About 40 years have passed since the classical glycosidase inhibitor nojirimycin was discovered from the cultured broth of the Streptomyces species. Since then, over 100 glycosidase inhibitors have been isolated from plants and microorganisms. Modifying or blocking biological processes by specific glycosidase inhibitors has revealed the vital functions of glycosidases in living systems. Because enzyme-catalyzed carbohydrate hydrolysis is a biologically widespread process, glycosidase inhibitors have many potential applications as agrochemicals and therapeutic agents. Glycosidases are involved in the biosynthesis of the oligosaccharide chains and quality control mechanisms in the endoplasmic reticulum of the N-linked glycoproteins. Inhibition of these glycosidases can have profound effects on quality control, maturation, transport, and secretion of glycoproteins and can alter cell–cell or cell–virus recognition processes. This principle is the basis for the potential use of glycosidase inhibitors in viral infection, cancer, and genetic disorders. In this review, the past and current applications of glycosidase inhibitors to agricultural and medical fields and the prospect for new therapeutic applications are reconsidered.
In 1968 the validamycin family was detected by the researchers of Takeda Chemical Industries in greenhouse assay for the treatment of sheath blight disease of rice plants caused by the fungus Rhizoctonia solani. Validamycin A (1, Figure 1), the major and most active component of the complex, was isolated from the cultured broth of Streptomyces hygroscopics var. limoneus (Iwasa et al., 1970, 1971; Horii et al., 1972). Takeda commercialized validamycin A for the control of the pathogen in rice and other plants as well as damping off diseases in vegetable seedlings. The structure of validamycin A contains the pseudodisaccharide, validoxylamine A (2, Figure 1), and an additional glucose unit. Validamycin A is not fungicidal or fungistatic but reduces the maximum rate of hyphal extension and increases hyphal branching without affecting the organism's specific growth rate (Nioh and Mizushima, 1974; Trinci, 1985; Robson et al., 1988). In 1987, validoxylamine A was found to be a powerful inhibitor of trehalases of R. solani and other organisms (Asano et al., 1987; Kameda et al., 1987). Furthermore, validamycin A was found to be much more readily taken up into the fungal cell than validoxylamine A and to be intracellularly converted into a powerful inhibitor, validoxylamine A, by β-glucosidase (Asano et al., 1987). Interestingly, validamycin A is readily transported into the fungal cell by the common uptake system with β-oligosaccharides, such as laminarioligosaccharides and cellooligosaccharides (Asano et al., 1991). Validamycin A is a prodrug that is converted within the fungal cell and suppresses breakdown of intracellular trehalose at a very low concentration of 0.1 µg/ml. Trehalose is well known as a storage carbohydrate, and trehalase plays the essential role in the transport of glucose in insects and fungi (Elbein, 1974; Defaye et al., 1980; Thevelein, 1984; Elbein et al., 2003). The pathogen R. solani contains the remarkable trehalose level of 10% (dry weight) (Asano et al., 1987) and is one of the most typical examples of a fungus that grows rapidly by transporting nutrients from the basal part to the tip part through long stretches of hypae (Misato et al., 1977). Validamycin A is practically nontoxic, and the mode of action described gives it a favorable biological selectivity because vertebrates do not depend on the hydrolysis of trehalose.
All powerful trehalase inhibitors reported to date are the pseudodisaccharide type of inhibitors, such as validoxylamine A (2, Figure 1), MDL 25637 (α-homonojirimycin-7-O-β-D-glucopyranoside) (3, Figure 1) (Anzeveno et al., 1989), trehazolin (4, Figure 1) (Ando et al., 1991), salbostatin (5, Figure 1) (Vértesy et al., 1994), and casuarine-6-O-α-D-glucopyranoside (6, Figure 1) (Kato et al., 2003). These compounds are powerful competitive inhibitors of porcine kidney trehalases, with Ki values in the nanomolar range (Salleh and Honek, 1990; Vértesy et al., 1994; Kyosseva et al., 1995; Kato et al., 2003). There are two subsites, one for catalysis and one for recognition, on the active center of porcine kidney trehalase (Asano et al., 1996). The extremely high affinity of pseudodisaccharide inhibitors derives from the synergistic interactions of an alkaloid unit and a sugar (or cyclitol) unit with two subsites (Asano et al., 1996). Trehalose is a blood sugar in insects and a major storage sugar in fungi and yeast. Therefore trehalase inhibitors are expected to have potential as insecticides or fungicides. In fact, validoxylamine A has an insecticidal activity when injected (10 µg) into tobacco cut worm larvae (Asano et al., 1990), and trehazolin also shows an insecticidal activity toward silkworm larvae (50–100 µg, Ando et al., 1995). However, these trehalase inhibitors are not practical as insecticides because they have poor percutaneous absorption through insect skin. The design of trehalase inhibitors or prodrugs with good percutaneous absorption will lead to practical use as insecticides.
