Portal de Periódicos da CAPES

Extensive Glycosphingolipid Depletion in the Liver and Lymphoid Organs of Mice Treated with N-Butyldeoxynojirimycin

1997; Elsevier BV; Volume: 272; Issue: 31 Linguagem: Inglês

10.1074/jbc.272.31.19365

1083-351X

Frances M. Platt, Gabriele Reinkensmeier, Raymond A. Dwek, Terry D. Butters,

Pharmacological Effects of Natural Compounds

The imino sugarN-butyldeoxynojirimycin is an inhibitor of the ceramide-specific glucosyltransferase that catalyzes the first step in glycosphingolipid biosynthesis. It results in extensive glycosphingolipid depletion in cells treated in vitro, without causing toxicity. However, we currently do not know the degree to which glycosphingolipids can be depleted in vivo in a mammalian species. We have therefore administeredN-butyldeoxynojirimycin long term to young mice and have found that glycosphingolipid levels are reduced (50–70%) in all tissues examined, without resulting in any overt pathology. When the lymphoid tissues from these mice were examined, they were found to be 50% acellular relative to non-lymphoid tissues. These data implicate a role for glycosphingolipids in the biology of the immune system or indicate an additional as yet unknown activity ofN-butyldeoxynojirimycin. Extensive glycosphingolipid depletion resulting from N-butyldeoxynojirimycin administration is therefore well tolerated in adult mice, and this compound may be in an invaluable tool for probing glycosphingolipid functions in vivo. In addition, this drug may be effective in clinical situations where glycosphingolipid depletion would be desirable, such as the in the treatment of the human glycosphingolipidoses. The imino sugarN-butyldeoxynojirimycin is an inhibitor of the ceramide-specific glucosyltransferase that catalyzes the first step in glycosphingolipid biosynthesis. It results in extensive glycosphingolipid depletion in cells treated in vitro, without causing toxicity. However, we currently do not know the degree to which glycosphingolipids can be depleted in vivo in a mammalian species. We have therefore administeredN-butyldeoxynojirimycin long term to young mice and have found that glycosphingolipid levels are reduced (50–70%) in all tissues examined, without resulting in any overt pathology. When the lymphoid tissues from these mice were examined, they were found to be 50% acellular relative to non-lymphoid tissues. These data implicate a role for glycosphingolipids in the biology of the immune system or indicate an additional as yet unknown activity ofN-butyldeoxynojirimycin. Extensive glycosphingolipid depletion resulting from N-butyldeoxynojirimycin administration is therefore well tolerated in adult mice, and this compound may be in an invaluable tool for probing glycosphingolipid functions in vivo. In addition, this drug may be effective in clinical situations where glycosphingolipid depletion would be desirable, such as the in the treatment of the human glycosphingolipidoses. Glycosphingolipids (GSLs) 1The abbreviations used are: GSL(s), glycosphingolipid(s); PDMP,Dl-threo-1-phenyl-2-decanoylamino-3-morpholino-1-propanol);NB-DNJ, N-butyldeoxynojirimycin;NB-DGJ, N-butyldeoxygalactonojirimycin; PBS, phosphate-buffered saline; PAGE, polyacrylamide gel electrophoresis; mAb, monoclonal antibody. are ubiquitous components of the plasma membranes of eukaryotic cells where they are believed to mediate a number of biological functions (1Hakomori S J. Biol. Chem. 1990; 265: 18713-18716Abstract Full Text PDF PubMed Google Scholar). Their biosynthesis and catabolism have been extensively studied (2Sandhoff K. van Echten G. Adv. Lipid Res. 1993; 26: 119-142PubMed Google Scholar, 3van Echten G. Sandhoff K. J. Biol. Chem. 1993; 268: 5341-5344Abstract Full Text PDF PubMed Google Scholar, 4Sandhoff K. Kolter T. Trends Cell Biol. 1996; 6: 98-103Abstract Full Text PDF PubMed Scopus (148) Google Scholar), as have the disease states that result from their incomplete catabolism within lysosomes (5Neufeld E.F. Annu. Rev. Biochem. 1991; 60: 257-280Crossref PubMed Scopus (487) Google Scholar). They are exploited as receptors by a number of micro-organisms (6Karlsson K.-A. Annu. Rev. Biochem. 