Skip Navigation


Human Molecular Genetics Advance Access originally published online on September 2, 2004
Human Molecular Genetics 2004 13(21):2709-2723; doi:10.1093/hmg/ddh281
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
13/21/2709    most recent
ddh281v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (19)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Matsuda, J.
Right arrow Articles by Kuroda, Y.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Matsuda, J.
Right arrow Articles by Kuroda, Y.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Human Molecular Genetics, Vol. 13, No. 21 © Oxford University Press 2004; all rights reserved

Mutation in saposin D domain of sphingolipid activator protein gene causes urinary system defects and cerebellar Purkinje cell degeneration with accumulation of hydroxy fatty acid-containing ceramide in mouse

Junko Matsuda1,*, Makiko Kido1, Keiko Tadano-Aritomi4, Ineo Ishizuka4, Kumiko Tominaga1, Kazunori Toida2, Eiji Takeda3, Kunihiko Suzuki5,6,{dagger} and Yasuhiro Kuroda1

1Department of Pediatrics, 2Department of Anatomy and Cell Biology and 3Department of Clinical Nutrition, The Institute of Health Bioscience, The University of Tokushima Graduate School, 3-18-15, Kuramoto-cho, Tokushima 770-8503, Japan, 4Department of Biochemistry, Teikyo University School of Medicine, Kaga 2-11-1, Itabashi-ku, Tokyo 173-8605, Japan and 5Department of Neurology and 6Department of Psychiatry, Neuroscience Center, University of North Carolina, Chapel Hill, NC 27599-7250, USA

Received July 11, 2004; Accepted August 23, 2004


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
The sphingolipid activator proteins (saposins A, B, C and D) are small homologous glycoproteins that are encoded by a single gene in tandem within a large precursor protein (prosaposin) and are required for in vivo degradation of some sphingolipids with relatively short carbohydrate chains. Human patients with prosaposin or specific saposin B or C deficiency are known, and prosaposin- and saposin A-deficient mouse lines have been generated. Experimental evidence suggests that saposin D may be a lysosomal acid ceramidase activator. However, no specific saposin D deficiency state is known in any mammalian species. We have generated a specific saposin D–/– mouse by introducing a mutation (C509S) into the saposin D domain of the mouse prosaposin gene. Saposin D–/– mice developed progressive polyuria at around 2 months and ataxia at around 4 months. Pathologically, the kidney of saposin D–/– mice showed renal tubular degeneration and eventual hydronephrosis. In the nervous system, progressive and selective loss of the cerebellar Purkinje cells in a striped pattern was conspicuous, and almost all Purkinje cells disappeared by 12 months. Biochemically, ceramides, particularly those containing hydroxy fatty acids accumulated in the kidney and the brain, most prominently in the cerebellum. These results not only indicate the role of saposin D in in vivo ceramide metabolism, but also suggest possible cytotoxicity of ceramide underlying the cerebellar Purkinje cell and renal tubular cell degeneration.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
The physiological functions of sphingolipids are drawing attention because of their newly found functions in variety of cellular processes including proliferation, differentiation, apoptosis and cellular senescence (1–3). Many of the lysosomal diseases are caused by abnormal catabolism of sphingolipids, resulting in severe neurological dysfunction (4). However, very little is known about the distinct pathophysiology of abnormal sphingolipid metabolism in the nervous system of individual lysosomal diseases.

Sphingolipid activator proteins (saposins A, B, C and D) are small homologous glycoproteins derived from a common precursor protein (prosaposin) encoded by a single gene (5). They are specifically required for in vivo degradation of sphingolipids with short carbohydrate chains. Human patients of the prosaposin deficiency and the specific saposin B or C deficiencies are known (6,7). In addition, we have previously generated a mouse model of prosaposin deficiency (8) and also have recently generated a saposin A deficient mouse line that exhibits the clinical/pathological/biochemical phenotype of a late-onset, chronic form of globoid cell leukodystrophy (9,10). This confirmed earlier suggestions and provided definitive evidence that saposin A is indispensable for normal catabolism of galactosylceramidase substrates and that normal cellular functions cannot be maintained in its absence, although human patients with specific saposin A deficiency have not yet been known. Although there is experimental evidence suggesting that saposin D is an acid ceramidase activator (5,11,12), it is unclear whether saposin D is indispensable for normal cellular function. No human patients or animal models with specific saposin D deficiency are known.

In order to clarify the physiological function of saposin D, we have generated and characterized a line of a specific saposin D mutant mouse. This mutant shows a complex clinical/biochemical/pathological phenotype in the cerebellum and kidney. In this report, we describe the phenotype of the specific saposin D–/– mouse and discuss the physiological functions of saposin D with respect to the in vivo ceramide metabolism, which may have a specific relevance to the pathogenesis of the unusual selective degeneration of cerebellar Purkinje cells.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 
Generation of the saposin D–/– strain
The targeting vector was designed to introduce a mutation in exon 13 that changes the 5th cysteine in saposin D to serine, resulting in destruction of one of the three disulfide bridges in saposin D (Fig. 1A). Correctly targeted embryonic stem (ES) cells, before and after removal of the neomycin resistance gene (Neo) by Cre recombinase, and genotypes of resultant mice were confirmed by appropriate DNA analyses (Fig. 1B and C). The allele with the C509S mutation in the saposin D domain generated stable prosaposin mRNA of normal length, as expected, in contrast to its absence in the prosaposin deficiency (Fig. 1D). It should be noted that two lines of saposin D–/– mice were generated originating from two different targeted ES cell lines. No phenotypic differences have so far been detected between the two mutant lines.



View larger version (40K):
[in this window]
[in a new window]
  		Figure 1. Saposin D targeting strategy and genotype analysis. (A) Schematic diagrams and partial restriction maps of the mouse prosaposin locus (wild-type), the saposin D targeting vector (containing exons 8–15) after mutating cysteine (Cys) at 509 to serine (Ser), and the mutant prosaposin allele. The targeting vector was constructed to introduce a mutation in exon 13 that changed the fifth of the six strictly conserved Cys residues to Ser, resulting in elimination of one of the disulfide bridges in saposin D. The vector sequence also introduced a new XhoI site. In addition to the mutation, a neomycin-resistance gene (Neo) flanked by the loxP sequences was present in the intronic sequence of the targeting vector. The Neo gene and one loxP site were subsequently removed by Cre recombinase before the ES cells were introduced into blastocysts. The numbered boxes represent the prosaposin exons. The bold line represents the 5'-flanking probe used to identify the targeted allele. The open boxes indicate either the MC1-Neo or the PGK-TK gene. B, BamHI; H, HindIII; E, EcoRI, S, SalI. (B) Southern blot analysis of tail DNA from wild-type, saposin D+/– and saposin D–/– F2 progeny. Genomic DNA was digested with BglII and XhoI and hybridized with the probe depicted in A. The expected sizes of the BglII–XhoI fragments are ~11.7 and ~10.8 kb in the wild-type and the recombinant allele, respectively. (C) PCR analysis of tail DNA from offspring of a heterozygous mating. The process generates 216 and 323 bp fragments from the wild-type and mutant alleles, respectively. (D) Semi-quantitative RT–PCR analysis of prosaposin mRNA expression. Stable prosaposin mRNA of normal size is expressed in wild-type, saposin D+/– and saposin D–/– F2 progeny mouse brain, while it is absent in prosaposin–/– mouse brain.