Chitin, linear polymers of β1,4-linked N-acetylglucosamine, occurs in fungi and many invertebrates, including insects. Chitin turnover is regulated by two enzymes, chitin synthase and chitinase. Polyoxins interfere with the fungal cell wall synthesis and the insect ecdysis in vivo by inhibiting chitin synthase (Cohen and Cashida, 1982; Gooday, 1989). Polyoxins are commercially produced via fermentation. Polyoxin B finds application against many plant fungal pathogens, and polyoxin D is marketed as the Zn salt to control rice sheath blight R. solani. Chitinase hydrolyzes chitin into oligomer of N-acetylglucosamine and is the key enzyme for the ecdysis of insects. This enzyme has also been proposed as an attractive target for the development of fungicides, insecticides, and antimalarials.
The first chitinase inhibitor, allosamidin (7, Figure 2), was isolated from the mycelial extract of Streptomyces sp. no. 1713 (Sakuda et al., 1987). Allosamidin is a potent inhibitor of insect chitinase, with an IC50 value in the submicrogram/ml range. When allosamidin at a dose of more than 4 µg was injected to the larvae of silkworm and armyworm, all larvae failed in larval ecdysis to the next stage and death followed (Sakuda et al., 1987). Allosamidin derivatives are able to kill Lucilia cuprina blowfly larvae in vitro systems after contact or feeding (Blattner et al., 1997). Quite recently, allosamidin has been found to be a competitive inhibitor of the fungal chitinase CiX1 from Coccidioides immitis, with a Ki value of 60 nM (Bortone et al., 2002). C. immitis, a pathogenic fungus that causes coccidioidomycosis or San Joaquin Valley fever (Drutz and Catanzaro, 1978), is most commonly found in the deserts of the southwestern United States and Central and South America. During the early 1990s, the incidence of coccidioidomycosis in California increased dramatically. About 60% of infections resolve without ever causing symptoms and are only recognized by a positive coccidioidin skin test. The major antigen for antibodies used in this test is C. immitis chitinase CiX1 (Yang et al., 1996). In the remaining 40% of patients, symptoms range from mild coldlike or flulike symptoms to severe pneumonia. Antifungal agents, such as amphotericin B and triazole compounds, are available for treating coccidioidomycosis. However, amphotericin B must be given intravenously and is considerably toxic. If allosamidin is effectively taken up into the fungal cell in vivo, it may become a reasonable therapeutic agent without side effects for coccidioidomycosis.
In the course of the screening for new insecticides, two new cyclopentapeptide chitinase inhibitors, argifin (8, Figure 2) (Omura et al., 2000) and argadin (9, Figure 2) (Arai et al., 2000), were found from the fungal cultures of Gliocladium sp. FTD-0668 and Clonostachys sp. FO-7314, respectively. Argadin potently inhibited the blowfly Lucilia cuprina chitinase, with IC50 values of 150 nM at 37°C and 3.4 nM at 20°C. This inhibitory potential was about 30-fold more potent than that of argifin and 9-fold weaker than that of allosamidin at 20°C (Arai et al., 2000). Interestingly, the structures of these cyclic peptides complexed with chitinase B from Serratia marcescens showed in detail how the peptide backbone and side chains of the inhibitors mimic the interactions of the enzyme with the substrate chitooligosaccharides (Houston et al., 2002). Both argifin and argadin interact with side chains (Asp-142, Glu-144, and Tyr-214) in the chitinase active site that are conserved completely and required for catalytic activity in family 18 chitinases (van Aalten et al., 2000, 2001; Papanikolau et al., 2001). Allosamidin inhibits chitinase B with a Ki value of 450 nM, whereas argifin (8) shows weaker inhibition (Ki = 33 µM) and argadin (9) shows about 20-fold stronger inhibition (Ki = 20 nM) than allosamidin (Houston et al., 2002). Argifin has a more open and oval shape than argadin, and there is only one intramolecular hydrogen bond in argifin, but four in argadin. These results may explain the 1000-fold greater inhibition by argadin over argifin. This interesting approach based on the structures of enzyme–inhibitor complexes leads to the development of more potent inhibitors with possible chemotherapeutical potential.
Digestive α-glucosidase inhibitors
In the 1970s, it was realized that inhibition of all or some of the intestinal disaccharidases and pancreatic α-amylase by inhibitors could regulate the absorption of carbohydrate and these inhibitors could be used therapeutically in the oral treatment of the non–insulin-dependent diabetes mellitus (type II diabetes). Bayer's researchers found that the Actinoplanes strain SE 50 yields a potent sucrase inhibitor, acarbose (10, Figure 3), which inhibits pig intestinal sucrase with an IC50 value of 0.5 µM (Schmidt et al., 1977). Acarbose was also effective in carbohydrate loading tests in rats and healthy volunteers, reducing postprandial blood glucose and increasing insulin secretion (Puls et al., 1977). After intensive clinical development, acarbose (Glucobay) was introduced onto the market in Germany in 1990 for the treatment of diabetes and has since been successfully marketed in Europe and Latin America. In 1996, acarbose was introduced onto the market in the United States under the brand name Precose. Thus acabose provided the physician with the first new therapeutic principle for the treatment of dibetes in nearly 40 years.
