cells. To gain insight into the mechanism of the slow appearance of enzyme activity, the appearance and disappearance kinetics of CRIM (i.e. enzyme protein) in liver and spleen was assessed by quantitative immunoblotting. The antibodies do not cross-react with endogenous mouse acid β-glucosidase or other proteins(Fig. 4B, lane 0). The CRIM in liver(Fig. 4) and spleen (Fig. 5) peaked at ≈10-20 min and 10 min, respectively. In liver the CRIM values at 10 and 20 min were not statistically different (p > 0.05), but they were in spleen (p < 0.001). In both tissues the initial disappearance of CRIM was rapid with t½ ≈ 1 h. The slower second phases of CRIM disappearance were similar in both organs with only low CRIM levels by 24 h.

Figure 4
figure 4

Time course for liver CRIM. CRIM was quantified densitometry scanning of immunoblots using rabbit anti-human acidβ-glucosidase antibody that does not cross-react with endogenous murine acid β-glucosidase. A reference standard of a known amount of human acidβ-glucosidase was used on each blot to normalize the results. InA, the means ± SE for duplicate samples from each of three mice are plotted. An immunoblot of liver samples (B) was typical of those used to quantitate the mass of human acid β-glucosidase in each sample. The 0.5* and 24* samples were from mice receiving a dose of 300 IU/kg of body weight. In each sample 6 μg of total protein were applied to the gels.

Full size image
Figure 5
figure 5

Time course for spleen CRIM. CRIM was quantified as inFigure 4, and the data points are as inFigure 4.

Full size image

To quantitate the relationship of enzyme protein and activity disappearances, the CRIM specific activities were determined in liver and spleen. This method provides an estimate of the catalytic rate constant of the enzyme [nanomoles of substrate cleaved/h/ng of CRIM (i.e. enzyme protein)]. When referenced to the enzyme in alglucerase (1.507 ± 0.085(SE) nmol/h/ng of acid β-glucosidase), the CRIM-specific activity indicates the degree of full activity that is retained by the administered enzyme. Theoretically, if the enzyme retains full activity per molecule during proteolysis in tissues, the CRIM-specific activity will remain at ≈1.5 because enzyme protein and activity will be lost in proportionally equal amounts. If additional processes also affect the enzyme i.e. denaturation or unfolding, the CRIM-specific activity will be <1.5;i.e. the enzyme will be partially active or some inactive undegraded enzyme protein will be present.

Figure 6 compares the CRIM-specific activity time courses in liver and spleen after a bolus dose of 60 IU/kg. In liver, the CRIM-specific activity peaked at 30 min (Fig. 6A) at a time when liver CRIM had fallen significantly (Fig. 4A). These values then remained relatively constant up to 4 h. The CRIM-specific activity in liver achieved about 70% of the theoretical maximal value. In spleen, the CRIM-specific activity peaked at 20-30 min and remain constant for about 4 h. The maximal CRIM-specific activity in spleen was 41% of that for the pure enzyme. The CRIM-specific activity values at 10 min show that the enzyme entering the liver and spleen have about 50 and 20%, respectively, of the activity of the administered enzyme. By 20-30 min this enzyme becomes more active in the liver (Fig. 6A) and maintains its activity in the spleen (Fig. 6B).

Figure 6
figure 6

Comparative time courses for liver (A) and spleen (B) CRIM-specific activity. The data were derived from the enzyme activity and quantitative CRIM data inFigures 2,and4 and5, respectively. The level of 1.5 nmol/h/ng of CRIM represents the full activity of purified acidβ-glucosidase purified from alglucerase (1.507 ± 0.268 nmol/h/ng of pure enzyme) under the assay conditions used in these experiments,i.e. 0.25% TX/0.25% TC, 0.04 mM citrate phosphate, pH 5.5, with 4 mM 4MU-β-Glc. The data points represent the means ± SE of duplicate assays from three experimental animals. Details for the first 4 h of the time courses are shown in the insets.