 
Clinical phenotype
Viable saposin D–/– mice were born with the frequency expected from the Mendelian single gene inheritance of autosomal recessive nature. Both male and female mice were fertile, and females were able to raise their offspring normally. Since there is experimental evidence suggesting that saposin D is an acid ceramidase activator (5,11,12), we hoped to see clinical phenotype similar to human acid ceramidase deficiency (Farber disease) (13). However, no skin nodules, joint involvements or hepatosplenomegaly, which are the hallmarks of human Farber disease, were observed. Unexpectedly, saposin D–/– mice developed progressive polyuria (more than 20 times normal) with polydipsia at around 2 months and some mice died of severe dehydration (Fig. 2A and B). Saposin D–/– mice began to show impairment in the motor coordination tasks around 4 months and apparent ataxia around 6 months suggesting cerebellar dysfunction. These clinical abnormalities could be demonstrated by paw prints and the runway test (Fig. 2C and D). Gait disturbance progressed and the majority of affected mice could hardly move at their terminal stage around 15 months. We estimate that the maximum life span of saposin D–/– mice is about 500 days.



View larger version (46K):
[in this window]
[in a new window]
  		Figure 2. Clinical phenotype of saposin D–/– mice. (A) Life span of saposin D–/– mice (n=13). Saposin D–/– mice developed progressive polyuria at around 2 months and ataxia at around 6 months. Although some mice died owing to dehydration at around 5 months, a large number of affected mice survived up to 15 months with severe neurological dysfunction. (B) Polydipsia and polyuria in saposin D–/– mice. Water intake: wild-type mice (n=5, median=10 ml, range=7–12 ml), saposin D–/– mice (n=10, median=17 ml, range=9–55 ml, P=0.027). Urine volume: wild-type mice (n=7, median=1 ml, range=0.2–2 ml), saposin D–/– mice (n=14, median=10 ml, range=2–40 ml, P=0.0003). The relative volume of urinary output/water intake: wild-type mice (n=5, median=15%, range=5–19%), saposin D–/– mice (n=10, median=42%, range=19–72%, P=0.0027). (C) Footprint pattern of 4-month-old wild-type mice (left), and saposin D–/– mice (right). Prints of the forepaws and hindpaws are red and blue, respectively. In contrast to the well-coordinated footprint patterns of the wild-type mouse, steps of the hind legs of saposin D–/– mouse are irregularly scattered. (D) Runway test of control and saposin D–/– mice. In contrast to the well-coordinated movement and almost no slips of the forepaw or hindpaw from the beam in the wild-type mice, saposin D–/– mice could hardly move on the beam and slipped frequently. In this figure, both hindpaws of the saposin D–/– mice are slipped off the beam.

 
Pathological phenotype
Histologically, kidneys of saposin D–/– mice showed various degrees of renal tubular degeneration. In contrast, renal glomeruli were relatively spared (Fig. 3). In their terminal stage at around 15 months, it was grossly apparent that the renal cortex was significantly thinned with an enormously dilated renal pelvis, indicative of hydronephrosis (Fig. 3A). In the liver, spleen and testis of saposin D–/– mice, no apparent pathological changes were observed (data not shown).



View larger version (114K):
[in this window]
[in a new window]
  		Figure 3. Histopathology of the kidney from saposin D–/– mice. (A) Macroscopically, 12-month-old saposin D–/– mice showed bilateral hydronephrosis. Histologically, kidneys of saposin D–/– mice (D and E) showed renal tubular degeneration with relatively spared glomeruli compared to wild-type mice (B and C). Hematoxylin and eosin staining. Scale bars are 5 mm in (A), 100 µm in (B and D) and 50 µm in (C and E).

 
The external aspects of the brain and the trigeminal and sciatic nerves of saposin D–/– mice were grossly indistinguishable from those of their wild-type littermates. In their terminal stage around 15 months, the cerebellum of saposin D–/– mice was smaller size than that of their wild-type littermates (data not shown). By basic histological analyses at around 3 months, the cerebellar cortex of saposin D–/– mice had the normal layering pattern—the outer molecular layer containing the Purkinje cell dendrites arranged in parasagittal fans, the single cell layer of the Purkinje cell bodies and the internal granule cell layer containing the small granule neurons. Loss of the Purkinje cells was first observed at around 4–5 months and progressed to the terminal stage. Majority of the cerebellar Purkinje cells disappeared at the terminal stage around 15 months (Fig. 4). At the terminal stage, conspicuous aggregations of apparent storage cells were observed around the blood vessels in the white matter of the cerebellum and the spinal cord (Fig. 4). These ‘storage’ cells contained periodic acid Schiff (PAS)-positive materials. In the spinal cord, the storage cells were more conspicuous on the dorsal side (Fig. 4).



View larger version (105K):
[in this window]
[in a new window]
  		Figure 4. Neuropathology of saposin D–/– mice. Upper column: cerebellum of wild-type (A), 3-month-old saposin D–/– (B), 6-month-old saposin D–/– (C) and 13-month-old saposin D–/– (D). Lower column: spinal cord of wild-type (E) and 13-month-old saposin D–/– [(G) higher magnification of (F)]. In contrast to the regularly lined Purkinje cells in the cerebellum of the 6-month-old wild-type mice (A) and the 3-month-old saposin D–/– mice (B), in the 5-month-old saposin D–/– mice, apparent loss of Purkinje cells were observed (C) and progressed to the terminal stage (D). Majority of the cerebellar Purkinje cells disappeared at 13-month-old saposin D–/– mice (D). At the terminal stage, conspicuous PAS-positive ‘storage cells’ were observed around the blood vessels in the white matter of the cerebellum (D). In the spinal cord, perivascular PAS-positive ‘storage cells’ were more conspicuous on the dorsal side of the white matter in saposin D–/– mice (F and G). LFB–PAS staining. Scale bars are 100 µm in (A–D) and 50 µm in (E–G).

 
Detailed analyses by immunohistochemistry revealed progressive and patterned loss of cerebellar Purkinje cells only in the saposin D–/– mice, never in the wild-type littermates (Fig. 5). In the immunostaining with an anti-calbindin D28k antibody as a Purkinje cell marker, the cerebellum of 10-month-old saposin D–/– mouse showed conspicuous patterned loss of the Purkinje cell bodies and dendrites in contrast to the regularly arranged Purkinje cell bodies and their fine axons and dendrites in the molecular layer both in the parasagittal and coronal sections of wild-type littermates. In the parasagittal sections, the survived Purkinje cells were concentrated in dorsal lobe vermis of the cerebellum (Fig. 5A–D). In the coronal sections, the survived calbindin immunopositive Purkinje cells aligned in stripes (Fig. 5E–H).



View larger version (99K):
[in this window]
[in a new window]
  		Figure 5. Calbindin immunostaining of the cerebellum of saposin D–/– mice. Calbindin immunohistochemistry of parasagittal cerebellar sections (A–D) and coronal sections (E–H) from 10-month-old wild-type mice (A, C, E and G) and saposin D–/– mice (B, D, F and H). In contrast to the regularly arranged Purkinje cells in the wild-type mice, saposin D–/– mice showed conspicuous patterned loss of cell bodies and dendrites. In the parasagittal sections, the surviving Purkinje cells were concentrated in the dorsal lobe vermis of the cerebellum. In the coronal sections, the surviving calbindin immunopositive Purkinje cells were aligned in stripes. M: molecular layer, P: Purkinje cell layer, G: granular layer. Scale bars are 1 mm in (A, B, E and F); 200 µm in (C and D), and 100 µm in (G and H).