1989; 58: 309-350Crossref PubMed Scopus (622) Google Scholar); however, their normal physiological functions remain largely obscure. GSL biosynthesis involves the co-ordinated action of multiple gene products (2Sandhoff K. van Echten G. Adv. Lipid Res. 1993; 26: 119-142PubMed Google Scholar). A key enzyme in this pathway is the ceramide-specific glucosyltransferase, which catalyzes the first step in the GSL biosynthetic pathway, the transfer of glucose to ceramide to form glucosylceramide (GlcCer). The sequential action of glycosyltransferases in the Golgi apparatus converts glucosylceramide into other neutral GSLs and gangliosides (2Sandhoff K. van Echten G. Adv. Lipid Res. 1993; 26: 119-142PubMed Google Scholar). The ceramide-specific glucosyltransferase has been cloned recently and is a unique gene product without substantial homology with other known glycosyltransferase genes (7Ishikawa S. Sakiyama H. Suzuki G. Kazuya I.-P. Hidari J. Hirabayashi Y. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 4638-4643Crossref PubMed Scopus (223) Google Scholar). Mutant B16 melanoma cells lacking this enzyme are not compromised in their growth or morphology, indicating that GSLs are not needed for membrane integrity and do not serve a basic housekeeping function, at least at the level of the single cell (7Ishikawa S. Sakiyama H. Suzuki G. Kazuya I.-P. Hidari J. Hirabayashi Y. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 4638-4643Crossref PubMed Scopus (223) Google Scholar, 8Ichikawa S. Nakajo N. Sakiyama H. Hirabayashi Y. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 2703-2707Crossref PubMed Scopus (142) Google Scholar). One approach to investigate the role of GSLs in intact organisms is to inactivate the glucosyltransferase either by gene disruption techniques or by using a specific inhibitor. It is not currently known the degree to which GSLs can be depleted in vivo in mammalian species without resulting in pathology (9Platt F.M. Butters T.D. Trends Glycosci. Glycotechnol. 1995; 7: 495-511Crossref Scopus (36) Google Scholar). The gene knockout approach would determine whether or not total inhibition of GSL synthesis is compatible with embryonic development. There is circumstantial evidence, such as the lack of human disease states that result from defects in GSL biosynthesis, which suggests that GSL expression may play a critical role in early mammalian development (4Sandhoff K. Kolter T. Trends Cell Biol. 1996; 6: 98-103Abstract Full Text PDF PubMed Scopus (148) Google Scholar). The complete lack of GSL biosynthesis during embryogenesis could therefore be potentially lethal. However, in studies of teleost development using Medaka fish in in vitro culture, extensive GSL inhibition resulting from treatment with the ceramide analogue PDMP did not disrupt development (10Fenderson B.A. Ostrander G.K. Hausken Z. Radin N.S. Hakamori S.-I. Exp. Cell Res. 1992; 198: 362-366Crossref PubMed Scopus (36) Google Scholar). The roles of GSLs during vertebrate development and their functions in the biology of the mature organism therefore remain to be elucidated. The alternative experimental approach which can be taken to explore GSL functions is to use specific inhibitors of enzymes required for the biosynthesis of GSLs. One advantage of this approach is that the consequences of partial GSL depletion can be investigated. An attractive target for enzyme inhibition is the ceramide-specific glucosyltransferase, which catalyzes the first step in the GSL biosynthetic pathway. We have recently identified two inhibitors of GSL biosynthesis that block the action of the ceramide-specific glucosyltransferase (11Platt F.M. Neises G.R. Dwek R.A. Butters T.D. J. Biol. Chem. 1994; 269: 8362-8365Abstract Full Text PDF PubMed Google Scholar, 12Platt F.M. Neises G.R. Karlsson G.B. Dwek R.A. Butters T.D. J. Biol. Chem. 1994; 269: 27108-27114Abstract Full Text PDF PubMed Google Scholar). These inhibitors are the imino sugarsNbutyldeoxynojirimycin (NB-DNJ) andN-butyldeoxygalactonojirimycin (NB-DGJ), a glucose and galactose analogue, respectively. In vitro these analogues result in GSL depletion in a wide range of human and murine cell lines (11Platt F.M. Neises G.R. Dwek R.A. Butters T.D. J. Biol. Chem. 1994; 269: 8362-8365Abstract Full Text PDF PubMed Google Scholar, 12Platt F.M. Neises G.R. Karlsson G.B. Dwek R.A. Butters T.D. J. Biol. Chem. 1994; 269: 27108-27114Abstract Full Text PDF PubMed Google Scholar). However, the extent to which they can reduce GSL levels in vivo is currently unknown, as is the extent to which GSL depletion can be tolerated in an intact adult mammal. The degree to which GSLs depletion is tolerated in vivo is also an important factor when considering GSL biosynthetic inhibitors as potential therapeutic agents for the treatment of the GSL lysosomal storage diseases, where in vivo GSL depletion is the primary clinical objective (9Platt F.M. Butters T.D. Trends Glycosci. Glycotechnol. 