Lack of iGb3 and Isoglobo-Series Glycosphingolipids in Pig Organs Used for Xenotransplantation: Implications for Natural Killer T-Cell Biology

Abstract

α-1,3-Terminated galactose residues on glycoproteins and glycosphingolipids are recognized by natural anti-α-1,3-galactose antibodies in human serum and cause hyperacute rejection in pig-to-human xenotransplantation. Genetic depletion of α-1,3-galactosyltransferase-1 in pigs abolishes the hyperacute rejection reaction. However, the isoglobotriosylceramide (iGb3) synthase in pigs may produce additional α-1,3-terminated galactose residues on glycosphingolipids. In both α-1,3-galactosyltranserase-1 knockout mice and pigs, cytotoxic anti-α-1,3-galactose antibodies could be induced; thus, a paradox exists that anti-α-1,3-galactose antibodies are present in animals with functional iGb3 synthases. Furthermore, iGb3 has been found to be an endogenous antigen for natural killer T (NKT) cells, an innate type of lymphocyte that may initiate the adaptive immune responses. It has been reasoned that iGb3 may trigger the activation of NKT cells and cause the rejection of α-1,3-galactosyltransferase-1-deficient organs through the potent stimulatory effects of NKT cells on adaptive immune cells (see ref.([20])). In this study, we examined the expression of iGb3 and the isoglobo-series glycosphingolipids in pig organs, including the heart, liver, pancreas, and kidney, by ion-trap mass spectrometry, which has a sensitivity of measuring 1% iGb3 among Gb3 isomers, when 5 μg/mL of the total iGb3/Gb3 mixture is present (see ref.([35])). We did not detect iGb3 or other isoglobo-series glycosphingolipids in any of these organs, although they were readily detected in mouse and human thymus and dendritic cells. The lack of iGb3 and isoglobo-series glycosphingolipids in pig organs indicates that iGb3 is unlikely to be a relevant immune epitope in xenotransplantation.

Figure 1

Linear ion-trap mass spectrometry detection of iGb3 and iGb4 in human dendritic cells. GSLs purified from human dendritic cells^[11] were subjected to linear ion-trap MSⁿ analysis as described. Left panels show MS¹ profile of iGb3 (A) and iGb4 (B) generated by a “precursor ion mapping method,”^[34,35] containing multiple lipoforms due to the variation of length and unsaturation of fatty acyl chains. Right panel shows representative MS⁴ profile of human dendritic cell iGb3 (A) and MS⁵ profile of iGb4 (B). Stars indicate signature ions generated from iGb3 (A) and iGb4 (B).

Figure 2

Absence of iGb3 and isoglobo-series GSLs in pig heart, kidney, liver, and pancreas as measured by ion-trap mass spectrometry. GSLs were extracted as described in the text. Permethylation was performed on GSLs to enhance the sensitivity of detection and allow better fragmentation assays. Ceramide, glucosylceramide (GlcCer), lactosylceramide (LacCer), Gb3, Gb4 (and nLc4), and Galα1,3nLc4 were detected. To detect whether the Hex-Hex-Hex-Cer (possibly representing Gb3 or iGb3) contains iGb3, we subjected all MS¹ ions to MS⁴ analysis (see below, Fig. 3). No characteristic fragment ions for iGb3 were found.

Figure 3

Absence of characteristic ions for iGb3 in MS⁴ analysis. Six molecular ions, 1215, 1243, 1271, 1299, 1313, and 1327, representing Hex-Hex-Hex-Cer with different lipid portions of the ceramide part, were subjected to MS⁴ analysis, and only characteristic ions for Gb3 were detected. Characteristic ions for iGb3, 211 and 371, were found in human monocyte-derived dendritic cells (Fig. 1A).

Figure 4

Absence of characteristic ions for iGb4 in MS⁵ analysis. Five molecular ions, 1460, 1488, 1502, 1516, and 1544, representing HexNAc-Hex-Hex-Hex-Cer with different lipid portions of the ceramide part, were subjected to MS⁵ analysis, and only characteristic ions for Gb4 were detected. Characteristic ions for iGb4, 343, 357, and 369, were found in human monocyte-derived dendritic cells (Fig. 1B).

Figure 5

Absence of iGb3 in pig organs as determined by anti-Galα3Gal antibody staining. Gb3 and iGb3 staining of neutral glycolipids isolated from pig kidney, pancreas, heart, and liver. A: The lipids were stained using anti-Gb3 (clone BGR-23, 51) as primary antibody and donkey antihuman IgG as second antibody, and 5 μg of GSLs were loaded with 5 μg of Gb3 as standard per lane of silica gel. B: The lipids were stained using AB human serum anti-Galα3Gal as primary antibody and donkey antihuman IgG as secondary antibody, and 5 μg of each lipid fraction was loaded per lane with 5 μg of iGb3 as standard. For both antibody stainings, GSLs were separated on HPTLC plates under the same conditions. The double-band appearance is due to structural heterogeneity in the ceramide part. The staining signal was specific when primary antibody was used at 0.5 μg/mL. At higher antibody concentrations, we observed a high background, consistent with studies published by other investigators,^[39] who could not show the difference in staining between iGb3 and Gb3 (negative control).

Figure 6

Synthesis of iGb3 and iGb4 by human iGb3 synthase. A: Synthesis of iGb3 by human iGb3 synthase. iGb3 was measured by linear ion-trap MS⁴ technology as described.^[35] Stars indicate signature ions generated from iGb3. Human iGb3 Syn: CHO cells transfected by human iGb3 synthase cloned in pIRES2-EGFP plasmid and sorted by EGFP expression. EGFP: CHO cells transfected by mock pIRES2-EGFP plasmid and sorted by EGFP expression. B: Synthesis of iGb4 in human iGb3 synthase0transfected CHO cells. iGb4 was measured by linear ion-trap MS⁵ technology as described.^[34] Stars indicate signature ions generated from iGb4. Human iGb3 Syn: CHO cells transfected by human iGb3 synthase cloned in pIRES2-EGFP plasmid and sorted by EGFP expression. EGFP: CHO cells transfected by mock pIRES2-EGFP plasmid and sorted by EGFP expression.