Full size image

The time courses after 30 min for CRIM-specific activities were somewhat different in liver and spleen (Fig. 6,A andB). After an initial rapid decrease (t½ ≈ 90 min) in CRIM-specific activity in liver, a low value was maintained until about 16 h. This result indicates that the early mechanism of enzyme turnover in liver includes rapid proteolysis and denaturation of the enzyme in this organ. The nearly constant CRIM-specific activity, but decreasing CRIM, from 4 to 16 h indicates a primarily proteolytic process during which enzyme activity and protein are lost at similar rates. In spleen, the CRIM-specific activity was constant until 4 h (p = 0.537) and then slowly decreased nearly to 24 h, i.e. proteolysis and denaturation of a partially functional enzyme was occurring. Similar studies could not be accomplished on lung, kidney, or bone marrow.

To assess the cellular distribution of the administered enzyme, immunofluorescence was done using monospecific anti-human acidβ-glucosidase antibody. At 60 IU/kg, the enzyme localized to sinusoidal lining cells throughout the hepatic lobule (Fig. 7a). Many of these cells had the histologic and staining characteristics of macrophages [F4/80+ cells, (data not shown)] or lining cells[endothelial cells, PECAM-1+ cells (data not shown)]. This was more apparent at 300 IU/kg when the entire sinusoidal lining was stained(Fig. 7b). Hepatocytes were not labeled as indicated by the dark nonfluorescent areas of the sections.

Figure 7
figure 7

Immunofluorescence microscopy of various tissues after the injection of alglucerase into mice (a-f) and a Gaucher disease patient (g-i). Panels a and c represent tissues after injection of 60 IU/kg of body weight and panels b, d, e, andf represent tissues after injection of 300 IU/kg of body weight. Human specific rabbit anti-human acid β-glucosidase antibody was used. Immunofluorescence with control serum or with antiserum to human acidβ-glucosidase of mouse tissue from uninjected animals were negative(insets in panels a and c). In liver(a and b) immunofluorescence was localized to nonparenchymal hepatic cells with the distribution of Kupffer cells, as well as tubular structures resembling blood vessels. The tubular appearance and localization to endothelial cells was more apparent with the injection of 300 IU/kg of body weight (panel b) indicating additional non-Kupffer cell delivery. The inset shows a lower power, different field of view. At the higher dose the linear appearance on cell surface, as well as intracellular spheroid particles, suggest that the presence of plasma membrane binding and lysosomal localization, respectively. In spleen (c andd) immunoreactive material was localized to nonlymphoid tissue (red pulp) and at 60 (panel c) 300 (panel d) IU/kg of body weight dosage, the white pulp (W) did not show immunofluorescence. In kidney (e and f) immunoreactive material was found in the mesangial cells of the glomeruli (G) and cortical tubular(CT) epithelial cells at 300 IU/kg of body weight (panele). The inset in panel e is a different field of view. Human enzyme also was evident as linear stained regions of the medulla(panel f) that resembled blood vessels, only at 300 IU/kg of body weight doses. In autopsy specimens of the spleen from the patient treated with alglucerase, small amounts of human acid β-glucosidase enzyme could be detected by immunohistochemical (panel g) and immunofluorescence(panel h) staining. In panel g, the fixed tissue of spleen, black cytoplasmic staining was present in only sinusoidal lining cells and scattered cells of the red pulp (×600). In panel h, immunofluorescence staining of snap frozen tissue suggests a similar distribution of administered acid β-glucosidase, but the loss of architecture resulting from the freezing did not permit identification of the cell type. In fixed tissue of the liver (panel i) the black cytoplasmic granules indicate the localization of the administered enzyme in crescent-shaped sinusoidal lining cells. The polygonal liver cells are nonreactive (×600)

Full size image

In the spleen, the immunoreactive material (60 IU/kg) showed a complex fragmentary linear pattern in red pulp only (Fig. 7c). This distribution was consistent with localization of F4/80+ or PECAM-1+ cells (data not shown). No staining was evident in the lymphoid cells of the white pulp. At 300 IU/kg the uptake pattern in the red pulp was accentuated (Fig. 7d). This indicated preferential distribution and uptake to macrophages and endothelial cells. In many of these cells of liver and spleen, the staining pattern was granular, indicating a cytoplasmic distribution that was consistent with lysosomal localization. Although this pattern was most evident in the liver sinusoidal lining cells, particularly, at the highest doses, staining also indicated a surface membrane localization (Fig. 7,b andd).

