Sex differences of urinary and kidney globotriaosylceramide and lyso-globotriaosylceramide in Fabry mice.

The aim of our study was to measure globotriaosylceramide (Gb3) and lyso-Gb3 levels by tandem mass spectrometry in the urine and kidney in Fabry (gla knockout) mice and wild-type controls. We found that urine Gb3 of male and female Fabry mice was higher than wild-type mice of the same sex but also significantly higher in male mice compared with females of the same genotype. In kidney tissue, sex and genotype-dependent differences in Gb3 levels paralleled those in the urine. Isoforms C16, C22:1, and C24OHA were particularly higher in males compared with females in both wild-type and Fabry mice. Similarly, kidney lyso-Gb3 concentrations were significantly higher in 12-month-old male Fabry mice than in their homozygous female counterparts. However, lyso-Gb3 was undetectable in wild-type mice of both sexes. α-Galactosidase A activity and mRNA levels in kidney were significantly lower in male wild-type mice compared with female mice. This study shows the sex differences in kidney and urine Gb3 and kidney lyso-Gb3 levels in both wild-type and Fabry mice, and it suggests that these male-female differences should be taken into consideration when using murine models for Fabry disease.

mouse gla were purchased from Applied Biosystems, Inc. (Foster City, CA). 18S rRNA was used as internal control and detected by TaqMan probe and primers (Applied Biosystems, Inc.).

Statistical analysis
For relative mRNA levels, statistical signifi cance was determined by Mann-Whitney test. Two-tailed Student's t -test was used for analy sis of other data. The data were presented as the mean ± SE.

RESULTS
Consistent with previous observations (14)(15)(16), urinary Gb 3 levels were signifi cantly higher in adult wild-type male mice compared with wild-type female mice ( Fig. 1A ). Urinary Gb 3 levels were lower in prepubertal (three-weekold) wild-type males [399 ± 31 ng/ml (n = 3)] compared with adult wild-type males, but they were similar to the urinary Gb 3 levels measured in one prepubertal female (255 ng/ml).
Similarly, urinary Gb 3 was signifi cantly higher in Fabry male mice compared with homozygous Fabry female mice ( Fig. 1A ). The urine Gb 3 level in Fabry homozygous females was signifi cantly higher than in wild-type females

Urine collection
Five to eight urine samples (collected separately from each mouse) were collected from each mouse group for analysis. All specimens were collected over a 24 h duration using polycarbonate metabolic cages (model 3600M021, Tecniplast, Montreal, Quebec, Canada). Urine samples were collected at room temperature and stored at Ϫ 80°C until analysis by mass spectrometry.

Analysis of kidney Gb 3 and lyso-Gb 3
Kidney tissue was harvested from 5-and 12-month-old mice. Gb 3 concentration in tissue was determined as described below.
Tissue Gb 3 quantifi cation. Liquid-liquid extraction with methyl tert-butyl ether followed by saponifi cation was performed on tissue homogenate corresponding to 200 mg of total protein. Each sample was reconstituted in water/methanol 20/80 (v/v), and 5 µl were injected into a UPLC-Xevo tandem mass spectrometry system (Waters Corp.). Separation was performed on a C18 Acquity BEH 100 × 1 mm, 1.7 µm column with a methanol/water and 0.1% formic acid gradient. The following transitions were monitored in MRM mode: m/z 1046 → 884 C16, 1074 → 912 C18, 1102 → 940 C20, 1128 → 966 C22:1, 1130 → 968 C22, 1156 → 994 C24:1, 1158 → 996 C24, 1174 → 1012 C24OH. A Gb 3 reference standard (Matreya) calibration curve in matching matrix prepared with the standard addition method was run in each assay for calculation of the concentration of total Gb3 in each sample. Gb 3 was calculated as the sum of the concentrations of each monitored isoform. The concentration of each isoform was determined by linear regression with the internal standard. Total Gb 3 was calculated using the software Targetlynx (Waters). Results were normalized to the amount of total protein (in milligrams) in each sample.

␣ -Galactosidase A assay
␣ -galactosidase A activity in kidney homogenates was determined by a fl uorimetric method as described previously ( 21 ).

