Background
In 2002 and 2003 miglustat was approved as an orphan drug by the Federal Drug Administration (USA) and the European Medicines Evaluation Agency, respectively, for the treatment of type 1 Gaucher disease, a lysosomal glycosphingolipid storage disorder. Miglustat can partially inhibit the biosynthesis of glucosylceramide, by inhibiting ceramide glucosyltransferase, and therefore can be used to reduce the glucosphingolipid levels in cells [
1]. Miglustat is tolerated in humans and at high doses in mouse disease models [
2,
3]. Apart from the ceramide-specific glucosyltransferase miglustat inhibits
in vitro a number of other enzymes: α-glucosidases I and II, sucrase, maltase, lysosomal glucocerebrosidase and the non-lysosomal glucosylceramidase [
4‐
7]. The pharmacokinetics of miglustat are relatively uncomplicated with no or very limited metabolism, no protein binding and excretion predominantly into the urine [
8]. Renal clearance in mouse and rat was greater than creatinine clearance, consistent with active renal secretion not seen in rhesus monkey, dog or man [
9].
Three weeks of oral administration of miglustat to male C57BL/6 mice are sufficient to render them infertile [
10]. The miglustat-induced infertility can be maintained for several months without giving rise to overt adverse effects, either physiological or behavioural [
11]. Full fertility is regained when the drug is withdrawn, even after 12 months of administration [
10,
11]. Spermatozoa released from the epididymis of miglustat-treated C57BL/6 males display a spectrum of abnormal head shapes. Acrosomal antigens are mostly absent or display irregular patterns. In addition, the mitochondria of these cells often have an aberrant morphology and are arranged in relatively short and wide mitochondrial sheaths. The motility of the affected spermatozoa is severely impaired. ICSI experiments with misshapen spermatozoa from treated C57BL/6 mice showed that miglustat does affect sperm morphology and physiology, but does not diminish the genetic potential of spermatozoa [
12].
The advanced clinical status of miglustat allowed the compound to be evaluated for its reproductive effects in a small number of normal healthy men. They received 100 mg of miglustat twice daily (on average 2 mg/kg/day) for 6 weeks, a similar dose as is given to Gaucher patients [
13]. During drug treatment and the following 12 weeks various sperm parameters, including morphology and capacity to undergo the acrosome reaction were measured. In the miglustat-treated men none of the reproductive features were significantly affected. The question is now what underlies the difference in the response to miglustat between men and male C57BL/6 mice. We have assessed the effects of miglustat on sperm morphology, particularly shape of nucleus and acrosome, in rabbits and in 18 additional mouse strains. We have found large differences in the sensitivity to the drug between mouse strains, and also found the rabbit to be insensitive. This suggests that the susceptibility to miglustat has a genetic basis. We have therefore investigated its mode of inheritance by analyzing the effects of miglustat on sperm morphology in interstrain hybrid mice that were generated from two strains of mice that differ profoundly in their response to the drug.
Methods
Animals
Mice of the following inbred strains were purchased from Harlan UK (Bicester, Oxfordshire, UK): A/J, BALB/c, C3H/HeN, CBA/Ca, C57BL/6J, C57BL/10ScSn, DBA/2, FVB/N, MRL/Mp, NZB and NZW. Male mice of the 129S1/SvImJ, AKR/J, C57BR/J, C57L/J, C58, MA/MyJ, SM/J and YBR/J strains were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). For the serum level experiments in mice C57BL/6J and FVB/N mice were purchased from Charles River Wiga GmbH (Sulzfeld, Germany).
Russenkaninchen rabbits strain (Crl:CHBB (HM)) were from Charles River Wiga GmbH (Sulzfeld, Germany). Male (2.4 – 3.2 kg, ca. 50 weeks) and female rabbits (2.6 – 3.0 kg,) were housed under 12 h light : 12 h dark (lights on 7:00 h) at 22°C and had access to standard chow (Ssniff, Soest, Germany) and water ad libitum., Only one strain of rabbits was used because of the high costs of miglustat. Also for this reason the Russenkaninchen strain was used because of the relatively small size of the animals of this strain.