At 60 IU/kg, kidney immunofluorescence was in the mesangial region of the glomeruli (G) and proximal cortical tubules (CT) of the kidney (Fig. 7e). At the highest doses, linear tubular structures, resembling blood vessels, were outlined in the medulla of the kidney (data not shown). Renal medullary epithelial cells did not stain positively (Fig. 7f). It should be noted that there was a large amount of immunoreactive material present in the cortical tubules of the kidney at 24 h after the largest doses of enzyme (Fig. 7e, inset). However, enzymatic activity was not detectible in these samples.

In bone marrow and lungs, very small amounts of immunoreactive material were detected. Only a few bone marrow mononuclear cells stained positively. In lungs, several isolated, but uniformly distributed, cells had positive immunofluorescence after alglucerase administration (data not shown). In the brain, no immunoreactivity was detected at any time point. Significantly, blood vessel endothelial cell staining was not observed in brain or lungs at any dose.

Studies in human tissues. In an effort to correlate the findings in mice with those obtained in Gaucher disease patients, we conducted enzymatic activity analyses, immunoblot, and immunohistochemical studies of siblings with Gaucher disease type 2. The first patient (case 1), who had received 2 mo of enzyme therapy, died at 9 mo due to complications of Gaucher disease type 2. The second patient (case 2) died after 15 mo of enzyme therapy. Only enzyme assays were accomplished in case 1, whereas more complete analyses were done in case 2. As shown in Table 1, the enzymatic activities were significantly above those in livers and spleens from untreated Gaucher disease patients. No or small increases in enzyme activity were evident in the lung and brain samples, respectively. The lack of enzyme activity in lungs was surprising because the alveolar spaces in cases 1 and 2 were completely obliterated by Gaucher cells(22). In addition, the liver and spleen in case 2 were essentially normal by histologic studies at the time of death(22). In contrast, case 1's liver and spleen were massively infiltrated by Gaucher cells. By immunoblot, small amounts of intact and some degraded endogenous enzyme could be detected in the liver and spleen from case 2, whereas none was apparent in lungs or brain (data not shown). By immunohistochemical staining using fixed or frozen tissue from case 2, small amounts of acid β-glucosidase could be detected in cells resembling reticuloendothelial cells of the spleen(Fig. 7,g andh), and in Kupffer cells and sinusoidal lining cells of the liver (Fig. 7i). The amount of immunohistochemical reactive material was low due, in part, to the extensive tissue treatment needed to remove interfering pigments. By immunofluorescence staining of frozen tissue, greater amounts of signal were obtained(Fig. 7h), but the morphology of tissue was severely disrupted. In the brain, small amounts of acid β-glucosidase could be detected in periadventitial cells in a nonuniform distribution (data not shown).

Table 1 Tissue levels of acid β-glucosidase activities in normal and Gaucher disease patients
Full size table

DISCUSSION

The availability of enzyme therapy for Gaucher disease type 1 has revolutionized medical care for afficted patients and families. Although nearly 900 patients are currently receiving this form of therapy, relatively little is known about the tissue distribution and fate of administered enzyme either in animal models(27, 28) or in patients(18, 22). Despite this paucity of data, substantial arguments have been made for a variety of dosing regimens based on clinical outcome using hepatic and/or splenic volume changes and hematologic improvements as the primary response indicators(1, 29). Because alglucerase (placental) and imiglucerase [recombinant(20)] are expensive drugs and the therapy for Gaucher disease type 1 is highly efficacious, it seemed prudent to evaluate its tissue and cellular distribution and half-lives as a basis for understanding its efficiency. Recent immunoelectronmicroscopic studies(27, 28) showed that much of the administered enzyme did not become membrane-bound within the lysosomes of Kupffer and other hepatic sinusoidal cells. The present studies show significant enzymatic and targeting inefficiencies for delivery of i.v. administered enzyme to macrophages.