Quantitative RT-PCR
Quantitative real time RT-PCR was performed as previously described ( 3 ). Predesigned TaqMan probe and primers for Fig. 1. Urinary and kidney Gb 3 and lyso-Gb 3 contents in mouse kidneys. A: Urine Gb 3 levels at 5 months of age. Urine Gb 3 levels were signifi cantly higher in males compared with females in both wild-type and Fabry mice (n = 5-6). The increment of Gb 3 levels in Fabry mice compared with wild-type mice was statistically significant in females but not in males.* P < 0.0001, wild-type female versus Fabry female. B: Kidney Gb 3 levels at 5 months of age for wild-type and Fabry and Fabry at 12 months of age. Kidney Gb 3 levels were markedly higher in males compared with females in both wild-type and Fabry mice. The increment of Gb 3 levels in Fabry mice compared with wild-type mice was statistically significant in females but not in males. * P < 0.01, wild-type female versus Fabry female at 5 months of age (n = 3); ** P < 0.03, Fabry male versus Fabry female at 12 months of age (n = 8). C: Lyso-Gb 3 in 5and 12-month-old wild-type and Fabry mice. F, female; M, male.
should be taken into consideration when using murine models for Fabry disease.
Our fi nding of markedly elevated urinary and kidney Gb 3 in mature male wild-type mice compared with females confi rms previous observations (14)(15)(16). This is a general murine phenomenon, as it was also present in B6CBA mice and other strains ( 15 ). In humans, however, no difference in urinary Gb 3 levels was found between male and female patients ( 19 ). The male-female difference in Gb 3 (and other related glycosphingolipids) level in mouse kidney is thought to be due to increased synthesis caused by testosterone in male mice ( 15,16 ). In this study, we found signifi cantly lower kidney ␣ -galactosidase A activity in male mice compared with females. This suggests that the lower ( Fig. 1A ). Although the urine Gb 3 level in Fabry hemizygous male mice was higher than in wild-type controls, this difference was not signifi cant ( Fig. 1A ).
A considerable amount of glycosphingolipid in mouse urine is excreted from the kidney, and glycosphingolipid content of urine is correlated with that in kidney tissues ( 14 ). We measured Gb 3 levels in mouse kidneys and found markedly elevated Gb 3 levels in males compared with females in both wild-type and Fabry mice ( Fig. 1B ). Fabry female mice had signifi cantly higher kidney Gb 3 levels compared with age-matched, wild-type females ( Fig. 1B ). In parallel to observations in urine Gb 3 , the increment of kidney Gb 3 in Fabry male mice compared with wild-type male mice at fi ve months of age was not statistically significant ( Fig. 1B ). We also measured lyso-Gb 3 . It was undetectable in male and female wild-type mouse kidney. However, lyso-Gb 3 was measurable in kidney of Fabry KO mice where it was found to be signifi cantly higher in 12-month-old male Fabry mice compared with homozygous female Fabry mice ( Fig. 1C ).
To determine whether there is any specifi c isoform that predominantly contributes to sex differences in kidney Gb 3 levels, the concentration of individual isoforms of Gb 3 were compared between males and females ( Fig. 2A , B ). In both wild-type and Fabry mice, isoforms C16, C22:1, and C24OHA were relatively more predominant in males compared with females ( Fig. 2A, B ). The concentration of each isoform was also compared between Fabry and wildtype mice ( Fig. 2C ). The fold increase in Fabry mouse kidney Gb 3 concentration compared with wild-type was greater in females than in males for all Gb 3 isoforms ( Fig. 2C ). This increase ranged from a 2-fold (C24OHA) to 10-fold (C24_B) greater concentration of Gb 3 in Fabry males compared with wild-type males, and from a 15-fold (C24OHB) to 58-fold (C22:1) greater concentration of Gb 3 in Fabry females compared with wild-type females.
Testosterone increases synthesis and excretion of glycosphingolipids in mouse kidney (14)(15)(16). Thus, different levels of testosterone in adult male and female mice can cause sex differences in kidney and urine Gb 3 . On the other hand, testosterone increases the synthesis of some lysosomal enzymes such as ␤ -glucuronidase and ␤ -galactosidase in murine kidney ( 22,23 ). To determine whether the lysosomal hydrolysis of Gb 3 contributes to sex differences in kidney Gb 3 , we measured ␣ -galactosidase A activity in kidney tissues. We found signifi cantly lower ␣ -galactosidase A activity in wild-type males compared with wild-type females ( Fig. 3A ). To determine whether this lower enzyme activity is due to lower expression level of the enzyme, mRNA level of ␣ -galactosidase A in kidney tissues was measured by quantitative RT-PCR. The results showed signifi cantly lower mRNA level in wild-type males compared with wild-type females ( Fig. 3B ).