1995; 7: 495-511Crossref Scopus (36) Google Scholar, 11Platt F.M. Neises G.R. Dwek R.A. Butters T.D. J. Biol. Chem. 1994; 269: 8362-8365Abstract Full Text PDF PubMed Google Scholar, 12Platt F.M. Neises G.R. Karlsson G.B. Dwek R.A. Butters T.D. J. Biol. Chem. 1994; 269: 27108-27114Abstract Full Text PDF PubMed Google Scholar). In this study we have administered NB-DNJ to young mice using a range of dosing regimes and monitored GSL biosynthesis.NB-DNJ was used, rather than the galactose analogueNB-DGJ, due to the fact that this compound is available in the multigram quantities required for this study. We have found significant GSL depletion in multiple tissues of mice treated withNB-DNJ without any overt toxicity, indicating that partial inhibition of GSL biosynthesis is well tolerated in vivo. The implications these data have for GSL function and the therapeutic administration of GSL inhibitors for treating GSL lysosomal storage disorders are discussed. NB-DNJ was a gift from Searle/Monsanto. Age- and sex-matched (6-week-old female) C57BL/6 mice were fed on a diet of powdered mouse chow (expanded Rat and Mouse Chow 1, ground, SDS Ltd., Witham, Essex, UK) containing NB-DNJ. The diet and compound (both as dry solids) were mixed thoroughly before use, stored at room temperature, and used within 7 days of mixing. Water was available to the micead libitum. The mice were housed under standard nonsterile conditions. The mice were maintained on 600 mg/kg/day per mouse for 50 days, and two animals were sacrificed from the control group (untreated, n = 10) and two from the experimental group (diet + NB-DNJ, n = 10). The remaining animals in the experimental group were then maintained on a 1200 mg/kg/day diet for a further 50 days when another two animals from each group were sacrificed. The remaining animals in the treated group were fed on a 1800 mg/kg/day diet for 20 days at which point all remaining animals from each group were sacrificed. In an independent study, animals were maintained on a 2400 mg/kg/day diet for 14 days. Blood samples were collected by cardiac puncture, the serum expressed, and stored at −20 °C prior to assay. Steady state serum levels ofNB-DNJ were determined by isolating compound from 50 μl of blood using AG-50W-X12 (H+ form) resin (200 μl of packed beads). Compound and an internal standard (N-hexyl-DNJ orN-propyl-DNJ) were eluted with a 4 m ammonia solution. Using radiolabeled compounds, greater than 99% of the compound was recovered from the resin. Detection of compound was achieved either by matrix-assisted laser desorption ionization time of flight mass spectrometry (MALDITOF) or by cation-exchange chromatography. For mass spectrometry (Finnigan MAT), an internal standard was included with normal serum samples to construct a standard curve with NB-DNJ concentrations ranging from 5 to 40 μg/ml of serum. Cation-exchange chromatography (Dionex BioLC) was performed by an isocratic elution of a CS-10 column with sodium sulfate containing 5% (v/v) acetonitrile. Quantitation of NB-DNJ was obtained by peak area estimation in comparison with the internal standard and applying experimentally derived response factors. Tissue samples were homogenized in water, lyophilized, 5 mg of dry weight extracted in chloroform:methanol (2:1, v/v) overnight at 4 °C, dried under nitrogen, resuspended in 20 μl of chloroform:methanol (2:1, v/v), and separated by TLC (chloroform:methanol:H20, 65:25:4). The TLC plate was air-dried, sprayed with α-naphthol (1% (w/v), in methanol) followed by 50% (v/v) sulfuric acid and heat-treated (80 °C for 10 min). The assay has been described previously (11Platt F.M. Neises G.R. Dwek R.A. Butters T.D. J. Biol. Chem. 1994; 269: 8362-8365Abstract