The uptake and degradation kinetics of administered enzyme activity was tissue-dependent. Also, the uptake and delivery of acid β-glucosidase to lysosomes is complex with the initial form of the enzyme having only about 50% of the CRIM-specific activity of the injected enzyme form. This increased to a maximum of ≈70% by 30 min in liver, but occurred when the total activity in the liver had already decreased to about 50%. After this initial phase, the enzyme in liver retained only partial catalytic function (≈10-20% of maximum) for about 16 h during which the enzyme protein was being proteolytically digested rather than denatured. These results imply that the majority of enzyme that enters the liver is incapable of maximal catalytic activity over a prolonged period. Similar mechanisms were operative in the spleen, but the losses of enzyme activity were more severe (>60%). However, the CRIM-specific activity of the enzyme delivered to the spleen was maintained for a much longer period than it was in the liver. These results suggest that the prolonged persistence of partially active administered enzyme forms in spleen may provide a greater degree of total hydrolytic activity for glucosylceramide per dose than obtained in the liver. This could be one basis for the relatively more rapid clinical responses of the spleen, than liver or bone marrow, to enzyme therapy(15–20, 30). The results of bone marrow activity assays indicate a prolonged latency period before the peak of enzymatic activity. In general, the endogenous enzyme activity in all mouse tissues precluded similar analyses of smaller doses of alglucerase.

Because our assay system is nonphysiologic, we are able to assess the maximal potential of the enzyme's catalytic activity within tissues. Our studies cannot determine if the enzyme was partially active or if some fraction was totally inactive. However, both mechanisms may occur because the exogenous enzyme can be denatured at physiologic pH and the partial recovery of intrinsic enzyme activity within tissues suggests a refolding. Inasmuch as the vast majority of enzyme was delivered to either Kupffer cells or endothelial cells within the liver, loss of catalytic power also cannot be accounted for by differential cellular uptake into hepatocytes and/or other nonspecific cells that could act as sinks for inactive enzyme. A significant amount of acid β-glucosidase was localized by immunofluorescence to the endothelial surface membrane at early time points. This finding suggests that the administered enzyme may be exposed in this microenvironment for a significant time to denaturing pH.

Because these studies were conducted in normal animals, we cannot address the relative efficacy of the enzymes and the potentially changing effects in diseased and normalizing tissues of affected patients who are being treated. However, the presence of endogenous mouse enzyme is important. All patients with Gaucher disease type 1 have residual enzyme activity and protein in their cells(25, 31, 32). The majority of these patients have mutant alleles that encode partially functional and, in the case of the frequent N370S allele, normally stable(33, 34) acid β-glucosidase. Thus, in patients the administered enzyme could compete with the mutant enzyme for available binding sites (protein and/or lipids) on the inner lysosomal membrane and endogenous activatorsi.e., saposin C(26, 35) and lead to diminished in situ enzyme function.

The appearance rate of the activity in whole tissues indicates dynamic changes of acid β-glucosidase in transit from the surface of the cells to the lysosome and onto an innerlysosomal membrane. This process is slow (at least several minutes) and may lead to a substantially lowered lifetime for the enzyme in the cells, because the partially unfolded or denatured enzyme forms may be more susceptible to continued denaturation and proteolysis than the native enzyme. This is suggested by the t½ ≈ 60 h of endogenously synthesized acid β-glucosidase in fibroblasts(31, 32) compared with thet½ values <1 h and 12-16 h for exogenous enzyme in murine liver and spleen. Also, the nonphysiologic nature of our enzyme assays may not accurately reflect the available enzyme activity for substrate cleavage within the various organs, but will only reflect the total and, probably maximal, amount of enzyme activity recoverable from those tissues. Consequently, if, as seems reasonable, binding to an inner lysosomal membrane surface is required for the in situ acid β-glucosidase function, the physiologically available enzyme for cleavage of endogenous substrate may be significantly less than suggested by our assays.