In 1984, the validamycin A–producing organism Streptomyces hygroscopicus var. limoneus was reported to coproduce valiolamine (11, Figure 3), which is a potent inhibitor of pig intestinal maltase and sucrase with IC50 values of 2.2 and 0.049 µM, respectively (Kameda et al., 1984). Numerous N-substituted valiolamine derivatives were synthesized to enhance its α-glucosidase inhibitory activity in vitro and the very simple derivative voglibose (AO 128) (12, Figure 3), which was obtained by reductive amination of valiolamine with dihydroxyacetone, was selected as the potential oral antidiabetic agent (Horii et al., 1986). Its IC50 values toward maltase and sucrase were 0.015 and 0.0046 µM, respectively. Voglibose (Basen) has been commercially available for the treatment of diabetes in Japan since 1994.
In 1966 nojirimycin (13, Figure 3) was discovered as the first glucose analog with the nitrogen atom in place of the ring oxygen (Inoue et al., 1966). Nojirimycin was first described as an antibiotic produced by Streptomyces roseochromogenes R-468 and S. lavendulae SF-425 and was shown to be a potent inhibitor of α- and β-glucosidases from various sources (Niwa et al., 1970). However, because this iminosugar with the hydroxyl group at C-1 is fairly unstable, it is usually stored as bisulfite adducts or it may be reduced by catalytic hydrogenation with a platinum catalyst or by NaBH4 to 1-deoxynojirimycin (DNJ) (14, Figure 3) (Inoue et al., 1967). DNJ was later isolated from the roots of mulberry trees and called molanoline (Yagi et al., 1976). DNJ is also produced by many strains in the genera Bacillus and Streptomyces (Schmidt et al., 1979; Murao and Miyata, 1980; Ezure et al., 1985). Despite the excellent α-glucosidase inhibitory activity in vitro, its efficacy in vivo was only moderate (Junge et al., 1996). Therefore, a large number of DNJ derivatives were prepared in the hope of increasing the in vivo activity. Thus miglitol (BAY m 1099) (15, Figure 3) was selected as the most favorable inhibitor out of a large number of in vitro active agents. Miglitol differs from acarbose (10, Figure 3) in that it is almost completely absorbed from the intestinal tract and may possess systemic effects in addition to the effects in the intestinal border (Joubert et al., 1987, 1990). In 1996, miglitol (Glyset) was granted market clearance by the U.S. Food and Drug Administration (FDA) and was introduced onto the market in 1999 as a more potent second-generation α-glucosidase inhibitor with fewer gastrointestinal side effects.
Current scientific evidence demonstrates that much of the morbidity and mortality of diabetes can be eliminated by aggressive treatment with diet, exercise, and new pharmacological approaches to achieve better control of blood glucose level. Furthermore, the possibility of preventing the onset of diabetes using dietary supplements and/or herbal medicines has attracted increasing attention. Salacia reticulata Wight, known as kothalahimbutu in Singhalese and distributed in Sri Lanka and Indian forests, has been used as a supplementary food in Japan to prevent obesity and diabetes. Traditionally, ayurvedic medicine advised that a person suffering from diabetes should drink water left overnight in a mug carved from kothalahimbutu wood. Salacinol (16, Figure 3) and kotalanol (17, Figure 3) have been identified as α-glucosidase-inhibiting components from the water-soluble fraction of the roots and stems of S. reticulata (Yoshikawa et al., 1997, 1998). The IC50 values of salacinol toward rat intestinal maltase, sucrase, and isomaltase are 3.2, 0.84, and 0.59 µg/ml, respectively. The inhibitory activities toward maltase and sucrase are nearly equal to those of acarbose (10, Figure 3) and that toward isomaltase is much more potent than that of acarbose. Kotalanol shows a more potent inhibitory activity than salacinol and acarbose toward surase. Furthermore salacinol has been found to more strongly inhibit the increase of serum glucose levels in sucrose-loaded rats than acarbose (Yoshikawa et al., 1998). The use of dietary supplement to prevent or treat diabetes will increase dramatically as knowledge about bioactive components of food in health increases.