All animals were housed and sacrificed in accordance with the UK Home Office Animals (Scientific Procedures) Act 1986 or in accordance with German legal requirements dependent on the place where experiments have been performed.
Fertility in rabbits
After passing a predosing phase of at least 2 weeks, 5 male rabbits were fed 5, 15 or 50 mg/kg/day miglustat (Toronto Research Chemicals, Toronto, Canada) for 8 weeks. Miglustat was added to the standard chow and pressed into pellets (Ssniff, Soest, Germany). Average drug uptake was 5.2 ± 0.6 mg/kg/day in the 5 mg/kg/day group, 15.3 ± 2.0 mg/kg/day in the 15 mg/kg/day group and 49.1 ± 3.8 mg/kg/day in the 50 mg/kg/day group. Control animals received standard chow without miglustat. Food intake was not affected by the presence of miglustat in the chow. Semen and blood samples were taken every 2 weeks. Blood was taken from the ear artery for measuring miglustat levels with LC/MS. Semen samples were collected according to [
14]. Eight weeks after beginning the treatment male rabbits of the control group and the 50 mg/kg/day group were caged together with one female each for determination of fertility.
Quality of semen/spermatozoa from rabbits
Semen parameters determined were as follows: volume, colour, pH, coagulation and sperm count. Spermatozoa were subjected to computer-assisted sperm analysis (CASA) (Hamilton Thorne Research, Beverly, MA, USA). Percentage motile spermatozoa was estimated manually by counting flagellating or non flagellating spermatozoa. For the analysis by CASA 10 minutes after collection spermatozoa were diluted 20-times in medium (NaCl 97 mM, KCl 3 mM, Na2HPO4 2.8 mM, MgCl2 4 mM, CaCl2 2 mM, Na-pyruvate 0.2 mM, glucose 5 mM, HEPES 10 mM, BSA 6 g/l, NaHCO3 25 mM, lactate 32.8 mM) and approximately 10 μl of the suspension were placed in a 20 μm deep sperm analysis chamber at 37°C (Standard Count Analysis Chambers, Leja, Nieuw-Vennep, Netherlands). Ten digitised film sequences were recorded (60 Hz). In general, 100 sperm tracks from the recording were analysed by CASA (Animal Motility, version 12.2). The settings used were as follows: frame rate: 60 Hz, number of frames analysed: 30, minimum contrast: 50, minimum size: 7, minimum number of detected data points for a track: 10. For counting the number of spermatozoa with aberrations air-dried sperm smears were made and stained with PNA/DAPI (lectin PNA from Arachis hypogaea, Alexa Fluor 488 conjugate, Molecular Probes, Oregon, USA and 4',6-Diamidino-2-phenylindole Dihydrochloride, Sigma Aldrich, Germany diluted in PBS (GIBCO)).
Drug administration and mating tests in mice
Miglustat (gift from Oxford GlycoSciences, Abingdon, UK/CellTech UK, Slough, Berkshire, UK, or purchased from Toronto Research Chemicals, Toronto, Canada), as a dry crystalline solid powder, was mixed thoroughly with powdered mouse chow (expanded rat and mouse chow 1, SDS, Witham, Essex, U.K.) and stored at room temperature. Male mice were fed powdered mouse chow with or without miglustat for 5 weeks, at indicated doses. Alternatively, the drug was administered via a mini-osmotic pump (Model 2004, pump rate 0.22 – 0.26 μl/h, Alzet, DURECT Corporation, California, USA) for 7 weeks. The mice were 6 weeks old at the start of miglustat treatment. Natural mating tests were performed as described by van der Spoel
et al [
10]. Briefly, each male mouse was caged with two or four female 10-week old C57BL/6 mice for nine days, after which the male was removed and the females were monitored for pregnancies and litter sizes.
