Failure of Alglucerase Infused into Gaucher Disease Patients to Localize in Marrow Macrophages

Ernest Beutler, Wanda Kuhl, and Luke M. Vaughan Department of Molecular and Experimental Medicine, The Scripps Research Institute, La Jolla, California, U.S.A.

ABSTRACT Background: Gaucher disease is a common glycolipid Gaucher disease immediately after enzyme infusion. storage disease, caused by a deficiency of lysosomal Marrow ,B-glucosidase activity and chitotriosidase (a ,3-glucosidase (glucocerebrosidase). Alglucerase is a form macrophage marker) was determined on all samples. of glucocerebrosidase enriched with terminal mannose Results: Even with the large bolus doses used for the moieties, so as to "target" the preparation to the high- treatment of Gaucher disease by some, scarcely any affinilty macrophage receptor in patients with Gaucher (-glucosidase activity was found in marrow samples; the disease. Our earlier in vitro studies indicated that alglu- amount of the enzyme was much less than would have cerase was bound by cells other than macrophages by a been anticipated had the enzyme been evenly distrib- widely distributed, low-affinity mannose receptor. uted to all body cells. Materials and Methods: Bone was removed at surgery Conclusions: Alglucerase is not targeted to marrow from six patients with Gaucher disease; in three cases, macrophages. Its unquestioned therapeutic effectiveness bone was obtainable both when the patient was un- must be due either to its activity at some site other than treated and after receiving an infusion of alglucerase. marrow macrophages or to the fact that the doses ad- Four samples of bone were obtained from patients with- ministered are so enormous that even a small fraction is out Gaucher disease and served as controls. A bone sufficient to achieve a therapeutic effect. marrow aspirate was obtained from another patient with

INTRODUCTION cally bound by the mannose receptor of macro- Gaucher disease is the most common glycolipid phages. storage disorder, with an estimated incidence of However, in studies in vitro we found that approximately 1:800 in the Ashkenazi Jewish alglucerase was bound by receptors found on all population. The disease phenotype is manifested cell types tested, and that the specific mannose in macrophages, which become engorged with receptor accounted only for a minute amount of glucocerebroside as a result of the glucocerebro- the binding observed (1). We now present in vivo sidase deficiency that causes this disease. data, obtained from six Gaucher disease patients Aside from allogeneic marrow transplanta- undergoing joint replacement and one patient tion, the only treatment effective in reversing whose marrow was obtained by aspiration, that glycolipid storage in Gaucher disease is the infu- show that far from being macrophage targeted, sion of glucocerebrosidase that has been modi- very little of the infused enzyme reaches the fied by removing outer sugars to expose inner marrow. mannose moieties. Such enzyme (alglucerase; Ceredase) has been designated as "macrophage targeted", because of the belief that it is specifi- MATERIALS AND METHODS Address correspondence and reprint requests to: Ernest Sources of Marrow Samples Beutler, Department of Molecular and Experimental Medi- cine, The Scripps Research Institute, 10666 North Torrey Bone specimens were obtained from six patients Pines Road, La Jolla, CA 92037, U.S.A. with type I Gaucher disease. Each was undergo-

