Distributions of Globotriaosylceramide Isoforms, and Globotriaosylsphingosine and Its Analogues in an α-Galactosidase A Knockout Mouse, a Model of Fabry Disease

Fabry disease is caused by deficient activity of α-galactosidase A (GLA) and characterized by systemic accumulation of glycosphingolipids, substrates of the enzyme. To gain insight into the pathogenesis of Fabry disease based on accumulated substrates, we examined the tissue and plasma distributions of globotriaosylceramide (Gb3) isoforms, and globotriaosylsphingosine (lyso-Gb3) and its analogues in a GLA knockout mouse, a model of Fabry disease, by means of liquid chromatography-mass spectrometry and nano-liquid chromatography-tandem mass spectrometry, respectively. The results revealed that the contents of these substrates in the liver, kidneys, heart, and plasma of GLA knockout mice were apparently higher than in those of wild-type ones, and organ specificity in the accumulation of Gb3 isoforms was found. Especially in the kidneys, accumulation of a large amount of Gb3 isoforms including hydroxylated residues was found. In the GLA knockout mice, the proportion of hydrophobic Gb3 isoforms was apparently higher than that in the wild-type mice. On the other hand, hydrophilic residues were abundant in plasma. Unlike that of Gb3, the concentration of lyso-Gb3 was high in the liver, and the lyso-Gb3/Gb3 ratio in plasma was significantly higher than those in the organs. The concentration of lyso-Gb3 was apparently higher than those of its analogues in the organs and plasma from both the GLA knockout and wild-type mice. This information will be useful for elucidating the basis of Fabry disease.


Sample preparation
Gb3 isoforms were extracted from three organs (liver, kidneys and heart), and plasma of the wild-type (three males and three females) and GLA knockout mice (six males and six females). The organ tissues (20-100 mg) were homogenized in methanol using 3 mm Zr beads with a homogenizer (Micro Smash MS-100R; TOMY DIGITAL BIOLOGY CO., LTD., Tokyo, Japan). Each homogenate including 5 mg tissue was diluted with 600 μl of methanol containing 500 ng Gb3(C17:0), as an internal standard. Then, 300 μl of chloroform and 100 μl of water were added, and the crude lipids were extracted. In the case of extraction from plasma, 50 μl of plasma was diluted with 100 μl of water, and then 600 μl of methanol containing 500 ng Gb3  (C17:0) and 300 μl of chloroform were added and mixed. Each mixture was centrifuged for 10 min at 14,000g, and then the supernatant was dried in an evaporator. Each residue was reconstituted in 200 μl of methanol and then used for conventional LC-MS/MS analysis. As standard samples, 100 μl of 0 to 100 μg/mL Gb3 (mixture of Gb3(C16:0) and Gb3 (various isoforms) = 1/1) in methanol was extracted following the same procedure.
Lyso-Gb3 and its analogues were also extracted from the organs and plasma of the wild-type (three males and three females) and GLA knockout mice (six males and six females). The homogenates including 5 mg tissue were diluted with 600 μl of methanol containing 1 ng lyso-Gb3-IS. Then, 300 μl of chloroform and 100 μl of water were added and mixed, and then the crude lipids were extracted. For extraction of the crude lipids in plasma, 20 μl aliquots of plasma samples were diluted with 100 μl of water. Then, 600 μl of methanol containing 1 ng lyso-Gb3-IS and 300 μl of chloroform were added and mixed. Each extract was centrifuged for 10 min at 14,000g, and then the supernatant was dried in an evaporator. Each residue was reconstituted with 1% H 3 PO 4 /MeOH, followed by transfer to an OASIS MCX cartridge (30mg, 60mm; Waters Corp., Milford, MA, USA). Extraction of lyso-Gb3 and its analogues by OASIS MCX was performed by the method described in our previous report [16]. As standard samples, 100 μl of 0 to 1500 ng/mL of lyso-Gb3 in methanol was extracted following the same procedure.

Measurement of Gb3 isoforms by means of LC-MS/MS
For measurement of Gb3 isoforms, LC-MS/MS analysis was performed according to a previous report [27]. For conventional-LC, an Ultimate 3000 (Thermo Fisher Scientific, Inc., Waltham, MA, USA) was used. Hypersil Gold (150 mm x 4.6mm, 3um particles; Thermo Fisher Scientific) was used for reversed phase based chromatographic separation. Five micro liters aliquots of samples were injected. The column oven was set at 40°C. Solvent A consisted of ACN/ MeOH/water (19/19/2), 20 mM ammonium formate and 5 mM formic acid, and Solvent B of IPA/water (100/1), 20 mM ammonium formate and 5 mM formic acid. A mobile-phase gradient was produced during a 35 min run: 0 min, 10% B; 25 min, 25% B; 25.1 min, 95% B; 30 min, 95% B; 30.1 min, 10% B; and 35 min, 10% B. The flow rate was 0.5 mL/min. A Q-Exactive mass spectrometer (Thermo Fisher Scientific) was used for the detection of Gb3 isoforms. Instrument calibration was performed before each analysis. The samples were injected into the mass spectrometer from 2 to 25 min by a switching valve. The targeted MS/MS analysis (HRPS) mode was selected for quantification of Gb3 isoforms. The ion spray voltage was set at 2500V in the positive ion mode. The first quadrupole was operated at 1.0 FWHM and the Orbitrap spectrometer at 70,000 FWHM. The AGC target value was set at 2E5, with a maximum injection time of 100 ms. The collision energy value was 40% for all the compounds of interest. The theoretical masses of Gb3 isoforms, which had various fatty acids linked with various sphingosine moieties corresponding to lyso-Gb3 and four analogues (Fig 1B), were targeted. The target masses and acquisition times are shown in Table 1. The precursor ions of Gb3 isoforms were sodium adducts. Fragment ions due to neutral loss of a single galactosyl fragment (162.05 Da) from the precursor ion were selected for quantification of Gb3 isoforms with our instrument ( Table 1). The calculation for measurement of Gb3 isoforms was performed using the Quan Browser software (Thermo Fisher Scientific). The total Gb3 concentration was calculated from the sum of all the Gb3 isoforms we had detected.