Full Text PDF PubMed Google Scholar). Briefly, splenocytes from treated and untreated mice were incubated with fluorescein isothiocyanate-conjugated cholera toxin B chain, for 30 min on ice in the dark. Quantitative flow cytometric analysis has been described elsewhere (11Platt F.M. Neises G.R. Dwek R.A. Butters T.D. J. Biol. Chem. 1994; 269: 8362-8365Abstract Full Text PDF PubMed Google Scholar) and was performed using a FACScan Cytometer (Becton Dickinson, Sunnyvale CA). Data on viable cells (propidium iodide dye exclusion) were collected on a 4-decade log10 scale of increasing fluorescence intensity. Splenocytes (3 × 107 mononuclear cells) from twoNB-DNJ-treated (2400 mg/kg/day) and two untreated mice were washed twice in PBS, resuspended in 1 ml of PBS containing 1 mm sodium metaperiodate, and left on ice for 5 min. The oxidation was stopped by the addition of 0.2 ml of PBS containing 0.1m glycerol. The cells were washed twice with PBS, resuspended in 0.5 ml of PBS, and tritiated sodium borohydride added (0.8 mCi; specific activity, 12 Ci/mmol) and incubated at room temperature for 30 min. The cells were washed three times with PBS and radiolabeled GSLs extracted from the cell pellet (2 × 1 ml of chloroform:methanol 2:1 (v/v) overnight at 4 °C and 2 h at room temperature). An equal volume of each extract (equivalent to 0.75 × 107 cells) was taken for analysis by high performance thin layer chromatography using a solvent of chloroform:methanol:0.5% CaCl2 (55:49:9, v/v/v). Radiolabeled GSLs were detected by fluorography and their migration compared with authentic GSL standards. The cell pellets remaining after the lipid extraction were resuspended in 200 μl of distilled water to which SDS was added (1% (w/v) final concentration) and dithiothreitol (0.5 mm, final concentration). The samples were heated to 95 °C for 5 min, spun at 13,000 × g for 5 min, and aliquots taken for radioactivity determination by scintillation counting. Protein content was determined using the BCA method (Pierce, Chester, UK) and 20-μg equivalents of protein analyzed by SDS-PAGE and fluorography before and after treatment with peptide:N-glycosidase F (New England Biolabs (Hitchin, UK) Ltd., 50,000 units/ml final concentration for 4 h at 37 °C). The following rat mAbs were used for this study and have been described previously (13Platt F.M. Cebra-Thomas J.A. Baum C.M. Davie J.M. McKearn J.P. Cell. Immunol. 1992; 143: 449-466Crossref PubMed Scopus (6) Google Scholar): 187.1, anti-mouse κ light chain; GK1.5, anti-CD4; anti-Ly-2, CD8 (Becton-Dickinson, Mountain View, CA); and 7.3, anti-Lyt-1. A goat anti-rat fluorescein isothiocyanate conjugate (Ortho) was used to detect anti-CD4, anti-κ and anti-Lyt-1 binding. The staining procedure and flow cytometry analysis have been described elsewhere (13Platt F.M. Cebra-Thomas J.A. Baum C.M. Davie J.M. McKearn J.P. Cell. Immunol. 1992; 143: 449-466Crossref PubMed Scopus (6) Google Scholar). Data on viable cells (propidium iodide dye exclusion) were collected on a 4-decade log10 scale of increasing fluorescence intensity. The percentage of cells staining with the mAbs was defined by the positioning of cursors on the basis of the unstained control (autofluorescent background) for directly conjugated mAbs or on the basis of the secondary only control with the nonconjugated mAbs. To monitor the overall well being of the mice treated with NB-DNJ, the mice were monitored daily, body weights recorded, and the effects ofNB-DNJ on growth rates determined (Fig.1). The doses of NB-DNJ given over the course of the study are indicated with arrows. Over the treatment period, and with the dosing regime given, the mice on theNB-DNJ containing diet grew slower than the untreated sex- and age-matched controls. The final body weights attained at the end of the study (118 days, 2400 mg/kg/day) were 15% lower in theNB-DNJ-treated group. When the mice were switched to a diet rich in glucose (65%, w/w), but lacking complex carbohydrates (AIN 76, SDS), the body weights remained 15% lower than the untreated controls. The mice had no diarrhea at any stage in the study, in contrast to mice given comparable or lower concentrations of NB-DNJ by oral gavage every 8 h. 2D. Semler, personal communication. The general appearance, activity levels, and