 
Further multiple immunostaining using confocal microscopic analysis indicated the selective loss of Purkinje cells of the saposin D–/– mice (Fig. 6). In contrast to the progressive degeneration of cerebellar Purkinje cells, the other surrounding GABA-neurons and astroglial cells including Bergmann glia, were well preserved. The density of GAD65/67 immuno-positive GABA-neurons in the molecular layer was similar to that of the wild-type (Fig. 6E and F). Bergmann glia are unipolar cerebellar astrocytes, whose radial fibers associate with developing granule cells and mature Purkinje cells. They line up systematically with Purkinje cells and play an important role for maintenance of the Purkinje cell. In the cerebellum of saposin D–/– mice, the S-100ß-immuno-positive Bergmann glia were well preserved and aligned regularly even in the segment where the Purkinje cells were absent (Fig. 6G and H). These findings collectively indicate selective death of the Purkinje cells in the saposin D–/– mice.



View larger version (82K):
[in this window]
[in a new window]
  		Figure 6. Selective Purkinje cell death in the cerebellum of saposin D–/– mice. Multiple immunostaining using confocal microscopic analysis with Hoechst (blue) (A), calbindin (FITC; green) (B), GAD65/67 (Cy3; red) (C) and the composite (D) in the cerebellum of 12-month-old saposin D–/– mice. A higher magnification of the composite immunostaining of saposin D–/– mouse (F) and of wild-type mouse (E). Double immunostaining with calbindin (Cy3) and S100ß (Cy5) in the cerebellum of wild-type (G) and saposin D–/– mice (H). In contrast to the loss of calbindin immunoreactive Purkinje cells in saposin D–/– mice, the density of GAD65/67 immunoreactive GABA neurons in the molecular layer was similar to that of the wild-type mice. In wild-type mice, the Purkinje cells and the Bergmann glia were well aligned, whereas the S100ß-immuno-positive Bergmann glia were well preserved and aligned regularly in the saposin D–/– mice, even in the segment where the Purkinje cells were absent (H). These findings indicate selective death of the Purkinje cells in the saposin D–/– mice. Arrows and arrowheads indicate the Purkinje cells and the Bergmann glia, respectively. P: Purkinje cell layer, G: GABA neurons, BG : Bergmann glia. Scale bars are 500 µm in (A–D), 50 µm in (E and F)and 100 µm in (G and H).

 
Biochemical phenotype
In view of the suggested function of saposin D as an activator for in vivo degradation of ceramide, we analyzed the brain, kidney and liver tissues from homozygous mutant mice at 40 days and 5, 10, 13 months for their lipid composition by thin-layer chromatography (TLC) and compared with those of wild-type littermates, saposin A–/– mice, and 39-day-old prosaposin–/–mice. An alkaline-resistant lipid, which moved more slowly than ceramides containing non-hydroxy fatty acids (NFA/d18:1), was detected in the kidney and brain of saposin D–/– mice. (Fig. 7A). By liquid secondary ion and electrospray ionization mass spectrometry, this lipid was identified as ceramide containing 2-hydroxy fatty acids and 4-sphingenine (HFA/d18:1), which was barely detectable in the wild-type mice (Fig. 7A). In the kidney of saposin D–/– mice, other molecular species, which consisted of HFA and 4-hydroxysphiganine (HFA/t18:0) (14,15) were also increased (Fig. 7A and B, Table 1). The structures of ceramides were further confirmed by low energy collision-induced dissociation of the [M–H]– ions, which produced product ions referring to the identities of the fatty acid substituent and of the sphingoid base (Figs 8 and 9) (16). Ceramides containing non-hydroxy fatty acids (NFA/d18:1), which accumulated in the tissue of human Farber disease (17–20), also increased in saposin D–/– mice but to a lesser extent (up to 220% in the kidney and 130% in the brain) (Fig. 7; Table 1). The accumulation of HFA-ceramide (HFA/d18:1) was already evident at 40-day in the kidney of saposin D–/– mice in a similar degree observed in the kidney of prosaposin–/– mice where the major accumulation was observed for non-hydroxy fatty acid-containing ceramides (Fig. 7B; Table 1). In the brain of saposin D–/– mice, the accumulation was first detected around 6–7 months and the accumulation was more prominent in the cerebellum than in the cerebrum (Fig. 7A and C). No accumulation was observed in the brain of saposin A–/– mice (Fig. 7C). Consistent with the absence of histopathological changes, there was no accumulation of HFA-ceramide in the liver (Fig. 7A; Table 1). No major abnormalities in the profile of other lipids including sulfatides and gangliosides were found in saposin D–/– mice, except for dihexosylceramide, which increased 2–3 times in the kidney (data not shown).



View larger version (61K):
[in this window]
[in a new window]
  		Figure 7. Thin-layer chromatography of lipids from the kidney, brain and liver of male mice. Solvent systems were chloroform–methanol 97 : 3 (v/v) followed by chloroform–methanol–acetic acid 47 : 1 : 2 (v/v/v) for brain and kidney, chloroform–methanol–acetic acid 47 : 1 : 2 (v/v/v) for liver. Detection was done by the 3% cupric acetate–phosphoric acid spray. Ceramide is abbreviated as Cer. St: authentic ceramide (ceramide natural mixture, bovine brain, Funakoshi), molecular species of which are identified as NFA/d18:1 (upper major band) and HFA/d18:1 (lower major band) by mass spectrometry. (A) Wet weight equivalents of 12 mg, 6 mg and 10 mg of lipids from the kidney, brain and liver, respectively, were applied on the plate without saponification. Lane 1: wild-type; 139 days; Lane 2: saposin D–/–, 160 days; Lane 3: saposin D–/–, 396 days; Lane 4: wild-type, 139 days; Lane 5: saposin D–/–, 160 days; Lane 6: saposin D–/–, 358 days; Lane 7: saposin D–/–, 396 days; Lane 8: wild-type, 236 days; Lane 9: saposin D–/–, 396 days. An alkaline-resistant lipid, which moved more slowly than ceramides containing non-hydroxy fatty acids (NFA/d18:1), was detected in the kidney and the brain, but not in the liver of saposin D–/– mice. The bands labeled by * are alkaline-labile and not detected after saponification. (B) Ten milligram wet weight equivalents from the kidney were applied on the plate. Lane 1: wild-type, 40 days; Lane 2: saposin D–/–, 40days; Lane 3: wild-type, 40 days; Lane 4: prosaposin–/–, 39 days. The accumulation of HFA/d18:1-ceramides was already evident at 40 days in the kidney of saposin D–/– mice to a similar degree observed in the kidney of prosaposin–/– mice, where the major accumulation was observed for NFA/d18:1-ceramides. (C) Ten milligram wet weight equivalents of saponified lipids from the brain were applied to the plate. Compared with the wild-type and saposin A–/– mice, saposin D–/– mice showed conspicuous accumulation of HFA/d18:1 in the cerebellum. The ages of the mice were: saposin D–/–, 397 days; saposin A–/–, 187 days and wild-type, 137 days. NFA: non-hydroxy fatty acids, HFA: 2-hydroxy fatty acids.

 

View this table:
[in this window]
[in a new window]
  		Table 1. Concentration of ceramides in the brain, kidney and liver of wild-type and mutant mice
 


View larger version (11K):
[in this window]
[in a new window]
  		Figure 8. Fragmentation scheme for ceramides. Symbols with script a, b and c designate fatty acid-related ions, sphingosine-related ions and common ions, respectively.