320 Copyright © 1995, Molecular Medicine, 1076-1551/95/$10.50/0 Molecular Medicine, Volume 1, Number 3, March 1995 320-324 E. Beutler et al.: Alglucerase Is Not Targeted to Marrow Macrophages 321 ing either surgical joint replacement or a revision RESULTS of a joint replacement. In the case of Patients 1, The results of the ,B-glucosidase activity measure- 4, and 6, the bone came from the femur. The ments are summarized in Tables 1 and 2. As sample from Patient 2 was from the knee, and expected, the activity in bone from Gaucher dis- the first (pretreatment) sample from Patient 3 ease patients ranged from one fifth to one tenth from the shoulder and the second, (post-treat- of that from normal subjects. After infusion of ment) sample from the acetabulum (assay on the alglucerase, only a very small increase in enzyme femoral head gave similar results). In three in- activity was found in each of the three patients, stances, we were fortunate enough to be able to much less than calculated even if the distribution obtain bone from the same patient on two occa- of the enzyme had been uniform to all body cells sions, one while receiving alglucerase therapy (see below). and one when not being treated with alglu- The chitotriosidase activity of normal mar- cerase. An iliac crest marrow needle aspirate was row exceeded that of peripheral blood lympho- obtained from a seventh patient with Gaucher cytes (N.V. = 1.1 U/g DNA) (unpublished), a disease and from two normal volunteers. Three finding consistent with that expected from a control surgical bone specimens were obtained macrophage-rich tissue. The activity of the en- from the hip and one from the lumbar spine zyme was greatly increased in most of the mar- from four patients without Gaucher disease un- row samples from patients with Gaucher disease, dergoing joint replacement surgery. indicating that Gaucher cells were probably present in the samples, since these cells are known to be very rich in chitotriosidase. Sample Preparation and Assay Procedures Surgical bone samples (Table 1) were packed in dry ice and transported to our laboratory, where they were stored in a -70°C freezer until DISCUSSION in assay. Each sample was homogenized a Or- Enzyme thoblend Bone Mill replacement therapy for glycolipid stor- (distributed by Zimmer, age diseases was proposed by De Duve Warsaw, Indiana), the bone chips were removed first (6) and has seemed a by centrifugation, and enzyme activity was de- particularly promising thera- peutic to the treatment of Gaucher dis- termined by the method of Raghaven et al. (2). approach The cells from marrow aspirate (Table 2) were ease because macrophages function to remove fractionated with Ficoll-Hypaque, and the adher- material from the circulating blood and, if gluco- cerebrosidase were introduced into the circula- ent cells from the interface fraction (3) were assayed for ,B-glucosidase activity using the tion, macrophage uptake might result in enzy- method of Raghaven et al. (2). DNA levels were matic removal of the storage lipid. Enzyme determined the method of Labarca and therapy was first attempted by Brady et al., who by Paigen infused unmodified enzyme into the circulation and was as (4), f3-glucosidase activity expressed in 1974 (7). Although it was claimed that there U/g DNA. Chitotriosidase was measured activity was a in a minor modification of the method of Hollak significant decrease the amount of glu- by cocerebroside in the liver and in et al. (5). erythrocytes, no therapeutic effect was achieved. We targeted the enzyme to macrophages by incorporating it into erythrocyte ghosts (8), as had first been proposed Alglucerase Infusions by Ihler et al. (9), and by coating the cells with Commercially available alglucerase (Ceredase; immunoglobulin to enhance uptake by macro- Genzyme Corp., Boston, MA, U.S.A.) was di- phages. This approach, used in seven patients, luted in physiological saline solution and infused proved very effective in targeting the enzyme to intravenously over a period of 2 hr into each of macrophages in vitro (10) and in vivo (1 1). In one five patients. In the case of three of the patients patient, a decrease in liver size was observed. (Patients 4, 5, and 7), the enzyme infusion was However, because the amount of enzyme we concluded immediately before entering the op- were able to purify in a research laboratory was erating room or performing the marrow aspira- very small and the efficiency of entrapment was tion; in the case of Patient 3, it was given 8 hr not very high, this approach seemed impractical before surgery; and in the case of Patient 6, it was and was abandoned. given 48 hr before surgery. Achord and Sly (12) suggested that the man- 322 Molecular Medicine, Volume 1, Number 3, March 1995

TABLE 1. The f8-glucosidase activity of bone removed from patients with and without Gaucher disease (GD) before and after infusion with alglucerase 13-Glucosidase Activity (U/g DNA) Chitotriosidase Untreated Treated Expecteda (U/g DNA)

Normal Controls (4) 9.3 ± 5.3 10.0 ± 7.00 GD Patient 1 0.8 16 GD Patient 2 1.5 64 GD Patient 3 0.8 7.7" >300 (48, 76)c GD Patient 4 0.7 3.2d >11.5 (110,129) GD Patient 5 1.5 2.6e >600 (215, N/A) GD Patient 6 2.Of >4.7 (N/A, 70)

Normal values are mean + 1 standard deviation. N/A, not available. aExpected values are those that would be achieved if the infused enzyme were uniformly distributed to all body cells and re- mained with a t,12 of 8 hr. bBone taken 8 hr after infusion of 60 U alglucerase/kg body wt. 'Values in parentheses are pre- and postinfusion values, respectively. dBone taken at end of 2-hr infusion of 1.15 U alglucerase/kg body wt. eBone taken at end of 2-hr infusion of 60 U alglucerase/kg body wt. fBone taken 48 hr after infusion of 60 U alglucerase/kg body wt.