The immunofluorescence studies in liver and spleen show the distribution of administered α-mannose terminated acid β-glucosidase to macrophages as well as to endothelial cells of hepatic and splenic blood vessels. Sato and Beutler(21) previously suggested the existence of nonspecific receptors on endothelial cells that might lead to uptake of substantial amounts of alglucerase(21). Our studies indicate the delivery of acid β-glucosidase to monocyte/macrophage derived cells and/or endothelial cells of hepatic and splenic sinusoids even at very high doses. This also confirms the specificity suggested by Furbishet al.(13). These studies support the contention of Sato and Beutler(21) and Murray et al.(27, 28), which had been previously shown(36), of the existence of endothelial cell receptors for internalization of α-mannosyl-terminated glycoproteins. This localization could be a source of inefficiency of enzyme therapy either by misdirecting the enzyme away from monocyte/macrophage derived cells or by continuing to expose the administered enzyme to serum pH that can irreversibly denature the enzyme. Curiously, even at high doses (300 IU/kg) of enzyme, endothelial cell immunofluorescence was not detected in the lung, the brain, or lymphoid areas of the spleen. Consequently, the receptors on endothelial cells of these organs may have difference specificities than those in the liver and spleen. Alternatively, their affinity may be so low that they do not participate significantly in the uptake and distribution of alglucerase. Although our data do not directly demonstrate intralysosomal localization of acid β-glucosidase in endothelial cells, the presence of cytoplasmic granules in mononuclear cells in the liver and spleen indicated the incorporation of the enzyme into this organelle of these cells.

Distribution data are available for only one Gaucher disease patient. They correspond to the results in the mouse. This patient was homozygous for the highly disruptive mutation, L444P, that leads to a highly unstable inactive protein with a half-life of less than 24 h in the cells(33, 34). The very low enzyme levels in this patient's tissues provided a nearly null background to assess the presence of the administered enzyme. Enzyme therapy in this patient maintained normalcy of her liver, spleen, and bone marrow over a period of about 15 mo(22). In addition, her progressive pulmonary disease was indistinguishable from that of her untreated brother, i.e. it was not prevented by alglucerase infusions, and her demise was directly attributable to her lung disease(22). The presence of enzyme activity in her liver and spleen, 48 h after an enzyme infusion, indicates the prolonged presence of exogenous activity in these tissues. Although the half-life of the enzyme in her tissues could not be determined, the peak of activity would be substantially higher than was evident from the autopsy studies. The immunologic studies indicate that the enzyme was localized primarily to macrophages of liver and spleen. The lack of endothelial immunohistochemical signals in this patient suggests that the enzyme may be rapidly turned over in this compartment, whereas enzyme may persist for substantial periods in the macrophage compartment after administration. The lack of enzymatic activity in the lungs, although low CRIM levels were detectable, suggests the possibility that enzyme can be delivered to lung alveolar macrophages via cells which are migrating to the alveolar space carrying the administered enzyme. However, this enzyme must be irreversibly denatured and/or degraded and has little, if any, enzymatic activity in the Gaucher cells in the alveolar space.

These data have significant implications for the ongoing development of enzyme therapy strategies for Gaucher disease type 1. In the initial phases of therapy, a variety of doses and dosing schemes have been used to effect improvement. Some of these schemes have been developed on the basis of dose response(19, 30) or as being modeled after the initial pivotal trial conducted by Barton et al.(15). Other schemes of high frequency, lower dosing of enzyme have been devised based on the reasonable presumption that continuous supply of enzyme in smaller amounts is probably more physiologic(6). However, with either dosing approach, clinical outcome has been the primary indicator for therapeutic efficacy and the potential cost differential is about 2-3-fold. This has created substantial problems for the comparative evaluation at the tissue level of these therapeutic regimens. What is quite clear is that the events after i.v. administration of the enzyme to therapeutic outcome are very poorly understood, including: 1) the stability of the enzyme in various compartments, 2) the directed targeting of the enzyme to various cell types, 3) the transit of the enzyme through the endosomal compartments from the surface to lysosomes, 4) the need to reattach to the inner-lysosomal membrane, and 5) the need to maintain the stoichiometry with the endogenous activator proteins (the saposins). The present studies show that the enzyme may be adversely altered during these intermediate steps and additional elucidation is needed to understand their highly complex nature. It appears likely that efforts directed to increasing the intracellular stability and lysosomal membrane binding of this enzyme could produce substantially greater therapeutic effects than presently attainable.