 


View larger version (23K):
[in this window]
[in a new window]
  		Figure 9. Negative ion electrospray ionization mass spectrometry of ceramides from the kidney of saposin D–/– mice (396 days). Full scan spectra of ceramides containing NFA/d18:1 (A), HFA/d18:1 (B) and HFA/t18:0 (C). Product ion (MS2) spectra of the [M–H]– ions of the 24:0/d18:1-ceramide at m/z 648 (D), 24h:0/d18:1-ceramide at m/z 664 (E) and 24h:0/t18:0-ceramide at m/z 682 (F). Similar spectra were obtained for the corresponding molecular species, NFA/d18:1 and HFA/d18:1, prepared from the brain, although the ceramide containing HFA/t18:0 was not detected in the brain. NFA: non-hydroxy fatty acids, HFA: 2-hydroxy fatty acids.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
MATERIALS AND METHODS
 REFERENCES
 
Earlier, the gene targeting technology allowed us to generate mouse models of prosaposin deficiency (8) and saposin A deficiency (9,10). Phenotype of the prosaposin-deficient mice closely mimicked that of human prosaposin deficiency, whereas the specific saposin A knockout mouse established saposin A as a physiologically essential activator of galactosylceramidase in the sense that normal metabolism of its substrates cannot be maintained in its absence. The saposin A–/– mice showed a clinical, pathological and biochemical phenotype that is qualitatively identical to that of the twitcher mutant due to a genetic galactosylceramidase defect. In the study described here, we asked an equivalent question as we asked when we generated the saposin A–/– mouse, "What would be the in vivo consequences of specific absence of saposin D?" There is a series of in vitro and cell culture evidence in the literatures suggesting that saposin D might be a physiological activator of lysosomal acid ceramidase (5,11,12).

We utilized the same experimental design as we used for the saposin A inactivation (9). The targeting construct had two features in addition to the standard inclusion of the Neo gene and thymidine kinase gene for double selection of correctly targeted ES cells; presence of the loxP sequences flanking the Neo gene within an intron for later removal of Neo, and a newly generated XhoI site for convenient genotyping. Even though the Neo gene was within an intron, we removed it because the presence of a foreign sequence of the size of the Neo gene would have been likely to result in an abnormal splicing and an unstable expression. We had considered application of Cre recombinase in vivo using Cre transgenic mice but eventually chose to remove Neo gene at the stage of targeted ES cells. Application of Cre recombinase using transient transfection of the Cre expression plasmid removed the Neo gene efficiently from ES cells. The single loxP sequence left in within the intron appeared to have no ill effect in the splicing judged from the size and stability of prosaposin mRNA.

The phenotype of saposin D–/– mice turned out to be much more complex than we had anticipated. Saposin D–/– mice were born and grew apparently normally. There were no subcutaneous nodules characteristic of human patients with Farber disease due to acid ceramidase deficiency (13). Phenotypic abnormalities did not become obvious until affected mice were well into their adulthood. Both males and females were fertile and females were able to raise their pups normally. However, two phenotypic abnormalities stood out that were relatively subtle but slowly progressive; the renal abnormalities and the central nervous system (CNS), particularly cerebellar abnormalities. At this time, the metabolic relationship between the renal and the CNS pathology is unclear.

Our expectation for a mouse model of human Farber disease due to saposin D deficiency was clearly optimistic. There are no skin lesions. Although ceramide, particularly HFA-ceramide, does accumulate abnormally, it is much lesser than the massive accumulation seen in human Farber disease, and perhaps more importantly, the accumulation is selective among different organs. Substantial accumulation of HFA-ceramide occurs in the brain and kidney but not in the liver. Cultured fibroblasts from saposin D–/– mice did not show accumulation of either ceramide (data not shown). Thus, the saposin D–/– mouse is not an appropriate model of human Farber disease. Perhaps the simplest explanation is that: (a) one or more of saposins other than saposin D also functions as ceramidase activator in vivo and that specific saposin D deficiency results in partial defect in degradation of ceramide and/or (b) saposin D may have more affinity with HFA-ceramide than NFA-ceramide and may be more indispensable for degradation of HFA-ceramide than NFA-ceramide. These explanations may be supported by the fact that the accumulation of NFA-ceramide is more prominent than that of HFA-ceramide in the kidney of prosaposin–/– mice (Fig. 7B; Table 1), which lack all saposins. Nevertheless, we believe that the unusual CNS pathology is the consequence of HFA-ceramide accumulation (discussed subsequently).

Recently, novel function of saposin C for lipid antigen presentation through CD1b to human T-cell was identified (21–23). Vaccaro and coworkers also demonstrated that saposin D, like saposin C, also strongly binds to membranes and that anionic phospholipids promote and modulate its interaction with lipid surfaces (24–26). On the basis of these pieces of in vitro evidence, the role played by saposin D may be more general than promoting sphingolipid degradation, e.g. the saposin D might also be a key mediator of the solubilization of intralysosomal/late endosomal anionic phospholipid-containing membrane. Further investigation to define the physiological function of saposin D itself in the cerebellum and kidney is of interest.

We had not anticipated the kidney involvement, including dramatic polydipsia, polyuria, dehydration and eventual hydronephrosis nor can we explain the metabolic basis for these observations. It can be assured, however, that the kidney abnormalities are also due to saposin D deficiency, because the phenotype is identical between the two separate lines of saposin D deficient mice generated from two different targeted ES cells. The wild-type or heterozygous littermates never showed the renal abnormalities. Both males and females exhibited basically similar kidney pathology. As it is known that there are gender differences in the lipid profile of the kidney of normal mice (27,28), we limited our studies of kidney lipid to males in this series. Experimental evidence exists regarding the physiological function of sphingolipids or ceramide in proliferation, apoptosis and transport abnormalities associated with cystic renal disease. Altered sphingomyelinase and ceramide expressions were demonstrated in ischemic and nephrotoxic acute renal failure (29–33). The pathogenetic mechanism of the renal abnormalities remains for future investigations.

Slowly progressive and specific degeneration of cerebellar Purkinje cells is clearly the most interesting observation in saposin D–/– mice. Subtle neurological signs of motor incoordination suggestive of cerebellar involvement appear at around 4 months. The neuropathological counterpart appears to be the specific disappearance of the cerebellar Purkinje cells with relative preservation of other cell types including GABA-neurons and the Bergmann glia (Fig. 6). The abnormal accumulation of HFA-ceramide is also most conspicuous in the cerebellum. Although very preliminary, saposin D might normally be highly localized in the Purkinje cells (as discussed subsequently). Purkinje cell death in the cerebellum has been observed in several spontaneous and experimental mouse mutants as well as in some human diseases (34). Niemann–Pick disease type C (NPC) is an inherited lysosomal disease characterized by hepatosplenomegaly and severe CNS deterioration (35). In a mouse model of NPC, progressive motor impairment and the corresponding degeneration of cerebellar Purkinje cells occur (36–38). The pattern of cerebellar Purkinje cell death in NPC mice is similar to that seen in saposin D–/– mice. In calbindin immunostaining in saposin D–/– mice, surviving Purkinje cells are arranged in longitudinal strips probably corresponding to the normal longitudinal cerebellar compartmentation (39,40). Although further quantitative analyses are needed, the Purkinje cells in the posterior vermis seemed to be relatively resistant. These spatial patterns of the Purkinje cell death suggest differential susceptibility of morphologically similar cells perhaps due to their different molecular phenotypes. Cultured cerebellar Purkinje cells have been shown to require ceramide for survival and dendritic differentiation (41). In the saposin D–/– mice, the Purkinje cells presumably degenerate owing to the impaired degradation of ceramide. Recently, a patterned expression of sphingosine kinase was identified in the cerebellar Purkinje cells (42). Some other Purkinje cell constituents may also exhibit similar distribution pattern and they may be responsible for the patterned death of the Purkinje cells of saposin D–/– mice.