condition of theNB-DNJ-treated mice were comparable with the untreated controls (data not shown), indicating that the animals toleratedNB-DNJ in their diet and were not overtly compromised by its administration. Serum levels of NB-DNJ were determined at all the dosing levels and were found to range in a dose-dependent fashion from 18 μm at 600 mg/kg/day to 57 μm at 2400 mg/kg/day (TableI).Table ISteady state serum levels of NB-DNJDoseSerum levelmg/kg/dayμm60018.2120030.6180043240056.8Serum levels of NB-DNJ were determined at each dose of compound on serum samples from at least three animals, and each serum sample was analyzed twice by high performance liquid chromatography or four times by mass spectrometry. The values are means and the variance was within ±10% for all samples evaluated. Open table in a new tab Serum levels of NB-DNJ were determined at each dose of compound on serum samples from at least three animals, and each serum sample was analyzed twice by high performance liquid chromatography or four times by mass spectrometry. The values are means and the variance was within ±10% for all samples evaluated. As a measure of cellular GSL levels, cell surface ganglioside levels were determined on splenocytes from treated and untreated mice, using fluoresceinated cholera toxin B chain as a probe for GM1 ganglioside (11Platt F.M. Neises G.R. Dwek R.A. Butters T.D. J. Biol. Chem. 1994; 269: 8362-8365Abstract Full Text PDF PubMed Google Scholar). The cells were assayed by flow cytometry and binding sites per cell quantitated using fluorescent microbead standards. The percent reduction in cholera toxin binding sites per cell was measured at each drug dose (Fig.2). At 600 mg/kg/day the splenocytes had 10% fewer cholera toxin binding sites per cell than the untreated aged matched controls. This increased to an approximate 35% reduction at doses in the range of 1200–1800 mg/kg/day and to 50% at the highest dose tested, 2400 mg/kg/day. These data indicate that GSL depletion occurs in vivo in response to NB-DNJ and is dose-dependent. Cell surface glycoconjugates were examined in more detail by incorporating a radiolabel into cell surface sialic acid residues. Mild periodate treatment of glycoconjugates results in the selective oxidation of sialic acid residues. The aldehydes formed are reduced with NaB3H4 to stable, tritium-labeled 5-acetamido-3,5-dideoxy-l-arabino-2-heptulosonic acid. Using the conditions described, this method specifically introduces a label into cell surface sialic acids, including those carried by gangliosides (14Gahmberg C.G. Andersson L.C. J. Biol. Chem. 1977; 252: 5888-5894Abstract Full Text PDF PubMed Google Scholar). The autoradiograph of the high performance TLC separation of solvent extracted GSLs from splenocytes from two untreated mice and two treated with NB-DNJ (Fig.3 A) showed a quantitative decrease (60–70% reduction by densitometry) in labeled species compared with those from untreated animals. There did not appear to be a qualitative difference in the gangliosides extracted from labeled cells from the two groups of mice, nor were there any changes in the relative abundance of individual gangliosides in individual animals. These data strongly suggest that the expression of cell surface gangliosides is decreased in a nonselective manner as a result of treatment with NB-DNJ. When the protein fraction from these samples was analyzed by SDS-PAGE and visualized by fluorography, no discernible differences could be detected in the profile, either before or after peptide:N-glycosidase F treatment, of sialic acid containing glycoproteins derived from untreated andNB-DNJ-treated animals (Fig. 3 B). Using the same enzyme incubation conditions, mouse IgM was completely de-N-glycosylated (results not shown). This indicated that α-glucosidase inhibition was not occurring at the serum levels achieved; hence, there was no apparent reduction in glycoproteins carrying sialylated N-glycans. To determine the degree of GSL depletion in a non-lymphoid tissue, liver was chloroform:methanol-extracted and total lipids analyzed by thin layer chromatography. A reduction in GSL levels in the mice treated withNB-DNJ was observed, including GM2 and the trihexoside species (Fig. 4). Sphingomyelin levels increased in NB-DNJ-treated