O-glycoforms of polymeric immunoglobulin A1 in the plasma of patients with IgA nephropathy are associated with pathological phenotypes

Background. Immunoglobulin A1 (IgA1) O-glycosylation plays an important role in the pathogenesis of IgA nephropathy (IgAN). However, variations in IgA1 O-glycoforms have not been explored. We aimed to investigate the IgA1 O-glycoforms in the hinge region (HR) of polymeric IgA1 (pIgA1) and then evaluate the association between IgA1 O-glycoforms and crescent formation in IgAN. Methods. The discovery cohort (Cohort 1) comprised 11 crescentic IgAN patients, 10 noncrescentic IgAN patients and 10 healthy controls and the validation cohort (Cohort 2) comprised 11 crescentic IgAN patients, 9 noncrescentic IgAN patients and 9 healthy controls. A total of 143 IgAN patients with different crescent percentages (Cohort 3) were also included. pIgA1 was purified from the plasma of the participants. The variation in O-glycoforms was evaluated by estimating the molecular weights of IgA1 hinge glycopeptides using reversedphase liquid chromatography and tandem mass spectrometry under electron-transfer/higher-energy collision dissociation fragmentation mode. Results. In the discovery cohort (Cohort 1), the number of N-acetylgalactosamine (GalNAc) bound to one HR was lower in IgAN patients. The proportions of GalNAc3 (defined as O-glycans bound to one HR at three sites) and GalNAc4 were highest in crescentic IgAN patients, followed by noncrescentic IgAN patients, and were lowest in healthy controls [GalNAc 3: 9.92 +/- 3.37% versus 6.65 +/- 1.53% versus 4.05 +/- 1.24% (P < 0.001); GalNAc4: 45.91 +/- 4.75% versus 41.13 +/- 2.95% versus 40.98 +/- 2.95% (P = 0.004), respectively]. The proportions of GaINAc5 and GalNAc6 were lowest in crescentic IgAN patients followed by noncrescentic IgAN patients and were highest in healthy controls [GalNAc5: 50.15 +/- 4.27% versus 47.92 +/- 4.09% versus 45.87 +/- 3.79% (P = 0.028); GalNAc6: 6.58 +/- 2.53% versus 6.04 +/- 1.35% versus 4.65 +/- 2.27% (P = 0.034), respectively]. These results were consistent in the validation cohort (Cohort 2). In another cohort with 143 patients with different crescent percentages (Cohort 3), the number of GalNAc in pIgA 1 decreased with an increasing percentage of crescents. Conclusions. The number of GalNAc in IgA1 HRs was lower in IgAN patients, especially in crescentic IgAN patients, and may be associated with a severe IgAN phenotype.