Glycogen phosphorylase inhibitors
In type II diabetes, hepatic glucose production is increased (Defronzo et al., 1992). A possible way to suppress hepatic glucose production and lower blood glucose in type II diabetes patients may be through inhibition of hepatic glycogen phosphorylase (Martin et al., 1991). In enzyme assays, 1,4-dideoxy-1,4-imino-D-arabinitol (D-AB1) (18, Figure 4), which was first isolated from the fruits of the legume Angylocalyx boutiquenus (Nash et al., 1985), was found to be a potent inhibitor of hepatic glycogen phosphorylase (Fosgerau et al., 2000). D-AB1 further inhibited hepatic glycogen breakdown in vivo and displayed an accompanying antihyperglycemic effect, which was most pronounced in obese mice (Fosgerau et al., 2000). In primary rat hepatocytes, D-AB1 was also shown to be the most potent inhibitor ever reported (IC50 = 1 µM) of basal and glucagon-stimulated glycogenolysis (Andersen et al., 1999). Recently the synthetic piperidine alkaloid isofagomine (19, Figure 4), (3R,4R,5R)-5-hydroxymethylpiperidine-3,4-diol, has been reported to potently inhibit hepatic glycogen phosphorylase with an IC50 value of 0.7 µM, and to prevent basal and glucagon-stimulated glycogen degradation in cultured hepatocytes with IC50 values of 2–3 µM (Jakobsen et al., 2001). However, its N-substitution always resulted in a loss of activity compared to the parent compound, and fagomine ((2R,3R,4R)-5-hydroxymethylpiperidine-3,4-diol) (20, Figure 4) was a weak inhibitor of this enzyme, with an IC50 value of 200 µM (Jakobsen et al., 2001). Inhibitors of glycogen phosphorylase would be a beneficial target to attack in the development of new antihyperglycemic agents because such inhibitors have not yet been on the market for the treatment of type II diabetes.
Processing α-glucosidase I inhibitors
The biosynthesis of the oligosaccharide chains of the N-linked glycoproteins is initiated by the cotranslational transfer of a common oligosaccharide precursor (Glc3Man9GlcNAc2) to the glycosylation site (Asn-X-Ser/Thr) of newly synthesized polypeptides in the endoplasmic reticulum (ER) (Elbein, 1987, 1991). The N-glycan chain is modified by a series of reactions within the ER and Golgi apparatus, including trimming by the action of specific processing α-glucosidases and α-mannosidases and elongation by the addition of fucose, galactose, N-acetylglucosamine, sialic acids, and sulfate catalyzed by glycosyltransferases and sulfotransferases. The viral envelope glycoproteins are often essential for virion assembly and secretion and/or infectivity. Compounds that interfere with the glycosylation processes of viral glycoproteins can be expected as antiviral agents. In fact, α-glucosidase inhibitors such as DNJ (14, Figure 3), N-butyl-DNJ (NB-DNJ) (21, Figure 5), castanospermine (22, Figure 5), and 6-O-butanoylcastanospermine (MDL 28574) (23, Figure 5) inhibit human immunodeficiency virus (HIV) replication and HIV-mediated syncytium formation in vitro (Karpas et al., 1988; Fu et al., 1993; Ficher et al., 1995; Taylor et al., 1991). The anti-HIV activity (EC50) determined by syncytial counts of compounds 14, 21, 22, and 23 are 560, 56, 29, and 1.1 µM, respectively, but these compounds are not as potent as zidovudine (EC50 = 0.1 µM) (Taylor et al., 1991). These sugar analogs showing anti-HIV activity have the common property that they are potent processing α-glucosidase inhibitors but not processing α-mannosidase inhibitors. The activities of these inhibitors toward processing α-glucosidase I correlate well with the antiviral effects (Taylor et al., 1991; Sunkara et al., 1989). The in vivo data obtained to date do not promise practical use of processing α-glucosidase I inhibitors as anti-HIV agents. Problems exist in achieving therapeutic serum concentrations of inhibitors needed to inhibit α-glucosidase I sufficiently and side effects such as diarrhea occur. With respect to diarrhea, the prodrug of NB-DNJ, glycovir (SC 49483) (24, Figure 5), was developed as a candidate anti-HIV agent. This prodrug is NB-DNJ tetrabutanoate, which will be converted into active NB-DNJ after it passes through the intestines, avoiding diarrhea (Cook et al., 1995). However, it is believed that morphologic changes in various tissue cells were the result of nonspecific inhibition of host α-glucosidases by the prodrug, causing clinically silent perturbation in host cell glycoprotein processing and/or glycoprotein transport (Khan et al., 1996).
In contrast to the heavily glycosylated HIV envelope glycoproteins, the envelope glycoproteins of hepatitis B virus (HBV) contain only two glycosylation sites (Mehta et al., 1998). However, the HBV glycoproteins are sensitive to inhibitors of the N-linked glycosylation pathway. In this virus, correct glycosylation appears to be necessary for processes involved in transport of the virus out of the host cell. In vitro treatment of HBV with NB-DNJ results in a high proportion of virus particles being retained inside the cells (Block et al., 1994). Recent work has shown that N-nonyl-DNJ (NN-DNJ) (25, Figure 5) reduced the viremia in chronically infected woodchucks in a dose-dependent manner (Block et al., 1998). NN-DNJ is 100–200 times more potent than NB-DNJ in inhibiting HBV in cell-based assays (Mehta et al., 1998). Furthermore NN-DNJ, compared with NB-DNJ, exhibits a prolonged hepatic retention of bovine viral diarrhea virus, a tissue culture surrogate of human hepatitis C virus (Zitzmann et al., 1999). A single drug against HBV and hepatitis C may be of great therapeutic value. However when processing α-glucosidase inhibitors are used as antiviral agents, it remains to be determined what effects occur on host cell glycoprotein processing and/or glycoprotein transport.