mice, presumably due to the increased levels of ceramide available to go down this biosynthetic pathway, due to inhibition of GSL biosynthesis. Removal of the drug from the diet for 2 weeks resulted in relatively normal liver GSL levels with sphingomyelin being present in quantities comparable with the untreated controls and the trihexoside species being detectable. GM2 levels were slower to recover, and this ganglioside was barely detectable after the 2-week period, following removal of NB-DNJ from the diet. GSL depletion was therefore occurring in livers of NB-DNJ-treated mice and was reversible following drug withdrawal. The fact that GM2recovery is slow may indicate possible long term effect of the drug. Spleens and thymuses removed from NB-DNJ mice were observed to be consistently smaller (on a wet weight basis) than those from untreated age matched controls. For example, spleens from 600–1800 mg/kg/day were 30% lighter than the untreated controls (±6%), and at 2400 mg/kg/day they were on average 55% lighter (±6%) than spleens taken from age-matched control mice. This was in contrast to organs such as liver and brain, which were comparable by wet weight, irrespective ofNB-DNJ treatment (data not shown). To investigate the nature of the reduction in lymphoid tissue size in response to NB-DNJ treatment, the 2400 mg/kg/day animals were investigated further. A cohort of animals were fed on aNB-DNJ-containing diet and changes in spleen and thymus size monitored over time (Fig. 5, Aand B). Both tissues reduced in size in response toNB-DNJ treatment, and maximal reduction in organ size occurred between 7 and 11 days following drug administration. The lymphoid tissues remained smaller as long as the animals were treated with NB-DNJ, but returned to normal size after 2 weeks following withdrawal of the drug from the diet (for example, spleen; Fig. 6).Figure 6Spleen weights following withdrawal ofNB-DNJ from the diet. Mice were treated withNB-DNJ (2400 mg/kg/day) and their spleen weights determined after 2 weeks. The drug was withdrawn from an age-matchedNB-DNJ-treated group of animals for a further 2 weeks and their spleen weights measured. The data are expressed as the means ± S.D. for an untreated age-matched control group of mice (−),NB-DNJ-treated mice (+), and the drug withdrawal group (+/−).View Large Image Figure ViewerDownload Hi-res image Download (PPT) The cellular compositions of spleen and thymus after 2 weeks on 2400 mg/kg/day NB-DNJ were examined by flow cytometry (TableII). The spleen contained a higher percentage of T cells relative to the frequencies observed in untreated animals (44% compared with 30%), and the B cell frequencies were reduced from 51% in untreated controls to 45% inNB-DNJ-treated mice. When drug was removed from the diet, within 2 weeks the relative proportions of T and B cells within the spleen returned to normal levels (Table II). The increased proportion of T cells within the spleen was due to an increased frequency of both CD4+ and CD8+ cells, with the greatest increase being in the CD8+ cells (9% positive in the untreated spleen to 16% positive in the NB-DNJ-treated spleen), while CD4+ cells accounted for 20% of cells in the untreated spleen and 31% in NB-DNJ-treated spleen. Histological examination of spleens from untreated and NB-DNJ-treated mice showed no gross morphological differences with normal follicular architecture maintained in spleens from NB-DNJ-treated mice (data not shown).Table IIEffect of NB-DNJ on the cellular compositions of mouse spleenNB-DNJPercent cells positiveLyt-1CD4CD8Kappa−30.75 ± 2.720.5 ± 2.79 ± 1.250.75 ± 4.3+44 ± 3.031 ± 2.816 ± 1.044.5 ± 0.5±33.5 ± 2.521.5 ± 3.59.5 ± 0.549 ± 8.0Immunophenotyping of splenocytes by flow cytometry from untreated andNB-DNJ-treated mice. Subset analysis was carried out according to "Materials and Methods" (n = 3 mice per group). Untreated mice (−), 2400 mg/kg/day for 2 weeks (+), 2400 mg/kg/day for 2 weeks then drug withdrawn from diet for 2 weeks (±). Open table in a new tab Immunophenotyping of splenocytes by flow cytometry from untreated andNB-DNJ-treated mice. Subset analysis was carried out according to "Materials and Methods" (n = 3 mice per group). Untreated mice (−), 2400 mg/kg/day for 2 weeks (+), 2400 mg/kg/day for 