The question remains as to whether the accumulation of hydroxyl fatty acid-containing ceramide is mechanistically responsible for the observed renal abnormalities and the Purkinje cell death. Ceramide generated by action of sphingomyelinase has been implicated to play a critical role in the apoptotic pathway (43,44). However, the mechanism of ceramide action as an essential death mediator has not been clarified completely (45,46). An acid ceramidase knockout mouse showed early embryonic lethality in homozygotes, suggesting the important role of ceramide in vivo (47). On the other hand, natural ceramide formed in the lysosomal compartment appears to be unable to escape out of the lysosome, which makes it difficult to understand how hydrophobic ceramide could activate protein targets present in the cytosol or other subcellular compartments (48). These findings strongly suggest that, in contrast to fluorescent derivatives, endogenous long-chain ceramide is unable to exit from lysosomes, therefore making the lysosome-sequestered ceramide unlikely to be a biomodulatory molecule. Burek et al. (46) stimulated cultured cells from patients with the Farber disease with various stress agents and observed no difference in the frequency of apoptosis from control cells. On the other hand, cell-permeable exogenous ceramides had stronger pro-apoptotic activity on the Farber cells than on the control cells. In addition, anti-CD95 induced apoptosis occurred more frequently in the Farber cells. Their data suggest that accumulation of ceramide due to defective acid ceramidase may not play an important role as a mediator in drug- and irradiation-induced apoptosis. Furthermore, the accelerated apoptosis of the Farber cells in response to CD95 ligation might represent its function as facilitator and amplifier of receptor signaling in lipid-rich microdomains. However, in accordance with a role in lipid-rich microdomains, ceramide may function as an amplifier in CD95-mediated apoptosis by altering membrane composition (46).

A point of uncertainties must be pointed out. The definitive proof that our mouse has specific deficiency of saposin D with other three saposins completely intact is not yet on hand because of the limitation in the present technology. Ideally, it should be demonstrated that only saposin D protein is absent and the other three are normally present by immunochemical means. Laboratory of Dr Grabowski kindly provided us with their rabbit antisera against mouse saposin D (49,50). This antibody gave punctuate staining of the Purkinje cells in wild-type mice, whereas staining was completely negative in the Purkinje cells of saposin D–/– mice (Fig. 10). This is clearly consistent with the specific saposin D deficiency in our mice and in any case positively demonstrates that the Purkinje cells of affected mice lack saposin D. However, we must view this finding as preliminary because the antibody also reacts with the precursor protein, prosaposin. The clear-cut results in the Purkinje cells could not be obtained in other cellular types in the brain. We can interpret this observation that other areas of the brain normally contain higher concentrations of prosaposin, which should also be present in the saposin D–/– mouse but the definitive proof must wait for an antibody specific to prosaposin. We, nevertheless, feel comfortable to present the mouse as saposin D–/– based on the following several pieces of circumstantial evidence. Equivalent disease-causing mutations that replace one of the strictly conserved cysteines in the saposin B and C domains with phenylalanine or serine are known among human patients. In these patients, it has been shown that the initial translation product is normally processed and only the specific mutated saposin is absent, whereas others are present and functionally intact (5). This was precisely our rationale to introduce this specific mutation into the saposin D domain. Our mouse does not show any of the clinical, pathological or biochemical phenotypes we would have expected if it were also deficient in any of the other three saposins.



View larger version (111K):
[in this window]
[in a new window]
  		Figure 10. Immunohistochemical detection of saposin D in the cerebellum. The immunostaining with an antibody against saposin D/prosaposin on coronal cerebellar sections of wild-type littermates [(A), 139 day, male] and saposin D–/– mice [(B) 132 day, male] revealed the clear immunoreactivity in the cytoplasm of the Purkinje cells and the Bergmann glia of the wild-type mice, but almost complete loss of its immunoreactivity in the Purkinje cells of saposin D–/– mice. Arrows and arrowheads indicate the Purkinje cell and the Bergmann glia, respectively. Saposin D/prosaposin: FITC (green); anti-S100ß:Cy5 (blue); propidium iodide: Cy3 (red). M: molecular layer, P: Purkinje cell layer, G: granular layer. Scale bars are 50 µm.

 
In conclusion, we have newly generated saposin D-deficient mice. They not only indicate the role of saposin D in in vivo ceramide metabolism but also suggest possible cytotoxicity of ceramide underlying the cerebellar Purkinje cell and renal tubular cell degeneration. The saposin D–/– mice should be a useful new model to study the physiological function of saposin D and ceramide.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
REFERENCES
 
Saposin D targeting vector and generation of the saposin D mutant mice
The targeting strategy that we have used was an approach basically similar to that we used to generate the saposin A–/– mice (9). The targeting vector was designed to introduce specific mutations (C509S) in the saposin D domain of the mouse prosaposin gene using Cre/loxP system. This amino acid substitution is expected to destroy one of the disulfide bonds in saposin D (GenBank: U57999) (51). We picked this mutation because equivalent mutations that abolish one of the disulfide bonds in the saposin B and C domains, respectively, are known to result in specific saposin B or C deficiency in humans. Mouse prosaposin gene was cloned from a 129Sv mouse genomic library (Stratagene, La Jolla, CA, USA) as described previously (8). To construct the targeting vector, three separate gene fragments were subcloned into pBluescript KS(+) (Stratagene); a 4.0 kb EcoRI–HindIII fragment containing exons 8 and 14 (subclone 3), 7.7 kb EcoRI–SalI fragment containing exons 8–15 (subclone 7) and a 1.2 kb HindIII–HindIII fragment containing exons 15 (subclone 8). Subclone 7 was used to introduce a fifth Cys to Ser substitution at amino acid position 509 in exon 13 by using the two-oligonucleotide PCR mutagenesis method. The mismatch oligonucleotide sets were: Primer 1; 5'-ATGGATCCTGGATTTGTCTCGAGTGTAAGCCGCGCCAGA-3' (containing the unique BamHI site) and Primer 2; 5'-GCTGACTCACGTCCCTGAGAAGGAA-3' (located outside the unique HindIII site). Underlined letters indicate the mismatched nucleotides. This mutation leads to the substitution of the fifth Cys and destruction of one of the disulfide bonds in saposin D, and simultaneously introduces a unique XhoI recognition site for convenient genotyping. The mutant PCR products were digested with BamHI and HindIII and replaced by the homologous fragment in subclone 3. The targeting construct was generated in the vector, OSdupdel, which contained MC1-neomycin resistance gene (Neo) flanked by two loxP sites and thymidine kinase (TK) under control of 3' phosphoglycerate kinase (PGK) (52). The homologous 4.0 kb EcoRI–SalI fragment from subclone 3 carrying the introduced C509S mutation was ligated to the SfiI–XhoI site of the vector upstream to the Neo selection cassette flanked by two loxP sites to form the 5' region of the homology. The homologous 1.2 kb HindIII–HindIII fragment was ligated to the HindIII site between the MC1-Neo and the TK–PKG genes to form the 3' homology region. The long- and short-homologous arms of prosaposin gene fragments were divided within an intron (Fig. 1). The targeting vector was linearized by NotI and electroporated into 2x107 ES cells derived from 129/SvEv strain, and stably transfected ES cell clones were isolated after double selection with G418 and Ganciclovir. Homologously recombined ES cell clones were identified by PCR using the primers from the 3' end of the Neo gene, 5'-CTTCTATCGCCTTCTTGACGAG-3', and just outside the short arm of the targeting vector, 5'-CTAGGTCCATTTAACTGCACAGG-3', indicated by arrows in Figure 1. A 1428 bp PCR product was generated from the recombinant ES cell clones and those were confirmed by Southern blot analysis using the combination of BglII and XhoI digestion with the 5'probe located outside of the targeting vector to identify correctly targeted ES cell clones before and after Cre-treatment, and to confirm subsequent mutant mice. Correctly targeted clones were transiently transfected with 20 µg of a CMV promoter–Cre expression, developed in the UNC Animal Models Core Facility, to remove the Neo cassette, and the result confirmed by PCR and Southern blot analysis. Positive mutant clones were used to produce chimeric animals by microinjection into C57BL/6J blastocysts. Male chimeras were then mated with female C57BL/6J mice and heterozygous F1 mice (saposin D+/–) were then intercrossed to generate saposin D–/– mice in a 129SvEv–C57BL/6J mixed background. Two independent lines of saposin D–/– mutants were generated from two separate targeted ES cells.