The influenza pandemic of 1918–19, known as Spanish flu that killed more than 20 million people, prompted a substantial research effort to find a preventive agent and/or cure for this oldest and most common disease. There are two major surface antigen proteins, hemagglutinin and neuraminidase (sialidase; NA), which are seen as spikes covering the surface of the viral particle (Winn and Westenfeld, 1995). In addition to these two proteins, influenza A virus has a small hydrophobic protein (M2) that functions as an ion channel selective for monovalent ions (Pinto et al., 1992). Until the approval of new agents (NA inhibitors) in 1999, there were two options available: vaccines and the antiviral drugs amantadine and rimantadine. The M2 inhibitors, amantadine and rimantadine, are both clinically effective for the treatment of influenza , but not influenza B. The use of these drugs are of limited usefulness because of the lack of activity against influenza B virus and the emergence of drug-resistant virus, often occurring as rapidly as 2 days after start of therapy (Hayden and Couch, 1992; Aoki, 1998).
NA performs a vital function in the final stage of the viral life cycle. NA-deficient influenza virus is still infective, but the budding virus particles form aggregates or remain bound to the infected cell surface (Liu et al., 1995). That is, the lack of NA makes it difficult for virus to travel from one cell to another. Hence, viral NA is a potential and promising molecular target for anti-influenza drugs. The first substrate-based NA inhibitor described was Neu5Acen (2-deoxy-2,3-dehydro-N-acetylneuraminic acid; DANA) (26, Figure 6) (Meindl and Tuppy, 1969). DANA is a reasonably potent inhibitor of influenza NA with a Ki value in the micromolar range, but it is not selective for influenza NA (Meindl et al., 1974; Holzer et al., 1993). Despite in vivo less efficacy of DANA (Palese and Schulman, 1977) and the possibility that such compounds may not be suitable candidates as anti-influenza drugs, considerable research effort has been devoted to the development of substrate-based NA inhibitors. Introduction of the trifluoroacetamido group in place of the acetamido of DANA to give 2-deoxy-2,3-dehydro-N-trifluoroacetylneuraminic acid (FANA) (27, Figure 6) resulted in a slight improvement in inhibitory activity (Ki = 0.8 µM) toward influenza NA (Palese et al., 1974), but it was ineffective in animal models of influenza infection (Palese and Schulman, 1977).
A breakthrough for rational drug design was certainly the deciphering of influenza NA's three-dimensional structure in 1983 (Varghese et al., 1983; Colman et al., 1983). Computational analysis of the catalytic site revealed that replacement of the C-4 OH group in DANA by an amino group, to give 4-amino-4-deoxy-DANA (28, Figure 6), would be beneficial (von Itzstein et al., 1993, 1996). This approach further predicted a higher affinity of the 4-deoxy-4-guanidino derivative of DANA, zanamivir (29, Figure 6), for the binding site (von Itzstein et al., 1993). Zanamivir (Relenza) inhibits NA of both influenza A and B strains in the subnanomolar range (Taylor and von Itzstein, 1994; von Itzstein et al., 1994) and received FDA approval in 1999.
There has also been considerable effort expended in the search for noncarbohydrate-based zanamivir analogs, resulting in the development of benzoic acid mimics of zanamivir (Singh et al., 1995; Williams et al., 1995; Chand et al., 1997; Sudbeck et al., 1997). Among benzoic acid derivatives, the most potent inhibitor was 4-(acetylamino)-3-guanidinobenzoic acid (BCX 140) (30, Figure 6), with an IC50 value of 2.5 µM (Singh et al., 1995). BCX 140 was oriented in the active site of the NA in a manner that was not predicted from the reported active site binding of zanamivir with NA (Chand et al., 1997). Unfortunately, BCX 140 did not show in vivo activity when dosed intranasally in a mouse model of influenza. Although it is not clear why BCX 140 was not active in vivo, it is possible that it may be removed rapidly from lung by metabolism and/or absorption (Chand et al., 1997).