Fluorescence microscopy of murine spermatozoa
Cauda epididymides were dissociated in M2 medium (Sigma) with forceps. Smears of spermatozoa were prepared as described by van der Spoel [
10] and stained by indirect immunofluorescence with monoclonal antibody Mab18.6 [
15], and Alexa488-conjugated goat anti-mouse IgG (Molecular Probes). Propidium iodide (1 μg/ml) was used as a nuclear stain. All antibodies were diluted in PBS containing 0.5% (w/v) BSA and 0.15 M glycine. Nuclear morphology and acrosomal staining were assessed using a Zeiss Axioskop 2 plus microscope. Images were acquired with a Zeiss 510 META confocal fluorescence instrument.
Western blotting
Mouse spermatozoa were released by gentle pressure from caudae epididymides into ice-cold PBS containing Ca
2+ and Mg
2+, and washed twice in cold 10 mM Tris-HCl, pH 8.0, 1 mM EDTA, and taken up in Laemmli sample buffer. Aliquots of lysed cells containing 5 μg protein were separated by SDS-PAGE and transferred to Immobilon-P polyvinylidene fluoride membranes (Millipore, Watford, UK). Blots were blocked in 10% milk powder in Tris-buffered saline containing 0.05% Tween-20 (TBST), and incubated with affinity-purified anti-bovine IAM38 antibodies [
16] (1:1000, overnight) or affinity-purified anti-VYK antibodies [
17] (1.2 ng/ml, overnight) in 2% milk powder in TBST. Anti-VYK is an antiserum raised against a synthetic peptide corresponding to residues 542–558 from murine sp56 [
17]. Anti-human cytochrome C oxidase subunit I (A6403, Molecular Probes) was used (1:500, 2 hr) to verify equal loading. Blots were washed in TBST, incubated with horse-radish peroxidase-conjugated anti-IgG antibodies (Vector Laboratories, Peterborough, UK), washed, and developed with a chemiluminescent substrate (ECL Advance Western Blotting Detection Kit; Amersham, Chalfont St.Giles, Buckinghamshire, UK).
Serum levels of miglustat
Serum level of miglustat was determined using liquid chromatography and mass spectrometry. For sample preparation acetonitrile (4-times the sample volume) including an internal standard (1 μM N-(4-Chlorophenyl)-2-[(4-pyridylmethyl)amino]benzamide) was added to the serum sample and mixed thoroughly. Samples were spun for 20 min at 4°C and > 2000 × g. The supernatant was removed by rotary evaporation and reconstituted with 20% acetonitrile solution. HPLC was performed on a Varian Polaris C18-A column (particle size: 3 μm, column dimensions: 2 × 50 mm, Varian, Palo Alto, CA, USA) at a flow rate of 0.3 ml/min. Eluents were 0.025% NH3·H2O in water with an increasing acetonitrile gradient (5–95% over 3 min). Eluates were analysed by tandem mass spectrometry.
Statistics
Quantitative data were statistically evaluated with Student's t-test (comparison between two groups) or one-way ANOVA with Tukey's post-hoc test (more than two groups) using GRAPHPAD INSTAT version 3.0a for Macintosh (GraphPad Software). Values were considered statistically significantly different when P < 0.05.
Discussion
We have evaluated the reproductive effects of miglustat in rabbits and in various strains of the C57 and Swiss/Castle lineages of inbred mice. In contrast to miglustat-treated C57 mice, fertility was not affected in the other mouse strains and the rabbit strain that have been studied. The contraceptive effect of miglustat seems therefore to be specific to the C57 strains.
