Enhanced macrophage uptake of synthetically glycosylated human placental beta-glucocerebrosidase.

In order to increase the delivery to macrophages of p-glucocerebrosidase (the enzyme deficient in macrophages from patients with Gaucher’s disease), we have coupled a synthetic triantennary mannosyl glycopeptide, trimannosyldilysine (Man3Lysz), to the native enzyme. The resulting derivative, Man3Lys2-p-glucocerebrosidase (containing 8 to 9 mol of Man3Lys2/67,000 molecular weight enzyme subunit), displays markedly higher affinity for the receptor-mediated glycoprotein uptake system in rat alveolar macrophages in vitro (K,,, of uptake = 0.22 X M, V,, = 0.86 pmol/h/5 X 10‘ cells, based on 67,000 g/mol of enzyme) than does the native enzyme (K, = 0.88 X lo-‘ M, v, = 0.42 pmol/h/5 X lo6 cells) which is still a ligand for this macrophage receptor. At an enzyme concentration of 0.7 p i , alveolar macrophages in vitro endocytose an amount of Man3Lysz-~-glucocerebrosidase qual to more than 11 times their endogenous p-glucosidase level per h. Macrophage-endocytosed native and Man3Lys2-/3-glucocerebrosidase are inactivated by a chloroquine-inhibitable process with half-lives of 1 and 1.5 h, respectively. Upon intravenous infusion into an anesthetized rat, native and Man3Lysz-~-glucocerebrosidase xhibited circulating half-lives of 18.5 and 8.5 min, respectively. A coinfusion of 3 mg of mannan (a specific blocker of the reticuloendothelial carbohydrate-specific uptake system) only slightly retarded the rate of plasma loss of native enzyme but dramatically reduced the rate for the derivatized enzyme (plasma tlIz = 34 min). The in vivo clearance studies also indicated the presence of a plasma inhibition of p-glucocerebrosidase enzyme activity in vitro which is reversible, noncompetitive, and nondialyzable and exhibits a K j of 0.6 pl of plasma with the synthetic substrate. By utilizing a more stable pH 6.0 enzyme infusion solution (versus pH 7.4) and correcting for the plasma inhibition, circulating levels of infused enzyme that were 90% of total activity injected could be accounted for 20 s after end of infusion, instead of as low as 4% of injected activity if these considerations are not heeded. The rate of in vivo plasma loss of infused native enzyme was not affected by coinfusions of 8 mg of asialofetuin or human milk lactoferrin and was only moderately affected by tying off liver circulation. These observations suggest hat the disappearance of infused circulating 8-glucocerebrosidase activity in vivo may be due to clearance mechanisms other than those spe-