Zanamivir must be administered intranasally or by inhalation because of low oral bioavailability due to the positive charge of the guanidino group. The carbocyclic zanamivir mimetic, oseltamivir carboxylate (GS 4071) (31, Figure 6), inhibits influenza NA with equal potency (IC50 = 1 nM) to zanamivir (Kim et al., 1997) and was also approved by the FDA in 1999. Oseltamivir carboxylate is more lipophilic than zanamivir and lacks a guanidino group but still suffers from low oral bioavailability due to a negatively charged carboxylate moiety (Li et al., 1998). Hence oseltamivir (GS 4104) (Tamiflu) (32, Figure 6), the ethyl ester of 31, was developed to improve oral bioavailability (Li et al., 1998). Oseltamivir is a prodrug metabolized by liver esterases to its active form 31 and was approved by the FDA in October 1999. Furthermore, the FDA has recently expanded the labeling for this compound to include the prevention of influenza. More recently, the tetrasubstituted cyclopentane derivative, peramivir (BCX 1812, RWJ 270201) (33, Figure 6), has been synthesized and found to be a potent inhibitor of influenza A and B NA, with IC50 values of <1 and <10 nM, respectively (Chand et al., 2001). These IC50 values are comparable or superior to those of zanamivir. BCX 1812 is orally active and is in Phase III clinical trials in North America and Europe. Thus protein crystals enable scientists to determine the three-dimensional structure of the enzyme and to build drugs designed to fit its active site. The success of the structure-based drug design through computational analysis provides encouragement for future efforts targeted at other diseases.
Molecular therapy for human genetic disorders
Ceramide glucosyltransferase (glucosylceramide synthase) inhibitors: substrate deprivation therapy for lysosomal storage diseases
Glycosphingolipid (GSL) storage diseases are relatively rare hereditary disorders that are severe in nature and frequently fatal (Neufeld, 1991). Possible strategies for the treatment of these lysosomal storage diseases include enzyme replacement therapy, gene therapy, substrate deprivation, and bone marrow transplantation. The only successful treatment for such diseases to date is the enzyme replacement therapy for patients with type 1 Gaucher disease (Barton et al., 1990). This disease results from mutations in the glucocerebrosidase gene, which leads to the storage of glucosylceramide (GlcCer). Genzyme introduced Ceredase in 1991 and its recombinant successor Cerezyme in 1994. In 2001, Fabrazyme was launched in the European Union for the treatment of Fabry diseases and has just received FDA approval in April 2003. Fabry disease is a rare hereditary disorder caused by the deficiency of the essential enzyme α-galactosidase A (α-Gal A), resulting in renal failure along with premature myocardial infarction and strokes (Brady et al., 1967; Desnick et al., 1995). The greatest problem in enzyme replacement is the cost, which can prevent many patients from obtaining treatment.
There is a new approach that may be generally applicable to the GSL storage disorders. As long as the biosynthesis of substrates continues, the pathological accumulation of undegraded substrates in the lysosomes proceeds. If substrate influx into the lysosomes could be reduced by inhibition of GSL biosynthesis, it should be possible to positively influence the severity as well as the onset of these diseases with the aid of inhibitors. Tifft and Proia (2000) have reviewed significant progress in the use of synthesis inhibitors as therapy for the GSL lysosomal storage diseases. There are two classes of GSL synthesis inhibitors in possible therapeutic agents for the treatment.
NB-DNJ (21, Figure 5) and N-butyl-1-deoxygalactonojirimycin (NB-DGJ) (34, Figure 7) were discovered to be specific inhibitors of glucosyltransferase-catalyzed biosynthesis of glucosylceramide (Platt et al., 1994a,b). Inhibitors with no effect on the maturation of N-linked oligosaccharides or on lysosomal glucocerebrosidase and with oral availability and low toxicity would be certainly a candidate toward the molecular therapy of glucosylceramide-based GSL storage disorders. Oxford GlycoSciences performed multiple clinical studies using NB-DNJ in type 1 Gaucher disease patients in the United Kingdom and Europe in 1999–2000. The result of this clinical trial indicates that NB-DNJ (OGT 918) is effective in depleting GSLs and leads to an improvement in clinical parameters measuring disease progression (Cox et al., 2000). Oxford GlycoSciences has filed a marketing approval application for Zavesca (formerly known as Vevesca or OGT 918) with the European Agency for the Evaluation of Medicinal Products. In 2002, Zavesca received a positive opinion recommending approval in the European Union for use in patients with mild to moderate type 1 Gaucher disease. On the other hand, the FDA rejected approval for Zavesca, indicated that Oxford GlycoSciences has not provided sufficient support for the safety and efficacy of the developmental compound and cited the necessity for the company to address these issues in further clinical trials. Because the drug's weakness as a monotherapy is pointed out, this drug's best hope seems to demonstrate clinical benefit in combination with Cerezyme. In combination with Zavesca, Cerezyme treatment could become much less expensive and less traumatic because is administered as an intravenous infusion and Zavesca is orally available as a pill.