Southern blot and PCR genotyping analysis.
We used 473 bp PCR fragments amplified by the primers (forward primer; 5'-GCTGGCCCTCTGATTTTAGTTAGT-3' and reverse primer; 5'-TACCCAGTGAGCTGTCCCTAAGA-3'), which contain intron 1, located outside of the targeting vector as a 5'probe to identify correctly targeted ES cell clones before and after Cre-treatment, and to confirm subsequent mutant mice (Fig. 1). An aliquot of 10 µg of amplified genomic DNA was digested with BglII and XhoI, and subjected to 0.8% agarose gel electrophoresis, which were then probed with DIG-labeled 5'probe as described previously (9). The presence of the XhoI site introduced by the targeting vector leads to a change of fragment length from 11.7 kb in wild-type allele to 10.8 kb in the mutant recombinant allele. Single digestion with BglII was also used to detect the correct Neo insertion and excision before and after the Cre-treatment.

Heterozygous F2 saposin D+/– mice on the 129SvEv–C57BL/6 mixed background were intercrossed to generate saposin D–/– mice. The genetic status for the saposin D mutation was determined by diagnostic PCR on genomic DNA extracted from clipped tails around 7 days postnatally. For genotype analysis of saposin D–/– mice, we designed two oligonucleotide primers to distinguish the mutant allele from the wild-type allele by PCR using genomic DNA extracted from the tip of the clipped tail. The oligonucleotide primers were designed to flank the retained one loxP site in intron 14 of the mouse prosaposin gene (sense, 5'-ATGGTCCCTGCCTCTGCACAAG-3' and antisense, 5'-GCTGACTCACGTCCCTGAGAAGGAA-3'), indicated by arrows (Fig. 1). The PCR product of the wild-type allele was a 216 bp fragment, whereas the PCR product of the mutant allele was a 323 bp fragment.

Semiquantitative RT–PCR.
Total RNA was extracted from the whole brain tissues from saposin D–/–, saposin D+/–, wild-type and prosaposin–/– mice. RT–PCR to see the expression of mouse prosaposin and glyceraldehydes-3-phosphate dehydrogenase (GAPDH) was evaluated by semi-quantitative RT–PCR as described previously (9).

Animal care
All animals were kept in a pathogen- and odor-free environment, which was maintained under 12 h light/dark cycle at ambient temperature (24±1°C), and food and water were available ad libitum. The present study was carried out in accordance with the Guidelines for the Care and Use of Laboratory Animal adopted by the Committee on Animal Research in The University of Tokushima, and also accredited by the Japanese Ministry of Education, Culture, Sports, Science and Technology. Every effort was taken carefully to minimize any pain or discomfort of animals used in all experiments.

Clinical observation
In order to follow the course of the disease, all mice were closely observed throughout their lives. Body weight was recorded once a day as an objective parameter for development and progression of the disease. Water intake and urine volume were measured using the metabolic cage (Nalge NUNC International, Tokyo, Japan) according to the manufacture's protocol (53). To determine the survival time, some mice were allowed to live as long as they could be maintained humanely according to the acceptable practice of laboratory animal care but without forced feeding or other extraneous interventions. More than 10 mice from each group were sacrificed at around 1, 2, 4, 6 and 12 months, for pathological and biochemical evaluation. Samples for biochemical analyses were stored at –80°C until analyses.

Motor coordination tests.
Footprint patterns of wild-type mice and saposin D–/– mice were analyzed using narrow tunnel (10 cm wide, 35 cm long and 10 cm high) with white paper at the bottom. Before traversing the tunnel, the hind- and forepaws of the animals were dipped in non-toxic blue and red ink, respectively (54).

The runway test was performed using narrow horizontally fixed beam (1 cm wide, 100  m long, held at a height of 40 cm from the table) (54). The animal was placed at one end of the beam and allowed to move toward the opposite end where an escaped platform was located. Slips of the forepaw and hindpaw were counted.

Histopathological analysis
Tissue preparation.
The mice were anesthetized with sodium pentobarbital and perfused through the left cardiac ventricle with physiological saline, followed by 4% paraformaldehyde in 0.1 M sodium phosphate buffer, pH 7.4 and immersed in the same fixative at 4°C overnight. Then the brain, spinal cord, trigeminal and sciatic nerves, liver, spleen, kidney and testis were dissected and processed for paraffin sections and immunohistochemical studies.

Histopathology and immunohistochemistry.
The paraffin sections of the central and peripheral nervous system tissues and liver, spleen, kidney and testis were stained with hematoxylin and eosin and with luxol fast blue/periodic acid Schiff (LFB–PAS).

For free-floating immunohistochemical studies, the brain was sliced in coronal and parasagittal direction at 50 µm thickness with a Microslicer (DTK-1000, Dosaka). The sections were pretreated with 3% H2O2 in 80% methanol for 30 min to reduce the endogenous peroxidase activities and to increase the penetration of antibodies. The sections were incubated in the primary antibodies for 1–3 days at 20°C, then in the second antibodies for 2 h. Primary antibodies used for the study were as follows: anti-calbindin D-28k (mouse monoclonal or rabbit polyclonal, 1 : 2000; Swant, Switzerland, a Purkinje cell marker), anti-GAD 65/67 (rabbit, 1 : 2000; Sigma, St Louis, MO, USA, a GABA-neuron marker) and anti-S100ß [mouse monoclonal (Sigma) or goat polyclonal (Santa Cruz, Santa Cruz, CA, USA), 1 : 2000, astrocyte markers]. The secondary antibodies used for the study were species specific secondary antibodies biotinylated or tagged with Cy-3, Cy-5 or FITC (Vector Laboratory, Burlingame, CA, USA). All the secondary antibodies were used at 1 : 200 dilution. Hoechst or propidium iodide (PI) staining was alternatively performed for DNA staining. For detection of the biotinylated secondary antibodies, we used the avidin–biotin complex (ABC) method with diaminobenzidin tetrahydroxychloride (DAB) for staining of single antibody (Vector). For multiple staining, with fluorolabeled secondary antibodies, confocal microscopic analysis was employed as described previously (55,56). Briefly, immunostained sections were mounted in Vectashield (Vector Laboratory) and examined with a confocal laser scanning microscope (CSLM; Bio-Rad Radiance 2000MP and 2000, Bio-Rad Laboratories, Hercules, CA, USA; mounted on a Nikon TE2000, E800) using laser beams of 345, 494, 554 and 649 nm for excitation with an appropriate filter set. The images were processed with Adobe Photoshop 6.0. The images were not digitally manipulated beyond the general adjustments.