In mice, the most profound effects of miglustat on the shape of sperm nuclei and on the presence of acrosomes on sperm heads were found in strains of the C57 family (the majority of sperm nuclei non-falciform dysmorphic, most spermatozoa without an acrosome), correlating closely with the contraceptive action in C57BL/6 mice. A milder category of sperm aberrations after miglustat treatment was found in most strains of the Swiss/Castle lineages (low to moderate frequencies of mild morphological abnormalities of falciform nuclei, most spermatozoa with acrosomes, a fraction of them with imperfections or aberrations). Clearly, this type of spermatozoal abnormalities is not of a severity that impairs the fertility of the animals, as the drug-treated 129S1/SvImJ, FVB/N and DBA/2 males had normal pregnancy rates in mating tests. We also observed an intermediate level of sperm aberrations, in miglustat-treated BALB/c, Ma/MyJ and AKR/J mice (a minority of spermatozoa with non-falciform nuclei, without acrosomes). This level of miglustat-induced spermatozoal abnormalities was not associated with infertility in the drug-treated Ma/MyJ and AKR/J mice. Finally, in one rabbit strain miglustat administration did not result in any observable changes in sperm phenotype.
What therefore underlies the differences in the effect of miglustat between various mouse strains and between species? One possibility is that the males of the Swiss/Castle strains and the rabbits could be less sensitive to the drug. That would imply that a higher dose of miglustat should result in a more severe phenotype in these strains and in rabbits. We have investigated this in FVB/N mice, by administering 150 and 600 mg/kg/day of miglustat. We found the same effects in males of this strain at both doses of the drug, not a C57-style sperm phenotype at the higher dose. This suggests that it is unlikely that the disparity in the imino sugar-response between Swiss/Castle and C57 strains is due to a lower drug sensitivity of the former strains, rather that the consequences of miglustat in these strains are qualitatively different. Secondly, the differences in reproductive outcome of miglustat treatment observed between mouse strains and species could be due to differential pharmacokinetics of the imino sugar (e.g. variations in the rate of renal excretion), resulting in different serum concentrations when administered at the same dose of drug. Clearly, this was not the case as serum levels were the same or even higher in the non-responding FVB/N strain (0.6 μM) and in rabbits (8.4 μM), compared to the highly responding C57BL/6 mice (0.5 μM). In men, the miglustat level was >4 μM in serum and 8 μM in seminal plasma [
13]
Recently, it has been speculated that the primary target of miglustat is β-glucosidase 2 (GBA2), because a deficiency in GBA2 results in the production of abnormal spermatozoa [
19,
20], and because GBA2 can be inhibited by compounds that are chemically similar to miglustat [
21]. In addition, GBA2 has been identified as being responsible for the non-lysosomal glucosylceramidase activity [
22], which can be inhibited by miglustat [
5]. To date it has not been reported whether miglustat inhibits GBA2
in vivo. Nevertheless, it is conceivable that the interaction of miglustat with its primary target (possibly GBA2) is the first step in a cascade of events that brings about the dramatic effects of miglustat on spermiogenesis in C57 strains. This cascade including downstream components needs to be considered when investigating the basis of the diversity in the reproductive consequences of miglustat in the various mouse strains and between species. Possibly, the direct effect of miglustat on the primary drug target, or its indirect impact on a component of the downstream pathway is strain/species-dependent. Studies are currently in progress focusing on the
in vivo biochemical consequences of miglustat administration, with particular emphasis on the glycosphingolipid pathway.
Our studies with the C57BL/6 × FVB/N interstrain hybrid mice provide an indication of the genetics of the sensitivity to miglustat. Whereas the majority of the fourth-generation hybrid mice responded to miglustat in a similar fashion as FVB/N mice, only very few displayed the C57BL/6 phenotype. About one-third of the interstrain hybrid mice simultaneously produced spermatozoa with an acrosome and a falciform nuclear shape, as well as acrosome-deficient spermatozoa with grossly aberrant nuclei. Remarkably, these two types of spermatozoa were produced in various proportions that had not been seen in either of the parental strains. Clearly, the sensitivity of spermatogenesis for miglustat is not inherited in a Mendelian fashion. Rather, the reproductive impact of miglustat, expressed as the percentage of acrosome-less spermatozoa with a non-falciform nuclear shape, appears to be a quantitative trait. It is therefore likely that multiple genes contribute to the sensitivity of spermatogenesis to miglustat. Similar observations have been made in the study of the susceptibility to lung injury by ozone. Mice of the A/J strain are susceptible in this respect, while C57BL/6 mice are resistant [
23]. Recombinant inbred strains generated from these two strains displayed many intermediate levels of ozone susceptibility, in a continuous range, indicating a multigenic trait [
24].