nose receptor of macrophages might provide a diately after enzyme infusion. The extracts were route for efficient delivery of enzyme in Gaucher rich in chitotriosidase activity, indicating that disease. Enhanced uptake of glucocerebrosidase they were rich in macrophages. After infusion of by murine liver was demonstrated to occur when alglucerase into five patients, three of whom had enzyme with increased exposed mannose con- assays performed when not receiving alglucerase tent was infused (13,15). Enzyme modified in therapy, an increase of ,B-glucosidase activity of this way was produced commercially as alglu- 1.7- to 9.6-fold were documented. This range cerase (Ceredase), and has been designated agrees with a single observation reported by Bar- "macrophage targeted". ton et al (16) of an 8-fold increase in 3-glucosi- We obtained bone removed during joint re- dase activity, but not with the report by Fallet et placement therapy from six patients with Gau- al (17) that there was an increase to approxi- cher disease and from a seventh patient on mately 12-fold of normal 19 hr after a dose of 30 whom marrow aspiration was performed imme- U alglucerase/kg and of 8-fold of normal 10 days after infusion of 60 U/Kg of alglucerase into Gau- cher disease patients. Indeed, even if the enzyme level were markedly increased after infusion of TABLE 2. The ,B-glucosidase activity of enzyme, with an apparent intracellular half-life adherent cells from bone marrow aspirates for exogenously introduced enzyme of about 8 removed from two normal volunteers and hr (1,18,19), it is difficult to understand how an from a patient with Gaucher disease immediately after infusion of 60 U alglucerase elevation could persist through 30 half-lives. per kg body wt Since the average DNA content of human tissues is 100 mg DNA/Kg body weight, uniform ,f-Glucosidase Activity distribution of 60 U/Kg of infused enzyme (U/g DNA) should have immediately increased the enzyme activity of the bone by 60 units of enzyme per Normal Controls 4.3, 6.2 100 mg of DNA, or 600 units of enzyme per gram GD Patient 7 0.9 of DNA. The infusion of 1.15 U/kg body wt would have increased the enzyme activity by E. Beutler et al.: Alglucerase Is Not Targeted to Marrow Macrophages 323