Statistical Analysis
The statistical significance test was performed with Stat Preclinica SAS 9.2 (Takumi Information Technology Inc., Tokyo, Japan). Data are expressed as means ± standard deviation. The differences in the Gb3 and lyso-Gb3 levels in organs and plasma between the GLA knockout mice and the wild-type ones were assessed by means of Student's t test.

Identification of MS peaks
The theoretical masses of Gb3 isoforms, having various fatty acids linked with various sphingosine moieties corresponding to lyso-Gb3 and its four major analogues (lyso-Gb3(-2), lyso-Gb3 (+16), lyso-Gb3(+18) and lyso-Gb3(+34)) [16], were targeted ( Fig 1B). Fragment ions due to neutral loss of a single galactosyl fragment (162.05 Da) from the precursor ion, not fragment ions from a ceramide moiety, were mainly observed with our instrument. Thus, we used them for quantification of Gb3 isoforms. We found that the ceramide moiety of Gb3 isoforms influenced the elution order of different molecular species, i.e., the molecules having short, unsaturated and hydroxylated fatty acids linked with sphingosine were eluted earlier (S1 Fig). From this, we determined the elution time corresponding to the MS peak of each Gb3 isoform. We set the acquisition time window for each compound of interest so as to obtain enough data points. The target masses and acquisition times are shown in Table 1. Some peaks were considered to be mixtures of two or more structural isomers having the same molecular mass.

Contents of total Gb3 and its isoforms in organs and plasma
As shown in Table 2, the concentration of Gb3 in the kidneys was apparently higher than those in the heart and liver in both the wild-type and GLA knockout mice. The mean Gb3 concentrations in all the organs and plasma of the GLA knockout mice were significantly higher than those in the wild-type mice (p<0.01). The distributions of Gb3 isoforms differs from organ to organ. Various Gb3 isoforms were observed especially in the kidneys (Fig 2), and kidney-specific Gb3 isoforms were hydroxylated. In the GLA knockout mice, the concentrations of hydrophobic isoforms (e.g., Gb3(d18:1-C24:0)) in the organs were higher than those of hydrophilic Gb3 isoforms (e.g., Gb3(d18:1-C16:0)). However, in the plasma, the concentrations of hydrophobic Gb3 isoforms were not much higher than those of hydrophilic Gb3 isoforms. On the other hand, in the wild-type mice, the concentrations of hydrophilic isoforms were apparently high in the plasma.

Contents of lyso-Gb3 and its analogues in organs and plasma
The lyso-Gb3 concentration was measured by means of nano-LC-MS/MS. Lyso-Gb3 in all the organs and plasma of the wild-type mice was successfully detected by means of this sensitive method ( Table 2). Unlike that of Gb3, the concentration of lyso-Gb3 was high in the liver but not in the kidneys. The mean lyso-Gb3 concentrations in all the organs and plasma of the GLA knockout mice were significantly higher than those in the wild-type mice (all, p<0.01). The lyso-Gb3 to Gb3 ratio in the plasma was higher than those in the organs in both the wild-type and GLA knockout mice. The concentration of lyso-Gb3 was higher than those of its analogues in all the organs and plasma from both the wild-type and GLA knockout mice. Among the lyso-Gb3 analogues, the concentrations of lyso-Gb3(-2) and lyso-Gb3 (+18) were higher than those of the others (Fig 3).

Discussion
To elucidate the basis of Fabry disease based on accumulated substrates of GLA, we examined the distributions of Gb3 isoforms and lyso-Gb3 and its analogues in the liver, kidneys, heart and plasma of 8-month-old GLA knockout mice and compared them with in the case of wildtype ones. It is reported that the Gb3 level of the organs in GLA knockout mouse increases up to 20-weeks of age and is stable after 20 weeks of age [26]. Therefore, 8-month-old GLA knockout mouse is well-suited for examining the glycosphingolipids distribution in the stationary phase. The results revealed that the contents of these substrates in the organs and plasma from the GLA knockout mice were apparently higher than those in the wild-type ones. Especially in the kidneys, which is one of the most affected organs in this disease, the accumulation of a large amount of Gb3 was found.
As to the distributions of Gb3 isoforms, organ specificity was found. In the kidneys, various kinds of hydroxylated residues were found in both the wild-type and GLA knockout mice. Such a characteristic distribution of Gb3 isoforms in the kidneys may be due to the organ-specific biosynthetic pathway for Gb3. In the organs of the GLA knockout mice, the proportion of hydrophobic Gb3 isoforms was apparently higher than that in the wild-type mice, suggesting that excessive accumulation of such hydrophobic Gb3 isomers affects the organ and tissues. On the other hand, in plasma, the The distributions of Gb3 isoforms in organs and plasma of the wild-type and GLA knockout mice. Numbers along the x-axis are elution orders from the RP column and correspond to numbers in Table 1. lyso-Gb3 analogues in plasma from Fabry patients [16]. However, the present study revealed that, in the GLA knockout mice, the concentrations of lyso-Gb3(-2) and lyso-Gb3(+18) were higher than those of the other lyso-Gb3 analogues. This would be due to a difference between humans and mice.
In this study, we elucidated the distributions of Gb3 isoforms and lyso-Gb3 and its analogues in wild-type and GLA knockout mice. This information will be useful for elucidating the basis of Fabry disease.