The prototype of another class of GSL synthesis inhibitors is a ceramide-based inhibitor. D-threo-1-phenyl-2-decanoylamino-3-morpholino-1-propanol (PDMP) (35, Figure 7) inhibits glucosylceramide synthase (GlcT-1) and produce the reversible depletion of cellular GSLs (Abe et al., 1992). PDMP is a more potent inhibitor of GlcT-1 than NB-DNJ but has an undesirable side effect of increasing intracellular ceramide levels, resulting in cytotoxicity (Abe et al., 1996). PDMP has been found to have two sites of action, one for the inhibition of GlcT-1 and one for the inhibition of 1-O-acylceramide synthetase (Abe and Shayman, 1998). Although the inhibition of GlcT-1 resulted in the cellular depletion of GSLs (Abe et al., 1995), that of 1-O-acylceramide synthetase resulted in the elevation of intracellular ceramide levels and inhibition of cell growth in vitro (Abe and Shayman, 1998; Lee et al., 1999). Subsequently, D-threo-1-phenyl-2-palmitoilamino-3-pyrrolidino-1-propanol (P4) (36, Figure 7), 4′-hydroxy-P4 (pOH-P4) (37, Figure 7), and 3′,4′-ethylenedioxy-P4 (EtDO-P4) (38, Figure 7) have been developed using PDMP as the lead compound (Lee et al., 1999). pOH-P4 and EtDO-P4 blocked GlcT-1 activity at concentrations that had little effect on 1-O-acylceramide synthetase activity and resulted in the inhibition of GSL synthesis in cultured cell without the elevation of intracellular ceramide levels and inhibition of cell growth (Lee et al., 1999). In 2001 Genzyme released preclinical data supporting Genz-78132 (EtDO-P4) as the second-generation substrate-deprivation agent. The company has reports that Genz-78132 is 100–5000 times more potent in vitro for substrate inhibition than the first-generation inhibitors NB-DNJ, NB-DGJ, and N-(5-adamantane-1-yl-methoxypentyl)-DNJ (AMP-DNJ) (39, Figure 7), and is at least 20 times more potent in vivo than AMP-DNJ, the most potent compound (Overkleeft et al., 1998). Although there is no supporting clinical data, Genz-78132 may allow the administration of small quantities of a drug to achieve the desired therapeutic effect, minimizing the potential for toxic side effects.
Lysosomal glycosidase inhibitors: chemical chaperone therapy for lysosomal storage diseases
The lumen of the ER provides a highly specialized compartment for the folding and oligomeric assembly of secretory proteins, plasma membrane proteins, and proteins destined for the various organelles of the vacuolar system. Their conformational maturation is a complex process determined not only by the amino acid sequence but also by post- and cotranslational modifications, by the intralumenal milieu, and by a variety of chaperones and folding enzymes (Gething and Sambrook, 1992; Helenius et al., 1992). The ER possesses efficient quality control mechanisms to ensure that transport is limited to properly folded and assembled proteins (Hurtley and Helenius, 1989). Recent experimental data show that some human genetic diseases are due to mutations in proteins that influence their folding and lead to retaining of mutant proteins in the ER and successive degradation (Bychkova and Ptitsyn, 1995; Welch and Brown, 1996).
Genetically inherited diseases are often characterized by specific point mutations or deletions that give rise to proteins that fail to achieve their properly folded state. In some cases, the mutations are relatively minor and result in the protein exhibiting only a partial loss of its normal activity. In the classic type of α1-antitripsin deficiency (the most common genetic cause of enphysema in adults and liver diseases in children) patients with the Z mutation secrete only about 15% of newly synthesized proteins into the extracellular fluid and the remainder of the mutant protein is retained in the ER (Teckman and Perlmutter, 1996). However, this secreted mutant protein retains about 80% of the functional activity of its wild-type counterpart (Bathurst et al., 1984; Ogushi et al., 1987). Although some of the retained mutant protein is eventually degraded, the remainder tends to form aggregates in the ER (Lomas et al., 1992; Yu et al., 1995). Over time the accumulation of the aggregate interferes with the normal functional activities of the hepatocyte.
Another example of a disease involving abnormal protein trafficking is cystic fibrosis (CF). CF is a human genetic disease caused by mutations in CF transmembrane conductance regulator (CFTR), which functions as a chloride channel in the plasma membrane (Sferra and Collins, 1993; Welsh and Smith, 1993). The most common mutation observed in patients with CF (over 70%) is a deletion of Phe-508 (ΔF508) of CFTR. This deletion leads to retention of the mutant in the ER, implicating the defect of its trafficking (Cheng et al., 1990). The defective folding in the ΔF508 CFTR mutant is temperature sensitive. The processing of the ΔF508 CFTR reverts toward that of wild type as the incubation temperature is reduced below 30°C. When the processing defect is corrected, cAMP-regulated Cl− channels appear in the plasma membrane (Denning et al., 1992). Hence, the major problem in α1-antitrypsin and CF appears to be a failure of the newly synthesized mutant proteins to exit from the ER and to move to their proper locale. These observations raise the possibility that a functional compound that can elicit the proper folding and trafficking of the mutant protein might prove to be effective strategy for the treatment of the genetic disorders.