11.5 U/g of DNA. Yet, the actual increase, even REFERENCES in the patients receiving the largest dosage, was 1. Sato Y, Beutler E. (1993) Binding, internal- less than 7 U/g DNA in each of the four patients. ization, and degradation of mannose-termi- Thus, as shown in the table, the amount of en- nated glucocerebrosidase by macrophages. J. zyme actually found in all of the treated patients Clin. Invest. 91: 1909-1917. was much less than would have been expected if 2. Raghavan SS, Topol J, Kolodny EH. (1980) the enzyme had been equally distributed to all Leukocyte beta-glucosidase in homozygotes body cells. In the case of Patient 4, who received and heterozygotes for Gaucher disease. Am. 60 U/kg, the amount actually found was only J. Hum. Genet. 32: 158-173. 0.2% of what might have been expected with 3. Beutler E, Kuhl W, Matsumoto F, Pangalis G. uniform distribution of enzyme to all body cells. (1976) Acid hydrolases in leukocytes and In the case of Patient 6, bone sampling was car- platelets of normal subjects and in patients ried out 48 hr after enzyme infusion. Intracellu- with Gaucher's and Fabry's disease. J. Exp. lar enzyme half-life has been estimated to be Med. 143: 975-980. approximately 8 hr. This would lead to an esti- mated increment of 4.7 U/g DNA had the en- 4. Labarca C, Paigen K. (1980) A simple, rapid, zyme been equally distributed to all body cells and sensitive DNA assay procedure. Anal. and disappeared with a tl12 of 8 hr. A total of only Biochem. 2: 344-352. 2 U/g very close to normal baseline was found. 5. Hollak CEM, van Weely S, van Oers MHJ, Marrow from bones surrounding the proxi- Aerts JMFG. (1994) Marked elevation of mal joints is rich in macrophages, as documented plasma chitotriosidase activity. A novel hall- here by the measurement of chitotriosidase ac- mark of Gaucher disease. J. Clin. Invest. 93: tivity, an enzyme that appears only as monocytes 1288-1292. mature into macrophages. The lack of ,B-glucosi- 6. De Duve C. (1964) From cytases to lyso- dase activity in this marrow after alglucerase in- somes. Fed. Proc. 23: 1045-1049. fusion indicates that this preparation is not only 7. Brady RO, Pentchev PG, Gal AE, Hibbert SR, not targeted to marrow macrophages as hoped, Dekaban AS. (1974) Replacement therapy but it is preferentially removed from the circula- for inherited enzyme deficiency. Use of pu- tion at other sites, very likely the vascular endo- rified glucocerebrosidase in Gaucher's dis- thelium or endothelium of the liver. It is con- ease. N. Engl. J. Med. 291: 989-993. ceivable that marrow macrophages are, in some 8. Dale GL, Beutler E. (1976) Enzyme replace- respect, unique and that preferential uptake oc- ment therapy in Gaucher's disease: A rapid curs in other organs, such as liver and spleen. high-yield method for purification of gluco- However, it is notable that the response to alglu- cerebrosidase. Proc. Natl. Acad. Sci. U.S.A. 73: cerase of marrow function is readily demon- 4672-4674. strated, and yet, very little of the infused enzyme 9. Ihler GM, Glew RH, Schnure FW. (1973) can be found there. Measurements in other or- Enzyme loading of erythrocytes. Proc. Natl. gans such as liver and spleen are not ethically Acad. Sci. U.S.A. 70: 2663-2666. feasible. The mechanism by which alglucerase is therapeutically effective is unclear. It is not mac- 10. Dale GL, Kuhl W, Beutler E. (1979) Incor- rophage targeted. poration of glucocerebrosidase into Gau- cher's disease monocytes in vitro. Proc. Natl. Acad. Sci. U.S.A 76: 473-475. 11. Beutler E, Dale GL, Kuhl W. (1977) Enzyme replacement with red cells. N. Engl. J. Med. ACKNOWLEDGMENTS 296: 942-943. We appreciate the cooperation of our patients in 12. Achord DT, Brot FE, Bell CE, Sly WS. (1978) making samples available to us and to Harlan C. Human beta-glucuronidase: In vivo clear- Amstutz, MD, C. A. Chalian, MD, and John P. ance and in vitro uptake by a glycoprotein Heiner, MD, for collecting and shipping bone recognition system on reticuloendothelial samples to us. This is Manuscript 8967-MEM cells. Cell 15: 269-278. from The Scripps Research Institute. This work 13. Furbish FS, Oliver KL, Zirzow GC, Brady RO, was supported by National Institutes of Health Barranger JA. (1984) Interaction of human Grants DK36639 and RR00833 and the Stein placental glucocerebrosidase with hepatic Endowment Fund. lectins. In: Barranger JA, Brady RO (eds). 324 Molecular Medicine, Volume 1, Number 3, March 1995

Molecular Basis of Lysosomal Storage Disorders. 17. Fallet S, Sibille A, Mendelson R, Shapiro D, Academic Press, New York, pp. 219-232. Hermann G, Grabowski GA. (1992) Enzyme 14. Doebber TW, Wu MS, Bugianesi RL, et al. augmentation in moderate to life-threaten- (1982) Enhanced macrophage uptake of ing Gaucher disease. Pediatr. Res. 31: 496- synthetically glycosylated human placental 502. beta-glucocerebrosidase. J. Biol. Chem. 257: 18. Pentchev PG, Kusiak JW, Barranger JA, et 2193-2199. al. (1978) Factors that influence the uptake 15. Murray GJ, Doebber TW, Shen TY, et al. and turnover of glucocerebrosidase and al- (1985) Targeting of synthetically glycosy- pha-galactosidase in mammalian liver. Adv. lated human placental glucocerebrosidase. Exp. Med. Biol. 101: 745-752. Biochem. Med. 34: 241-246. 19. Beutler E. (1993) Gaucher disease as a par- 16. Barton NW, Brady RO. (1994) Gaucher dis- adigm of current issues regarding single gene ease: Skeletal responses to enzyme replace- mutations of humans. Proc. Natl. Acad. Sci. ment therapy. Gaucher Clin. Perspect. 2: 8-11. U.S.A. 90: 5384-5390.

Contributed by E. Beutler on December 28, 1994.