DISCUSSION
This study shows the sex differences in kidney and urine Gb 3 levels and kidney lyso-Gb 3 levels in both wild-type and Fabry mice and suggests that these male-female differences of kidney phenotype in Fabry mice including potential sex differences in these phenotypes is in progress.
The proportion of testosterone-induced Gb 3 to total Gb 3 , its chemical properties, and its biological role in Fabry mouse kidney are not clear. However, experimentally, this form of Gb 3 complicates the relationship between ␣ -galactosidase A activity and Gb 3 level in kidney of male Fabry mice. Our study suggests that if one looks at Gb 3 levels in kidney or urine as an outcome parameter for preclinical proof-of-concept studies evaluating new therapeutic approaches, female homozygous mice should be the better model to use because it is more straightforward and may give more sensitive and clearer results.
␣ -galactosidase A activity in male mouse kidney may also contribute to the higher Gb 3 . The lower enzyme activity is due, at least in part, to lower transcription level of the gene. Sex differences in mRNA level of ␣ -galactosidase A suggest the possibility that testosterone downregulates this gene in mouse kidney.
Owing to the enzyme defi ciency, Fabry mice show signifi cant Gb 3 accumulation in multiple organs including the kidney ( 7,8 ). However, our study demonstrated the presence of sex differences in kidney and urine Gb 3 and lyso-Gb 3 levels in Fabry mice. The higher Gb 3 (and lyso-Gb 3 ) content in hemizygous male Fabry mice than in homozygous female Fabry mice may be due to the portion of testosterone-induced Gb 3 in males. It is interesting that in kidney, Gb 3 in male Fabry mice cannot be cleared completely by intravenously infused ␣ -galactosidase A or other forms of therapy ( 8,9,13,24,25 ). Thus it seems that the testosterone-induced Gb 3 is resistant to replenishment of the defi cient enzyme. Besides the sex difference in total Gb 3 levels in kidney, isoform analysis revealed that the relative abundance of each isoform is also sex specifi c. Isoforms C16, C22:1, and C24OHA were signifi cantly higher in males in both wild-type and Fabry mice, suggesting that these isoforms may be testosterone inducible. Ioannou et al. reported that one of the two Gb 3 bands of Fabry mouse kidney extract in thin layer chromatography does not respond to administered enzyme ( 9 ). It is possible that the three isoforms described above may be the major components of the enzyme-resistant Gb 3 band. In addition, there is an apparent decrement of male/female ratio of most Gb 3 species in older Fabry mice (12 months compared with 5 months of age) ( Fig. 2B ). This is probably due to a relatively more rapid Gb 3 accumulation in the female Fabry mice ( Fig. 1B ).
Despite the signifi cant accumulation of Gb 3 and lyso-Gb 3 in kidney of Fabry mice shown in this study and others ( 17 ) and the pathological changes in kidney tissues ( 8,10 ), there are no obvious clinical abnormalities such as proteinuria and renal failure, which are common in Fabry patients, reported in Fabry mice. A systematic characterization Fig. 3. Enzyme activity and mRNA level of kidney ␣ -galactosidase A at 5 months of age. A: Kidney ␣ -galactosidase A activity. The enzyme activity was signifi cantly lower in male wild-type mice compared with female wild-type mice. B: mRNA level. mRNA level of alpha-galactosidase A in kidney was signifi cantly lower in male wildtype mice compared with female mice.