Lipid analyses
Lipid extraction from tissues and thin-layer chromatography (TLC).
Tissues (brain, liver and kidney) from male mice were homogenized with ice-cold water at 20% of concentration by weight in an all-glass Potter–Elvehjem homogenizer. Initial extraction with chloroform–methanol was done as described previously (8,57). Alternatively, tissues were extracted with 19 volumes of chloroform–methanol 2 : 1 (v/v) and with 10 volumes of chloroform–methanol–water 60 : 120 : 9 (v/v/v) successively (28,58). The pooled total lipid extract was analyzed by TLC as described subsequently. A part of the brain total lipid extract was fractionated to neutral and acidic fractions using the reverse phase column essentially according to Kyrklund's method (59) (Bond Elute C-18; 3 ml/500 mg, Varian Inc., Palo Alto, CA, USA). Aliquots of the neutral lipid fraction were subjected to the mercuric chloride-saponification procedure in order to remove essentially all glycerophospholipids (60). TLC was done with Merck high performance TLC plates (Silica gel 60, Merck, Darmstadt, Germany), with the solvent system, chloroform–methanol–acetic acid 47 : 1 : 2 (v/v/v). For analysis of the total lipid extract, the plate was first developed with chloroform–methanol 97 : 3 (v/v), thoroughly dried in vacuo, and then developed with chloroform–methanol–acetic acid 47 : 1 : 2 (v/v/v) to avoid interference from the large amount of other lipids. The total lipid extract was also analyzed by two-dimensional TLC using the solvent systems, chloroform–methanol 93 : 7 (v/v) (first direction) and chloroform–methanol–acetic acid 47 : 1 : 2 (v/v/v) (second direction). The bands were visualized with 3% cupric acetate–phosphoric acid and determined by densitometry (CS-9000, Shimadzu, Kyoto, Japan) (28,58) by comparison with known amounts of authentic ceramide (ceramide natural mixture, bovine brain, Funakoshi, Tokyo, Japan).

Identification of ceramides by mass spectrometry.
Ceramide bands on the TLC plate were identified by negative-ion liquid secondary ion mass spectrometry (61). Each band developed by two-dimensional TLC and stained with primuline reagent was transferred to a polyvinylidene difluoride (PVDF) membrane by iron-blotting (Far-eastern blotting) (62), and the band on the membrane was excised and placed on a mass spectrometry probe tip with triethanolamine as the matrix. Spectra were recorded at an accelerating voltage of 8 kV, with a scan rate of 5 s/decade, and at a resolution of 1000–2000 on a Concept IH mass spectrometer (Shimadzu-Kratos, Kyoto, Japan) (63).

Identification for purified samples was also performed by negative-ion electrospray ionization mass spectrometry using an LCQ DECA ion trap mass spectrometer (ThermoFinnigan, San Jose, CA, USA) (64). The total lipid extracts prepared from the kidney and brain (0.1 g wet tissue) were separated by TLC as described earlier. The plate was sprayed with primuline reagent. The band corresponding to each molecular species of ceramides was scraped and the lipid was eluted with chloroform–methanol 85 : 15 and 3 : 1 (v/v). Solvent volumes were then adjusted to the ratio for the standard Folch partition (chloroform–methanol–0.88% KCl 8 : 4 : 3, v/v/v). The lower phase was dried under a gentle stream of nitrogen and redissolved in 0.2 ml of chloroform–methanol 2 : 1 (v/v). An aliquot was diluted with 10–20 volumes of methanol and was directly infused into the ion source at a flow rate of 3 µl/min. The heated capillary was set at 250°C and the spray voltage was set at 5.0 kV. The sheath gas flow rate was set to 50 in arbitrary units. Low energy collision-induced dissociation was carried out on the molecular-related ions (MS2) using helium gas present in the ion trap. The relative collision energy used ranged from 35 to 40%.


   ACKNOWLEDGEMENTS
 
The authors thank Dr Nobuyo Maeda for providing us with the vector (OSdupdel), Dr Randy J. Thresher at the University of North Carolina Animal Models Core Facility for helping with ES cell handling and Dr Kazunori Ishimura at the University of Tokushima for his encouragement. The authors also thank Ms Akiko Yamazaki for her assistance in animal care. This work was supported by the Roscoe Brady Lysosomal Storage Diseases Fellowship from National Organization for Rare Disorders, Inc. (NORD) to J.M. and by the Japan Society for the Promotion of Sciences, Grant-in-aids for Scientific Research (16591032 to J.M., 1550239 to K.T. and 14370247 to Y.K.).


   FOOTNOTES
 
* To whom correspondence should be addressed. Tel: +81 886337135; Fax: +81 886318697; Email: junko{at}clin.med.tokushima-u.ac.jp

{dagger} Present address: Institute of Glycotechnology, Future Science and Technology Joint Research Center, Tokai University, Hiratsuka 259-1292, Japan. Back


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 REFERENCES
 

  1. Hakomori, S. (2000) Cell adhesion/recognition and signal transduction through glycosphingolipid microdomain. Glyco. J., 17, 143–151.

  2. Anderson, R.G.W. and Jacobson, K. (2002) A role for lipid shells in targeting proteins to caveolae, rafts, and other lipid domains. Science, 296, 1821–1825.[Abstract/Free Full Text]

  3. Hering, H., Lin, C.C. and Sheng, M. (2003) Lipid rafts in the maintenance of synapses, dendritic spines, and surface AMPA receptor stability. J. Neurosci., 23, 3262–3271.[Abstract/Free Full Text]

  4. Hers, H.G. (1966) Inborn lysosomal disease. Gastroenterology, 48, 625–633.[ISI]

  5. Sandhoff, K., Kolter, T. and Harzer, K. (2001) Sphingolipid activator proteins The Metabolic and Molecular Basis of Inherited Disease. In Scriver, C.R., Beaudet, A.L., Sly, W.S., Valle, D., Childs, B. and Vogelstein, B. (eds). McGraw-Hill, New York, pp. 3371–3388.

  6. Christomanou, H., Aignesberger, A. and Link, R.P. (1986) Immunochemical characterization of two activator proteins stimulating enzymic sphingomyelin degradation in vitro: absence of one of them in a human Gaucher disease variant. Biol. Chem. Hoppe-Seyler, 367, 879–890.[ISI][Medline]

  7. Stevens, R.L., Fluharty, A.L., Kihara, H., Kaback, M.M., Shapiro, L.J., Marsh, B., Sandhoff, K. and Fischer, G. (1981) Cerebroside sulfatase activator deficiency induced metachromatic leukodystrophy. Am. J. Hum. Genet., 33, 900–906.[ISI][Medline]

  8. Fujita, N., Suzuki, K., Vanier, M.T., Popko, B., Maeda, N., Klein, A., Henseler, M., Sandhoff, K., Nakayasu, H. and Suzuki, K. (1996) Targeted disruption of the mouse sphingolipid activator protein gene: a complex phenotype, including severe leukodystrophy and wide-spread storage of multiple sphingolipids. Hum. Mol. Genet., 5, 711–725.[Abstract/Free Full Text]