2 weeks then drug withdrawn from diet for 2 weeks (±). The thymuses from animals treated with NB-DNJ were also examined. Two-color flow cytometry after 7 days of NB-DNJ treatment revealed that, in contrast to the untreated thymus, which has a typical composition of approximately 5% CD4−/CD8−, 80% CD4+/CD8+, 10% CD4+/CD8−, and 5% CD4−/CD8+ cells, the NB-DNJ-treated mice had an increased frequency of single positive cells. For example the thymuses from NB-DNJ-treated mice contained approximately 7% CD4−/CD8−, 65% CD4+/CD8+, 18% CD4+/CD8−, and 10% CD4−/CD8+ cells (Fig.7). There was therefore an increase in the proportion of single CD4+ or CD8+ cells at the expense of CD4+/CD8+ double positive cells. As GSLs are not essential components of mammalian cells at the single cell level (8Ichikawa S. Nakajo N. Sakiyama H. Hirabayashi Y. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 2703-2707Crossref PubMed Scopus (142) Google Scholar), their functions must reside in more complex aspects of mammalian biology. If these functions are to be elucidated, methods and reagents are required that will permit GSL manipulationin vivo as well as in vitro. Although we can inhibit GSL biosynthesis in vitro using a number of pharmacological agents, compounds are required that are well tolerated long term in vivo. There are currently two main classes of GSL biosynthesis inhibitor that have been characterized to date which specifically target the ceramide-specific glucosyltransferase. The first class of inhibitors described were the ceramide analogues of which the prototypic compound is PDMP (15Vunnam R.R. Radin N.S. Chem. Phys. Lipids. 1980; 26: 265-278Crossref PubMed Scopus (119) Google Scholar). By virtue of ceramide mimicry these compounds prevent the glucosylation of ceramide by acting as reversible mixed-type inhibitors (16Inokuchi J.-I. Radin N.S. J. Lipid Res. 1987; 28: 278-565Abstract Full Text PDF Google Scholar). However, the highly hydrophobic nature of these compounds makes their long term administration to animals problematic (9Platt F.M. Butters T.D. Trends Glycosci. Glycotechnol. 1995; 7: 495-511Crossref Scopus (36) Google Scholar). The second group of GSL biosynthesis inhibitors are the N-alkylated imino sugars, such as the glucose analogue NB-DNJ. The nonalkylated parental compound (DNJ) does not inhibit the ceramide-specific glucosyltransferase, and a minimal alkyl chain length is required to render these compounds inhibitory, with butyl and hexyl chains being optimal (11Platt F.M. Neises G.R. Dwek R.A. Butters T.D. J. Biol. Chem. 1994; 269: 8362-8365Abstract Full Text PDF PubMed Google Scholar). The alkyl chain may render these compounds short chain ceramide mimics or serve to target these compounds to the membrane environment where the glucosyltransferase resides (9Platt F.M. Butters T.D. Trends Glycosci. Glycotechnol. 1995; 7: 495-511Crossref Scopus (36) Google Scholar). The alkylated imino sugars have a major advantage over PDMP and related compounds in that they exhibit minimal cytotoxicity in vitro, even at doses in excess of 2 mm (17Fischer P.B. Colin M. Karlsson G.B. James W. Butters T.D. Davis S.J. Gordon S. Dwek R.A. Platt F.M. J. Virol. 1995; 69: 5791-5797Crossref PubMed Google Scholar). They therefore have the potential to be well tolerated in vivo. In this study we have investigated whether or not we can achieve GSL depletion in vivo in healthy mice treated orally withNB-DNJ and whether GSL depletion is in any way deleterious to the animal. When mice were treated with various doses of NB-DNJ, serum levels in the range of 18–57 μm were achieved. Similar serum levels (steady state trough level of approximately 20 μm) were achieved in humans during the evaluation of this compound as an antiviral agent when patients were treated with 43 mg/kg/day (18Fischl M.A. Resnick L. Coombs R. Kremer A.B. Pottage J.C. Fass R.J. Fife K.H. Powderly W.G. Collier A.C. Aspinall R.L. Smith S.L. Kowalski K.G. Wallemark C.-B. J. Acquired Immune Defic. Syndr. 1994; 7: 139-147PubMed Google Scholar). The pharmacokinetics of NB-DNJ are 2 orders of magnitude poorer in mouse relative to human necessitating high dosing regimes in the mouse to achieve serum levels in the range required to achieve GSL biosynthesis inhibition (5–50 μm) (9Platt F.M.