Studies on residual α-Gal A activity of mutant enzymes in many Fabry patients revealed that some had kinetic properties similar to those of normal α-Gal A but were significantly less stable (Remeo et al., 1975). For example, purified mutant α-Gal A with Q279E mutation detected in Fabry patients with the cardiac variant form had the same Km and Vmax as the normal enzyme. However, the mutant enzyme lost most of its catalytic activity after incubation at pH 7.0 at 37°C for 30 min, whereas the normal enzyme was relatively stable under the same conditions (Ishii et al., 1993). Both mutant and normal enzymes were stable at pH 5.0 at 37°C for 30 min. Furthermore the intracellular mutant protein formed aggregates in the ER and was quickly degraded (Ishii et al., 1996). These results suggest that the enzyme deficiency in this mutant is caused primarily by abortive exit from the ER.
Recently, 1-deoxygalactonojirimycin (DGJ) (40, Figure 8), a potent competitive inhibitor of α-Gal A, effectively enhanced the mutant enzyme activity in lymphoblasts established from Fabry patients with R301Q or Q279E mutation (Fan et al., 1999). Furthermore, oral administration of DGJ to transgenic mice expressing a human mutant α-Gal A (R301Q) substantially elevated the enzyme activity in the major organs. DGJ seems to assist the successful transport of the mutant enzyme from the ER to the Golgi apparatus and its correct targeting to the lysosome. It appears that by occupying the catalytic site of the mutant enzyme, DGJ stabilizes its conformation, thus allowing resumption of the glycoprotein processing leading to maturation of the enzyme. Thus a functional compound that can elicit the correct folding of a mutant protein may serve as a specific chemical chaperone for the mutant protein to promote the successful escape from the ER quality control mechanisms (Fan et al., 1999). To establish the concept of using competitive inhibitors as specific chemical chaperones, a number of naturally occurring and chemically synthesized DGJ derivatives were tested for intracellular enhancement of a mutant α-Gal A activity in Fabry lymphoblasts. DGJ, α-homogalactonojirimycin (HGJ) (41, Figure 8) (Martin et al., 1995), α-homoallonojirimycin (HAJ) (42, Figure 8) (Asano et al., 1997; Martin et al., 1999), and β-1-C-butyl-DGJ (CB-DGJ) (43, Figure 8) (Ikeda et al., 2000) were inhibitors of α-Gal A with IC50 values of 0.04, 0.21, 4.3, and 16 µM, respectively, and addition of DGJ, HGJ, HAJ, and CB-DGJ at 100 µM to culture medium of Fabry lymphoblasts increased the intracellular α-Gal A activity by 14-, 5.2-, 2.4-, and 2.3-fold, respectively (Asano et al., 2000). These results suggest that more potent inhibitors are more effective chemical chaperone for the mutant enzyme and Fabry disease.
The chemical chaperone therapy established in Fabry disease is also applicable to the other type of lysosomal storage diseases. DGJ (40, Figure 8) and NB-DGJ (34, Figure 7) are also moderate inhibitors of lysosomal β-galactosidase, with IC50 values of 25 µM (Tominaga et al., 2001). Both compounds restored mutant enzyme activities in cell lines of β-galactosidosis obtained by introducing mutant human β-galactosidase cDNAs into knockout mouse fibloblasts (Tominaga et al., 2001). However, a survey of more effective compounds is needed for clinical application of this method to β-galactosidosis because relatively high concentrations (0.5–1 mM) were required for mutant enzyme activation.
Gaucher disease is a lysosomal storage disorder caused by deficient lysosomal β-glucocerebrosidase activity. Quite recently, it has been found that NN-DNJ (25, Figure 5) is a potent inhibitor of this enzyme, with an IC50 value of 1 µM, and that the addition of subinhibitory concentration (10 µM) of NN-DNJ to a fibroblast culture medium leads to a twofold increase in activity of mutant (N370S) enzyme (Sawkar et al., 2002). Because clinical data indicate that a small increase in enzyme activity may be effective in treating disease, this twofold increase in activity caused by NN-DNJ chemical chaperoning may be clinically useful. A number of different chemical chaperones that enhance stability of mutant proteins and enable their trafficking to the proper locale would be excellent candidates for a new molecular therapy of human genetic disorders.
The alteration of glycosidase activity by inhibitors in vivo is of great interest because of the involvement of glycosidases in a wide range of anabolic and catabolic processes, such as digestion, lysosomal catabolism of glycoconjugates, biosynthesis, ER quality control, and ER-associated degradation of glycoproteins. As reviewed, glycosidase inhibitors could have many kinds of beneficial effects as agrochemicals and therapeutic agents, such as as antifungal agents, insecticides, antidiabetics and antiobesities, antivirals, and therapeutic agents for some genetic disorders. The success of NA inhibitors as anti-influenza drugs was derived from rational designs based on detailed knowledge of the highly conserved active site of influenza virus NA and the interaction with the natural substrate, sialic acid, through computational analysis. In addition, in vitro and in vivo studies have established the NA inhibitors as effective and safe antiviral drugs. This method will surely become a primary approach for developing other therapeutic agents. There is no doubt that many more glycosidase inhibitors of practical use remain to be discovered.
Faculty of Pharmaceutical Sciences, Hokuriku University, Ho-3 Kanagawa-machi, Kanazawa 920-1181, Japan