  9. Matsuda, J., Vanier, M.T., Saito, Y., Tohyama, J., Suzuki, K. and Suzuki, K. (2001) A mutation in the saposin A domain of the sphingolipid activator protein (prosaposin) gene causes a late-onset, slowly progressive form of globoid cell leukodystrophy in the mouse. Hum. Mol. Genet., 10, 1191–1199.[Abstract/Free Full Text]

  10. Matsuda, J., Vanier, M.T., Saito, Y., Suzuki, K. and Suzuki, K. (2001) Dramatic phenotypic improvement during pregnancy in a genetic leukodystrophy: estrogen appears to be a critical factor. Hum. Mol. Genet., 10, 2709–2715.[Abstract/Free Full Text]

  11. Azuma, N., O'Brien, J.S., Moser, H.W. and Kishimoto, Y. (1994) Stimulation of acid ceramidase activity by saposin D. Arch. Biochem. Biophys., 311, 354–357.[CrossRef][ISI][Medline]

  12. Klein, A., Henseler, M., Klein, C., Suzuki, K., Harzer, K. and Sandhoff, K. (1994) Sphingolipid activator protein D (sap-D) stimulates the lysosomal degradation of ceramide in vivo. Biochem. Biophys. Res. Commun., 200, 1440–1448.[CrossRef][ISI][Medline]

  13. Moser, H.W., Linke, T., Fensom, A.H., Levade, T. and Sandhoff, K. (2001) Acid ceramidase deficiency: Farber lipogranulomatosis. The Metabolic and Molecular Basis of Inherited Disease. In Scriver, C.R., Beaudet, A.L., Sly, W.S., Valle, D., Childs, B. and Vogelstein, B. (eds). McGraw-Hill, New York, pp. 3573–3588.

  14. Karlsson, K.-A. (1970) On the chemistry and occurence of sphingolipid long-chain bases. Chem. Physics Lipids, 5, 6–43.[CrossRef][ISI][Medline]

  15. Samuelsson, B.E. (1982) Regional distribution of glycosylceramide-sulphates in human kidney. Lipids, 17, 160–165.[CrossRef][ISI][Medline]

  16. Hsu, F.F. and Turk, J. (2002) Characterization of ceramides by low energy collisional-activated dissociation tandem mass spectrometry with negative-ion electrospray ionization. J. Am. Soc. Mass Spectrom., 13, 558–570.[CrossRef][ISI][Medline]

  17. Sugita, M., Connolly, P., Dulaney, J.T. and Moser, H.W. (1973) Fatty acid composition of free ceramides of kidney and cerebellum from a patient with Farber's disease. Lipids, 8, 401–406.[CrossRef][ISI][Medline]

  18. Samuelsson, K. and Zetterstöm, R. (1971) Ceramides in a patient with lipogranulomatosis (Farber's disease) with chronic course. Scand J. Clin. Lab. Invest., 27, 393–405.[ISI][Medline]

  19. Moser, H.W., Prensky, A.L., Wolfe, H.J., Rosman, N.P., Carr, S. and Ferreira, G. (1969) Farber lipogranulomatosis: report of a case and demonstration of an excess of free ceramide and ganglioside. Am. J. Med., 47, 869–890.[CrossRef][ISI][Medline]

  20. Prensky, A.L., Ferreira, G., Carr, S. and Moser, H.W. (1967) Ceramide and ganglioside accumulation in Farber's lipogranulomatosis. Proc. Soc. Exp. Biol. Med., 126, 725–728.

  21. Winau, F., Schwierzeck, V., Hurwitz, R., Remmel, N., Sieling, P.A., Modlin, R.L., Porcelli, S.A., Brinkmann, V., Sugita, M., Sandhoff, K. et al. (2004) Saposin C is required for lipid presentation by human CD1b. Nat. Immunol., 5, 169–174.[CrossRef][ISI][Medline]

  22. Zhou, D., Cantu, C., III, Sagiv, Y., Schrantz, N., Kulkarni, A.B., Qi, X., Mahuran, D.J., Morales, C.R., Grabowski, G.A., Benlagha, K. et al. (2004) Editing of CD1d-bound lipid antigens by endosomal lipid transfer proteins. Science, 303, 523–527.[Abstract/Free Full Text]

  23. Kang, S.J. and Cresswell, P. (2004) Saposins facilitate CD1d-restricted presentation of an exogenous lipid antigen to T cells. Nat. Immunol., 5, 175–181.[CrossRef][ISI][Medline]

  24. Ciaffoni, F., Salvioli, R., Tatti, M., Arancia, G., Crateri, P. and Vaccaro, A.M. (2001) Saposin D solubilizes anionic phospholipid-containing membranes. J. Biol. Chem., 276, 31583–31589.[Abstract/Free Full Text]

  25. Lefrancois, S., May, T., Knight, C., Bourbeau, D. and Morales, C.R. (2002) The lysosomal transport of prosaposin requires the conditional interaction of its highly conserved D domain with sphingomyelin. J. Biol. Chem., 277, 17188–17199.[Abstract/Free Full Text]

  26. Ciaffoni, F., Tatti, M., Salvioli, R. and Vaccaro, A.M. (2003) Interaction of saposin D with membranes: effect of anionic phospholipids and sphingolipids. Biochem. J., 373, 785–792.[CrossRef][ISI][Medline]

  27. Hey, J.B. and Gray, G.M. (1970) The effect of testosterone on the glycosphingolipid composition of mouse kidney. Biochim. Biophys. Acta, 202, 566–568.[Medline]

  28. Tadano-Aritomi, K., Hikita, T., Fujimoto, H., Suzuki, K., Motegi, K. and Ishizuka, I. (2000) Kidney lipids in galactosylceramide synthase-deficient mice: absence of galactosylsulfatide and compensatory increase in more polar sulfoglycolipids. J. Lipid Res., 41, 1237–1243.[Abstract/Free Full Text]

  29. Deshmukh, G.D., Radin, N.S., Gattone, V.H., II and Shayman, J.A. (1994) Abnormalities of glycosphingolipid, sulfatide, and ceramide in the polycystic (cpk; cpk) mouse. J. Lipid Res., 35, 1611–1618.[Abstract]

  30. Zager, R.A., Conrad, D.S. and Burkhart, K. (1998) Ceramide accumulation during oxidant renal tubular injury: mechanisms and potential consequences. J. Am. Soc. Nephrol., 9, 1670–1680.[Abstract]

  31. Zager, R.A., Conrad, S., Lochhead, K., Sweeney, E.A., Igarashi, Y. and Burkhart, K.M. (1998) Altered sphingomyelinase and ceramide expression in the setting of ischemic and nephrotoxic acute renal failure. Kidney Int., 53, 573–582.[CrossRef][ISI][Medline]

  32. Ueda, N., Camargo, S.M.R., Hong, X., Basnakian, A.G., Walker, P.D. and Shah, S. (2001) Role of ceramide synthase in oxidant injury to renal tubular epithelial cells. J. Am. Soc. Nephrol., 12, 2384–2391.[Abstract/Free Full Text]

  33. Malik, R.K., Thornhill, B.A., Chang, A.Y., Kiley, S.C. and Chevalier, R.L. (2001) Renal apoptosis parallels ceramide content after prolonged ureteral obstruction in the neonatal rat. Am. J. Physiol. Renal Physiol., 281, 56–61.

  34. Sarna, J.R. and Hawker, R. (2003) Patterned Purkinje cell death in the cerebellum. Prog. Neurobiol., 70, 473–507.[ISI][Medline]