The observation that the effect of a drug, a toxicant, or gene ablation is dependent on genetic background is not uncommon, and has been seen in many fields, including reproductive biology. For example, low-dose cadmium causes necrosis in the seminiferous epithelium in DBA/2, but not in C57BL/6 mice [
25]. The resistance to this heavy metal is an autosomal recessive trait that is inherited in a Mendelian fashion [
26]. In three distinct lines of knockout mice, the deficiency results in male sterility in a 129/Sv background, but not in mixed C57BL/6 × 129/Sv animals [
27‐
29]. Germ cell depletion 2 (
gcd2), a chemically induced recessive mutation, affects almost all males when homozygously present in a CAST/Ei mice, but only a minority of mice against a C57BL/6 background [
30]. Also, the male infertility, seen in the leptin-deficient
ob/
ob mice that are on a C57BL/6 background, is rescued when the deficiency is bred into the BALB/c strain [
31,
32]. In addition, experimentally induced cryptorchidism has a negative impact on spermatogenesis in most strains, but not in AKR/N and MRL/Mp mice [
33].
It will be interesting to determine which genetic factors determine the susceptibility to miglustat, not the least because the capacity of the drug to specifically interfere in acrosome biogenesis is exceptional (even when manifest in only a limited number of mouse strains). Various approaches can be followed in this endeavour, starting with a genomic evaluation of the panels of the C57BL/6 × FVB/N interstrain hybrid mice and of the 19 mouse strains, in which the compound has been tested. These studies are currently in progress.
Acknowledgements
This work was funded by NIH HD HD45861 (ACS and FMP), the Oxford Glycobiology Institute (CMW) and Schering AG, Germany. We thank Ulrike Efiloglu, Thomas Kuhles, Tanja Rose and Astrid Seltz for their competent technical assistance and Hatu Lam and Bernd Woicke for the work on LC/MS, and Raymond Dwek and Terry Butters (Oxford Glycobiology Institute) for acquisition of funding and encouragement. We are also grateful for the excellent support of Dave Smith, Denise Jelfs and colleagues. We also express our gratitude to Harry Moore (University of Sheffield, Sheffield, UK) for the provision of the anti-acrosomal monoclonal Mab18.6. We acknowledge the Wellcome Trust Centre for Human Genetics (WTCHG, University of Oxford) for access to the confocal microscopy system and the staff of the WTCHG Microscopy Core for providing very helpful technical support. We also are indebted to Richard Mott and Ioannis Ragoussis (WTCHG), for the advice on analysis of strain-dependent features in mice, and to George Gerton (University of Pennsylvania Medical Center, Philadelphia, PA) for kindly providing the anti-VYK antibodies, and last (but not least) to Richard Oko (Queen's University, Kingston, ON, Canada) for providing the anti-IAM38 antibodies.
Competing interests
WB, MF, UV, EL and UG are employees of Schering AG, Germany. FMP and ACS are in receipt of a research grant from Schering AG, Germany.
Authors' contributions
WB participated in the conception and design of the study, in the execution of experiments, analysis of experimental data, preparation of the data for publication, and in the writing of the manuscript. CMW participated in the execution of experiments, in the analysis of experimental data. MF, UV, EL and UG participated in the conception and design of the study and revised the manuscript critically. SB performed and analyzed western blotting experiments. FMP participated in the design of the study, and in writing of the manuscript. ACS participated in the conception and design of the study, in the execution of experiments, analysis of experimental data, preparation of the data for publication, and in the writing of the manuscript. All authors read and approved the final manuscript.